+ All Categories
Home > Documents > DDMRP - Demand Driven Material Requirements Planning

DDMRP - Demand Driven Material Requirements Planning

Date post: 11-Sep-2021
Category:
Upload: others
View: 5 times
Download: 0 times
Share this document with a friend
417
Transcript
Page 1: DDMRP - Demand Driven Material Requirements Planning
Page 2: DDMRP - Demand Driven Material Requirements Planning

DemandDrivenMaterialRequirementsPlanning(DDMRP)

Page 3: DDMRP - Demand Driven Material Requirements Planning

DemandDrivenMaterialRequirementsPlanning(DDMRP)

CarolPtakandChadSmith

INDUSTRIALPRESS,INC.

Page 4: DDMRP - Demand Driven Material Requirements Planning

IndustrialPress,Inc.32HavilandStreet,Suite3

SouthNorwalk,Connecticut06854Tel:203-956-5593,Toll-Free:888-528-7852

E-mail:[email protected]

LibraryofCongressCataloging-in-PublicationDataNames:Ptak,CarolA.,author.|Smith,Chad,1971–author.Title:Demanddrivenmaterialrequirementsplanning(DDMRP)/CarolPtakandChadSmith.Description:SouthNorwalk,Connecticut:IndustrialPress,Inc.,2016.|Includesbibliographicalreferencesandindex.Identifiers:LCCN2016023395 (print) |LCCN2016026863 (ebook) | ISBN9780831135980(hardcover: alk. paper) | ISBN 9780831193843 (eBook) | ISBN 9780831193850 (Epub) |ISBN9780831193867(EMobi)Subjects:LCSH:Materialrequirementsplanning.|Manufacturingresourceplanning.|Productioncontrol—Dataprocessing.|Inventorycontrol—Dataprocessing.Classification:LCCTS161 .P7892016 (print) |LCCTS161 (ebook) |DDC658.5/03—dc23LCrecordavailableathttps://lccn.loc.gov/2016023395

ISBNprint:978-0-8311-3598-0ISBNePUB:978-0-8311-9385-0ISBNeMOBI:978-0-8311-9386-7ISBNePDF:978-0-8311-9384-3

Copyright©2016byIndustrialPress,Inc.Allrightsreserved.

Thisbook,oranypartsthereof,withtheexceptionofthosefiguresinthepublicdomain,maynot be reproduced, stored in a retrieval system, or transmitted in any form without thepermissionofthecopyrightholders.

SponsoringEditor:JudyBassCopyEditor:JudyDuguid

Compositor:PatriciaWallenburg,TypeWritingCoverDesigner:JanetRomanoMurray

industrialpress.comebooks.industrialpress.com

10987654321

TheDemandDriven Institute logo isa trademarkof theDemandDriven Institute.All rightsreserved.TheCertifiedDemandDrivenPlannerlogoisatrademarkoftheInternationalSupplyChainEducationAlliance(ISCEA).Allrightsreserved.

Page 5: DDMRP - Demand Driven Material Requirements Planning

Contents

ForewordDefinitionsinThisBookIntroductionAbouttheAuthorsAcknowledgments

PART1PERSPECTIVE

CHAPTER1

PlanningintheNewNormalTheMaterialRequirementsPlanningRevolutionEvidenceofaProblem

ReturnonAssetPerformanceDegradationWork-AroundProliferationTheInventoryBimodalDistribution

TheNewNormalSummary

CHAPTER2

TheImportanceofFlowPlossl’sFirstLawEstablishingFlowastheFoundationRelevantInformationandMaterialsTheBullwhipEffect

Page 6: DDMRP - Demand Driven Material Requirements Planning

Summary

CHAPTER3

MaterialRequirementsPlanningintheNewNormalWhatIsMRP?DistortionstoRelevantInformation

DemandSignalInputNervousnessTheWeeklyBucketFlatteningtheBillofMaterial

DistortionstoRelevantMaterialsCommonCauseVariationDelayAccumulation

Amplifying the Distortions to Relevant Information and Materials—BatchingPolicies

Summary

CHAPTER4

UnlockingaSolution—ThePowerofDecouplingDecouplingDecouplingPointBuffersSummary

PART2BECOMINGDEMANDDRIVEN

CHAPTER5

SupplyOrderGenerationandExecutionfortheNewNormalMRPVersusLean—WhatCanWeLearn?

DependenceVersusIndependenceSupplyOrderGeneration(PlanningVersusExecution)LeanandTechnology

DemandDrivenMaterialRequirementsPlanningTheHistoryof“DemandDriven”Position,Protect,andPull

Page 7: DDMRP - Demand Driven Material Requirements Planning

Summary

CHAPTER6

StrategicInventoryPositioningPositioningFactors

CustomerToleranceTimeMarketPotentialLeadTimeSalesOrderVisibilityHorizonExternalVariabilityInventoryLeverageandFlexibilityCriticalOperationProtection

ApplyingthePositioningCriteriaANewFormofLeadTimeAdvancedInventoryPositioningConsiderationsDistributionPositioningConsiderationsSummary

CHAPTER7

StrategicBuffersInventory:AssetorLiability?IntroducingDecouplingPointBuffers

TheGreenZoneTheYellowZoneTheRedZone

BufferProfilesFactor1:ItemTypeFactor2:LeadTimeFactor3:Variability

IndividualPartAttributesPartAverageDailyUsagePartLeadTimePartMinimumOrderQuantityPartLocation

CalculatingReplenishedPartBufferLevelsandZonesTheGreenZoneTheYellowZone

Page 8: DDMRP - Demand Driven Material Requirements Planning

TheRedZoneContinuingwithCompanyABC

IterationsCalculatingReplenishedOverrideBuffersCalculatingMin-MaxBuffersSummary

CHAPTER8

BufferAdjustmentsRecalculatedAdjustmentsPlannedAdjustmentFactors

DemandAdjustmentFactorZoneAdjustmentFactorLeadTimeAdjustmentFactor

Summary

CHAPTER9

DemandDrivenPlanningTheShifttoActualDemandTheNetFlowEquation

QualifyingOrderSpikesSupplyOrderGenerationBasedonNetFlowPositionSimulatingDDMRPSupplyOrderGenerationCalculatingAverageOn-HandInventory

AverageOn-HandRangeAverageOn-HandTargetAverageOpenSupplyOrders

DecoupledExplosionHybridModelSupplyOrderGenerationPrioritizedShareSupplyOrderConsideration

DiscountOptimizationFreightOptimizationCoverageOptimization

Min-MaxSupplyOrderGenerationCompletingtheCompanyABCExampleSummary

Page 9: DDMRP - Demand Driven Material Requirements Planning

CHAPTER10

DemandDrivenExecutionBufferStatusAlerts

ChallengingPrioritybyDueDatePlanningVersusExecutionDisplayCurrentOn-HandAlertProjectedOn-HandAlert

SynchronizationAlertsMaterialSynchronizationAlertsLeadTimeAlert

Summary

CHAPTER11

DDMRPImpactsontheOperationalEnvironmentDDMRPStrategicBufferCriteria

TheDecouplingTestTheBidirectionalBenefitTestTheOrderIndependenceTestThePrimaryPlanningMechanismTestTheRelativePriorityTestTheDynamicAdjustmentTest

DDMRPVersusSafetyStockandOrderPointSafetyStockandtheBufferCriteriaOrderPointandtheBufferCriteria

DDMRPImpactsonSchedulingDDMRPandMasterProductionSchedulingAssumptionsDDMRPShopFloorSchedulingImplicationsFiniteSchedulingwithDDMRP?AdditionalSchedulingSequenceImpacts

DDMRPandWIPPriorityManagementSummary

CHAPTER12

DDMRPMetricsandAnalyticsMeasuringRelevantInformation(SignalIntegrity)

Page 10: DDMRP - Demand Driven Material Requirements Planning

MeasuringDecouplingPointIntegrityOutlyingEventReports

MeasuringVelocityDrivingImprovementinDDMRPSummary

CHAPTER13

TheDemandDrivenOrganizationTheDemandDrivenAdaptiveSystemDemandDrivenSalesandOperationsPlanningConventionalSalesandOperationsPlanningTheFiveStepsofDemandDrivenSalesandOperationsPlanning

1.StrategicBusinessManagementDirectionandReview2.IntegratedReconciliation3.ManagingthePortfolioandNewActivities4.ManagingDemand5.ManagingSupply

DemandDrivenSalesandOperationsPlanningProjectionsWorkingCapitalSpaceCapacityProjectedOrderFrequency

SummaryContributionofDickLing

CHAPTER14

ImplicationsforTechnologyOperations and Information Technology—Two Ships Diverging in the

Night?DDMRPSoftwareComplianceCriteria

PART3APPENDICES

APPENDIXA

AnMRPExample

Page 11: DDMRP - Demand Driven Material Requirements Planning

TheScenarioSimulatingtheScenario

APPENDIXB

SimulatingDDMRPBuffersAbouttheSimulationWidgetSimulationResultsGazoonkSimulationResultsSimulatingtheImpactofMinimumOrderQuantitiesSummaryAbouttheAuthor

APPENDIXC

ApplyingDDMRPtotheApparelRetailEnvironmentTheNeedforaRetailApplicationofDDMRPARetailApparelDDMRPExampleSpecialCharacteristicsandChallengesoftheApparelRetailEnvironmentTheProposedModelRetailDDMRPBufferZoneConsiderationsRealizedResultsAbouttheAuthor

APPENDIXD

DemandDrivenMRPDictionary

APPENDIXE

DDS&OPChecklist

EndnotesReferencesIndex

Page 12: DDMRP - Demand Driven Material Requirements Planning

Foreword

When Iwas asked to helpwrite the chapter on S&OP for this book, Iwascomplimented and eager to make a contribution. With the development ofDDMRPandnowthewritingofthisbook,PtakandSmithhavebrokenthroughcommonpracticetobringcommonsensetosupplychainmanagement.

If your company is facing variability and uncertainty across your supplychainand the future looks little like thepast, then thisbookholds theanswer.DDMRP represents the future of planning in today’s complex and volatilesupplychains.Inherentflawsinthetraditionalplanningapproachesareexposedand resolved for today’s complex adaptive supply chains. With the DemandDrivenAdaptiveSchemaandthepivotalpositionofDemandDrivenSalesandOperationsPlanning,thisisnotjustabetterwaytoplan;itisabetterwaytorunan organization in today’s hypercompetitive environment. Operations andstrategycannoweasilyandrealisticallybeconnectedbi-directionally,allowingbothtoadapttocriticalchangesforthebestreturnonshareholderequity.

Thisbookistheultimatereferenceforthisnewwayoflifeacrossadynamicadaptivesupplychain.

DickLingS&OPConsultantAuthor of OrchestratingSuccess

Page 13: DDMRP - Demand Driven Material Requirements Planning

DefinitionsinThisBook

Thisbookwilluse twosourcesofdefinitions.Allknownandaccepted termsthat are not new with the advent of Demand Driven Material RequirementsPlanning(DDMRP)willbedefinedusingdefinitionsfromthefourteentheditionof the APICS Dictionary. The authors thank APICS for its support of thisproject. Since 1957, APICS has been the premier professional association forsupplychainandoperationsmanagementand the leadingproviderof research,education, and certification programs that elevate supply chain excellence,innovation,andresilience.

FortermsthatarenewwiththeadventofDDMRP,theauthorshavecreatedadictionaryspecifictoDDMRP.ThisdictionarycanbefoundinAppendixDofthis book. Translated versions of this dictionary inmultiple languages can befoundinthedownloadsectionatwww.demanddriveninstitute.com.

Page 14: DDMRP - Demand Driven Material Requirements Planning

Introduction

This isnotabookabout the intricaciesof traditionalMaterialsRequirementsPlanning (MRP). In 2011, at the request of McGraw-Hill, the authors wroteOrlicky’s Material Requirements Planning, third edition. That book was 542pagesandprovidedanexpansiveviewofconventionalplanningtacticsthatwereborninthe1950s,codifiedinthe1960s,andcommercializedinthe1970s.Thatbookalsodevotednearly100pagestoanemergingalternativemethodofformalplanning and execution—Demand Driven Material Requirements Planning(DDMRP).Thisbookisentirelyaboutthatalternativemethod.

ThisbookwillprovideanextensiveblueprintforDDMRP.Itistheauthors’intentionthatthistextwillservethesamepurposeasthefirstMRPbookwrittenby Joe Orlicky in 1975 by ushering in a new era in planning and executionmethodology,rules,andtools.Thebodyofknowledgeofthisalternativemethodisadvancing rapidlyandgainingacceptanceworldwide.Theauthorshope thatthis book will open the door for many other books on specific aspects,applications,andextensionsofthismethod.

Page 15: DDMRP - Demand Driven Material Requirements Planning

AbouttheAuthors

[email protected].

ChadSmith

ChadSmith is thecoauthor (withCarolPtak)of the thirdeditionofOrlicky’sMaterial Requirements Planning and the coauthor (with Debra Smith) ofDemandDrivenPerformance:UsingSmartMetrics.He is a cofounder of andpartner in the Demand Driven Institute, an organization dedicated toproliferatingdemanddrivenmethodsthroughouttheworld.Mr.Smithservesasthe Program Director of the International Supply Chain Education Alliance’sCertifiedDemandDrivenPlannerProgram.

In 1997,Mr. Smith cofounded ConstraintsManagementGroup (CMG), aservicesandtechnologycompanyspecializingindemanddrivenmanufacturing,materials, and project management systems for midrange and largemanufacturers. He served as Managing Partner of CMG from 1998 to 2015.Clients,pastandpresent, includeUnilever,LeTourneauTechnologies,Boeing,Intel, EricksonAir-Crane, Siemens, IBM,TheCharlesMachineWorks (DitchWitch), and Oregon Freeze Dry. Mr. Smith is also a certified expert in alldisciplinesoftheTheoryofConstraints,studyingdirectlyunderthetutelageof

Page 16: DDMRP - Demand Driven Material Requirements Planning

thelateDr.EliGoldratt.Chad Smith makes his home in Wenatchee, Washington, with his wife,

Sarah,andtwodaughters,SophiaandLily.

CarolPtak

Carol Ptak is currently a partner with the Demand Driven Institute and waspreviously a Visiting Professor and Distinguished Executive in Residence atPacificLutheranUniversity.Beforegoing toacademiashewasVicePresidentandGlobalIndustryExecutiveforManufacturingandDistributionIndustriesatPeopleSoft,whereshedevelopedtheconceptofdemanddrivenmanufacturing.Ms. Ptak spent four years at IBMCorporation, culminating in the position ofGlobalSMBSegmentExecutive.

A leading authority in the use of ERP and supply chain tools to driveimproved bottom-line performance, Ms. Ptak has an expertise that is wellgrounded in four decades of practical experience as a successful practitioner,consultant,andeducatorinmanufacturingoperations.Herpragmaticapproachtocomplex issues and her dynamic presentation stylemake her a person in highdemandworldwide on the subject of how to leverage these tools and achievesustainablesuccess.

Ms. Ptak holds an MBA from Rochester Institute of Technology andcompletedtheEMPOprogramatStanfordUniversity.Sheisafrequenteducatorattheuniversitylevelandpresentsatmanykeytechnicalconferencesaroundtheworld,includingconferencesinSouthAfrica,France,Israel,Australia,Ireland,andtheNetherlandsand11APICSInternationalConferences.Sheistheauthorofnumerousarticlesand thebooksOrlicky’sMaterialRequirementsPlanning,thirdedition,withChadSmith;MRPandBeyond;ERP:Tools,Techniques,andApplications for Integrating the Supply Chain; Theory H.O.W.: HowOrganizations Could Work with Harold Cavallaro; and Necessary but NotSufficientwithEliGoldrattandEliSchragenheim.TogetherwithDeanGilliam,

Page 17: DDMRP - Demand Driven Material Requirements Planning

sheupdatedQuantumLeap,originallywrittenbyJohnConstanza.Ms.Ptakhaslent her name to the internationally coveted Ptak Prize for Supply ChainExcellencethat isawardedannuallybyISCEA,theInternationalSupplyChainEducationAlliance.

Ms.Ptak is certified throughAPICSat the fellow level (CFPIM)andwascertified in Integrated Resource Management (CIRM) with the first groupinternationally. Ms. Ptak was the President and CEO of APICS for the year2000.Prior to her election asAPICSPresident, she served at theSociety in avarietyofpositions.

Page 18: DDMRP - Demand Driven Material Requirements Planning

Acknowledgments

IfIhaveseenfurther,itisbystandingontheshouldersofgiants.SIRISAACNEWTONTOROBERTHOOKE,FEBRUARY5,1676

This book is truly built on the shoulders of many giants. From the originalwork of the practitionerswho developedMRP, including JoeOrlicky,GeorgePlossl,Richard(Dick)Ling,andOllieWight,tothegreatthinkersbehindLean,SixSigma,andTheoryofConstraints—TaiichiOhno,W.EdwardsDeming,andEliyahu Goldratt. The authors have stood on the shoulders of these giants tounitethesedifferenttheoriesandmethodologiesandtakealeapforward—intoafutureofplanningwith relevantvisibility thatmitigates thevolatile,uncertain,andvariableworldthatseemsimpossibletoplan.Wehaveknownmanyofthesegiantspersonallyandwishtoexpressourappreciationtothem.

CollectivelytheauthorswouldliketothanktheInternationalSupplyChainEducation Alliance (ISCEA) and members of the Demand Driven InstituteGlobal Affiliate Network for a great partnership in bringing demand drivenconcepts to the mainstream throughout the world. Additionally, the authorswouldliketothankvariousmembersoftheAPICScommunityfortheiramazinginput and support in trying to restore the promise and effectiveness of formalplanning.ThosepeopleincludeKeithLaunchbury,KenTitmuss,BobReary,andAbeEshkenazi.

Theauthorswouldliketopointoutparticularindividualswhohavemadealasting contribution to the demand driven body of knowledge. These peopleincludeGregCass,DebraSmith,ErikBush,DavidPoveda,DickLing,PaddyRamaiyengar,KirkBlack,CarolineMondon,andLaurentVigouroux.

Theauthorswouldliketohighlightafewindividualsandorganizationsthat

Page 19: DDMRP - Demand Driven Material Requirements Planning

have been instrumental in spreading the DDMRPmessage. CarolineMondonandtheFapicsorganizationledanamazingchargeinFrance.PhilippeBornert,BernardMilian,andthewholeAgileateamhavebeeninstrumentalinsupportingthatchargeinFrance.KenTitmussandtheSAPICSorganizationbroughttheseconcepts to Africa. David Poveda was instrumental in starting a massiveproliferationinSouthAmerica.

ChadSmithwouldlikethankhiswife,Sarah,andtwodaughters,SophiaandLily, for putting upwith the prolonged absences and locked office door. Thesupportand loveof these threepeoplehaskepthimgoing.Additionally,ChadwouldliketothankCarmineMainieroandNickMantenutofortheirdedicationinbringingdemanddrivenconceptstoaglobalgiantintheworldoffast-movingconsumer goods; what a learning experience! Chad would also like toacknowledgetheteamatDemandDrivenTechnologiesanditsCEO,ErikBush,in believing in bringing real and sustainable results to customers.ChadwouldadditionallyliketothanktheteamatConstraintsManagementGroup,LLC,foranamazing journey fornearly20years.Finally,Chadwould like to thankhispartnerandcoauthorCarolPtakforaveryrewardingandfulfillingpartnership.

CarolPtakwouldliketothankherhusband,Jim,fortheunderstandingandthesupporttokeepgoingandtothankherparents,DorothyandBud,whotaughtherfromtheyoungestage thatshewas limitedonlybyher imagination.Carolwould especially like to thank Chad Smith for an incredible experience andpartnership—far beyond any imagination. Chad opened all our eyes to thedeepertruthofanewworldofplanning.Ithasbeenanhonorandaonce-in-a-lifetimeexperience.

Page 20: DDMRP - Demand Driven Material Requirements Planning

DemandDrivenMaterialRequirementsPlanning(DDMRP)

Page 21: DDMRP - Demand Driven Material Requirements Planning

PART1

Perspective

Page 22: DDMRP - Demand Driven Material Requirements Planning

CHAPTER1

PlanningintheNewNormal

To truly understand where industry is today, it is necessary to discuss thehistory behind conventional planning. Where did it come from? What did itreplace? What circumstances was it developed for? Is it still relevant andappropriatefortheenvironmentoftoday?

TheMaterialRequirementsPlanningRevolution

TodaymostmidrangeandlargemanufacturingenterprisesthroughouttheworlduseaplanningmethodandtoolcalledMaterialRequirementsPlanning(MRP).This method and tool was conceived in the 1950s with the increasingavailability,promise,andpowerofcomputers.Computersallowedforrapidandcomplexcalculationsaboutwhat andhowmuchwasneeded tobebought andmadegivenademandinput.

Themore complex the products, themore powerful the promise ofMRP.TheAPICSDictionary1definesMRPas:

A set of techniques that uses bill of material data, inventory data, and themasterproductionscheduletocalculaterequirementsformaterials. Itmakesrecommendations to release replenishment orders for material. Further,because it is time-phased, it makes recommendations to reschedule openorderswhenduedatesandneeddatesarenotinphase.Time-phasedMRPbeginswiththeitemslistedontheMPSanddetermines(1)thequantityofallcomponentsandmaterialsrequiredto fabricatethose itemsand(2) thedatethat the components and material are required. Time-phased MRP isaccomplished by exploding the bill of material, adjusting for inventoryquantities on hand or on order, and offsetting the net requirements by theappropriateleadtimes.(p.103)

By 1965 the modern acronym “MRP” was in existence. Then in 1972

Page 23: DDMRP - Demand Driven Material Requirements Planning

capacityreconciliationwasincorporatedintoMRP.Thiswascalledclosed-loopMRP.Theyear1980 saw the significant incorporationof cost accounting intoMRP, transforming it into a system known as Manufacturing ResourcesPlanning (MRP II). Finally, by 1990, as client-server architecture becameavailable, MRP II had evolved into Enterprise Resources Planning (ERP).Throughout this progression the definition of the MRP portion of theinformationsystemhasremainedunchanged.

WhilethisisnotabookaboutMRP,abasiclevelofunderstandingofMRPwillbehelpfultothereader.Thisbasicexplanation,andevenademonstrationofMRP,isincludedinChapter3andAppendixA,respectively.

PerhapsthemostrecognizedleaderoftheMRPchargewasJoeOrlicky.His1975 seminalworkMaterialRequirementsPlanning:TheNewWayof Life inProductionandInventoryManagementprovided theblueprintandcodificationof MRP that is still the standard today. Consider that when this book waswritten, only 700 companies or plants in the world had implemented MRP,almostalllocatedintheUnitedStates:

Asthisbookgoesintoprint,therearesome700manufacturingcompaniesorplants that have implemented, or are committed to implementing, MRPsystems. Material requirements planning has become a new way of life inproduction and inventorymanagement, displacing oldermethods in generaland statistical inventory control in particular. I, for one, have no doubtwhateverthatitwillbethewayoflifeinthefuture.(p.ix)

MRP did become the way of life in manufacturing. The codification andsubsequent commercialization of MRP fundamentally changed the industrialworld, and it did so relatively quickly.Orlicky, alongwith others at the time,recognizedtheopportunitypresentedbychangingmanufacturingcircumstancesandtheinventionofthecomputerthatenabledaplanningapproachneverbeforepossible:

Traditional inventorymanagement approaches, in pre-computer days, couldobviously not go beyond the limits imposed by the information processingtoolsavailableatthetime.Becauseofthisalmostallofthoseapproachesandtechniquessufferedfromimperfection.Theysimplyrepresentedthebestthatcould be done under the circumstances. They acted as a crutch andincorporatedsummary,shortcutandapproximationmethods,oftenbasedontenuousorquiteunrealisticassumptions,sometimesforce-fittingconcepts torealitysoastopermittheuseofatechnique.

Thebreakthrough,inthisarea,liesinthesimplefactthatonceacomputer

Page 24: DDMRP - Demand Driven Material Requirements Planning

becomes available, the use of such methods and systems is no longerobligatory. Itbecomes feasible tosortout, revise,ordiscardpreviouslyusedtechniques and to institute new ones that heretofore it would have beenimpracticalorimpossibletoimplement.Itisnowamatterofrecordthatamongmanufacturing companies that pioneered inventory management computerapplications in the 1960s, themost significant results were achieved not bythosewhochosetoimprove,refine,andspeedupexistingprocedures,butbythosewhoundertookafundamentaloverhauloftheirsystems.(p.4)

Inhisbook,Orlickymadethecaseforafundamentalreexaminationofhowcompanies planned and managed inventory and resources. This case was socompellingthattheconceptsthathebroughttothetableproliferatedthroughoutthe industrial world within two decades. That proliferation remains largelyunchanged in the present. Today we know that nearly 80 percent ofmanufacturingcompaniesthatbuyanERPsystemalsobuyandimplement theMRPmoduleassociatedwiththatsystem.

Perhaps themost interesting and compelling part of the passage from theoriginalOrlickybookisthesentencethatisitalicized.Thiswassimplycommonsense that was easily demonstrable with the results of precomputer inventorymanagement systems. Yet could this same description be applied to thewidespread use of MRP today? Could it be that conventional planningapproachesandtoolsare:

Actingasacrutch?Incorporatingsummary,shortcut,andapproximationmethodsbasedontenuousassumptions?

Force-fitting concepts to reality so as to permit the use of atechnique?

In the authors’ 60+ years of combinedmanufacturing experience across awidearrayof industries, theanswer isa resoundingyes toall thesepoints.Bytheendofthisbook,thereaderwillalsobeabletounderstandwhytheanswerisyestoallthesepoints.Indeediftheanswerisyestothesepoints,thereshouldbeevidence to support the assertion thatMRP systems are not living up to theirbilling—thattheyareinfactguiltyaschargedinthepreviousthreebulletpoints.

Beforewereviewtheevidence,let’sstartwithtwobasicobservationsaboutrules:

Observation1.Mostrulesarelifelimited.Rulesareinstitutedmost

Page 25: DDMRP - Demand Driven Material Requirements Planning

oftenbasedonassumptionsabouttheenvironmentatthetimetheyaremade.Rulesareoftenmadetoaccommodatecertainlimitations.When those assumptions or limitations change, the rules must bereexaminedtodeterminewhethertheyarestillappropriate.Souder’slaw states that “repetition does not establish validity.” Simplycontinuing to do something that has always been done does notdefinewhether it is or ever has been the appropriate thing to do.Worse yet, the longer the repetition, the more invalid orinappropriatetherulemaybe.

Observation 2. “Optimizing” inappropriate rules iscounterproductive. Attempts and investment meant to enable oraccelerate compliance to rules that are inappropriate can bedevastating toanorganization. If therule isnotonly inappropriatebutalsodamaging,thentheorganizationisatrisktodothewrongthingsfaster.

EvidenceofaProblem

There are three areas that point to major issues with the rules and tools ofconventionalplanningfeaturingMRP.

ReturnonAssetPerformanceDegradation

As described above, the United States led the adoption of manufacturinginformation systems starting with MRP in the 1960s. These systems areexpensivetopurchase,toimplement,andtomaintain.Thevalueoftheseformalplanning systems has always been based on the ability to better leverage theassets of a business. Did the widespread adoption of MRP and subsequentinformationsystemsenabletheU.S.economytobettermanageassets?

In late2013DeloitteUniversityPress releaseda reportbyJohnHagel III,John Seely Brown, Tamara Samoylova, and Michael Lui that is quite eye-opening when considered against the progression and adoption rates ofinformationsystems.2Figure1-1isachartfromthereportthatdepictsthereturnonassetperformanceoftheUnitedStateseconomysince1965.

There is a steady decrease in return on assets for theU.S. economy from1965to2012.Furthermore,during this timeperiod thesamereportshowsthatlaborproductivity (asmeasuredbyTornqvist aggregation)more thandoubled!Whatismostinterestingaboutthisgraphicinrelationtoinformationsystemsis

Page 26: DDMRP - Demand Driven Material Requirements Planning

that by 1965we had themodern acronymMRP, butmassive proliferation ofinformationsystemsdidnotoccuruntilafter1975and,inparticular,after1980withMRPII.

Obviously there are many factors at play with this decrease in return onassets,but thisreportwouldcertainlyleadonetorealizethat theimpactof thewidespreadadoptionofMRP,MRPII,andERPsystems(atleastintheUnitedStates) has not significantly helped companies manage themselves to betterreturnsonassetperformance.Indeed,whenthisdeclineistakenincombinationwith the increase in labor productivity, it actually suggests that companies areacceleratingtheirmistakes.

But this is just one point of data, a high-level viewwithmany unrelatedfactors contributing to these effects. What additional evidence indicts theefficacyoftheconventionalplanningapproach?

FIGURE1-1ReturnonassetpeformancefortheU.S.economy

Work-AroundProliferation

Inadditiontoexaminingtheperformanceofanentireeconomyoveraperiodoftime, next examine the day-to-day actions of the people chargedwithmakingdecisions about how to utilize assets. One hallmark of supply chains is thepresenceofsupplyorders.Supplyordersarethepurchaseorders,stocktransferorders, and manufacturing orders that dictate the flow and activities of any

Page 27: DDMRP - Demand Driven Material Requirements Planning

supplychain.Theverypurposeofaplanningsystemistoultimatelydeterminethetiming,

quantity, and collective synchronization of the supply orders up, down, andacrossthelevelsofthenetwork.Insidemostmanufacturerstherearetierswithinthe planning system where stock transfer orders could prompt manufacturingorders that in turn would prompt purchase orders. Additionally, within mostsupplychains thereare tiersofdifferentplanningsystemsateachorganizationlinked togetherby theseordersandcommunicating through thesesupplyordersignals. For example, purchase orders from a customer can prompt stocktransfersormanufacturingordersatsuppliers.

Perhaps thebiggest indictmentof just how inappropriatemodernplanningrules and tools are can be observed in how frequently people choose toworkaroundthem.Thetypicalworkaroundinvolvestheuseofspreadsheets.Dataareextracted out of the planning system and put into a spreadsheet. The data arethenorganizedandmanipulatedwithinthespreadsheetuntilapersonalcomfortlevel is established. Recommendations and orders are then put back into theplanningsystem,essentiallyoverridingmanyoftheoriginalrecommendations.

ConsiderpollingonthissubjectbytheDemandDrivenInstitutefrom2011to 2014. With over 500 companies responding, 95 percent claim to beaugmenting theirplanning systemswith spreadsheets.Nearly70percent claimthese spreadsheets areused to aheavyormoderatedegree.The resultsof thispolling are consistent with other surveys by analyst firms such as AberdeenGroup.Thisrelianceonspreadsheetshasoftenbeenreferredtoas“Excelhell.”Validation for this proliferation can be easily provided by simply asking themembersofaplanningandpurchasingteamwhatwouldhappentotheirabilitytodotheirjobiftheiraccesstospreadsheetsweretakenaway.

But why have planners and buyers become so reliant on spreadsheets?Becausetheyknowthatiftheystayedcompletelywithintherulesoftheformalplanning system, approving all recommendations, it would be very careerlimiting.Tomorrow theywouldundoor reversehalf the things theydid todaybecause MRP is constantly and dramatically changing the picture. Thisphenomenon,knownas“nervousness,”isexplainedinChapter3.

Sowhatdotheydoinstead?Theyworkaroundthesystem.Theyeachhavetheirownwaysofworkingwithtoolsthattheyhavecraftedandhonedthroughtheir years of experience. These ways of working and tools are highlyindividualizedwith extremely limited ability to be utilized by anyone but theoriginator.This isadifferent, informal,highlyvariable,andhighlycustomized

Page 28: DDMRP - Demand Driven Material Requirements Planning

setofrules.Worseyet,thereisnooversightorauditingoftheseside“systems.”Thereis

no “vice president of spreadsheets” in any company the authors have everworkedinorvisited.Everyonesimplyassumesthatthepeoplewhocreatedthesespreadsheets built andmaintain themproperly.Consider an article in theWallStreetJournal’s“MarketWatch”in2013:

Close to 90%of spreadsheet documents contain errors, a 2008 analysis ofmultiple studies suggests. “Spreadsheets, even after careful development,contain errors in 1% or more of all formula cells,” writes Ray Panko, aprofessorof ITmanagementat theUniversityofHawaiiandanauthorityonbadspreadsheetpractices.“Inlargespreadsheetswiththousandsofformulas,therewillbedozensofundetectederrors”(JeremyOlshan,April20,2013)

Asanexampleofhowdisastrousspreadsheeterrorscanbe,considertheroleaspreadsheeterrorplayed ina$6billiondisaster for JPMorgan in2012.Thefollowing is an excerpt from the zerohedge.com article “How aRookieExcelErrorLedJPMorgantoMisreportItsVaRforYears”3:

Just under a year ago,when JPMorgan’s LondonWhale trading fiascowasexposedasmuchmorethanjusttheproverbial“tempestinateapot,”Morganwatchers were left scratching their heads over another very curiousdevelopment:thedramaticsurgeinthecompany’sreportedVaR,whichasweshowed last June nearly doubled, rising by some 93% year over year, aglaringcontrasttowhattheotherbankswerereportingtobedoing.

Specifically, we said that “in the 10-Q filing, the bank reported a VaR of$170millionforthethreemonthsendingMarch31,2012.Thiscomparedtoatiny$88millionforthepreviousyear.”JPM,whichwasdesperatetocoverupthismodellingsnafu,keptmumandshedaslittlelightontheissueaspossible.In itsownwordsfromtheQ1201210-Qfiling:“the increaseinaverageVaRwas primarily driven by an increase in CIO VaR and a decrease indiversificationbenefitacross theFirm.”And furthermore: “CIOVaRaveraged$129millionforthethreemonthsendedMarch31,2012,comparedwith$60million for the comparable 2011 period.The increase inCIO average VaRwasduetochangesinthesyntheticcreditportfolioheldbyCIOaspartof itsmanagementofstructuralandother risksarising fromtheFirm’son-goingbusinessactivities.”Keeptheboldedsentence inmind,becauseasitturnsoutitisnothingbutaeuphemismfor,drumroll,epic,amateurExcelerror!

Howdoweknowthis?WeknowitcourtesyofJPMorganitself,whichintheverylastpageofitsJPMtaskforcereporthadthistosayonthetopicofJPM’sVaR:

“...adecisionwasmadetostopusingtheBaselII.5modelandnottorely

Page 29: DDMRP - Demand Driven Material Requirements Planning

onitforpurposesofreportingCIOVaRintheFirm’sfirst-quarterForm10-Q.Followingthatdecision,furthererrorswerediscoveredintheBaselII.5model,including, most significantly, an operational error in the calculation of therelativechangesinhazardratesandcorrelationestimates.Specifically,aftersubtracting the old rate from the new rate, the spreadsheet dividedbytheir sum instead of their average, as the modeler had intended. Thiserror likely had the effect ofmuting volatility by a factor of two and oflowering the VaR.... it also remains unclear when this error wasintroducedinthecalculation.”

In other words, the doubling in JPM’s VaR was due to nothing but thediscoverythatforyears,someonehadbeenusingagrosslyincorrectformulain theirExcel,andasaresultmisreporting theentire firmVaRbya factorofnearly 50%! Somuch for the official JPM explanation in its 10-Q filing thatsomewhatconvenientlymissedtomentionthat,oops,wemadearookie,firstyearanalysterror.(TylerDurden,February2,2013)

Perhaps a more interesting question is why are personnel allowed to usethesead-hocapproaches?Fromadataintegrityandsecurityperspective,thisisanightmare.Italsomeansthatthefateofthecompany’spurchasingandplanningeffectivenessisinthehandsofafewessentiallyirreplaceablepersonnel.Thesepeoplecan’tbepromotedorgetsickorleavewithoutdireconsequencestothecompany. This alsomeans that due to the error-prone nature of spreadsheets,globallyonadailybasistherearealotofwrongsignalsbeinggeneratedacrosssupply chains.Wouldn’t it be somucheasier to justwork in the system?Theanswer seems so obvious. The fact that reality is just the opposite shows justhowbigtheproblemiswithconventionalsystems.

Tobefair,manyexecutivesaresimplynotawareofjusthowmuchworkisoccurring outside the system.Once they become aware, they are placed in aninstant dilemma. Let it continue, thus endorsing it by default, or forcecompliance to a system that your subject-matter experts are saying is at bestsuspect?Thechoice isonlyeasy the first timeanexecutiveencounters it.Theauthorsof thisbookhave seencountlessexamplesof executivesattempting toend the ad hoc systems only to quickly retreat when inventories balloon andservicelevelsfalldramatically.Theymaynotunderstandwhat’sbehindtheneedforthework-arounds,buttheynowknowenoughtosimplylooktheotherway.SotheymaketheappropriatenoisesabouthowtheentirecompanyisonthenewERPsystemanddownplayjusthowmuchadhocworkisreallyoccurring.

TheInventoryBimodalDistribution

Page 30: DDMRP - Demand Driven Material Requirements Planning

Another piece of evidence to suggest the shortcomings of conventionalMRPsystems has to do with the inventory performance of the companies that usethese systems. To understand this particular challenge, consider the simplegraphicaldepiction inFigure1-2. In this figureyou seea solidhorizontal linerunninginbothdirections.Thislinerepresentsthequantityofinventory.Asyoumove from left to right, the quantity of inventory increases; right to left thequantitydecreases.

FIGURE1-2Taguchiinventorylossfunction

Acurveddottedlinebisectstheinventoryquantitylineattwopoints:

PointA, the pointwhere a companyhas too little inventory.Thispoint would be a quantity of zero, or “stocked out.” Shortages,expedites,andmissedsalesareexperiencedatthispoint.PointAisthepointatwhichthepartpositionandsupplychainhavebecometoobrittleandareunabletosupplyrequiredinventory.Plannersorbuyers that have part numbers past this point to the left typicallyhavesalesandoperationsscreamingatthemforadditionalsupply.

PointB,thepointwhereacompanyhastoomuchinventory.Thereis excessive cash, capacity, and space tied up in working capital.PointBisthepointatwhichinventoryisdeemedwaste.Plannersorbuyers thathavepartnumberspast thispoint to the right typicallyhavefinancescreamingatthemformisuseoffinancialresources.

Ifweknowthat these twopointsexist, thenwecanalsoconclude that foreachpartnumber,aswellas theaggregate inventory level, there isanoptimalrangesomewherebetweenthosetwopoints.Thisoptimalzoneislabeledinthemiddle and colored green.When inventorymoves out of the optimal zone ineitherdirection,itisdeemedincreasinglyproblematic.

Page 31: DDMRP - Demand Driven Material Requirements Planning

Thisdepiction is consistentwith thegraphicaldepictionof a loss functiondeveloped by the Japanese business statisticianGenichi Taguchi to describe aphenomenonaffectingthevalueofproductsproducedbyacompany.Thismadeclear theconcept thatqualitydoesnotsuddenlyplummetwhen,for instance,amachinist slightly exceeds a rigid blueprint tolerance. Instead “loss” in valueprogressivelyincreasesasvariationincreasesfromtheintendednominaltarget.

The same is true for inventory. Chapter 2 will discuss how the value ofinventoryshouldberelatedtotheabilityofinventorytohelppromoteorprotectflow. As the inventory quantity expands out of the optimal zone and movestowardpointB,thereturnonworkingcapitalcapturedintheinventorybecomesless and less as the flowofworking capital slowsdown.The converse is alsotrue:as inventoryshrinksoutof theoptimalzoneandapproacheszeroor less,thenflowisimpededduetoshortages.

When the aggregate inventory position is considered in an environmentusing traditionalMRP, there is frequently a bimodal distribution noted.Withregardtoinventory,abimodaldistributioncanoccurontwodistinctlevels:

1.Abimodaldistributioncanoccuratthesingle-partleveloveraperiodof time, as a part will oscillate back and forth between excess andshortage positions. In each position, flow is threatened or directlyinhibited.Thebimodalpositioncanbeweighted towardone sideortheother,butwhatmakesitbimodalisaclearseparationbetweenthetwogroups—thelackofanysignificantnumberofoccurrencesinthe“optimalrange.”

2. The bimodal distribution also occurs across a group of parts at anypoint in time.At anyonepoint,manypartswill be in excesswhileother parts are in a shortage position. Shortages of any parts areparticularly devastating in environmentswith assemblies and sharedcomponents because the lack of one part can block the delivery ofmany.

Page 32: DDMRP - Demand Driven Material Requirements Planning

FIGURE1-3Bimodalinventorydistribution

Figure1-3isaconceptualdepictionofabimodaldistributionacrossagroupofparts.Thebimodaldistributiondepictsalargenumberofpartsthatareinthetoo-little range while still another large number of parts are in the too-muchrange.TheYaxisrepresents thenumberofpartsatanyparticularpointonthelossfunctionspectrum.

Not only is the smallest population in the optimal zone, but the time anyindividualpartspendsintheoptimalzonetendstobeshort-lived.Infact,mostpartstendtooscillatebetweenthetwoextremes.Theoscillationisdepictedwiththesolidcurvedlineconnectingthetwodisparatedistributions.Thatoscillationwilloccur every timeMRP is run.At anyone time, anyplannerorbuyer canhavemanypartsinbothdistributionssimultaneously.

This bimodal distribution is rampant throughout industry. It can be verysimplydescribedas“toomuchof thewrongand too littleof the right”at anypointintimeand“toomuchintotal”overtime.Inthesamesurveynotedearlier,taken between 2011 and 2014 by theDemandDriven Institute, 88 percent ofcompanies reported that they experienced this bimodal inventory pattern. Thesamplesetincludedover500organizationsaroundtheworld.

Three primary effects of the bimodal distribution are evident in mostcompanies:

1. High inventories. The distribution can be disproportionate, asmanyplanners and buyers will tend to err on the side of toomuch. Thisresults in slow-moving or obsolete inventory, additional spacerequirements, squandered capacity and materials, and even lowermarginperformanceasdiscountsarefrequentlyrequiredtoclearout

Page 33: DDMRP - Demand Driven Material Requirements Planning

theobsoleteandslow-movingitems.2.Chronicandfrequentshortages.Thelackofavailabilityofjustafewparts can be devastating to many manufacturing environments,especiallythosethathaveassemblyoperationsandcommonmaterialorcomponents.Thelackofanyonepartwillblockanyassembly.Thelackofcommonmaterialorcomponentswillblock themanufactureof all parent items calling for that common item. This means anaccumulationofdelays inmanufacturing, latedeliveries,andmissedsales.

3.Highbimodal-relatedexpenses.Thiseffecttendstobeundermeasuredand underappreciated. It is the additional amount ofmoney that anorganization must spend in order to compensate for the bimodaldistribution. When inventory is too high, third-party storage spacemayberequired.Wheninventoryistoolow,premiumandfastfreightare frequently used to expedite material. Overtime is then used topushlateordersthroughtheplant.Partialshipmentsaremadetogetthe customers some of what they ordered but with significantlyincreasingfreightexpenses.

Why the bimodal distribution occurs is explained in Chapter 3. It is acombinationofbasicMRPtraits,thetypeofdemandsignalthatistypicallyusedin conjunctionwithMRP, and the complex volatile supply chain environmentwithinwhichcompaniesnowmustoperate.

TheNewNormal

ExperiencedplanningandpurchasingpersonnelknowthatiftheysimplyfollowwhatMRP recommends, they will be in big trouble. Shortages will increase.Excess inventory will increase. Expedites will increase. Intuitively, plannersunderstand that materials and inventory management, under conventionalpractices,places them inano-win situation.Whathappened to thepromiseofMRPasverbalizedbyJoeOrlickyinthebeginningofthischapter?Theanswerisexceedinglysimple:theworldchangedandMRPdidnot.

The circumstancesunderwhichOrlicky andhis cadredeveloped the rulesbehindMRPhavedramaticallychanged.Customertolerancetimeshaveshrunkdramatically,drivenbylowinformationandtransactionalfrictionlargelyduetotheInternet.Customerscannoweasilyfindwhat theywantataprice theyarewillingtopayandgetitinashortperiodoftime.

Page 34: DDMRP - Demand Driven Material Requirements Planning

Ironically, the planning complexity is largely self-induced in the face ofthese shorter customer tolerance times. Most companies have made strategicdecisionsthathavedirectlymadeitmuchhardertodobusiness.Productvarietyhasrisendramatically.Supplychainshaveextendedaroundtheworlddrivenbylow-costsourcing.Productcomplexityhasrisen.Outsourcingismoreprevalent.Productlifeanddevelopmentcycleshavebeenreduced.

Add on top of this an increased amount of regulatory requirements forconsumer safety and environmental protection, and there are simply morecomplexplanningandsupplyscenariosthaneverbefore.Thecomplexitycomesfrommultipledirections:ownership,themarket,engineeringandsales,andthesupply base.While this complexity has risen, the potential of technology hasprogressed and accelerated. The lack of significant financial return ontechnology investments would strongly suggest that this potential, up to thispoint,haslargelybeensquandered.

Figure1-4istakenfromDemandDrivenPerformance:UsingSmartMetricsbyDebraSmithandChadSmith.Thefigureshowsthetremendousdifferenceinsupplychaincircumstancesbetween1965and2015.

Page 35: DDMRP - Demand Driven Material Requirements Planning

FIGURE1-4Changingsupplychaincircumstances

FromDebraSmithandChadSmith,DemandDrivenPerformance:UsingSmartMetrics,McGraw-Hill,2013,p.9.

Summary

WeappeartohavecomefullcircleasMRP,accordingtoobservable,prevailing,andwidespreadeffectsacross theworld,nowappears tobeguiltyof thesamedeficiencies as the techniques that preceded it. Software is simply a tool thattranslatesandreinforcesrulesintoaroutine.Iftherulesbehindthesoftwareareinappropriate and outdated, then the rules must change before the tools canchange. In recent years, however, industry and software providers haveattempted to combat increasing complexity with more sophisticated softwareapplications,applicationswiththeoldrulesstillembeddedattheircore.Theneteffect is that we have improved the efficiency of doing the wrong orinappropriatethings.Moneyandenergyspenttooptimizeantiquatedruleswithincreasinglysophisticatedtoolsarewasteful,distractive,andcounterproductive.Given the current world of increased variability and volatility, conventionalplanning logic now requires a fundamental overhaul. The authors think JoeOrlickywouldagree.

The authors’ self-imposed mission was to stand on the shoulders of JoeOrlicky’s incredible vision in order to see further.This bookproposes elegantand intuitive alternative planning rule sets to address the volatile twenty-first-century landscape. Complexity cannot be combated with more complexity.Effective and simplified rules and subsequent tools are necessary for acompany’s resources to work more closely in alignment with the market,enabling a demand driven world. There can be no more lip service to smallincremental changes thatmay ormay not improve a company’s performance;concrete and proven tactics are required that drive sustainable bottom-lineresults.Wheretostart?

Page 36: DDMRP - Demand Driven Material Requirements Planning

CHAPTER2

TheImportanceofFlow

Tounderstandwhyprecisehigh-poweredtools likeMRParenot livinguptotheir potential as well as the direction for a potential solution, let’s start at afundamental level. All for-profit entities have the same objective: to driveshareholder equity. Thus the rules and toolswithin a for-profit entitymust bealigned to that objective. There is in fact a fundamental principle that alignsbusinessrulesandtoolstothatobjective.

Plossl’sFirstLaw

Manufacturing comprises a bewildering and distracting variety of products,materials, technology,machines, and people skills that obscure the underlyingelegance and simplicity of it as a process.The essence ofmanufacturing (andsupplychainingeneral)istheflowofmaterialsfromsuppliers,throughplants,throughdistributionchannels,tocustomers;theflowofinformationtoallpartiesaboutwhatisplannedandrequired,whatishappening,whathashappened,andwhatshouldhappennext;andtheflowofcash.

An appreciation of this elegance and simplicity brings us towhatGeorgePlossl(afoundingfatherofMRPandauthorofthesecondeditionofOrlicky’sMaterialRequirementsPlanning)articulatedasthefirstlawofmanufacturing:

All benefits will be directly related to the speed of flow of information andmaterials.

“Allbenefits”isquiteanencompassingphrase.Itcanbebrokendownintocomponents that most companies measure and emphasize. All benefitsencompass:

Service. A system that has good informational andmaterial flow

Page 37: DDMRP - Demand Driven Material Requirements Planning

produces consistent and reliable results. This has implications formeeting customer expectations, not only for delivery performancebutalso forquality.This isespecially true for industries thathaveshelf-lifeissues.

Revenue.Whenservice isconsistentlyhigh,market share tends togrow—or,ataminimum,doesn’terode.

Quality.Whenthingsareflowingwell,fewermistakesaremadethatarecausedbyconfusionandexpediting.

Inventories.Purchased,work-in-process(WIP),andfinishedgoodsinventories will be minimized and directly proportional to theamountoftimeittakestoflowbetweenstagesandthroughthetotalsystem.Thelesstimeittakesproductstomovethroughthesystem,the less the total inventory investment. The simple equation isThroughput * lead time =WIP. Throughput is the rate at whichmaterialisexitingthesystem.Leadtimeisthetimeittakestomovethroughthesystem,andWIPis theamountof inventorycontainedbetween entry and exit. A key assumption is that the materialentering the system is proportionate to the amount exiting thesystem.ThebasisforthisequationisthequeuingtheoryknownasLittle’slaw.

Expenses.Whenflowispoor,additionalactivitiesandexpensesareincurred to close the gaps in flow. Examples would be expeditedfreight, overtime, rework, cross-shipping, and unplanned partialships.Mostoftheseactivitiesareindicativeofaninefficientoverallsystem and directly cause cash to leave the organization. Thesetypes of expenses were described in Chapter 1 in relation to thebimodaldistribution.

Cash.Whenflowismaximized,thematerialthatacompanypaidforis converted to cash at a relatively quick and consistent rate.Thismakescashflowmucheasier tomanageandpredict.Additionally,theexpedite-relatedexpensespreviouslymentionedareminimized,limitingcashleavingtheorganization.

What happens when revenue is growing, inventory is minimized, andadditional and unnecessary ancillary expenses are eliminated? Return oninvestment(ROI)movesinafavorabledirection.Thusthefundamentalprinciple

Page 38: DDMRP - Demand Driven Material Requirements Planning

is established that the rules and toolsof abusiness shouldbebuilt around theprotectionandpromotionofflow.

EstablishingFlowastheFoundation

It isdifficult tofoster theflowofrelevant informationandmaterials throughasystemwhen the components of the system cannot relate their actions to thatflow.It’sbecomeclichétosaythatourorganizationshave“silos.”Thosesilostypically result in friction, conflict, and communication difficulties betweenfunctions. This is because we tend to control segments of our organizationsthrough differentmetrics. Figure 2-1 lists an organization’s primary functionsandtherespectiveprimaryobjectivesandexamplemetricstoaccomplishthoseobjectives.

The actions that each of these functionsmight take tomeet their primaryobjectivesandmetricsoftencomeintoconflict.Asanexample,sales typicallyhasadifferentprimarymetricthanoperations.Itcanfrequentlybethecasethatwhen operations looks tomaximize its primarymetric, itmay compromise orjeopardizetheprimarymetricofsalesandviceversa.Whenqualitymaximizesitsdepartmentalmetric,thenoperationsmightbeadverselyaffected.

Yet we have already established that when a system flows well, service,revenue, quality, inventories, expenses, and cash are all better controlled. AlltheseelementsdirectlyprotecttheprimaryobjectivesofthefunctionsinFigure2-1.Butifflowisnotmadevisibleandincorporatedintotheroutineandmetrics,thenhowcanitpossiblybeprotected?Flow,ifencouraged,measured,andmadeproperly visible, can align all these objectives with the system goal ofmaximizingreturnonshareholderequity.

FIGURE2-1Organizationalfunctions,objectives,andmetrics

Thus,aligningthefunctionstothepromotionandprotectionoftheflowcan

Page 39: DDMRP - Demand Driven Material Requirements Planning

be the bridge between local actions and the global benefits. Furthermore, thisalignment should significantly raise the quality and timing of relevantinformationandcorrespondingrelevantmaterialsinasystem.

Additionally, the protection and promotion of flow is a unifying conceptwithin major process improvement disciplines and their respective primaryobjectives.Dr.EliyahuGoldratt,theinventoroftheTheoryofConstraintshadaprimary objective of driving system throughput. This was accomplished by afocusontotalsystemflow.LateinDr.Goldratt’slife,hiswritingsbecameveryspecific about the interdependence between the Theory of Constraints andTaiichiOhno’sworkwiththeToyotaProductionSystem(TPS)andflow.MostintheWestmightsaythatthegoalofTPSandanyLeansystemistoeliminatewaste.When things flowwell, there is indeed lesswaste.ButTPS is not justabout waste elimination. When Ohno’s writings are examined closely, itbecomesevidentthattheprimarygoalwasinfactflowasdescribedinhisRiverProductionSystemforFlow.Additionally,thequalitymovementdrivenbyDr.W.EdwardsDemingandhis14points forqualityheavily reliedon flow.Theneedforflowisobviousinthisframeworksinceimprovedflowresultsfromlessvariability.

Any discussion or time spent on ideological battles between thesedisciplines is a completewasteof timeandquite frankly, boring.Focusingonflowisaboutachievingacommonobjectivethroughacommonstrategybasedon common sense (also leveraging physics, biology, economics, andmanagementaccounting).

Goldratt, Ohno, and Deming did not invent the concept of flow. Theirdisciplines simply built off the works and concepts of industrial giants thatchanged manufacturing forever and gave birth to the corporate managementstructure in use today. To these industrial pioneers the concept of flow wassimplycommonsense.Theseindustrialpioneersinclude:

Frederick Taylor, a founding father of operations management.Taylordevelopedtheprocessesfortimestandards,productroutings,tools,methods,and instructionsaswell asvariablecosting systemand standard variance analysis. He developed the concept ofplanningasanactualbusinessfunction.

Henry Ford, a founding father ofmass production.TheprocessesusedinearlyFordproductionwerebasedonthefactthattheslowesttaskgovernsflowandthatwhenthereissynchronizationofactivity

Page 40: DDMRP - Demand Driven Material Requirements Planning

to, through, and from those tasks, total system speed and velocityareprotected.Fordwaswellknownforhisfocusonthevalueof“nowaittime.”

F.DonaldsonBrown,afoundingfatherofmanagementaccounting.During his time at DuPont, Brown developed theDuPont ROI inadditiontocost,volume,profitanalysis,andflexbudgeting.Browndefined relevant information for decision making and pioneeredmarket segmentation at scale—all of which was based on afoundationofthepromotionofflow.

RelevantInformationandMaterials

Yet there is an important caveat toPlossl’s first law that becomes crucial andcentraltodeterminingwhetherflowtranslatestobetterROIperformance.Ithasalreadybeenhintedatseveral times in theprecedingtext.ThegreatbasketballcoachJohnWoodensaid,“Nevermistakeactivityforachievement.”Acompanycannot just indiscriminatelymovedataandmaterialsquickly throughasystemand expect to be successful. Today organizations are frequently drowning inoceans of data with little relevant information and large stocks of irrelevantmaterials(toomuchofthewrongstuff)andnotenoughrelevantmaterials(toolittleoftherightstuff).Whenthisoccurs,thereisadirectandadverseeffectonreturnoninvestment.

Thustheflowofinformationandmaterialsmustberelevanttotherequiredoutputormarketexpectationofthesystem.Toberelevant,boththeinformationandmaterialsmustsynchronizetheassetsofabusinesstowhatthemarketreallywants;nomore,noless.Havingtherightinformationisaprerequisitetohavingtherightmaterials.Withthisismind,Plossl’sfirstlawcanbeamendedto:

Allbenefitswillbedirectlyrelatedtothespeedofflowofrelevantinformationandmaterials.

But in thehighlycomplexandvolatileNewNormal, is itevenpossible topromoteandprotecttheflowofrelevantinformation?Whatstandsintheway?

TheBullwhipEffect

A massive amount of research and literature has been devoted to thephenomenonknownasthebullwhipeffect.However,verylittle, ifany,ofthatbodyof knowledgehas been related specifically to the objective of protecting

Page 41: DDMRP - Demand Driven Material Requirements Planning

andpromotingtheflowofrelevantinformationandmaterial.APICSDictionarydefinesthebullwhipeffectas:

An extreme change in the supply position upstream in a supply chaingenerated by a small change in demand downstream in the supply chain.Inventorycanquicklymovefrombeingbackorderedtobeingexcess.This iscaused by the serial nature of communicating orders up the chain with theinherenttransportationdelaysofmovingproductdownthechain.Thebullwhipcanbeeliminatedbysynchronizingthesupplychain.(p.19)

Thisdefinitionclearlydealswith importantpointsdiscussedearlier in thisbook.“Inventorycanquicklymovefrombeingbackorderedtobeingexcess”isdescriptive of the oscillation effect that occurs with the bimodal distribution.Additionally, this definition deals with both information and materials.“Communicating orders up the chain” is the information component, while“movingproductdownthechain”isthematerialscomponent.

Inthisrespect, thebullwhipisreallythesystematicbreakdownofrelevantinformationandmaterialsinasupplychain.Figure2-2isagraphicaldepictionofthebullwhipeffect.Thewavyarrowmovingfromrighttoleftisthedistortiontorelevantinformationinthesupplychain.Thearrowwavegrowsinamplitudeinordertodepictthatthefartherupthechainyougo,themoredisconnectedtheinformation becomes from the origin of the signal, as signal distortion istransferred and amplified at each connection point. An MRP characteristicknownasnervousnesscombinedwithbatchingpracticescreatesthistransferandamplification,respectively.BothareexplainedinChapter3.

In thefigure, thewavyarrowmovingfromleft to right is thedistortion inrelevantmaterialsinthesupplychain.Thewavegrowsinamplitudefromlow-level suppliers to the end item producer (OEM) to show the accumulation ofdelaysthatoccurduetochronicshortagesandlateshipments.Thistransferenceand amplification occurs due to batching practices and the inherentsynchronization problems associated with the probability of simultaneousavailability;bothareexplainedinChapter3.

Itcouldandshouldbearguedthattheprevalenceofthebullwhipeffectisafourth indicatorofconventionalplanning logicdeficiency.Chapter3describeshowthislogic,drivenbyonekeyattribute,directlyleadstothebullwhipeffect.

Page 42: DDMRP - Demand Driven Material Requirements Planning

FIGURE2-2Thebullwhipeffect

Summary

The flowof relevant informationandmaterials is the fundamentalprinciple toachievesustainablesuccessacrossthesupplychain.Istheconceptofpromotingflowdifficult forpeople tograsp?Titansofearly industry likeHenryFord,F.DonaldsonBrown,andFrederickTaylorallunderstoodthisimportanceandbuilttheirmodelsaroundit,modelsthatprovidedthebackboneofmoderncorporatestructure.LaterthoughtleaderssuchasPlossl,Ohno,Deming,andGoldrattbuiltentire methodologies around the concept. The concept is a basic tenet ofmanagementaccounting.

Theconceptisalsointuitive.Ingeneral,mostpeoplewithinanorganizationseem to intuitively grasp why flow is so important. Yet there is a significantamount of evidence to suggest that most companies are incapable of reallymanaging their assets with this fundamental principle. Next, the planningsystemsinusethroughouttheworldtoplanandmanagetheuseoftheseassetsareexaminedtodiscoverwhythisisthecase.

Page 43: DDMRP - Demand Driven Material Requirements Planning

CHAPTER3

Material Requirements Planning in the NewNormal

AsdiscussedinChapter2,theprimaryenemyoftheprotectionandpromotionofthe flow of relevant information and materials is the bullwhip effect. Thebullwhip effect exists largely due to the characteristics and configurations ofconventionalplanning systemsutilizingMRP.This chapterwilldescribe thesecharacteristics and configurations and highlight one key attribute as a coreproblem.

WhatIsMRP?

TheAPICSDictionarydefinesMaterialRequirementsPlanning(MRP)asa:

A set of techniques that uses bill of material data, inventory data, and themasterproductionscheduletocalculaterequirementsformaterials.(p.103)

MRPisessentiallyacalculationhub.Themasterproductionschedulefeedsdemand signals to MRP, which in turn creates a synchronized list of supplyordersbasedon current inventory records (onhandandonorder) andproductstructure (bill of material). The supply orders have date and quantityrequirements that define the elements of that synchronizationplan.Thesedateand quantity requirements are then fed to a manufacturing execution system.Theyareturnedintotransferorderstodistributionsites,manufacturingorderstobescheduledontheshopfloor,andpurchaseorders toberelayedtosuppliers.Figure3-1showsthisconventionalplanningapproach.

TherequirementstorunMRParesimpleandstraightforward:

Themasterschedulemustbestatedintermsofthebillofmaterial.Uniqueitemnumbersexistforeveryitem.

Page 44: DDMRP - Demand Driven Material Requirements Planning

Thebillofmaterialexistsatthetimeofplanning(productstructurefile).

Inventoryrecordsareavailableforallitems(inventoryrecordfile).

Whentheserequirementsareimplementedinthecomputersystem,thentheMRPbatchprogramcanberun.However,tobeconsideredaClassAuserortoexpectsomekindofreasonableresultfromthecomputersystem,thefollowingassumptionsaremade:

FIGURE3-1Theconventionalplanningschema

Filedataare100percentaccurateandcomplete.Leadtimesarefixedandknown.Everyinventoryitemgoesintoandoutofstock.Thereisfullallocation;noorderisstartedunlessallthecomponentsareavailable.

Componentsarediscrete—thingscanbecountedandmeasured(no“useasrequired”).

There isorder independence,whichmeansthateveryordercanbestartedandcompletedonitsown.

MRPwasahugeleapforwardbecauseforthefirsttimewhatwasrequiredcouldbecalculatedbasedonwhatwasalready therecomparedwithwhatwasneeded,withthenetresulttimephased.TheobjectiveofMRPwastopreciselytime-phase the requirements and replenishments to dramatically reduce

Page 45: DDMRP - Demand Driven Material Requirements Planning

inventoryfromthepreviousorderpointapproachwheresomeofeverythingwaskeptaroundall the time.Thisability tocalculatedependentdemand throughabill of material was a significant development. It was no longer necessary toforecast dependent demand—it could be calculated based on the expecteddemandfortheparentpart.APICSdefinesdependentdemandas:

Demandthatisdirectlyrelatedtoorderivedfromthebillofmaterialstructureforother itemsorendproducts.Suchdemandsare thereforecalculatedandneednotandshouldnotbeforecast.(p.46)

MRP evolved because of the advent of the computer, and the age ofmarketing in the 1950s introducedmore product variety and complexity thanwas managed previously. Order point (the previous method of materialsmanagement) clearly could not affordably handle these new requirements. TounderstandhowplannersdealwithMRPonadailybasis,refertoAppendixA,where a simulated environment demonstrates the day-to-day difficultiesassociatedwithMRP.

Yetevenifacompanyhas100percentoftherequirementsand100percentof the assumptions validated, the conventional planning approachwill still beineffective.Theremainderofthischapterwillexplainwhy.

DistortionstoRelevantInformation

Theconventionalplanningapproachactuallycreatesthebullwhipeffectanditsinherentdistortions to the flowof relevant informationandmaterials.Someoftheways inwhichconventionalplanningcreates thebullwhip is related to themanner in which convention chooses to useMRP. Other contributions to thebullwhip are related to hard-coded traits inMRP systems.All of these issues,however,arerelatedtoonekeyandfundamentalattributeofMRP.

DemandSignalInput

MRP is essentially a calculator. It needs three basic inputs to perform itscalculation. One of those inputs is “demand.” Different demand inputs willproducedifferentoutputs.TheAPICSDictionarydefinesdemandas:

Aneedforaparticularproductorcomponent.Thedemandcouldcomefromany number of sources (e.g., a customer order or forecast, an interplantrequirement, a branch warehouse request for a service part or themanufacturingofanotherproduct.(p.44)

Page 46: DDMRP - Demand Driven Material Requirements Planning

By this definition, demand can be broken down into two different types:forecasted and actual. Both of the following definitions are from the APICSDictionary:

Forecast. An estimate of future demand. A forecast can beconstructed using quantitative methods, qualitative methods, or acombinationofmethods,anditcanbebasedonextrinsic(external)orintrinsic(internal)factors.Variousforecastingtechniquesattempttopredictoneormoreofthefourcomponentsofdemand:cyclical,random,seasonal,andtrend.(p.68)

Actualdemand.Actualdemandiscomposedofcustomerorders(andoften allocations of items, ingredients, or raw materials toproduction or distribution). Actual demand nets against or“consumes” the forecast, depending upon the rules chosen over atime horizon. For example, actual demand will totally replaceforecast inside the sold-out customer order backlog horizon (oftencalled the demand time fence) but will net against the forecastoutsidethishorizonbasedonthechosenforecastconsumptionrule.(p.4)

ThetypeofdemandthatischosentodrivetheMRPcalculationisaprimarydeterminant of how much relevant information can be produced from MRP.Remember, the flow of information and materials must be relevant to therequired output ormarket expectation of the system. To be relevant, both theinformationandmaterialsmust synchronize theassetsofabusinesswithwhatthemarketreallywants;nomore,noless.

A hard-coded trait of MRP is that with a given demand signal, MRP isdesignedtonetperfectlytozero.Youmakeexactlywhatyouneedwithoutanyexcess.InthisregarditcouldbearguedthatMRPistheperfectJITsystem.Ifthe demand signal is perfectly accurate, then the MRP calculation will beperfectly accurate.Given that themath allowsno tolerance for error, it seemsobviousthatMRPshouldonlybegivenasaccurateasignalaspossible.

With that inmind, should the demand input toMRP bewhat a companythinks the market wants to buy or what the customers actually want to buy?Which will produce a more relevant result? As described in the definition ofactualdemandaswellasFigure3-1, theconventionalapproachcombinesbothtypesofdemand.Forecastisusedtocreateplannedorders,andthendemandis

Page 47: DDMRP - Demand Driven Material Requirements Planning

adjusted as the picture becomes clearer with actual orders. Why is thisproblematic?

Therearethreetruthsaboutforecasts:

1. All forecasts start out with some inherent level of inaccuracy. Anypredictionaboutthefuturecarrieswithitsomemarginoferror.ThisisespeciallytrueinthemorecomplexandvolatileNewNormal.

2. Themore detailed or discrete the forecast is, the less accurate it is.Thereisdefinitelyadisparityintheaccuracybetweenanaggregate-level forecast (all products or parts), a category-level forecast (asubgroup of products or parts), and a SKU-level forecast (singleproductorpart).

3.Themoreremoteintimeorfartheroutforecastsgo,thelessaccuratetheyget.Predictingtheweathertomorrowismuchmoreaccuratethanpredictingtheweather52daysfromtoday.Yes,historycanbeusedasabasisforaprediction,but themarginofpotentialerror ismuchhigher.Itisnotuncommonthatinmanyindustriestheaccuracyofaforecastcandropbelow10percentbeyond90daysattheSKUlevel.

Todaymanyforecastingexpertsadmitthat70to75percentaccuracyisthebenchmark for the SKU level. Figure 3-2 is the results of a 2012 surveyconducted by forecastingblog.com showing the reported forecast error ratesacrossvariousindustriesattheSKUlevel.

Unfortunately,whenyoustartaserial,complex,andinterdependentprocesswithanerror-proneinput,theresultingoutputintegritymustbesuspect.Plannedorders are derived from these forecasts, and very real commitments of cash,capacity,andmaterialsaredirectlyderived fromaprediction that is subject tovaryingdegreesofinaccuracy,sometimeswithextremelysignificantdegreesofinaccuracy.

As timeprogresses, the demandpicture changeswith the incorporation ofactualdemand,MRPisrerun,andsubsequentchangesoccur.Theresultisthatweendupwith things thatwedonotneedanddesperatelyexpedite thingswehavejustdiscoveredthatwedoneed.Thesearethethreeeffectsofthebimodaldistribution.Thusthebimodaldistributionstartswiththeuseofplannedordersbasedonthatforward-lookingforecast.

Page 48: DDMRP - Demand Driven Material Requirements Planning

FIGURE3-2Averageforecastaccuracyacrossindustries

RohanAsardohkar,August22,2012,http://www.forecastingblog.com/?p=423

This is aknownandaccepted routine inmost industriesdespite thewasteand performance erosion associatedwith it.Whywould industry intentionallysabotageperformancebyusinganinputwithknowninaccuracytodriveactivityandcommitmentswhen there isanobviousalternative?Whynot justuseonlysalesorders?

Themostaccurateformofdemandinputisasalesorder.Asalesorderisastatedintentionandcommitmenttobuyfromanactualcustomerintermsofbothquantityandtime.Itisessentiallyanuncashedcheck.Inthiswayitisahighlyaccurateandrelevantpieceofinformation.Thereshouldbenodebatethatsalesorders are an order ofmagnitudemore accurate than planned orders. Sowhydon’tcompaniessimplyloadonlysalesordersintoMRP?

UsingMRPwithonlysalesorders,however,assumessomething thatdoesnotexistintoday’sNewNormal—enoughtime.AbasicattributeofMRPistonettozeroacrosstheentirenetworkofdependencies.ThismeansthatMRPbydefinitionmakesallactivitiesdependentoneachother.Thus,inorderforMRPtobethatperfectJITsystem,theremustbesufficienttimetoprocureandmakeeverything to the stated demand—called “cumulative lead time” (the longeststatedchainoftimeinthebillofmaterialincludingpurchasingleadtime).

This means that customer tolerance time would have to be equal to orgreater than the cumulative lead time. Today’s supply chains, however, are

Page 49: DDMRP - Demand Driven Material Requirements Planning

characterized by shorter and shorter customer tolerance times and extended,elongated, and increasingly complex supply chains. There simply is notsufficientvisibilitytosalesorderssoonenoughtoproperlyplanforthemusingconventionalMRP.Figure3-3conceptually shows thedisparitybetweenwhencompaniesgainvisibilitytosalesorders(actualdemand)versusthetimethatittakestoprocureandproducetheproduct(thetimeframeinwhichMRPmakesitcalculations).

FIGURE3-3Manufacturingandprocurementtimesversussalesordervisibility

FIGURE3-4Planninghorizondepiction

WithMRP’scharacteristicofmakingeverythingdependent,theonlywaytofindenough time is toattempt topredictwhatactualdemandwill look likeso

Page 50: DDMRP - Demand Driven Material Requirements Planning

that an organization can attempt to ensure that the necessary materials areavailableinquantityandtimeasthemarketplacesitssalesorders.A“planninghorizon”extendsintothefuturefarenoughtocoverthecumulativeprocurementandmanufacturing cycle. Figure3-4 shows the planning horizon covering thecumulativeprocurementandmanufacturingcycleintheexample.

This explains the need to loadMRPwith demand that is largely derivedfroma forecast and then tomake adjustments close in as sales orders becomevisible. Planned orders for end items are launched at the beginning of theplanning horizon. The longer the procurement and manufacturing cycle, thelonger theplanninghorizonmustbe.The longer theplanninghorizon, the lessaccurate the planned orderswill be. The less accurate the planned orders, themorecoursecorrectionsarerequired.ThisconstantsetofcorrectionsbringsustoanotherinherenttraitofMRPcalled“nervousness.”

Nervousness

MRP’s nature of making everything dependent creates nervousness.NervousnessisthecharacteristicinanMRPsystemrelatedtochangesinparentdemand transferring down and across bills ofmaterial.TheAPICSDictionarydefinesnervousnessas:

ThecharacteristicinanMRPsystemwhenminorchangesinhigherlevel(e.g.level 0 or 1) records or the master production schedule cause significanttimingorquantitychangesinlowerlevel(e.g.5or6)schedulesororders.(p.86)

Figure3-5 illustrates theconceptofnervousness.The figure illustrates theproduct structure for an end item called FPA.A timing or quantity change inFPA ripples down through the entire product structure, causing timing andquantitychangesateverycomponentpositionasthesystemconstantlystrivestonettozero.Thedottedcurvedarrowsdepictthatchange.Thiscreatesaconstantseriesofactionmessagesforplannersandbuyerstoreviewandinterpret.

ThechallengeofsystemnervousnesshasbeenknownsincetheearliestdaysofMRP.However, the system nervousnesswasmanageable since plansweredoneoncepermonth.Conceptslikefirmplannedorders,thedemandtimefence,andthemasterproductionscheduleweredevelopedtomanagethenervousness.But the complex and volatile environment characterized by the New Normalmakestheissueabiggerchallenge.GiventhenatureofMRPtomakeeverythingdependent, the onlyway to stop nervousness is tomake no changes.Yet that

Page 51: DDMRP - Demand Driven Material Requirements Planning

would mean significant service challenges, as the forecasted orders will vary(manytimesdramatically)fromwhatthemarketwillreallydesire.Whatcanbedone to limit the impact of nervousness? MRP users are forced intocompromisesinordertoslowdowntherateofchanges.

FIGURE3-5Nervousnessillustrated

TheWeeklyBucket

Inmostconventionalenvironments,planningoccursinweeklybuckets.Thisisadirect effect of the nervousness discussed above—nervousness that is directlyrelatedtotheuseofplannedorderswithMRP.PlanningorganizationsknowthatiftheyranMRPdaily,orworseyetinrealtime,theresultingnervousnesswouldcreatechaos.Theamountofactionflagsandmessagesontheplanningscreenswouldbeoverwhelming.

Instead, aweekly interval is typically used to calm thewaters on a dailylevel.This, however, comes at a price. First, it forces the planning horizon toextend even further (one week). This has a direct correlation to the level ofsignal inaccuracy at the end of the horizon. Second, it creates a latency thatalmost guarantees that the level of change between MRP runs will bedramatically larger. Instead of lots of little changes on a daily basis, there aremassivechanges(andsignaldistortions)onaweeklybasis.

Figure3-6depicts thedifferences innet change impactbetweendailyandweeklyMRPruns.TheupperlefthandbarchartdepictsMRPruneachday.Thelevelofeachchangeisrelativelysmallbuteachchangeripplesthroughalllowerdependencies.Thebarchart in theupper rightportionof thegraphicdepictsa

Page 52: DDMRP - Demand Driven Material Requirements Planning

weekly MRP run. Days 1–7 are stable (no change) yet Day 8 introduces asignificant change (40) that will ripple through the environment. The relativedifference in changes is depicted in the chart in the lower left corner of thegraphic.

PlanningorganizationsarestuckbetweenthesetwohardplacesbecauseofMRP’shard-codedtraitofmakingeverythingdependent.

FlatteningtheBillofMaterial

AnotherwaytocombatnervousnessistoreducethenumberofconnectionsthatMRPseesandcalculatesagainst.Onewaytoaccomplishthisisto“flatten”thebill of material of a product. MRP from a planning and synchronizationperspective then becomes blind to intermediate components. Figure 3-7illustrates the difference between a full product structure (on the left) and aflattened one (on the right). The flattened structure has eliminated allintermediatepositions.

While thisreduces theamountofchanges to intermediates(since therearenone) and this reduces the total number of action flags, does it producemorerelevant information or actually distort the picture further? The key to morerelevant information is not to simply ignore dependencies. When we ignorecritical dependencies, we risk oversimplification. Oversimplification means tosimplifytothepointoferror,distortion,ormisrepresentation.

Thebillofmaterial filesused inaplanningsystemshouldreflecthowtheproductisactuallymade.DramaticallyflatteningbillsliketheexampleinFigure3-7 effectively ends any capability to provide visibility to and plan forsynchronization between the finished and purchased part levels. The price forthis is paid by the manufacturing floor as scheduling and schedule executionbecomeanorderofmagnitudemoredifficult.

Page 53: DDMRP - Demand Driven Material Requirements Planning

FIGURE3-6DailyversusweeklyMRPruns

FIGURE3-7Theflatteningofabillofmaterial

AllofthesefactorscombinetomeanthatMRPisproducingplans:

Withhighdegreesofknownerror(forecastinput)Inaconstantstateofchange(nervousness)Withadegreeoflatency(weeklybucket)Thatmaymisrepresenttheenvironment(flattenedbillsofmaterial)

Thismeans that theverynatureofMRPcombinedwith theway that it is

Page 54: DDMRP - Demand Driven Material Requirements Planning

typically used inevitably leads to distortions to relevant information.Furthermore,allofthesedistortionstorelevantinformationhavebeenrelatedtoonesingleattributeofMRP.Haveyouspottedityet?

DistortionstoRelevantMaterials

Thenextconsideration is thesupplyportionof thebullwhip—thedistortionofrelevantmaterials.Asmentioned previously,MRP creates a synchronized andpreciseplanatalllevelsofthebillofmaterialbasedonitsrequiredinputsandassumptions. This planwill happen only if everything in the entire dependentnetworkgoespreciselyaccordingtoplan.Inalmosteverymodernenvironment,thisisanimpossibilityfortworeasons.

CommonCauseVariation

First,thereisabasicandinherentlevelofvariabilityinanyenvironment,evenonedeemedtobeincontrol.Demingcalledthenormalorrandomvariationthatoccurs in processes “common cause variation.”Normal or randomoperationalvariabilityresultsinaprocessthatmaybestatisticallywithincalculatedcontrollimitsbutstillvaryingbetweenthoselimits.Reducingthegapbetweenthelimitsisaworthygoal.Theeliminationofthegapisanimpossibility—itwouldrequireeveryprocesstobeperfect.

DelayAccumulation

We know that any process cannot be perfect. The collective effect of thisimperfection must be examined. Figure 3-8 appeared in the first and thirdeditions of Orlicky’s Material Requirements Planning. The figure has threecolumns. The first column is the number of components required to make aparentitem.Thesecondtwocolumnsaredifferentlevelsofaveragecomponentavailability. The left column assumes all components have 90 percentavailability, whereas the right column assumes 95 percent availability. Forexample,aparent itemwith4components thataverage90percentavailabilityhasa65.6percent(.9×.9×.9×.9)chancethatallcomponentswillbeavailablesimultaneouslywhenrequired.Aparentitemthathas10componentsthathavean average of 95 percent availability will have a 59.9 percent chance that allcomponentswillbesimultaneouslyavailablewhenneeded.

Figure 3-9 shows how less than perfectmaterial availability results in anerosion of the probability that all materials will be present when needed.

Page 55: DDMRP - Demand Driven Material Requirements Planning

Remember,MRPassumesfullallocation—noordershouldbestartedunlessallthecomponentsareavailable.Infact,evenifmanycomponentshaveextremelyhighvariabilityor arrive early, theparentorder release is still at themercyofanyonemissingcomponent.

Figure 3-9 illustrates an environment inwhich four of thematerials havehigh availability while one component has low availability on average.Components1,3,4,and5haveextremelyhighaverageavailability(95percent,98percent,97percent,and99percent,respectively.Component2,however,hasa relatively low average availability level (72 percent). The impact thatcomponent2hasontheoverallprobabilitythatallcomponentswillbeavailablewhen required is significant; that probability drops to 64.4 percent. Thistranslatestodelaysintheplannedrelease.

Page 56: DDMRP - Demand Driven Material Requirements Planning

FIGURE3-8Probabilitiesofsimultaneousavailability

FIGURE3-9Oneproblematicmaterial

Page 57: DDMRP - Demand Driven Material Requirements Planning

Thusasimpleruleemergeswithregardtodependentstructuresthatcontainintegrationpointsrequiringsimultaneousinputstoadvancetothenextstageofthestructureorplan.ThisisavaliddescriptionoftheplansthatMRPgenerates.Thissimpleruleis“delaysaccumulate,whilegainsdonot.”

Figure 3-10 conceptually illustrates this effect. A dependent structure isvisible at the bottom of the graphic. In this case that dependent structure is asynchronizedplanbaseduponproductstructure.Thereareconcurrentpathsandintegration points culminating in a finished item (FPA). Above the structurethereisagraphicaldepictionofdelayaccumulation.Thearrowsteadilyrisesasactivityprogresses through theplannedbuild.Thearrow’spositionat anyoneplacedepictsbothhowfaralong theplannedactivitypath thebuild is (Xaxiscorrespondingtothestructure)andtheamountofaccumulateddelay(Yaxis).

Thiseffectisonlypartiallyimpactedbysignalaccuracy.Inotherwords,thedemand signal could be perfect, but delay accumulation will still affect theenvironment if normal and randomvariation exist in the resources required toexecute those signals. This delay accumulation results in an effect that isfrequently referred to as “supply continuity variability.” This forces twoprofoundrealizations:

1.FromanexecutionperspectiveMRPwillnevercreatearealisticplaninenvironmentsofevenmoderatecomplexity.

2.Any true solution to the bullwhip effectmust addressboth demandsignaldistortionandthematerialsupplydistortion(supplycontinuityvariability).

Page 58: DDMRP - Demand Driven Material Requirements Planning

FIGURE3-10Illustratingdelayaccumulation

Amplifying the Distortions to Relevant Information and Materials—BatchingPolicies

Thedistortion to relevant information andmaterial inherent in the bullwhip isamplifieddue tobatchingpolicies.BatchingpoliciesaredeterminedoutsideofMRPandaretypicallyformulatedtoproducebetter-unitizedcostperformanceoraredue toprocess restrictionsor limitations.Batchingpoliciesdictate thewaythat MRP must perform its calculation (demand signal distortion) as well asinfluence theway inwhichmaterials progress through a supply chain (supplycontinuityvariability).

ThebatchingpoliciesthatdictatetheMRPequationincludeorderminimum—the amount that must always be ordered; order maximum—the largestquantitythatcanbeassignedtoanorder;andordermultiple—arulethatgovernsordering between the minimum and the maximum. The order minimum andmaximumshouldbeevenlydivisiblebytheordermultiple.

An example: an intermediate component can have an order minimum of100, a multiple of 50, and a maximum of 500. This means that if theintermediatecomponenthasaparentdemandof102pieces,aminimumof150(theminimumplusthenextmultiple)ofthecomponentmustbeorderedtocoverthat demand.At some point later if the parent requirement changes to 99, theintermediatecomponentrequirementdropsto100.Theparentchangedby3;the

Page 59: DDMRP - Demand Driven Material Requirements Planning

componentchangedby50.Theeffectofthiscomplicationisdevastatinginanyenvironmentwhereorderingpolicies,particularlyminimumsandmultiples,aredramaticallydifferentateachlevelofthebillofmaterial.

Figure3-11isanexampleof thedemandamplificationinamorecomplexenvironment.Anend item (FPA)has three components.All three componentshaveminimums andmultiples assigned.A demand of 115 for FPAwill yielddemands of 150 for Intermediate Component A (ICA), 250 for IntermediateComponentBand200forIntermediateComponentC.

Batchingpracticescandramaticallyaffectthewaythatmaterialmovesinasupply chain, contributing to or amplifying the accumulation of delays. Forexample, delay accumulation could occurwhile an orderwaits on a truck forotherorderstofillupthetruck.Thetransportationbatchingpolicydictatesthatonlyfulltrucksareallowed.

The logic and policies behind batching policies can be very problematic.Mostbatchesareheavilyinfluencedbyanemphasisonprotectingunitcostandhave no consideration for flow. That emphasis on unit cost actually furtherdistorts the flow of relevant information throughout most companies. Theassumption that driving to unit cost performance equates to the best returnoninvestmentperformanceisunequivocallyandmathematicallyprovenfalse.Yetindustryignoresthisfacteveryday.Thissubject,however,istechnicallyoutsidethescopeofthistext.Foranin-depthlookatthisissue,refertoDemandDrivenPerformance:UsingSmartMetricsbyDebraSmithandChadSmith.

FIGURE3-11BatchingcomplicationstoMRPsupplyordercalculations.

Summary

Are the challenges described in this chapter and Appendix A unknown toseasoned planning professionals? Absolutely not. These challenges are wellknownandcommon.Theyexplaintheexistenceofthepoorassetperformance,thework-arounds, the bimodal inventory distribution, and the bullwhip effect.Additionally,theyleavemostplanningorganizationsinahugedilemma:utilize

Page 60: DDMRP - Demand Driven Material Requirements Planning

MRPor ignore it. The answer to this dilemma is almost always the same; doboth.Most organizations are simultaneously ignoring and utilizingMRP. Justhow much ignoring and utilizing is something that tends to be specific toorganizationalfunctionsandtheindividualusers.Therehastobeabetterway.

MRP enabled organizations to quickly calculate and synchronize totalrequirements given a set of demand inputs. Thiswas of particular importancewhenthecompanyhadadeepbillofmaterialormanysharedcomponents.ThewholepurposeofMRPwastosynchronizeconnectionsanddependencies.IntheNew Normal there are undoubtedly more connections and dependencies thanever.ThusMRPshouldbemorerelevanttodaythanever.YetMRPisfailingintheNewNormal.

MRP’s role in the modern supply chain is significant. Even in the NewNormal, the heart of every supply chain ismanufacturing, and at the heart ofmanufacturing isMRP—it is the conductor of the supply order symphony inevery supply chain. Each node in the supply chain has an MRP systemsupportingadifferentmanufacturingoperation.Therefore,aprimary limitationofanysupplychainwillbehowwellMRPsystemsperformnotjustindividuallyateachnodebutalsocollectivelythroughouttheweb.

If industrywantsmore agilemanufacturing and supply chains that protectandpromote the flowof relevant informationandmaterials, then industrywillneedamoreagileformofMRP.Asevidencedinthischapter,companiescannotsimply expect to implement conventional MRP better to get the necessaryprotectionandpromotionofflow.ThefirstbuildingblockofamoreagileformofMRPwillbeexplainedinthenextchapter.Thisbuildingblockwillmitigateif not largely eliminate the bullwhip effect by simultaneously addressingbothdemandsignaldistortionandmaterialsupplydistortionbydealingwiththecoreproblemdrivingthebullwhipeffect.Thisbuildingblockiscalled“decoupling.”

Page 61: DDMRP - Demand Driven Material Requirements Planning

CHAPTER4

Unlocking a Solution—The Power ofDecoupling

Thischapterestablishesaprimarysolutionelement toeliminate thebullwhipeffectandcreateaframeworkforaprovenandpracticalmethodofplanningandexecutionfortheconditionsoftheNewNormal.

Chapter 3 described how the conventional planning approach featuringMaterial Requirements Planning (MRP) directly leads to the distortions ofrelevantinformationandmaterialsthatcomprisethebull-whipeffect.Figure4-1summarizes the connection between MRP’s core trait of making everythingdependent (our previously alluded to core problem) and the distortions torelevantmaterialsandinformation.Theboxesatthetipsofthearrowsareeffectsoftheboxesatthetailofthearrow.

At thebottomofFigure4-1 there isa roundedboxwith thewords“MRPtreatseverythingasdependent.”Therearetwoprimarypathsthatleadfromthisbox.Thefirstpathhastodowithdistortionstorelevantinformation.Thatpathisnotedwithdashedroundedboxeswithnofill.ThispathshowsthatsinceMRPtreats everything asdependent thenmanufacturing andprocurement cycles aresimplytoolongtorespondtoactualdemand.Thisforces theuseofforecasteddemand which means the initial signal is in error by definition and that thedemandsignalswillchangeastheincorporationofactualdemandorchangestoforecast occur. This triggers nervousness which creates constantly changingsignals or leads to distortive behaviors to compensate for the nervousness(weeklybucketsand/orBOMflattening).

Figure4-1culminateswithaneffectofdistortionstorelevantmaterials.Ofcourse,itwillbeverydifficulttohavethe“therightmaterialattheneededtime”if relevant information is distorted. But even if relevant information was not

Page 62: DDMRP - Demand Driven Material Requirements Planning

distorted,ifdemandwasknownandaccurateanddidnotchange,theeffectthat“the right material is not ready at needed time” would still exist. This is thesecond path depicted by the shaded boxes to the right. Since MRP treatseverythingasdependent,thenallofthetimingandquantityrequirementsinitsplans are subject to those dependencies. Chapter 3 shows how dependentnetworkssufferperformanceerosion.AnMRPplan,evenwithperfectdemandinformation,will onlybe realistic if everythinggoes exactly according toplanwithnovariation.

This core problem of MRP has remained in place in large part becausecalculationdependencywasdevelopedastherealpoweroftheMRPtool.Ifthisdependencycalculationwas removed, then the truevalueof theMRP toolhasalsobeenremoved.YetasdescribedinChapter3andinFigure4-1,thistraitistheprimaryculpritincreatingthetransferenceandamplificationsofvariabilitytotheflowofrelevant informationandmaterials.Failingtodealwiththis traitand its effects will guarantee that system flow and return on investmentperformancewillbesubpar.

Page 63: DDMRP - Demand Driven Material Requirements Planning

FIGURE4-1Thecoreproblemofthebullwhip

Ifthetransferandamplificationofvariabilityintheformofdemandsignaldistortion and supply continuity is thebiggest enemy to system flow, thenwehave to design supply chain capability that stops ormitigates the transfer andamplificationofvariabilitythroughthesystem.Buthowtodothat?Theanswercannotbe“guessbetter”or“eliminateallvariability.”Industryhastriedthatfordecades,spendingfortunesonreengineeringeffortsandexpensivesoftwareonlytoseetheproblempersist.

Decoupling

Theaccumulationandimpactofsupplyanddemandvariabilityistheenemyofflow.Variabilitycanbesystematicallyminimizedandmanaged,butvariabilitywill never be eliminated. The onlyway to stop nervousness and the bullwhipeffect is to stop variation from being passed between the linked parts of thesupplychaininbothdirections.

Thisisaccomplishedthroughaconceptcalled“decoupling.”APICSdefinesdecouplingas:

Creating independence between supply and use of material. Commonlydenotes providing inventory between operations so that fluctuations in theproductionrateofthesupplyingoperationdonotconstrainproductionoruseratesofthenextoperation.(p.43)

Decoupling disconnects one entity from another. This isolates events thathappeninoneentityorportionofasystemandkeepsthemfromimpactingotherentitiesorportionsofthesystem.Thinkofdecouplingasifitwereafirewallinabuilding, automobile, or information systemor a breakwall aroundboats in amarina.

Theconceptofdecouplingprovidesthefundamentalbreakfromconventionthat is needed tomitigate variability.Decoupling breaks the direct connectionbetweendependencies.Theplacesatwhich thesystem isdecoupledarecalled“decouplingpoints.”APICSdefinesdecouplingpointsas:

The locations in theproductstructureordistributionnetworkwherestrategicinventory is placed to create independence between processes or entities.Selectionofdecouplingpointsisastrategicdecisionthatdeterminescustomerleadtimesandinventoryinvestment.(p.43)

Page 64: DDMRP - Demand Driven Material Requirements Planning

Decouplingisnotanewidea.TheconcepthasbeenaroundformanyyearsbutwithnopracticalwaytoimplementitinMRP.MRPwasdesignedwiththeexplicit intentionof tightlycouplingeverythingso thatpreciseequationscouldbe performed in order to synchronize the environment. Limited amounts ofdecoupling canoccur inMRP, but onlywith complicationswhere costs likelyoutweightheirbenefits(thisisdiscussedfurtherinChapter9).

Figure4-2isbasedonFigure3-3anddepictsthedependentviewofanMRPsystemandtheaccumulateddemandsignaldistortion(theupperarrowmovingrighttoleft)andthesupplycontinuityvariability(thelowerarrowmovinglefttoright).There is nodecoupling; thus thedistortion toboth relevant informationandmaterialsaccumulatesthroughthesystem.

FIGURE4-2TheMRPapproach

Decouplingpointsrepresentaplace todisconnect theeventshappeningonone side from the events happening on the other side. They delineate theboundariesofindependentlyplannedandmanagedhorizons.Thedeterminationof theirplacement isnot tobe taken lightly—it isastrategicdecision thatwilldramatically affect how the system operates and how effective the overallsystemwillbe.

DecouplingPointBuffers

Page 65: DDMRP - Demand Driven Material Requirements Planning

For the decoupling points tomaintain their decoupling effect, theremust be alevelofprotectionthatabsorbsdemandandsupplyvariabilityatthesametime.This level of protection is a concept called “decoupling inventory.” APICSdefinesdecouplinginventoryas:

An amount of inventory kept between entities in a manufacturing ordistribution network to create independence between processes or entities.Theobjectiveofdecouplinginventoryistodisconnecttherateofusefromtherateofsupplyoftheitem.(p.43)

Decouplingpointinventoryinthisbookwillbereferredtoasa“decouplingpoint buffer” or simply “buffer.” Decoupling point buffers are quantities ofinventoryor stock that aredesigned todecoupledemand fromsupply.Buffersarecommonlyamountsofinventorythatwillprovidereliableavailabilitytotheconsumersof thestockwhileat thesame timeallowingfor theaggregationofdemand orders, creating amore stable, realistic and efficient supply signal tosuppliersofthatstock.

Figure4-3depictsthesamesystemasFigure4-2butwithdecouplingpointbuffersinplace.Theplacementofdecouplingpointbuffers(representedasthetieredbucketiconsinthedependentstructure)createsindependentplanningandexecution horizons. These horizons are indicated by the dotted lines withroundedterminalpointsoneachside.Demandandsupplyvariabilityisstoppedfrom further accumulation at those terminal points. This is represented by thewall-likeiconslabeled“breakwall.”Thismeansthattheuseofdecouplingpointbuffers addressesboth componentsof thebullwhip at the same timeand fromthesameplace;itisabidirectionalsolution.

Decoupling buffer placement has huge implications for lead times. Bydecouplingsupplying lead times fromtheconsumptionsideof thebuffer, leadtimesare instantlycompressedbetweenbuffersand to thecustomer.This leadtime compression has immediate service and inventory implications. Marketopportunitiescanbeexploited,whileworkingcapitalrequiredinbuffersplacedathigherlevelsintheproductstructurecanbeminimized.

Furthermore,Figure4-3revealsanadditionalleadtimecompressionbenefitduetodecoupling:itsimpactonrelevantinformation.AsdiscussedinChapter3,MRPusersareforcedtomakecommitmentstoademandsignalthatissubjecttovaryingdegreesoferror(forecastedorders).Whilethereisanalternativemuchmore accurate demand signal (sales orders), MRP’s inability to decouplepreventstheexclusiveuseofthatsignal.

Page 66: DDMRP - Demand Driven Material Requirements Planning

FIGURE4-3Asystemwithdecouplingpointbuffers

YetwhatFigure4-3showsisthatwhenadecouplingpointbufferisplacedinside the sales order visibility horizon, it will allow for the system toexclusivelyusethataccuratedemandinput.Wehaveeffectivelyfoundthetimethatwebelievedwe lacked that forced theuseof forecastedorders in the firstplace.When decoupling point buffersmatch the sales order visibility horizon,thedemandvariabilityisreduced.

Summary

Decoupling simply makes sense given the basic circumstances that we facetoday.Wehaveextendedsupplychainsgloballyandmadethemmorecomplexandfragile.Theselongerandmorecomplexsupplychainsaresubjecttomuchhigherlevelsofvariabilityandaremuchhardertoplan.Breakingdependenciesin key places will dramatically simplify the planning equation and providesshorterhorizonswithmuchmorerelevantinformation.

Theconceptofdecouplingposesanironicsituation.Inordertopromoteandprotecttheflowofrelevantinformationinasystem,youmuststrategicallyandpurposefully slow or interrupt flow at certain critical points. The size of thedecouplingpoint buffers defines the lengthof the slowdownor interruption atthesecachingpoints.

Unfortunately,conventionalplanningsystemsaredesigned topositionandthen manage decoupling points. The very basic foundation of Material

Page 67: DDMRP - Demand Driven Material Requirements Planning

RequirementsPlanningwastomakeeverythingdependentandonlyorderwhatwasneeded,whenitwasneeded,inamathematicallypreciseway.Decouplingcreates a position of independence. The inability to decouple is the primaryculpritbehindthebullwhipeffectandisamajorimpedimenttosystemflow.

Decouplingisthekeythatunlocksadecades-oldstrugglewithconventionalplanning approachesutilizingMRP, a struggle that is becomingmore acute intheNewNormal. Itallowsadoor toopen toaplacewheredailyplanningcanbecome obvious, intuitive, and beneficial for supply order generation andmanagement.What isneeded isasystematicapproachforutilizingdecouplingthatfundamentallyanswersthesekeyquestions:

Where to place these decoupling points? The answer is neither“everywhere” nor “nowhere.” The answer is simply stated as“somewhere.” But how to find that best somewhere? PlacingdecouplingpointswillbethesubjectofChapter6.

How to size the protection at the decoupling point? In order tomaintain the integrityof thedecouplingpoint, thebuffersmustbesizedinrelationtothespecificattributesoftheparts,planning,andexecutionhorizons they are protecting.Thiswill be the subject ofChapter7and8.

Howtomaintain thatprotection?Supplyordersmustbegeneratedandmanaged inaway thatkeeps thepointsproperlysuppliedandintact.ThesetechniqueswillbeexploredinChapters9and10.

Page 68: DDMRP - Demand Driven Material Requirements Planning

PART2

BecomingDemandDriven

Page 69: DDMRP - Demand Driven Material Requirements Planning

CHAPTER5

SupplyOrderGeneration andExecution fortheNewNormal

At this point we are at a crossroads.We can continue to struggle with ourconventional planning systems, or we can seek a break from convention, analternativedesignedfortheNewNormal.Thatalternativedesignmustpromoteandprotecttheflowofrelevantinformationandmaterials.Itmustsystematicallybreak down the distortions to demand signals and material supply thatcharacterize thebullwhipeffect throughtheeffectiveuseofdecouplingpoints.Butwheretogofromhere?

MRPVersusLean—WhatCanWeLearn?

The basic elements of this alternative design can be better understood byexploring a chronic conflict between two camps—the believers in LeanmethodologyandthebelieversinMRP.

Figure5-1 illustrates a side-by-side comparison of conventionalMRP andLean approaches. On the MRP side, forecasted demand feeds a masterproductionschedule(MPS).TheMPScreatesastatementofwhatwillbebuilt.This is then fed to MRP. MRP then explodes through the bill of material,creating synchronized supply orders (date and quantity) as dictated by theproductstructure.Safetystockisoftenusedatthefinishedandpurchasedpartslevelsinordertoabsorbvariability.

The Lean approach establishes kanban positions, which are independentinventorypositions typicallyplacedateachresourceposition.Thekanbansaresizedaccordingtoarequiredtakttimerate.Thisratecanbeestablishedthrougha forecast or past consumption. The kanbans are connected with “loops” thatprovideeasy-to-interpretsignalsforeachpositiontoproduceornotproduce.A

Page 70: DDMRP - Demand Driven Material Requirements Planning

“supermarket”canbeplacedat the intermediateorpurchasedcomponent levelthat produces the same easy-to-interpret signal as the kanban. The differencebetweenthesupermarketandthekanbanissimplythatthesupermarketisatthepart(productstructure)levelandthekanbanisplacedattheresourcelevel.

Many Lean implementations attempt to abandon the formal planningapproachofMRPbecauseitisseenasinappropriate,transactionintensive,non-valueadded,evenantithetical,towhatLeanistryingtoaccomplish.Thiscausestremendousfrictionbetweenplanningpersonnelandthosepushingtoeliminatethese systems. Lean facilitators and advocates often see MRP as anovercomplicatedandwastefuldinosaurthatsimplydoesn’tworkinacustomer-centricworld.

FIGURE5-1MRPandLeancomparison

Planning personnel, however, see it in a completely different way. Theybelieve that without the ability to see and synchronize complex and dynamic

Page 71: DDMRP - Demand Driven Material Requirements Planning

environments,criticalblindspotswill exist in theplanningprocesswhichwilllead to shortages, expedites, and even excessive inventory positions tocompensate. They see the simple pull approach for managing materials andinventory as a gross oversimplification for the complex planning and supplyscenariosthatarethenormintoday’smorevolatileenvironment.

What if both camps are right? What if in many environments today thetraditionalMRPapproachistoocomplexandtheLeanapproachistoosimple?Wherewould that leavesupplychainmanagement? Itwouldcreatea situationwherecompaniesoscillatebetween the twooptions,dependingon thepoliticalwind employing a constantly changing and unsatisfactory number of work-arounds and compromises. Executives get frustrated, in-fighting escalates,efforts are sabotaged, more money and time are spent, and improvementsdeterioratetolipservice.

Einsteinoncesaid,“Any intelligent foolcanmake thingsbiggerandmorecomplex. It takes a touch of genius—and a lot of courage—to move in theopposite direction.” He also said, “Everything should be made as simple aspossible,butnotsimpler.”Howprophetic.

Can traditionalMRPbeoverlycomplex?Withoutadoubt.Mostpeople inmanufacturingcompaniesdon’tevenfullyunderstandwhattheplanningsystemisorhow itdoeswhat itdoes.Everyday,plannersaredrowning inoceansofdata and action messages. The hard-coded rules are rooted firmly in the old“push and promote”methodology that makesMRP ill-suited to today’smorevolatileandservice-orientedworld.Furthermore,“fixing”or“cleaningup” thesystem seems to be a never-ending, transaction-intensive, and expensivejourney;theendofwhichistoalwaysbepreciselywrong.

Can Lean be an oversimplification? When it comes to materials andinventoryplanning,theanswerinmanyenvironmentsisyes.Oversimplificationis defined as “To simplify to the point of causing misrepresentation,misconception, or error.”1 By failing to provide visibility to criticaldependencies and relationships with regard to supply, demand, on-handinventory, andproduct structure, theLean tool set canattempt tooversimplifymanyenvironments.Thelarger,morecomplex,andvariabletheseenvironmentsare, the more likely that Lean’s simple kanban controls and lack of materialplanningcreateanoversimplifiedapproach.

Yet there is one thing they actually agree on: a common objective. Bothcamps can agree that flow is paramount. Now more than ever, a decisivecompetitiveadvantagecanbeachievedbycompanieswithahighdegreeofflow

Page 72: DDMRP - Demand Driven Material Requirements Planning

throughandtotheircustomers.Chapter2highlightstheneedforandbenefitsofflow.The better the flow of relevant information andmaterials, the better theservice levels and use of working capital. The better the service levels andworkingcapital,thebetterthebottomline.

Do MRP advocates disagree with this? Certainly not! Materials andprocessesthatflowreliablyaretheeasiesttoplanandmanage.Havingtherightthingsattherighttimeisthekeytoflow.MRP’sentirereasonforexistencewastoattempttosynchronizeenvironmentstohavetherightthingsattherighttime.ButdoesconventionalMRPhavedeficienciesthathurtflow?Undeniablyyes!

DoLeanadvocates agreewith theneed for flow?Absolutely. Informationandmaterialsthatflowreliablygenerateconsiderablylesswaste.ButdoesLeanhave a complete tool set for fully protecting and improving flow at the plant,enterprise,andsupplychainlevelinamorecomplexandvolatileworld?Thereseemstobesomethingmissing.

Soifthereisacommonobjective,whyarethesecampslockedinachronicconflict?Asdiscussedpreviously,MRPisaperfectjust-in-timesystemthatnetstozeroinventory.ThissoundsincrediblycompatiblewithLean’sapproach.Yetthe conflict persists because they represent diametrically opposed approacheswithregardtotwocriticalfactors.Thesecriticalfactorsareessentiallytwosidesof the same coin but are worth discussing specifically. Any solution mustinvolveaddressingthisinherentoppositioninthesetwoareas.

DependenceVersusIndependence

Figure 5-2 illustrates the area of this particular conflict. At the top of thestructure is thecommonobjectiveofprotectingandpromotingflow.MRPandLeanhavedifferentcriticalneedsinordertoaccomplishthisobjective,andeachmethodhasaspecificattributedesignedtosecureitsrespectivecriticalneed.

Page 73: DDMRP - Demand Driven Material Requirements Planning

FIGURE5-2ThedependenceversusindependenceconflictbetweenMRPandLean

We will explore the MRP side first. MRP advocates understand that theprotection and promotion of the flow of relevant information cannot occurwithouttheabilitytosynchronizecomplexanddynamicenvironments.Inordertoaccomplish this synchronization,MRP,asdiscussed inChapters3 and4, ishard coded to obey the dependencies defined by product structure. A changeanywherecreateschangeeverywhere.MRPwasdesignedinthiswayinordertomake sure that the operating environment could understand the impact ofchangesastheyoccur.

OntheLeansideweseethattheprotectionandpromotionofflowrequiresthatresourceshaveclearsignalstooperateby.Whensignalsbecomeconfusingorconflicting,aresource’sabilitytodeterminewhatiscorrectbreaksdown.Toomany points of data or constantly changing signals create that confusion orconflict.ThusLeanmakeseverythingindependent.Resourcesonlyneedtolooktooneplace(thekanbanthattheyfeed)todetermineiftheyshouldproduceornotproduce.Itisliterallythatsimple.

Making everything dependent versus making everything independent iscertainlymutuallyexclusive.Infact,bothsides’attributesbreakdowntheotherside’scriticalneed.Bymakingeverythingdependent,MRPcreatesanincrediblyconfusing set of constantly changing and conflicting signals. By makingeverything independent, an environment loses the ability to synchronize tochanges that can and will occur. This is particularly true in environments

Page 74: DDMRP - Demand Driven Material Requirements Planning

characterized by heavy demand fluctuations, long lead time parts, sharedresourcebases,andproductinnovation.

Is there an alternative that can have the benefits of dependence(synchronization) and the benefits of independence (clear signals) at the sametimewithoutconflict?

SupplyOrderGeneration(PlanningVersusExecution)

Figure5-3 illustrates theareaofanotherconflict.Onceagain,at the topof thestructure is the common objective of protecting and promoting flow. MRP’sneed remains unchanged from the previous conflict: synchronize complex anddynamicenvironments.Leanhasadifferentcriticalneed.

Both sides have a different attribute designed to secure their respectivecriticalneedwithregardtosupplyordergeneration.Onecouldarguethatthisissimplyanextensionof thedependenceversus independenceconflict.That isavalidargument,butthereisanadditionallevelofinsightthatcouldbebroughttobearbydiscussingthemseparately.

AsdescribedindepthinChapter2,MRPistypicallyloadedwithforecasteddemand in order to attempt to synchronize the long manufacturing andprocurement cycleswith anticipateddemand.Thishappenswell in advanceofconsumption. The Lean side, however, seeks to protect and promote flow bypacingtoactualdemandbecausetheinherentinaccuracywithforecastsdirectlyimpedes flow. Resources are squandered on things that are overproduced andoverorderedinadvance,whileexpeditesmustaccommodatethethingsthatwereunderproducedorunderorderedasthepicturebecomesclearer.Theonlywaytotrulyknowifdemandisreal isafter ithasoccurred.Consumptionisdefinitiveproofofdemand.

Aswiththepreviousconflict,theattributesseemtobemutuallyexclusive.WhenMRPgeneratessupplyorderswellinadvanceofanticipatedconsumption,it loses the capability to pace to actual demand at least by the amount of theforecast error. The longer the planning horizon, the greater the forecast error.When we generate supply orders at the execution level, there is a delay inresponding to significant changes, as the supply orders must make their waythroughtheentireconnectivestructureonelevelatatime;thereisrudimentarybutextremelyslowsynchronizationatbest.

Page 75: DDMRP - Demand Driven Material Requirements Planning

FIGURE5-3ThesupplyordergenerationconflictbetweenMRPandLean

Isthereanalternativethatcanhavethebenefitsofsupplyordergenerationat the planning level (synchronization) and the benefits of supply ordergenerationattheexecutionlevel(pacetoactualdemand)?

Itisvitaltounderstandthatallthecriticalneedsarerequiredtoprotectandpromote flow. Planning and execution systems must pace to actual demand,provide clear signals for resources, and synchronize complex and dynamicenvironments.

Focusing on only one critical need and discounting the others almostguarantees challenges to flow. Indeed that seems descriptive of the impassebetween the MRP and Lean worlds. When it comes to the protection andpromotion of the flow of relevant information and materials, both Lean andMRP have weaknesses in today’s more volatile and complex environments.Lean’srelianceonindependentreplenishmentkanbanswithlittletonovisibilityorconnectivityattheplant,enterprise,andproductstructurelevelisaproblemfortheprotectionandpromotionofflow.Buttheantiquatedandcomplexrulesof conventionalMRP that govern demand and supply order generation createunrealistic,constantlychanging,andgenerallyconfusingplansandschedules.

Toprotectandimproveflow,ablendofsimplevisiblepullsignalsandthecomputational and connective power of technology is necessary. This isn’t acompromise for the two sides to live in peace; it must be a harmoniousintegrationwherebothsides’criticalneedsareincorporatedtocreateastronger

Page 76: DDMRP - Demand Driven Material Requirements Planning

solution for the protection and promotion of flow. And it must be practical,consistent,andscalable!

Whatifthereisawaytodefineasolution(rulesandtools)thatisnotoverlycomplex or overly simple? What if there is a way to take key and relevantaspectsofbothpointsofviewandcreateanelegantblueprintthatwillworkforand enhance both sides’ objectives? This solution must include a level ofsophistication that can provide more visibility and synchronization from aplanning and execution perspective while at the same time pace to actualdemand and promote simple, clear, and highly visible signals across theenterprise. This solution is called Demand Driven Material RequirementsPlanning(DDMRP).

LeanandTechnology

Lean advocates often get accused of being anti-technology, but do Leanadvocatesreallywantmanufacturingcompaniestoentirelyabandonthepromiseof technology? The answer should be yes when that technology is wasteful,confusing, and not reflective of reality, when it force-fits concepts so as tosimplypermittheiruse.Unfortunately,thishasbeenthesituationforquitesometime with regard to traditional MRP and DRP (distribution requirementsplanning)systems.

Point 8 of the Toyota Production System states, “Use only reliable,thoroughlytestedtechnologythatservesyourpeopleandprocesses.”2Untilnowtheprevailingmaterialsandinventoryplanningandexecutiontechnology,whilethoroughly tested, have been largely inappropriate to serve the people andprocesses in companies transforming to a leaner approach. Chapter 3 clearlydemonstratedthatpoint.

Yet the proliferation and sustainability of Lean implementations has beennegativelyimpactedbythelackofappropriatesupplychainmaterialsplanningand execution technology. Many well-respected manufacturing analysts haveconcluded that there is tremendous potential for the incorporation of betterplanning and visibility software into Lean implementations. ManufacturingneedsLeanmethodstostaycompetitiveinthemorecomplexenvironmentofthetwenty-first century. Lean needs an effective customer-centric planningapproachtobringthatvisiontoreality.

Whatiftherewereanappropriatetechnology?Whatifareliable,thoroughlytested method for a customer-centric pull-based planning and execution ofsupplychainmaterialswithhighdegreesofvisibilitycouldbeintroducedtothe

Page 77: DDMRP - Demand Driven Material Requirements Planning

MRPworld?UnderthatconditionLeanandMRPwouldbothfindaneffectivesolution.ThismethodisDemandDrivenMaterialRequirementsPlanning.

DemandDrivenMaterialRequirementsPlanning

ThissectionwillservetointroduceDDMRP—itsbasicfoundationanditsmajorcomponents.Butfirstitmayhelptounderstandwhattheterm“demanddriven”reallymeans and thehistorybehind it.The termwasoriginallydefinedas theabilityto“sensechangingcustomerdemandandadaptplanningandproductionwhilepullingfromsuppliersallinrealtime.”

TheHistoryof“DemandDriven”

The termwaspioneeredbyPeopleSoft in2002whileCarolPtakwas thevicepresident of manufacturing and distribution industries.When Oracle acquiredPeopleSoftin2003,thetermwaslargelyabandoned.Itwasthenresurrectedin2007 byAMR. In 2010AMRwas acquired byGartner, andGartner used thetermaspartofwhatitcalledits“DemandDrivenValueNetwork”approach.

In2011thethirdeditionofOrlicky’sMaterialRequirementsPlanning(Ptakand Smith) introduced the initial blueprint for Demand Driven MaterialRequirementsPlanningasanalternativeformalplanningandcontrollogic.Theyear2011alsomarkedthefoundationoftheDemandDrivenInstitutebyCarolPtakandChadSmith.TheDemandDrivenInstitutehaspublishedseveralwhitepapersandcasestudiesontheDDMRPtopic.Arepositoryofcasestudiesandwhite papers on DDMRP is available free athttp://www.demanddrivenworld.com.

In2012theDemandDrivenInstitutepartneredwiththeInternationalSupplyChain Education Alliance (ISCEA) to offer the Certified Demand DrivenPlanner (CDDP) program. The CDDP program was designed to provideconsistent global standards for theDDMRPapproach and to teach and certifypractitionersinthosestandards.From2012to2015over1,000peopletooktheCDDPprogramonsixcontinents.

In2013DemandDrivenPerformance:UsingSmartMetricswaswrittenbyDebraSmithandChadSmith.Thisbookextendedthetermacrosstheenterpriseintofinance,scheduling,shopfloorcontrol,andstrategy,effectivelydefiningtheDemandDrivenOperatingModel. Thiswill be defined and discussed later inthischapter.

In 2016, with Demand Driven Performance: Using Smart Metrics as aguide, theDemandDriven Instituteand ISCEAreleased theCertifiedDemand

Page 78: DDMRP - Demand Driven Material Requirements Planning

DrivenLeader(CDDL)program.TheCDDLprogramwasdesignedtoprovideconsistentglobalstandardsfortheDemandDrivenOperatingModel(includingDDMRP)andtoteachandcertifypractitionersinthosestandards.

Position,Protect,andPull

Theoriginaldefinitionof“demanddriven” stillworks in today’smorematureandlargerdemanddrivenbodyofknowledge.Additionally,thismaturationandexpansionhasprovidedclarityonpreciselywhatdemanddrivendoesnotmean.Itdoesnotmean“make toordereverything.” Itdoesnotmean“put inventoryeverywhere.” It does not mean “forecast better.” Becoming demand drivenrequires a fundamental shift from the centrality of supply- and cost-basedoperational methods (commonly referred to as “push and promote”) to acentralityofactualdemand-andflow-basedmethods(commonlyreferredtoas“position,protect,andpull”).Theterm“actualdemand”isextremelyimportantin distinguishing it from a rebranded and somehow superior forecastingapproach.

DemandDrivenMaterialRequirementsPlanning isa formalmulti-echelonplanning and execution method to protect and promote the flow of relevantinformation and materials through the establishment and management ofstrategicallyplaceddecouplingpointstockbuffers.DDMRPhasroots inmanyconventional methods. Figure 5-4 shows the methodological foundation forDDMRP.

DDMRPcombinessomeofthestillrelevantaspectsofMRPandDRPwiththepullandvisibilityemphasesfoundinLeanandtheTheoryofConstraintsandthe variability reduction emphasis of Six Sigma. Do these methods all justnaturallyfusetogether?No.Ataminimum,asnotedearlierinthischapter,thereareconflictsbetweenLeanandMRP.AsimilarconflictoccurswithMRPandtheTheoryofConstraints.ThefinalcomponentofthisfusionrequiresafewkeyinnovationsthatareuniquetoDDMRP.

Page 79: DDMRP - Demand Driven Material Requirements Planning

FIGURE5-4ThemethodologicalfoundationofDDMRP

Demand Driven Material Requirements Planning has five sequentialcomponents.Figure5-5 illustrates these components, their sequence, and howtheyrelatetothemantra“position,protect,andpull.”ThesefivecomponentsarerespectivelyfeaturedinsequentialchaptersstartingwithChapter6.

The first three components essentially define the initial and evolvingconfiguration of a Demand Driven Material Requirements Planning Model.Strategic inventorypositioningwilldeterminewhere thedecouplingpointsareplaced. Buffer profiles and levels will determine the amount of protection atthose decoupling points. Dynamic adjustments define how that level ofprotection flexes up or down based on operating parameters,market changes,andplannedorknownfutureevents.

The fourth and fifth elements define the actual operational aspects of aDDMRPsystem:planningandexecution.InDDMRP,demanddrivenplanningis theprocess bywhich supplyorders (purchaseorders,manufacturingorders,andstock transferorders)aregenerated.Visibleandcollaborativeexecution istheprocessbywhichaDDMRPsystemmanagesopensupplyorders.

DDMRP is at theheart of theDemandDrivenOperatingModel,which isdefinedthisway:

Demand Driven Operating Model—a supply order generation, operationalscheduling,andexecutionmodelutilizingactualdemand incombinationwithstrategicdecouplingandcontrolpointsandstock,time,andcapacitybuffersinordertocreateapredictableandagilesystemthatpromotesandprotectstheflow of relevant information and materials within the tactical relevantoperational range (hourly, daily, and weekly). A Demand Driven OperatingModel’s key parameters are set through the Demand Driven Sales and

Page 80: DDMRP - Demand Driven Material Requirements Planning

Operations Planning process to meet the stated business and marketobjectiveswhileminimizingworkingcapitalandexpedite-relatedexpenses.

FIGURE5-5ThefivecomponentsofDDMRP

Figure5-6depictstheDemandDrivenOperatingModelschema.Modelandpartparameters,commonlyreferredtoas“mastersettings,”aresuppliedbytheDemandDrivenSalesandOperationsPlanning(DDS&OP)processtoDemandDrivenMRP.ThesesettingswillbethoroughlydescribedinChapters6,7,and8.DDS&OP is the subject ofChapter13. Thesemaster settings configure theDDMRPsystemandarecombinedwithinventoryandproductstructurerecordsand actual demand to generate supply orders. If these supply orders aremanufacturing orders, they are sent to scheduling. If the supply orders arepurchaseorstocktransferorders,theygodirectlyintoexecution.

ThisbookprovideslimitedcontentonDemandDrivenCapacitySchedulingor its respective master setting inputs. What content there is on this topic isfound in Chapter 11. Additionally, the execution aspects in this book areconfined to the execution elements of DDMRP—the management of opensupplyorders.Forin-depthcontentonDemandDrivenCapacitySchedulingandrelatedshopfloorexecutiontactics,refertoSmithandSmith’sDemandDrivenPerformance: Using Smart Metrics. Figure 5-7 depicts the aspects of theDemand Driven Operating Model covered in the subsequent chapters of thisbook.

Page 81: DDMRP - Demand Driven Material Requirements Planning

FIGURE5-6DemandDrivenOperatingModelschema

FIGURE5-7Theemphasisofthisbook

Page 82: DDMRP - Demand Driven Material Requirements Planning

Summary

ThischaptercombinedwithChapters3and4has laid thebasic foundationforthe critical elements of a new supply order generation method for the NewNormalcalledDemandDrivenMaterialRequirementsPlanning.

WhatdoweknowsofarabouttherequirementsforDDMRP?

1. It should be based on the protection and promotion of the flow ofrelevant informationandmaterials.Thisconnects it todrivingbetterreturnsoninvestment.ThiswasexplainedinChapter3.

2. Itmustallowfordecoupling inorder tomitigatedemandsignalandsupplycontinuityvariabilityaswellas tocompress lead times.ThiswasexplainedinChapter4.

3.Itshouldusethemostrelevantdemandinformationavailable—actualdemand.TheproblemsassociatedwithforecasterrorweredescribedinChapter3andpreviouslyinthischapter

4. Itmust provide easy-to-interpret signals for all resources. Thiswasdescribedpreviouslyinthischapter.

5. It must provide for a way to synchronize complex and dynamicenvironments.

Page 83: DDMRP - Demand Driven Material Requirements Planning

CHAPTER6

StrategicInventoryPositioning

At the riskofoversimplifying theeveryday tasksofbuyersandplanners,weshouldunderstandthattheyareconstantlydealingwithtwoquestionsofsupplyorder management. The two questions are these: How much and when?Hundreds ifnot thousandsofbookshavebeenwrittenaboutawidevarietyoftechniquesandtrickstoattempttoanswerthesequestions.

Thequestionofhowmuchisaquestionconcerningquantity.Plannersandbuyersarecontinuallyvalidating,verifying,andsupplementinghowmuchtheyreallyneedversuswhatMRPistellingthem.Thequestionofwhenissimplyaquestionoftiming.Plannersandbuyersarecontinuallyvalidating,verifying,andsupplementingwhen they reallyneed thingsversuswhatMRP is telling them.Thisisaconstantlychangingseriesofwronganswersassystemnervousnessandthebullwhipimpacttheenvironment.

Thustheirdailyobjectivedegeneratestosimplybeinglesswrong.Theyareconstantlychallengedabouthowtheyhistoricallyansweredthesequestionsandwhy things are not available in the time or quantity that they are needed. Acommonpractice is for theplannersorbuyers tosavescreenshotsof theMRPsysteminordertocreateadefenseforwhytheydidwhattheydidandwhentheydidit.Afrustratingsituationindeed.

Perhaps all this activity and series of constantly dissatisfactory answers isnot related to the questions, how much and when? Perhaps it is first andforemostrelatedtoourfailuretoaskamorefundamentalquestion.

As discussed earlier, the key to protecting and promoting the flow ofrelevantinformationrequirestheuseofdecouplingpoints.Decouplingenablesabidirectional benefit—it mitigates the demand signal distortion (relevantinformation)andsupplycontinuityvariability(relevantmaterials)inherentinthebullwhip effect. But this raises a question—where should these decoupling

Page 84: DDMRP - Demand Driven Material Requirements Planning

points be placed within a supply chain or organization to maximizeeffectiveness?

PositioningFactors

Mostorganizationsarecompletelyunprepared todealwith thisquestion.First,they lack the knowledge, comprehension, or even capability to even ask thesimplequestion,“Where?”Eveniftheydoasktherightquestion,theylacktheabilitytoeffectivelyanswerthatquestion.

Thus the first component of Demand Driven Material RequirementsPlanningisdeterminingwherethedecouplingpointsandtheirrespectivebuffersshould be placed. This component becomes the cornerstone of the DemandDrivenOperatingModeldiscussedinChapter5.Theselectionofthesepointsisastrategicdecisionthatimpactstheperformanceofthesupply-demandnetworkinmanyregards:service,workingcapital,expedite-relatedexpenses,cashflow,andultimatelyreturnoninvestment.

Chapters1to4createdthein-depthcaseaboutwhythequestionof“where”must be asked. This chapter focuses on how to properly answer this questionthroughtheconsiderationofsixkeyfactors.

CustomerToleranceTime

This is the time the typical customer is willing to wait before seeking analternative source.Customer tolerance time also can be referred to as demandleadtime.AccordingtoAPICS,demandleadtimeis:

Theamountoftimepotentialcustomersarewillingtowaitforthedeliveryofagoodoraservice.Syn:customertolerancetime.(p.45)

Determining this lead time often takes the active involvement of sales andcustomerservice.

MarketPotentialLeadTime

This lead time will allow an increase of price or the capture of additionalbusiness through either existing or new customer channels. Determining thisleadtimetakestheactiveinvolvementofsalesandcustomerservice.Beawarethat there could be different stratifications of market potential lead time. Forexample, a one-week reduction in lead timemay only result in an increase inorders, whereas a two-week reduction in lead time could result in both an

Page 85: DDMRP - Demand Driven Material Requirements Planning

increase in orders and a potential price increase on some of those orders.Properlysegmentingthemarketwillmaximizethepossiblerevenuepotentialforthe company and provide excellent revenue growth control. This is aconsideration in Demand Driven Sales and Operations Planning, covered inChapter13.

SalesOrderVisibilityHorizon

Thesalesordervisibilityhorizonisthetimeframeinwhichwetypicallybecomeawareofsalesordersoractualdependentdemand.Inretailsituations,customersdonot issueasalesorder toa shop inadvanceofgoing to theshop.Thus thesales order visibility horizon in this situation is zero. In most manufacturingscenarios, however, there are sales orders conveyed in advance of expectingreceipt of the item. Often the sales order visibility either matches or exceedscustomertolerancetime.Thelongerthevisibilitytosalesorders, thebetter thecapabilityoftheenvironmenttoseepotentialspikesandderiverelevantdemandsignal information. In many cases relevant requirements are obscured fromplanners because all demand (including planned orders based on forecast andsafety stock requirements) is aggregated together for aggregate planningpurposes.

ExternalVariability

Externalvariabilityconsidersbothdemandandsupplyvariability.

VariableRateofDemandThis refers to the potential for swings and spikes in demand that couldoverwhelm resources (capacity, stock, cash, etc.). This variability can becalculatedbyavarietyofequationsordeterminedheuristicallybyexperiencedplanning personnel. As noted in the APICS Dictionary, “Mathematically,demandvariabilityoruncertaintycanbecalculatedthroughstandarddeviation,mean absolute deviation (MAD) or variance of forecast errors.” If the datarequired formathematicalcalculationdonotexist, companiescanalsouse thefollowingcriteria:

High-demand variability. Products and parts that are subject tofrequentspikeswithinthecustomertolerancetime.

Medium-demand variability. Products and parts that are subject tooccasionalspikeswithinthecustomertolerancetime.

Page 86: DDMRP - Demand Driven Material Requirements Planning

Low-demand variability. Products and parts that have little to nospikeactivity.Thedemand is stablewithin thecustomer tolerancetime.

VariableRateofSupplyThis is the potential for and severity of disruptions in sources of supply orspecificsuppliers.Thiscanalsobereferredtoassupplycontinuityvariability.Itcan be calculated by examining the variance of promise dates versus actualreceipt dates. When first considering the variable rate of supply, the initialvariances can be caused by critical inherent flaws in the MRP system.Additionally, thosedatesoften shiftdue toother shortcomingsassociatedwiththewayMRP is employed rather thanbecauseof the supplier capability.Anycritical supplier of a major manufacturer will know exactly which day itscustomer regenerates itsMRP. These suppliers will see a flurry of additionalorders, canceled orders, and changes to orders (quantity, specification, andrequestdate).

Ifthedatarequiredformathematicalcalculationdonotexist,thefollowingheuristicscanbeused:

Highsupplyvariability.FrequentsupplydisruptionsMediumsupplyvariability.OccasionalsupplydisruptionsLowsupplyvariability.Reliablesupply

InventoryLeverageandFlexibility

Thereareplacesintheintegratedbillofmaterial(BOM)structure(matrixbillofmaterial) or the distribution network that provide a company with the mostavailableoptionsaswellasthebestleadtimecompressiontomeetthebusinessneeds. Within manufacturing, these places are typically represented by keypurchasedmaterials,subassemblies,andintermediatecomponents.Thisbecomesmore critical in environments with BOMs that are deeper and more complex(broader)andhavemoresharedcomponentsandmaterials.Thisconceptwillbeexploredindetaillaterinthischapter.

CriticalOperationProtection

Similar to how variability can impact a bill of material, the longer andmorecomplex the routing structure and dependent chain of events (including

Page 87: DDMRP - Demand Driven Material Requirements Planning

interplanttransfers),themoreimportantitcanbetoprotectidentifiedkeyareas.Thesetypesofoperationsincludeareaswherethereislimitedcapacity,orwherequality can be compromised by disruptions, or where variability tends to beaccumulated or amplified. In Lean, these areas might be referred to aspacesetters. In the Theory of Constraints, they can be referred to as drums.Whatever manufacturing or operational methodology a company ascribes to,theseresourcestypicallyrepresentcontrolpointsthathaveahugeimpactonthetotal flowor velocity that a particular plant, resource, or area canmaintain orachieve.

TheprecedingsixfactorsmustbeappliedsystematicallyacrosstheentireBOM,routing structure, manufacturing facilities, and supply-demand network todetermine the best decoupling positions for purchased, manufactured, andfinisheditems(includingserviceparts)inordertoprotectandpromotetheflowofrelevantinformationanddrivereturnoninvestmentperformance.

ApplyingthePositioningCriteria

Asanexample,letusapplythesesixfactorstoarelativelysimpleenvironment.Inourexample,onlytwofinishedproductsaremade.Figure6-1showsthebillofmaterialforthetwoproducts:FPEandFPF.

Thenumbers in thecircles represent themanufacturingorpurchasing leadtime in days for each discrete part number. For instance, FPE takes 2 days tomakewhenallcomponentsareavailable,and204Phasapurchasingleadtimeof20days.

For each part number in this example, there are three relevant lead times.ThesearedescribedintheAPICSDictionaryas:

Manufacturing lead time (MLT):The total time required tomanufactureanitem, exclusive of lower level purchasing lead time. For make-to-orderproducts, it is the length of time between the release of an order to theproduction process and shipment to the final customer. For make-to-stockproducts, it is the length of time between the release of an order to theproduction process and receipt into inventory. Included here are orderpreparation time, queue time, setup time, run time, move time, inspectiontime,andput-awaytime.(p.98)

Page 88: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-1ProductstructuresforFPEandFPF

Cumulative lead time (CLT): The longest planned length of time toaccomplish theactivity inquestion. It is foundby reviewing the lead time foreach bill of material path below the item; whichever path adds up to thegreatestnumberdefinescumulativeleadtime.(p.38)

Purchasing lead time (PLT): The total lead time required to obtain apurchased item. Included here are order preparation and release time;supplier lead time; transportation time; and receiving, inspection, and put-awaytime.(p.142)

Considering these definitions, for FPE the manufacturing lead time is 2days,whilethecumulativeleadtimeis26days(20-daypurchasingleadtime+4days manufacturing lead time for 101 + 2 days manufacturing lead time forFPE). In the case of FPF, the manufacturing lead time is 3 days, while thecumulativeleadtimeis27(20-daypurchasingleadtime+4daysmanufacturingleadtimefor101+3daysmanufacturingleadtimeforFPF).

Toproperlyapplythesixfactors,wewillneedadditionalinformationabouttheenvironment.Figure6-2showstheproductandroutingstructureofbothFPEandFPFtogether.A“routing,”asdefinedbyAPICS, is“informationdetailingthemethodofmanufactureofaparticularitem.Itincludestheoperationstobeperformed,theirsequence,thevariousworkcentersinvolved,andthestandardsfor setup and run.” Together, the BOM and the routing paint a relativelycomplete picture of the view needed to consider positioning for this scenario.Note thatno run ratesandsetup timeshavebeendefined,as thesewillnotberelevantforthissimpleexample.

Once a part 205P is introduced to the manufacturing process, it is runthroughaseriesofresources(A>B>C>D)andcombinedwithaconverted

Page 89: DDMRP - Demand Driven Material Requirements Planning

204PatresourceZ.Part204Pisrunthroughaseriesofresources(B>C>E>F). Resource Z is an assembly operation and the final step in producingintermediate part 101. This conversion process (from 204P and 205P to 101),assuming concurrent activity across paths, takes four days on average. Thus101’smanufacturingleadtimeisfourdays.

FIGURE6-2ProductandroutingstructureforFPEandFPF

ResourceZisa“convergentpoint.”Aconvergentpointisanyplacewhereroutinglegscometogether.AsdiscussedinChapter3,thesepointsofintegrationoccurmostoftenwheresignificantdelaysaccumulatebecauseallpartsmustbepresent for the resource to perform its operation. Resource Z requires aconverted204P fromresourceFandaconverted205P fromresourceDat thesametimeandquantity.ThismakeresourceZacandidateforaresourcethatwewouldliketoprotectasmuchaspossible—acriticaloperation.

Part101 isa“pointofdivergence.”Adivergentpointmeans thatpart101can be directed into differentmanufacturing paths culminating in various enditems.Adivergentpointrepresentsacommitment thatcannotbepracticallyorcost-effectively reversed.An examplewould be the introduction of a sheet ofsteelintoafabricationprocess.Oncethesheetiscut,theoptionsavailabletouseitarenarrowedsignificantly.Thusthedecisiontocutitprecludesitfrombeingusedinmanyotherways.

For thisexample,part101 isdirected to resourcesSandT toeitherbegintheprocesstoconvertittoFPEorbecombinedwiththepurchasedpart102Ptobe finished into an FPF. The conversion into FPE takes two days, and theconversion to FPF, a more complicated build, takes three days. Thus themanufacturingleadtimeistwodaysforFPEandthreedaysforFPF.

Whencheckingwithsalesandcustomerservice,wefindthatthecustomertolerancetimeforbothproductsisatthreedays.FPFhaslowervolumes,asitis

Page 90: DDMRP - Demand Driven Material Requirements Planning

ahigher-endproduct,butthemarketexpectsitwithinthesametimeframeasthelower-endproductFPE.Additionally,saleshasindicatedthattherearefrequentopportunities in themarket forFPE towinquick-turnbusiness.Customersarenotinclinedtopaymorefortheitems,butthevolumewoulddefinitelyincreasewiththecapabilitytooffersame-dayfulfillment.Finally,withtheexceptionofquickturn requests, this company typically receives sales orders at least threedays in advance forbothproducts.Occasionally there canbe largeorders,butthoselargerorderstendtohaveatleasttwoweeksofsalesordervisibility.

Whencheckingwithpurchasing,wediscoverthatthesuppliersfor204Pand205Phavedecentreliability.Occasionaldisruptionsdohappen,butoverallbothhave performedwell over the last year. The supplier for 102P, however, is adifferent story.This supplier isnotorious for latedeliveriesandevenroutinelyproduces suspect quality. Figure 6-3 summarizes the positioning criteriainformationforthisexample.

FIGURE6-3Exampledecouplingpointpositioninganswers

Based on these answers, how should decoupling point positioning beapproachedinthisenvironment?Theimpactofeachofthecriteriaonthemodelisconsidered:

Customer tolerance time. Three days makes it a requirement toconsiderdecouplingattheenditemor101and102Plevels.Todoanything less will require making product to some sort ofanticipatedsignalorforecastandincurthenegativesassociatedwiththat.

Market potential lead time. The opportunity for FPE suggests a

Page 91: DDMRP - Demand Driven Material Requirements Planning

benefitfordecouplingandstockingatFPE.Theadditionalvolumeorcustomerscouldprovideprofitablerevenuegrowth.

Saleordervisibilityhorizon.Decouplingatthefinishedgoodsor101and102Plevelswouldallowtheenvironmenttopacetoactualsalesorders. This is the most relevant demand signal assuring thealignmentofourresourcestoactualrequirements.

Externalvariability.Demandvariabilitydoesnotseemtobeahugeissue—large orders are typically known in advance. Supplyvariability is an issue for 102P. Stocking at 102P would seemprudent.

Inventory leverageand flexibility.Decoupling and stocking at 101would allow the common component to flow to the end items asrequired.

Criticaloperationprotection.Whilethesuppliersfor204Pand205Pare reliable, decoupling those positions would provide as muchprotection to resource Z as possible from a product structureperspective.

In consideration of these answers to the positioning criteria, Figure 6-4showsamodelforthisenvironment.

FIGURE6-4Decouplingpositionsbasedonpositioningfactoranswers

Thekeyelementsandbenefitsofthismodelinclude:

TheFPEstockpositionallowsforquick-turnbusinesstobesatisfied.Thisallowsforanincreaseinsalesrevenue.

TheFPEstockpositionisminimizedduetotheshortleadtimefromthedecouplingpointat101.

Page 92: DDMRP - Demand Driven Material Requirements Planning

FPF can move to an assemble-to-order strategy as the lead time(three days) and customer tolerance times (three days) arecompatible.Achievingthisleadtimereliablyshouldbepossibleforthree reasons. First, 101 and 102P are available as needed,decoupling lead time from the front part of the manufacturingprocessandsupplier,respectively.Second,demandvariabilityisnotan issuewith this product, as large orders are typically known inadvance. And third, the buffer at FPE minimizes short-rangecapacitycontentioninresourcesSandTthatcouldaffecttheabilitytoconsistentlyachievethethree-dayleadtimeforFPF.

Thedecouplingpointsat204Pand205Pallowsuppliervariabilitytobeisolatedfromtheconcurrentmanufacturingprocessesinfrontofresource Z, thus minimizing as much as possible from a productstructureperspectivethevariabilityexperiencedatresourceZasanassembly operation.More can be done to protect resource Z, butthose options are outside the scope of decoupling pointconsiderations.Forexample,a timebuffercanbeused inadvanceofresourceZinordertoallowforcomponentstobesynchronizedeffectively.This,however,isattheschedulingandexecutionleveloftheDemandDrivenOperatingModel.

Additionally, we can use this example to illustrate the disadvantages ofoverflattening bills of material. Figure 6-5 assumes that the intermediatecomponentisremovedfromtheproductstructureofbothFPEandFPF.

Figure6-5showstheimpactondecouplingpointpositioning.Byremoving101fromtheproductstructure,weexpandthemanufacturingleadtimesforbothFPEandFPFfromtwoandthreedaystosixandsevendays,respectively.Sincethoseleadtimesarewellbeyondcustomertolerancetime,itforcesbothfinishedpositionstobestocked.Thosestockrequirementswillberelativetothatlongerlead timeaswell.Additionally,welose the inventory leverageof thecommonitem, resulting in higher finished goods inventory levels. Finally, losing thecapabilitytodecoupleat101meansthatthereisalongersequenceofactivitytoconvert rawmaterials into finished items. The variability is illustrated by thewavy linegrowing in amplitude inFigure6-6.This increasingvariabilitymaycreatetheneedtoholdadditionalstockattheenditemlevel.

Page 93: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-5FlattenedFPEandFPFbillsofmaterial

Thisexamplealsoserves tohighlight thesystemiceffectofusingmultipledecouplingpoints together.Whentherearemultiple tiersofdecouplingpoints,there are bidirectional benefits to each of them. Figure 6-7 depicts therelationship between tiers of decoupling points. The decoupling pointsessentiallyprotecteachother.Forexample,criticalsubcomponentsareprotectedfromdemandvariabilitybytheenditemdecouplingpoint,whilesubcomponentdecoupling points protect the end item decoupling point from long lead timesandlargeaccumulationsofsupplyvariability.

FIGURE6-6Examplewithflattenedbillofmaterial

Page 94: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-7Illustratingthebenefitsoftiersofdecouplingpoints

ThepreviouspositioningexamplebringsustoacriticalrealizationpointthatwillhavesignificantimplicationsonhowDemandDrivenMRPreallyworks.Byhighlighting the importance of tiers of decoupling points and leveraginginventory, this example has led us to another important impact of usingdecouplingpoints—anentirelynewformofleadtime.

ANewFormofLeadTime

The concept and necessity of decoupling was introduced and explained inChapter4.Decouplingallowsfordemandsignaldistortionandsupplycontinuityvariability tobe simultaneously combated.Decouplingpoints are theplaces atwhichwewage thatbattle.Theuseofdecouplingpoints also results inanewtypeofleadtimethatmustbeunderstoodandcalculatedinordertobeableto:

CompressleadtimestorequiredrangesDeterminerealisticduedateswhenneededSetdecouplingpointbufferlevelsproperlyFindhigh-valueinventoryleveragepointsfordecoupling

The previous positioning example began by using manufacturing andcumulativeleadtimesasfactorsindeterminingtherightpositionfordecoupling.Withintheexample,however,anewleadtimecanbeconceptualized,onethatcancreatetremendousopportunityformorecomplicatedmanufacturingentitiesandisarequirementforproperdecouplingpointbuffer-levelcalculations.Thesemore complex manufacturing entities have depth, breadth, and overlap withregard to their product structures that allow for this opportunity. The word“depth”impliesmorelevelstoaproductstructure.“Breadth”impliesmorelegs

Page 95: DDMRP - Demand Driven Material Requirements Planning

totheproductstructure.And“overlap”impliessharedcomponentsormaterialsacrossproductstructures.

In order to demonstrate this new lead time,wewill take a single productstructureforaproductcalledFinishedProductD(FPD).ThatproductstructureisdepictedinFigure6-8.Thenumbersinthecirclesrepresentthemanufacturingleadtimesfrommakeitemsandpurchasingleadtimesforpurchaseditems.Forexample,themanufacturingleadtimeforpart208is5days.Thepurchasingleadtimefor412Pis45days.Notethatthisproductstructurehasfourlevels(depth)andtwomajorlegs(breadth)beginningwithcomponents208and210.Thereisno overlap, as we have not introduced any additional bills of material; thatcircumstancewillbecoveredindepthlaterinthischapter.

FIGURE6-8FPDproductstructurewithleadtimes

WhenperformingcalculationsformanufactureditemsinproductstructuresliketheoneseeninFigure6-8,conventionalMRPsystemsonlyrecognizetwoformsofleadtime:manufacturingleadtimeandcumulativeleadtime.Oncetheconceptofdecouplingisembraced,theseleadtimesbecomerealisticinextremesituationsonly.Inmorecomplexmanufacturingoperations,thesetwoextremesrarelyexistinreality.

InFigure6-9weseetheFPDproductstructurewithsomedecouplingpointsinserted.Thedecouplingpointsare representedby the stripedbucket icons. Inthis example FPD, 208, 311P, 410P, and 412P are selected as decoupled andstockeditems.Whythosepointswerechosenfordecouplingisimmaterialtothe

Page 96: DDMRP - Demand Driven Material Requirements Planning

discussion of problems associated with manufacturing and cumulative leadtimes.

Let’s examine the problem associated with manufacturing lead time firstgiventhecircumstancesdepictedinFigure6-9.Manufacturing lead time is thetime it takes tomanufacture the part exclusive of lower-level lead times.Thislead time assumes that all components will be available on the parent orderreleasedate.Thepathwayfromthedecouplingpointsat410Pand412Pthrough310and210tothecompletionofFPDisasequenceofdependenteventssubjectto the accumulation of variability. This variability is represented by theincreasinglylargewavylinemovingupthroughtheproductstructureinFigure6-10.Thatpathwayisanunprotectedsequenceofevents.Thustheassumptionthat210willbereadypreciselywhenneededforthereleaseoftheFPDorderissuspect at best. This makes the use of manufacturing lead time anunderestimationofthetimerequiredtoaccomplishtheproductionofFPD.

FIGURE6-9FPDproductstructure,leadtimes,anddecouplingpointpositions

Page 97: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-10VariabilitypassedintheFPDproductstructure

An alternative way to plan FPD in MRP would involve the use ofcumulative lead time. Cumulative lead time is the longest sequence in theproductstructuredefinedintime.Thisassumesthatnocomponentsareavailableuponorderrelease.InFigure6-11FPD’scumulativeleadtimeisdepictedasthebolded path terminating in either 410P or 412P.The length of that path is 52days.

FIGURE6-11FPA’scumulativeleadtimechain

Page 98: DDMRP - Demand Driven Material Requirements Planning

When the decoupling points are consideredwith FPA’s product structure,the immediate problem with using cumulative lead time can be observed. Inshort,cumulativeleadtimeassumesnodecoupling;yet410Pand412P(aswellas208and311P in theotherpath)aredecoupled.Since thesecomponentsaredecoupled, it can be reasonably assumed that they are available upon parentorder release. This fact makes the use of cumulative order lead time a grossoverestimationofleadtimewhendecouplingpointsarepresent.

Anewformofleadtimeisemergingwiththeuseofdecouplingpointsformanufacturedparts.Thisnewformofleadtimeonlyassumesavailabilityofthecomponentonparentorderreleaseatdecouplingpoints.Thisnewformofleadtimeiscalleddecoupledleadtime(DLT).Itcanbedefinedas:

The longest cumulative coupled lead time chain in a manufactured item’sproductstructure.Itisaformofcumulativeleadtimebutislimitedanddefinedbytheplacementofdecouplingpointswithinaproductstructure.

DLT is calculated by summing all themanufacturing and purchasing leadtimesinthatchain.Thedecoupledleadtimealwaysincludesthemanufacturinglead timeof theparent.Anyparent itemwith at least one coupled componentwillalwayshavealongerdecoupledleadtimethanitsmanufacturingleadtime.

Thedecoupled lead timepathforFPDisdepicted inFigure6-12. It is theboldedlarge-dashedpathconnectingFPDand310.Thelengthofthedecoupledleadtimechainissevendays.Itiscalculatedbyaddingthemanufacturingleadtimefor310(fourdays)tothemanufacturingleadtimefor210(twodays)tothemanufacturing lead time for FPD (one day). Figure 6-12 also depicts anotherimportant elementwhenusingdecouplingpoints.The intermediate component208(whichisadecoupledposition)nowhasitsowndecoupledleadtimechain.Thatpathisdepictedbytheboldedsmall-dashedpathconnecting208to401P.Thatdecoupledleadtimeis19days.

Page 99: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-12FPD’scompressedleadtimechainand208’sdecoupledleadtimechain

Figure6-13showsthedecoupledleadtimeforintermediatecomponent208being compressed from 19 days to 9 days by decoupling the purchased part401P.

Thedecoupledleadtimeissimplyaqualifiedcumulativeleadtimeconcept.Asconventionalmaterialplanningsystemswerenotdesignedtousedecouplingpoints, this form of lead time has remained hidden or at best obscured fordecades. By using the DLT information, planners can now determine morerealisticdates for thereplenishmentofapartand the inventory levels requiredfor the decoupling point buffers.Of course, using this approach requires eachdiscrete manufactured part number to have a manufacturing lead time orproduction lead time defined, and those lead times should be as accurate aspossible.

Page 100: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-13Decoupling401Pcompresses208’sdecoupledleadtime

Sofar thevalueofDLTisshownindeterminingmorerealistic lead timesandbufferlevelswiththeuseofdecouplingpoints.Nextlet’sturnourattentiontotheuseofdecoupledleadtimetohelpfindnewopportunitiesfordecouplingpointsaswellastheeliminationofnon-value-addedstockpositions.

AdvancedInventoryPositioningConsiderations

Understandingdecoupled lead timeopensadoor formoreadvanced inventorypositioning analyses for environments in which there are deeper and broadermaterial structures andwhere shared components exist across those structures.Manycompanieshavemanydifferentproducts,eachwithitsownuniquebillofmaterial. Many companies have a significant amount of shared parts orcomponents—places where the bills of material essentially overlap. This factcombinedwith the concept of decoupled lead time exposes opportunities thathaveremainedelusiveformostofthesecompanies.

Sucha company is illustrated inFigure6-14.CompanyABCmakes threeseparateenditems:FPA,FPB,andFPC.Theseitemseachhaveauniquebillofmaterial, and there are also shared components. Note that all parents arecurrentlystockedaswellassomeofthepurchasedparts.

Finding the right opportunities for additional decoupling point placement,however,requiresanoldtoolthathasbeenaroundnearlyaslongasMRP.Thisconceptiscalledamatrixbillofmaterial.AccordingtoAPICS,amatrixbillofmaterialis:

Page 101: DDMRP - Demand Driven Material Requirements Planning

A chartmade up from the bills ofmaterial for a number of products in thesameorsimilarfamilies.Itisarrangedinamatrixwithcomponentsincolumnsand parents in rows (or vice versa) so that requirements for commoncomponentscanbesummarizedconveniently.(pp.103–104)

FIGURE6-14CompanyABCenvironmentwithenditemparentsthatsharecomponents

Figure6-15isthematrixbillofmaterialforthethreeproductsillustratedinFigure6-14.Parent items are displayed along the top (columnheaders),whilecomponentsaredisplayedalongtheside(rowheaders).Notethatacomponentwilloftenappearasbothachildandaparent.Component201isanexampleofthis. It is a child to FPA and intermediate components 101 and 102 but is aparent to intermediate component 301 and purchased part 302P. The shadedparents(FPA,FPB,andFPC)andcomponents(302P,305P,402P,403P,404P,410P, 411P,) represent parts that are currently decoupled or stocked.Furthermorethenumberswithinthegridarenotthequantityperparent.Thosenumbersonlyrepresent thenumberof times thataspecificconnectionappearsacrossallbillsofmaterial.Forexample,theconnectionbetweenthe201parentand301componentoccurs inall threebillsofmaterial in thisenvironment,sothereisanumber3inthatbox.

Clearly,thisisaverysimpleexample.ForcompaniesthathavehundredsofenditemswithdeepBOMsandmanysharedcomponents,thematrixBOMcangetquitecomplexandverylarge.Thisfactalonemeantthetoolwasneverreallyinwidespreaduse,particularlyinearlierdayswithlimitedcomputingpower.

Amatrix BOM is a much broader picture than a where-used report. Thewhere-usedreport isorientedtoaparticularcomponent toseewhichparents itgoesintoanditsusageperparent.TheAPICSDictionarydefinesawhere-usedlistas:

Page 102: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-15InitialmatrixBOMforCompanyABC

A listing of every parent item that calls for a given component, and therespectivequantityrequired,fromabillofmaterialfile.(p.190)

Awhere-usedlistisonlyaveryspecificsliceofamatrixBOM.Thematrixbillofmaterialshowsallconnectionsbetweenallparentsandallcomponentsinan environment. What the matrix bill of material does not show is whichconnectionsreallymatterthemost.Togainvisibilityofthoseplacesthematrixbill of material and the concept of decoupled lead time must be used incombination.Figure6-16showsthethreeproductswiththedecoupledleadtimechainsbolded.ThedecoupledleadtimechainforFPAis20daysterminatingat401P.Thedecoupled lead timechain forFPB is23days terminating in401P.ThedecoupledleadtimechainforFPCisalso23daysterminatingat401P.

Immediately we can begin to conceptualize the value provided by thisperspective. Since the end items’ parents (FPA, FPB and FPC) are stocked,consideringadditionaldecouplingpositionsontheDLTchainwillimmediatelycompress the parent’s lead times. This means that the level of each parent’sbufferwill bepositively affected.Themore the compressionof lead time, themorethecompressionofbufferstock.ThisrelationshipwillbeexploredindepthinChapter7.

If theparent itemswerenotstocked, thevalueofdecouplingtheleadtimechainscanstillbeconceptualized.Withoutstockingtheparentparts,FPA,FPB,

Page 103: DDMRP - Demand Driven Material Requirements Planning

andFPCwouldhaveleadtimesof20,23,and23days,respectively.DecouplingontheirDLTchainswillcauseadirectcompressionoftheleadtimethatcanbeofferedtothemarket.Anycompressiontoanonstockedpart’sleadtimemustbeevaluatedagainstthecustomertolerancetimeandmarketpotentialleadtimes.

Thelackofvalueprovidedbydecouplingcomponentandpartpositionsthatdo not lie on the decoupled lead time chain also becomes clear. For example,parts203,207, and202donot lieon thedecoupled lead timechainsofFPA,FPB,andFPC,respectively.Thusdecouplingandbufferingthosepositionswillprovidenobenefittotheparentandwouldrequireaninfusionofworkingcapitaltofundthebufferstockanddelivernotangiblebenefit.

Figure6-17showsthematrixbillofmaterialforthisenvironmentwiththedecoupled lead timechainshighlighted.Within thegrid,a shadedboxdenotesthattheconnectionslieonadecoupledleadtimechain.Forexample,component201liesonallthreedecoupledleadtimechains.

FIGURE6-16Decoupledleadtimechainsdefined

Page 104: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-17InitialmatrixBOMwithDLTchainshighlightedatCompanyABC

Two questions arise when looking at this view. First, where should weconsider decoupling on the DLT chains? Second, how can we effectivelyevaluate the potential financial impact of decoupling components on theDLTchains?

In answering the first question, it is possible to narrow it down to threecomponents.Thematrixbillofmaterial shows that components201,301, and401Pareon thedecoupled lead timechainsof allparents.Thusdecouplingatanyof thesepointswillhaveat leastsomevaluetoallparents.Butwhichwillprovidethemostvalue?

Another relevant factor requires an understanding of one of the criticalpositioningfactorsdefinedpreviouslyinthischapter—customertolerancetime.Whatwedonotknowyetaboutthisexampleisthecustomertolerancetimesforeachenditem.Whensalesorcustomerserviceisaskedwhatthemarketexpectsinturnaroundtime,theansweristhreedaysforallitems.Thatmeansthatwhenan item isordered, itmustbe shipped to the customerwithin threedays.Thisbringsanewperspectivetotheanalysis.Wemustdeviseapositioningstrategythat allows us to meet the customer tolerance times for each product whileeffectivelyleveragingandminimizinginventoryinvestment.

Whenexaminingthematrixbillofmaterial,animmediatequestionmustbeasked:Isitpossibletocompressanyparent’sleadtimetothepointthatitdoes

Page 105: DDMRP - Demand Driven Material Requirements Planning

nothavetobestocked?Iftheanswerisyes,thenwewouldbeabletopotentiallymovethoseparents toanassemble-to-ordersituation,allowingthecompanytostockcommoncomponentsandletthemflowtorequiredparentsasneeded.Thiscouldpotentiallyprovidethebestleverageforthosecommoncomponentswhileeliminatingtheneedforfinishedstockforsomeenditems.

FIGURE6-18FPA’sshiftingdecoupledleadtimechain

One end itemdoes provide the potential to build a positioning strategy inwhich it its finished goods stock can be eliminated. When examining FPA’sproduct structure, we can see that buffering 201 will decouple a significantamountofleadtime(19days).When201isdecoupled,however,thedecoupledleadtimechainshiftstoapathterminatingin303.ThisisillustratedinFigure6-18.

This isone important lessonaboutdecoupled lead timechains.These leadtime chains can be very dynamic based on decoupling position decisions. Forenvironments with broader and deeper product structures, the shift will oftencause an entirely different leg of the product structure to lie on the decoupledleadtimechain.ThisisthescenariowithFPAwhendecoupling201.ThisshiftisnowvisibleinthenewmatrixbillofmaterialseeninFigure6-19.

Decoupling 201 also has implications for FPB and FPC. There is animmediatecompressionofleadtimeandshifttotheirrespectivedecoupledleadtimechains.Figure6-20depictstheshiftsacrossallparentitemsaccomplishedby buffering 201. FPB’s decoupled lead time is reduced to 9 days (from 23days),andFPC’sdecoupledleadtimeisreducedto8days(from23days).

Decoupling 201 does not yet achieve the objective of reducing FPA’s

Page 106: DDMRP - Demand Driven Material Requirements Planning

decoupled lead timewithin the customer tolerance time.Another stepwill berequired. Now that the FPA’s decoupled lead time chain has shifted, wewillneedtoconsideranotherdecouplingpointat203.Decoupling203wouldallowtheleadtimeofFPAtodroptooneday—wellwithincustomertolerancetime.Figure6-21depicts thenewdecoupled lead timechainforFPAwithboth201and203decoupled.Notethat203nowhasitsowndecoupledleadtimechain,asithasatleastonecoupledcomponent.

Figure 6-22 is an updated matrix bill of material showing the impact ofdecoupling203.TheshadingofFPAhasbeen removed, indicating it isnowanon-decoupledpart.There areonly two remainingdecouplingpoints availablethatimpactallthreeproductstructures:components301and401P.Inrealitytheimpact of the two components is limited to only one product structure’sdecoupledleadtime—theproductstructurefor201.Butsince201isinvolvedonall three end item product structures, the matrix bill shows 301 and 401P asimpactingallthreeonadecoupledleadtimechain.

FIGURE6-19Theupdatedmatrixbillofmaterialwith201decoupled

Page 107: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-20FPA,FPB,andFPCnewDLTwith201bufferinplace

FIGURE6-21FPAwithaone-dayleadtime

Page 108: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-22Updatedmatrixbillofmaterialwith203decoupled

Component 401P would seem to provide the best candidate since it willdecouple external variability from the environment. Figure 6-23 shows theimpactto201ofbuffering401P.Itsdecoupledleadtimedropsfrom19daysto9days.

FIGURE6-23201’snewdecoupledleadtime

FIGURE6-24Matrixbillofmaterialwith401Pnowbuffered

Figure6-24showsthematrixbillofmaterialafter401Phasbeenbuffered.It is evident that 301 might still provide an opportunity for additional

Page 109: DDMRP - Demand Driven Material Requirements Planning

compression to 201. That compression is minimal and would still require astocking commitment. Furthermore, we can see that the remaining stockedparents (FPB andFPC) have nothing in common to compress their respectivedecoupledleadtimes.

Figure6-25depictsallthebufferedpositionsafterthispositioningexercise.Now there is the opportunity to eliminate a finished stock position (FPA) anddramaticallycompressthedecoupledleadtimesforFPBandFPCaswellasthecommoncomponent201.Whatisnotansweredyetisthefinancialimpacttothebusinessofmakingthesemoves.Thiswillrequireanunderstandingofhowthedecoupling point buffers are sized and the amount of averageworking capitalthatwill be contained in them. Thiswill be covered inChapter 7,where thisexamplewillcontinue.

FIGURE6-25ThefinalbufferpositionsforCompanyABC

DistributionPositioningConsiderations

One final aspect of positioningmust be explored. So farwe have focused ondecoupling point positions within a manufacturing environment where rawmaterials are converted into sellable products. In distribution there is noconversion. Distribution is about aligning finished product to best meetconsumption.Inmostlargecompanieswithbothmanufacturinganddistributionaspects, there is constant tension amongplanning,manufacturing, distribution,sales,andlogistics.

Mostdistributionnetworkshaveregionalorlocalwarehousesholdingstock.These locations are constantly attempting to balance between the criticalrequirementtohavewhatthemarketrequireswithinthetimeframeitrequiresit(usually instantly)and theneed to turnorconvert inventory intocashorprofit(not have toomuch inventory).To understand decoupling point positioning in

Page 110: DDMRP - Demand Driven Material Requirements Planning

the context of distribution, an appreciation for what must be decoupled isnecessary.

Let’sfirstconsiderdemandvariability.Makingtherightdecisions(whetherpredictionsoractions)isinherentlyeasierwhenthefactorsbeingconsideredaremore stable or known. Usually the biggest form of instability for distributionnetworksisdemandvariability.Forwardlocationsinadistributionnetworkcanoftenbeseenasthefrontlineinthebattleagainstdemandvariability.Ifactionsare not taken to mitigate this form of variability, shortages, expedites, andinventory imbalancewill occur (distortions to the flow of relevantmaterials),andthevisibilitytorelevantinformationwillbeobscured.

Figure 6-26 depicts a simple but typical distribution environment. Asourcing unit or manufacturing plant is feeding a network of four regionalwarehouses. Each regional warehouse experiences a specific level of demandvariability for any specific distributed item. The demand variability is muchhigherateachofthesediscretedistributionlocationsthanatthesourcingunitforthe same time period. The law of total variance means that aggregating thedemandvariabilityfromtheremotelocationscreatesanaturalsmoothingeffectatthesourcingunit.

FIGURE6-26Atypicaldistributionnetwork

A simple experiment will prove this key point about demand variabilitysmoothing asmore events are aggregated. For eachwarehouse location a fairdiceisrolledtosimulatedailydemand.Twodiewereusedforwarehouses1and

Page 111: DDMRP - Demand Driven Material Requirements Planning

3,whileasingledicewasusedforwarehouses2and4.Thedifferentnumberofdie being used is meant to bring some additional reality to the scenario bysimulating warehouses that might do higher levels of volume for a specificproduct. Figure 6-27 illustrates the results of the experiment. Each day thesourcingunit’sdemandisthesummationofallwarehousedemand.Thecolumnlabeled“ADU”istheaveragedailydemandoverthe14-daytimeframeateachlocation.

Now when the results for each location are charted, a picture similar toFigure6-27results.Figure6-28 charts thedaily demand for all locations.Thesmoothingeffectcanbeeasilyobserved.

This smoothingeffect isnot justobservable throughsimple linegraphs. Itcanbecalculatedmathematicallybyusing thecoefficientofvariance formula,which calculates the normalized measure of dispersion of a distribution. Theequationforthecoefficientofvariationisthestandarddeviationdividedbythemean. It is also known as relative standard deviation. Figure 6-29 depicts thecoefficientofvariationforeachlocationintheexperiment.Thecolumnlabeled“ADU”servesas themean for each location.Thecolumn labeled“SD” is thestandard deviation at each location. Finally, the column labeled “CV” is thecoefficientofvariation.

Thismakesacompellingcasethatthebestplaceinadistributionnetworktomitigateandmanagedemandvariabilityisatapointofaggregationwherethereislessinherentrelativevolatility.Yetthismathematicalfactseemstobelostonthepeopleandorganizationsrunningthevastmajorityofdistributionnetworks.Manydistributionnetworks aredesigned andmanaged in away that prohibitsthemfromtakingadvantageofthisconcept.

FIGURE6-27Resultsofthedistributiondiceexperiment

Page 112: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-28Dailydemandchartedforalllocations

FIGURE6-29Thecoefficientofvariationateachlocation

Figure6-30depictsthestructureofmostdistributionnetworksinwhichthevastmajorityofinventoryispushedoutclosesttothepointofconsumption.Theregionalwarehouseshavedifferentlevelsofinventoryofthefiveproductsthataredistributed.Whydoesthispusheffectoccur?Thereareseveralpredominantassumptionsorconditionsbehindthepushincluding:

1. An attempt to optimize freight costs. Many distribution networks’primary metric is transportation spend efficiency. Attempting tominimizethetransportationcostperunitoftenleadstolargeshipmentquantities (full trucks)within thenetworkcontainingproduct that isnotreallyrequiredattheforwardlocation.

2. An attempt to optimize sourcing unit costs. The sourcing unit istypicallymeasuredonsome formofunitizedcost.Smaller runsandadditionalsetupsdirectlyharmperformanceonthosemetrics.Biggerbatches are the rule, and that inventory has to go somewhere sincerarelyistherespaceatthesourcingunittostoreit.

Page 113: DDMRP - Demand Driven Material Requirements Planning

3. Location assumption. Many in an organization, especially in sales,believethatlocatingthemajorityofinventoryclosest tothepointofconsumptionhasthegreatestpotentialtomeetalldemand.

4.Ascarcitymindset.Whencapacityorsupplyisperceivedtobescarce,overorderingwill often occur by regionalwarehouses attempting toprotecttheiraccesstoproduct.

5. Space limitations at the sourcing unit. If the sourcing unit is notcapableofstoringinventory,itmustbesentouttothenetwork.

OneobviousconclusionthatFigure6-30revealsisthatwithoutanystockatthesourcingunit,thereisnowaytodecouplethesourcingunit’sleadtimefromthe transportation lead time to the warehouse. That means that the regionalwarehouses, theplaces thatexperience thehighest levelofdemandvariability,mustaccountforamuchlongerleadtimeaswellasplantproductionvariabilitywhenordering.Tosimplify,thereisamuchlongerandmorevariableleadtimegoingtoamuchmorevariablepointofdemand.

Thisisarecipeforpoorperformanceunderanysetofcircumstances.Whattobeorderedandwhentoorderitbecomesaguessinggame,asthislongerleadtime forces a longer planning horizon to be used. Additionally, the plant’scapacityissubjecttothevariabilityofdemandfromthedistributioncenters.

FIGURE6-30Inventoryplacementinatypicaldistributionsituation

Page 114: DDMRP - Demand Driven Material Requirements Planning

This situation often creates circumstances in which one region inevitablydoes not have enough, whereas others have too much. The result is cross-shipmentsbetweenthedistributioncenters,missedpotentialsales,andexpeditesplacedbackintotheplant’smanufacturingschedule.Astheseeffectsoccur,theycompromisemanyoftheassumptionsbehindtheoriginalnetworkdesign:

1. Freight costs are far from optimized as cross-shipping or“rebalancing” causes additional and often expensive transportationspend.

2.Frequentbreak-instotheschedulecausedbyemergencyorderswreakhavoconthesourcingunit’sscheduleandoperatingmetricsandfuelthebullwhipeffect.

3.Shortagesmeanmissedsalesopportunities.4.Shortages,longleadtimes,andthesourcingunit’slackofagilitycanreinforcethescarcitymindset,leadingtofurtherdistortiontoaccuratedemandsignals.

Intheaggregate,thesystemoftenhasenoughinventory;itisjustlocatedatthewrongplace.Iftheinventorywerebetteralignedinthefirstplace,itwouldallowfor:

BetterdemandcoverageforallpointsofconsumptionTheminimizationoreliminationofcross-shipping The removal of the sourcing unit’s lead time and variability fromregionalstock-levelconsiderations

Minimizationof thesourcingunit scheduledisruptions that reduceavailablecapacityandcomplicateplanningandmetrics

But how to achieve this better alignment? The answer is mathematicallyobvious. This better alignment is created by decoupling at the point ofaggregation that we saw in the previous experiment in Figure 6-29. Thedecoupling “hub” should be created as close to the sourcing unit as possible.Holding inventory closer to the source actually protects the largest portion ofpotentialconsumptionfortheleastamountofinventory.AdistributionnetworkisshapedlikeaV.InFigure6-30,onceinventoryispushedouttoRegion1,itisunavailabletotheotherregionsintheimmediateterm.Inthelongerterm,itcanbemadeavailable,butonlythroughcostlyrealignmentactivity.

Page 115: DDMRP - Demand Driven Material Requirements Planning

Realigningtheinventorytolocatethemajorityofitat thisdecouplinghubwill accomplish the desired benefits. Figure 6-31 is a realigned distributionnetworkshowinganinventoryhublocatedatthesourcingunitandsmallpointsofinventoryattheregionalwarehouses—the“spokes”ofthenetwork.

Theinsertionofthisdecouplinghuborcentralbufferhasmanybenefits:

1.Itprotectsregionallocationsbyensuringareliablepipelineofsupplydefinedbytransportationtimeonlyasopposedtoplantleadtimeplustransportationtime.Thismaximizesavailabilitywiththeleastamountoftotalinventoryrequirement.

2.Thehubessentiallyeliminatescross-shipments.Whycross-shipwhenyou can simply ship from the hub? Additionally, a strategy can beemployedtoshipbetweenhubandspokethatcanensureefficientuseof freight resources. This strategy, called prioritized share, will becoveredindepthinChapter9.

FIGURE6-31Decoupleddistributionnetwork

3.Theplanninghorizonisdramaticallycompressedatthemostvariabledemandpoint(theregion),resultinginmorerelevantdemandsignalsandaminimizationofthebullwhipeffect.

4.Thehubalsoallowsconsumptionandcorrespondingresupplysignalsto the plant to be naturally consolidated into batches that are still

Page 116: DDMRP - Demand Driven Material Requirements Planning

sensiblefrombothaplantcapacityandacostperspective.5.Thedecouplinghubalsoallowsthemanufacturingfacilitytoscheduleclose in time to theactual centralbuffer requirements rather than toarbitrary frozen schedule horizons that further limit flexibility andcreatefrustrationforsales

Yet this idealconfigurationseen inFigure6-31maynotbeachievable formanydistributionnetworks.Forexample,awholesaledistributorwillprobablynotbeabletoforceallitssupplierstoholdcentralstocksavailableondemandtotheir regional locations. Additionally, space restrictions may exist at sourcingunits that will prohibit implementation of a full decoupling hub. Differentoptionswill need to be explored to accomplish the samedecoupling hub-and-spokemodelgivendifferentcircumstances.

First,let’sexplorethecaseofawholesaledistributor.Wholesaledistributorsbuyandreceiveshipments frommanydifferentsourcingunits,manyofwhichmight be remote to the distributor’s facilities or each other. Figure 6-32illustrates justsuchasituation.Thisnetworkhas fourdistributionpoints.Eachpoint receives separate shipments fromsuppliers.Obviously this canpresent ahugechallengewith regard to supplierminimumorderquantities, freight cost,andspaceutilization.

FIGURE6-32Atypicalwholesaledistributor’snetwork

Page 117: DDMRP - Demand Driven Material Requirements Planning

As seen before, longer lead times and supply variability are going toinherently more variable points. Furthermore, the minimum order quantitiesimposedbysupplierscanbemuchlargerthantheindividualdistributionpoint’simmediaterequirements.Thiscancausedelaysinorderingthatleaddirectlytoshortages,oritcancauseextremeexcessinventorypositionsasthewarehouseshave to take in months of supply to meet the minimum order size. Spacebecomeslimitedandfurtherrestrictsorderingcapabilities.

One option is that every distribution network with more than onedistribution point can convert to the hub-and-spoke model in order to takeadvantage of its benefits. Moving to a hub-and-spoke configuration is not atrivialchange inmostcases. In theshort term, itmayrequire thecreationofahub through either an additional facility or the conversion of an existingwarehouse.Figure6-33showsthewholesaledistributionnetworkseeninFigure6-32 converted to a hub-and-spoke configuration. In this case one of thewarehouses(Region4)hasbeenconvertedintothehub.

Page 118: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-33Awholesaledistributor’snetworkwithahub-and-spokeconfiguration

Someconsiderationsinconvertingoneregiontoahubare:

1.Demandvolume.Ifthispositionhasthelargestcustomerandvolumebase, then the stock in the central hubwill onlybemovedonce forthiscustomerbase.

2. Proximity to suppliers. If this location is in close proximity to itssuppliers, the central storesmight beminimized due to shorter leadtimesandlessvariability.

3.Proximitytoqualitytransportationlanes.Ifthislocationliesclosetomajor transportation lanes, then favorable freight rates (for hub-to-spokeruns)mightbenegotiated.

4.Availablespace.Ahubwilltendtoholdthemajorityofinventory.Ifalocationhasalargerphysicalfootprintrelativetotheotherlocations,thenitmightmakeabetterhubcandidate.

Inadditiontothepreviouslydescribedbenefitsofthehubandspoke,therearesomeadditionalbenefitswithregardtothisexample:

1. External variability ismitigated. The central hub decouples supplierlead times, variability, and minimum order quantity requirementsfrom interfering with the ability of the spoke to service its uniquecustomerbaseanddemandvariability.

2.Spaceisbetterutilized.Asmentioned in thefirstbenefit, thehubatRegion 4 protects Regions 1, 2, and 3 from supplier lead times,variability,andminimumorderquantity requirements.Nowthe leadtimeandvariabilityare isolated to the transportation timeandorderfrequency between the hub and spoke. This will dramaticallycompressthetotalamountofspaceneededatthespokes.

3. Purchasing and in-bound freight canbe better leveraged.Withoneposition ordering against a single large stocking point, ordering isconsolidatedforpotentialvolumeandfreightdiscounts.

4.Ahubisbetterabletofulfilllarge-quantityorders.Largeorderscanbefulfilled and shipped directly from the larger hub position. Thisposition, by being larger, can better absorb these spikes withoutexpedites,backorders,andcrossships.

Page 119: DDMRP - Demand Driven Material Requirements Planning

Most detractors of this configuration will claim that this requiresmovinginventorytwiceforsalesinRegions1,2,and3.ButinthetypicalconfigurationinFigure6-32,theprevailingcross-shipsalsorepresent“doublehandling.”Yetthepointisstillvalid,asthemodelessentiallycallstomovethisinventorytwice.Thismustbeconsideredagainstbetteravailabilityanddiscountpotentialandthedecreasedworkingcapital,space,andcross-shiprequirements.Additionally,asnotedearlier,aconceptcalledprioritizedshareisexploredinChapter9thatwillhelpmakethemostofthefreightspendbetweenthehubandspokes.

Let’s expand this concept to a geographically larger distribution network.Eachregionalwarehousehasseveralsuppliers thatarelocatedrelativelycloserto it than other regional ware-houses. Figure 6-34 shows this distributionnetwork.Inthiscase,eachregionalwarehouseisreceivingshipmentsfromeachsupplier.The suppliers are representedby the factorywith a letter designation(A–L). There are 12 primary suppliers for the network, and the productsproduced by each are available in all regional locations. For example, thesuppliers of products A, B, and C are in relative proximity to the Region 2warehouse.Thecircleswith“C” in themaresimplymeant to show their localcustomerbase.

To apply the hub-and-spoke configuration to this network requires the“multi-hub”concept.AsFigure6-35shows,themulti-hubconfigurationallowseachwarehouse tobebothahubandaspoke.Eachwarehousewillserveasahubfor theproductsproducedin itsrelativegeographicproximityandaspokefortheproductsproducedinrelativeproximitytootherregionalwarehouses.Forexample,Region2willbethehubfortheproductsproducedbysuppliersA,B,andC,anditwillserveasaspokefortheproductsproducedbytheotherninesuppliers.

Page 120: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-34Supplierproximitytodifferentregionalwarehouses

Page 121: DDMRP - Demand Driven Material Requirements Planning

FIGURE6-35Multi-hubconfiguration

This network design brings with it all the benefits of the hub-and-spokeconfiguration plus one additional transportation-related benefit. Thetransportation lanes between each warehouse are now bidirectional. Forexample, trucks leave Region 3 full of G, H, and I products and return fromRegion1withK,J,andLproducts.Thisisatransportationsystemwithabuilt-inbackhaulsituationthatallowsacompanytonegotiatemorefavorablefreightratesormoreefficientlyuseitsowntrucks.

Oneadditionalconfigurationshouldbeconsidered.Itisessentiallyapartialhub-and-spoke and is known as the “hybrid.” The hybrid model can be usedwhenthereislimitedspaceatasourcingunit,disallowingthedeploymentofafullhubatthatlocation.Thehybridmodelfocusesondecouplingthevariabilitybetween the sourcing unit and the distribution network associated with slow-movingitems.Slow-movingitemsareofparticularconcernsincetheirminimumquantity requirements in relation to their usage rates often create significantimbalanceinthenetworkandschedulingdifficultiesfortheplant.

Figure 6-36 illustrates this hybrid approach. By establishing a hub forslower-moving items, the distribution network has a steady supply of slow-

Page 122: DDMRP - Demand Driven Material Requirements Planning

movingitemswithoutthespacepenaltiesofstoringthoseitemsforlongperiodsofsupplyattheregionallevel.Additionally,thehuballowsthesourcingunittorun a more consistent order cycle of faster-moving items while occasionallyscheduling reasonable batches of slow-moving items when the hub actuallyneedsresupply.Thefaster-movingitems(products1,2,and3)aresentouttothenetworkastheyareproducedatthesourcingunit.Loadsaresupplementedwithrequiredquantitiesofslower-movingitems(products4and5)foreachregion.InFigure 6-36 a manufacturing run of product 1 has just been produced by thesourcing unit and is being sent out to each of the regional warehouses. Eachtruckadditionallyhasquantitiesofproducts4and5foreachwarehouse.

FIGURE6-36Thehybridconfiguration

The hybrid is a compromise and brings with it some additionalconsiderations.

One question that arises is, How should slower-moving items bedetermined?Slowmoverswillbeidentifiedthroughtheuseofa“flowindex.”TheconceptoftheflowindexiscoveredinChapter12.

Other questions include: How should slow movers be fit into the

Page 123: DDMRP - Demand Driven Material Requirements Planning

manufacturing schedule? What if a particular fast-moving production run isinsufficient to deal with all immediate network demand?What if a particularfast-moving production run is above the immediate network demand?” Thesequestionswillbeansweredafter theactualsupplyordergenerationmechanismofDDMRPisexplainedinChapter9.

Summary

This chapter has provided an in-depth exploration of the first component ofDemand Driven Material Requirements Planning—inventory positioning.Inventory positioning is about answering themost primary question related topromoting and protecting the flowof relevant information andmaterials.Thatquestionis,Wheretodecouple?Asseeninthischapter,answeringthisprimaryquestion is a strategicprocesswithconsiderations that impact thebreadthofasupplychain.Thusplannersandbuyerscannotanswerthequestionalone.Thisis a cross-functional and strategic effort that should be discussed withrepresentatives from most aspects of the organization (operations, logistics,sales,purchasing,andfinance).ThisprocessisfurtherdescribedinChapter13inthesection“DemandDrivenSalesandOperationsPlanning.”

Page 124: DDMRP - Demand Driven Material Requirements Planning

CHAPTER7

StrategicBuffers

ThesecondstepofDDMRPisthemechanismthatallowsadecouplingpointtostaydecoupled—abuffer.Inthiscasethebufferwillbealevelofstockthatiscarefullysizedandmaintained.Thischapterfocusesonthesizingconsiderationsofstrategicinventorybuffers.First,acriticalquestionaboutinventorymustbeconsidered.

Inventory:AssetorLiability?

In order to better understand how to determine the protection levels ofdecouplingpositions,firstaquestion:Isinventoryanassetoraliability?Thereseemtobetwoprevailingandconfusinganswers.

Accordingtothecompanybalancesheet,inventoryisanasset.Fordecadesmany large companies have played paper games with regard to inventory.Despite having no demand, many companies continued to build inventory,realize the accounting value-add from that inventory, and declare accountingprofits against it. In the process, companies are drained of cash and may godeeply intodebt,butaccording togenerallyacceptedaccountingprinciples thecompanywasprofitable.

Today,withtheproliferationofmethodologiessuchasLeanandtheTheoryof Constraints, in addition to the global economic meltdown of 2008–2010,fewercompaniescanaffordtoplaythesegames.Cashisanimportantfocus,andWallStreetalsohasbecomeawareoftheruseandthepenaltiesassociatedwithtoomuchinventory.Buthasthependulumswungtoofartheotherdirection?

Edictstoslashinventoriescanbeextremelyharmful.Forexample,let’stakethe caseof adirectorofpurchasingwho is given an edict todramatically andimmediatelyreduceinventoryonpurchaseditems.Whenherstaffdoesthefullanalysison the inventory, shediscovers that thevastmajorityof thematerials

Page 125: DDMRP - Demand Driven Material Requirements Planning

andpackaginginventoriesareonslow-movingorobsoleteitems.Withregardtopurchasedmaterials, large-quantity buysweremade at significant discounts todriveapositivepurchasepricevariancebuthave resulted inbloatedpositions.Furthermore,toexacerbatethesituation,engineeringmadeasignificantmaterialrevision tomany of the higher-moving products, further eroding the usage ofseveralmaterialswithhighstockpositions.Attheircurrentrate,thesematerialswillnotdrainoutforoveroneyear.Thesituationwithpackaginginventoriesisevenworse.Marketing,withoutmuchnotification,made significantpackagingchanges in order to support a rebranding effort. Now the old packaging isessentiallyuseless,butfinanceisbalkingatwritingitoffdueto the impactontheprofitandlossstatement.

What can this director do when most of the inventory is not moving orunusable yet is under direct orders to reduce inventory immediately?There isreallyonlyoneoptiontoreduceinventoryinaveryshortperiodoftime.Thatisto cancel or defer open supply orders on thematerials and packaging that areactuallymoving. In just a few shortweeks thismanagerwillmake averybigdentininventoryvalue.

Will this director be praised? Stocks of critical, fast-moving items willdeplete rapidly. Shortages in materials and packaging will block or delay themanufacturingschedule.Servicelevelswillbeadverselyimpacted.Materialandpackagingexpediteswillrisedramatically.Flowwillbereducedtoatrickle,andoperationsandsaleswillscreamloudly.

Furthermore, metrics that focus on inventory turnover can be extremelydistortive.Oneofthemostimportantperformancemeasuresoftheoverallhealthof amanufacturing business is thought to be inventory turnover ratios. In theUnited States before 1980, these ranged from below one to as high as six.Management thought itwasdoingwell to increase the figureby50percent inone year.Manywho did so simply fell back to previous ratios the next year,indicating successful crisis management and actions but no permanent orsustainableimprovementinperformance.Manyfoundthatloweringinventoriesharmedcustomerservice,whichimpactedrevenueandcausedhighercosts.

Withtheadventofaleaneremphasis,inventoryturnscommonlyincrease—sometimesevendramatically.But is inventory turnoveralwaysan indicatorofexcellentoroptimalinventoryperformance?Highturnsandhighshortagescanandoftendocoincide.Thishappens frequentlywhencompanies leanout theirinventorytoomuchandessentiallymaketheirsupplychainstoobrittle.Theyarenotimproving,theyareputtingthemselvesatacompetitivedisadvantage.

Page 126: DDMRP - Demand Driven Material Requirements Planning

Thus we can conclude that true agility is not synonymous with zeroinventories.Thekey toeffectively leveraging theworkingcapitalandcapacitycommitment inherent in inventory is to find the places where inventory canmake the biggest positive impact and therefore provide the greatest return.Inventory can decouple otherwise dependent events so that the cumulativeeffectsofvariationarenotpassedoramplifiedbetweenthedependencies.Thusinventory can be a break wall against the variability experienced from eithersupply (externallyand internally)ordemandvariability.However,aswithanybreakwall,itiseffectiveonlyifitisplacedandsizedproperly.

There are two prerequisites for inventory to be a true asset to a business:placementandsizing.Fundamentally, theword“asset”hasaspecificmeaning.Businesses have expectations from the assets on their balance sheet; inparticular, theyexpecta returnon thoseassets. Inventory shouldbe treatednodifferently.Yethowcanwecalculatearateofreturnonsomethingwhensimplyhaving more of it distorts the financial picture or conversely slashing it anddrivingittoolowcompromisesourabilitytodecouplevariationandrespondtocustomerdemands?

Perhaps the value of these critical inventory placements can be calculatedbasedonsomethingthatweknowdirectlyconnectstoreturnoninvestment—theflow of relevant information and materials. As noted in Chapters 1–3,conventional planning systems typically result in a bimodal distribution inrelationtoinventorylevels.Manypartsareoverstocked,whileatthesametimesome are understocked. Whether overstocked or understocked, there is abreakdown in the flow of relevant information andmaterials. Thismeans thatwithregardto inventory, there isarangeinwhichinventory(assumingit is inthe right position) is truly an asset, andwhen outside of that range, inventorytrulybecomesaliability.

Whenacompanyhastoomuchinventory(overages),weknowthatexcesscash,capacity,materials,andspacearerequired.Obsolescencerisksarehigher.Discounts to liquidate stock cause losses and potentially cannibalize otherhigher-margin sales. Additionally, work-in-process levels might be higher,expanding lead times beyond the customer tolerance time and hurting sales.Fromaflowperspectivethisiscertainlylessthanoptimal.

Whenacompanyhastoolittleinventory,weknowthatchronicandfrequentshortages prevail, resulting in scheduling delays, missed sales opportunities,costlyexpedites,additionalfreightaspartialshipmentsaremade,andovertimeemployed. Once again, from a flow perspective this is certainly less than

Page 127: DDMRP - Demand Driven Material Requirements Planning

optimal.Thismeansthatinventoryisanassetsomewherebetweenthesetwopoints

—anominalpoint.Furthermore,itmeansthatthereisalossfunctionthatoccursineitherdirectionfromthenominal.Figure7-1 illustrates this lossofvalueaswemovetowardtheextremesoftoolittleortoomuchandoutsideofanoptimalrange.

Ifwe can find a simpleway to calculate this optimal range, thenwe canjudge inventory performance against it. If inventory positions are frequentlyoutside this range in either direction, then we know that there are potentialimprovementstobemade.Additionally,overthecourseoftimewecanchangetheparametersof the range through improvement activities (eithermaking therange smaller and/or shifting it to the left). In order to begin this journey,wemust explore how to calculate the size of the necessary protection at adecouplingpoint.

FIGURE7-1Theinventoryvalueloss(Taguchi)functionillustrated

IntroducingDecouplingPointBuffers

Theprotectionatthedecouplingpointiscalledabuffer.BuffersaretheheartofaDDMRPsystemandservethreeprimarypurposes:

Shockabsorption.Dampeningbothsupplyanddemandvariabilitysignificantly reduces or eliminates the transfer of variability thatcreates nervousness and the bullwhip effect. This was covered inChapters4and6.

Leadtimecompression.Bydecouplingsupplierleadtimesfromtheconsumptionsideofthebuffer,leadtimesareinstantlycompressed.ThiswascoveredinChapter6.

Page 128: DDMRP - Demand Driven Material Requirements Planning

Supplyordergeneration.All relevant demand information, supplyinformation,andon-handinformationarecombinedatthebuffertoproduce a “net flow” equation that determines supply ordergeneration. The buffers are the heart of the planning system inDDMRP.ThisiscoveredinChapter9.

The word “buffer” implies something substantive enough to be able toaccomplish thesepurposesyetnot too large soas to impede flow.Thismeansthereneedstobeapracticalwaytocalculatewhatthelevelofprotectionshouldbewiththenominalrangeandspecificationlimitsinmind.

DDMRP employs three types of stock buffering methods at decouplingpoints. The type of method used is based on whether a part is classified“replenished,” “replenished override,” or “min-max.” The bulk of this chapterwillfocusonthereplenishedpartclassification,asitisthepredominantmethodusedinaDDMRPsystem.

1. Replenished parts. Replenished parts use strategic and dynamicdecoupling point buffers. These parts are managed by a dynamicthree-zonecolor-codedbuffersystemforplanningandexecution.Thebuffer levels are calculated by a combination of globally managedtraitsrelativetothebufferprofileintowhichthepartfallsandafewcritical individual part attributes. These factors are adjusted withindefinedintervals.

2.Replenished override parts.Replenished override parts are strategicand static decoupling point buffers. These parts are managed by astatic three-zone color-coded buffer system for planning andexecution(asopposedtocalculatedanddynamicfor thereplenishedparts). Parts are assigned to this category when there are definedlimitations (space,process related, and/or cash)ordictated levelsofinventory (customer agreements, policy restrictions, etc.) within theplanningenvironment.Withoutthedynamicnatureofthebuffer,thecolor-codingsystembecomesthatmuchmoreimportantforplannerstoprioritizeplanning-andexecution-relatedactivity.

3.Min-maxparts(MM).Themin-maxdesignationisfornonstrategicandreadilyavailablestockedpartsandstock-keepingunits(SKUs).ThereisstillarolefortraditionallydefinedMMtacticsinDDMRP.APICSdefinesmin-maxas:

Page 129: DDMRP - Demand Driven Material Requirements Planning

Atypeoforderpointreplenishmentsystemwherethe“min”(minimum)istheorderpointand the “max” (maximum) is the “orderup to” inventorylevel.Theorderquantity isvariableand is the resultof themaxminusavailable and on-order inventory. An order is recommended when thesumof theavailableandon-order inventory isatorbelow themin. (p.105)

Min-max buffers in DDMRP are managed by a simpler two-zonecolor-coded system that canbedynamicallyalteredor adjusted in thesamewayasreplenishedparts.

All these parts’ buffer levels are determined by summing the zones thatcomprisethem.Replenishedandreplenishedoverrideutilizethreezones,whilemin-maxutilizesonlytwozones.Zonesarestratificationsorlayersinthebufferthatservespecificpurposesandhaveuniquecalculations.

TheGreenZone

Thegreenzoneistheheartofthesupplyordergenerationprocessembeddedinthe buffer. It determines average order frequency and typical order size. Thegreen zone is determined by one of three factors.Whichever factor yields thegreatestnumberdeterminesthesizeofthegreenzone.Inthiswayitrepresentsaconservativeviewwithregardtorecommendedandaverageorderfrequency.

TheYellowZone

Theyellowzoneistheheartoftheinventorycoverageinthebuffer.Theyellowzone is always calculated as 100 percent average daily usage (ADU) ×decoupledleadtime(DLT).

TheRedZone

The red zone is the embedded safety in the buffer. The higher the variabilityassociatedwiththepartorSKU,thelargertheredzonewillbe.Calculatingtheredzoneisaccomplishedwiththreesequentialequations.

Figure 7-2 represents a summary of the purposes and calculations for aDDMRP buffer. Performing the calculation for each of these zones isaccomplished through a combination of a grouping assignment (called bufferprofiles)andindividualpartattributes.

BufferProfiles

Page 130: DDMRP - Demand Driven Material Requirements Planning

Abufferprofile is agroupingofparts thathave similar characteristics.Bufferprofiles allow for the practical and effective global management of massivequantitiesofstrategicallydecoupledparts.Obviously,manydifferentmaterials,parts,andenditemsbehavedifferently.Conversely,manybehaveverymuchinthe samemanner. Buffer profiles are families or groups of parts forwhich itmakes sense to devise a set of rules, guidelines, and procedures that can beapplied the sameway to allmembers of a given buffer profile. Devising andrevising rules, guidelines, and procedures for hundreds or thousands of partsindividuallywouldbeoverwhelming.

FIGURE7-2Bufferzonesandpurposes

Thesefamiliesshouldnotbeconfusedwiththetraditionalnotionofproductormarketing families,which tend to be components or end items grouped bylike characteristics in terms of physical configuration ormarkets.With bufferprofiles,thefamilialconnectionismadebasedonthreespecificfactors.

Factor1:ItemType

Item type becomes the primary designator for globally managing families ofparts. The groupings will be made by determining whether an item ismanufactured (M), purchased (P), or distributed (D).The reasons to group bythesedesignationsare:

Responsibility. Companies often designate the control of thesedifferentitemtypestodifferentpeopleorgroups.

Intuition.Knowledgeaboutaspecificpartisfrequentlylimitedtothespecificgroupthatcontrolsthoseparts.

Page 131: DDMRP - Demand Driven Material Requirements Planning

Organizational control. There is often a varying degree of directorganizational control over these different item types. Companieswill tend to have more direct control over something containedwithin their facilities. The amount of control that extends topurchased and distributed items often depends on the verticalintegrationoftheenterprise.

Categorical differences. Relative lead time horizons can be verydifferent among these item types. Short lead times for purchaseditems could be up to a week. Short lead times for manufactureditemscouldbeonetotwodays.

The three item types—manufactured, purchased, and distributed—aretypically theminimumnumber of item-type designations a large supply chainentityshouldhave.Therecanbeothersifapplyingtheabovecriterialeadsustothe creation ofmore. For example, in some environments a distinction can bemade between end item manufactured items and intermediate manufactureditemswithregardtotheabovecriteria.Thismaycallforadifferentclassificationcalledintermediate(I).Thereisanexamplelaterinthechapterthatwillhaveanintermediatecategory.

Factor2:LeadTime

Leadtimeissegmentedintoat least threecategories:short,medium,andlong.Thesedesignationsarerelativetothecompany’sspecificenvironmentandparttype. Typically, there is a large distribution spread in the size of lead timesassociated with purchased parts. This spread could be anywhere from almostzeroleadtimeforon-sitesupplier-managedinventorytoleadtimesmeasuredinmonthsoryears.

Purchased parts that are reliably received with very short lead times aretypically not candidates for strategic replenishment designation. Little benefitcanbegainedfromtheadditionalmanagementoftheseparts.Figure7-3detailsthe distribution of lead times for purchased parts identified for strategicreplenishment inasampleenvironment (CompanyXYZ).Outof100parts,37areintheshortleadtimegroup,30inthemediumleadtimegroup,and33inthelongleadtimegroup.

Therearedifferingcircumstancesthatdictatewhattheparametersdefiningshort,medium, and longwithin anyparticular environmentwill be.Typically,the division point will come down to a comfort level for the buyers in that

Page 132: DDMRP - Demand Driven Material Requirements Planning

environment.Laterinthischaptertheleadtimedesignationinfluenceonbufferlevelsandzonesisdiscussed.

Manufactured parts have three types of calculated lead times that can beevaluatedinordertodeterminewhatisshort,medium,andlong.AsdiscussedinChapter 6, two of these lead times, manufacturing lead time (MLT) andcumulativeleadtime(CLT),areproblematicwhendecouplingpointsareinuse;they are either an underestimation or an overestimation, respectively. To thisextent,decoupledleadtimeshouldbeusedtodeterminewhatisshort,medium,and long. Figure 7-4 represents an example of short, medium, and longdesignations against the distribution of manufacturing parts chosen forreplenishment.Note that the lead time definitions of short,medium, and longdiffersignificantlyfromthepurchasedpartdefinitionsinFigure7-3.

FIGURE7-3ThedistributionofleadtimecategoryassignmenttopurchasedpartsforCompanyXYZ

FIGURE7-4ThedistributionofleadtimecategoryassignmenttomanufacturedpartsforCompanyXYZ

[Theleadtimecategorywillthenbeusedtosupplya“leadtimefactor”toparts within a profile. Figure 7-5 is a table of recommended lead time factorranges assigned to thedifferent lead timecategories.The lead time factor is apercentageofADUwithin thedecoupled lead timeof thepart.This lead timefactorwillimpactgreenandredzonecalculationsforeverystrategicpartwithin

Page 133: DDMRP - Demand Driven Material Requirements Planning

acertainprofile.Notice that the longer the lead timeof the part, the smaller the lead time

factor should be. A smaller lead time factor produces a smaller green zonecalculation.Sincethegreenzonedeterminesaverageordersizeandfrequency,asmallerleadtimefactorwillleadtosmallerandmorefrequentorders.Thismayseemcounterintuitiveformanyplannersandbuyers,buttheDDMRPapproachforces as frequent ordering as possible for long lead time parts (until theminimum order quantity or an imposed order cycle becomes a constrainingfactor).

Thisis indirectoppositiontothewaythatmanypurchasedlongleadtimepartsareoftenhandled.Typically,longleadtimepartsalsorepresentpersistentproblems and shortages. It is not uncommon that buyerswill buy double andtripletheorderminimumrepresentingmonthsormoreofsupplyjusttonothavetodealwiththesepartsasfrequently.Thisisespeciallytrueiftherehasbeenashortage in the recentpast.There is anold saying inpurchasing,“Buydoubleandstayoutoftrouble.”Ofcourse,muchofthisbehaviorisdirectlyassociatedwith the deficiencies of conventional MRP previously described. Once thesedeficiencieshavebeenaddressed,thenthebehaviorstocompensateforthemcanbereexamined.

DDMRP is about creating and protecting the flow of information andmaterials.For long lead timeparts,DDMRP isattempting tocreatea frequentdemandsignalrelatingtoactualneedandacorrespondingsupplying“pipeline”delivering a steady stream of supply orders. Figure 7-6 depicts the differencebetweenlargeinfrequentordersandasteadierstreamofsmallermorefrequentorders.

FIGURE7-5Recommendedleadtimefactorranges

Page 134: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-6Largeinfrequentversussmallerfrequentorders

Withlarge infrequentorders,adisruptionofaboat,port, truck,etc.,coulddisrupttheentireinboundsupply.Withsmallermorefrequentorders,therisktoall the inbound supply is significantly less. Additionally, there can be a cashflow advantage to paying smaller,more frequent invoices as opposed to largeinfrequentinvoices.

Factor3:Variability

Variabilityassignmentisthenextlevelofassignment.Ataminimum,variabilitycan be divided into three segments—high, medium, and low—with the twodimensionsofdemandandsupplyvariability.

Demand variability is the potential (frequency and size) for spikes indemand with regard to a particular part or SKU number. The variabilitydesignationcanbecalculatedbyavarietyofequationsordeterminedbyrulesofthumbwithintuitiveplanningpersonnel.

Heuristically, companies can use the following segmentation for demandvariability:

High demand variability. This part is subject to frequent spikeswithintheleadtime.

Mediumdemandvariability.Thispartissubjecttooccasionalspikeswithinleadtime.

Lowdemandvariability.Thisparthaslittletonospikeactivity—itsdemandisrelativelystable.

SupplyvariabilityisthepotentialforandseverityofdisruptionsinsourcesofsupplyforthispartorSKUnumber.Thiscanbecalculatedbyexaminingthe

Page 135: DDMRP - Demand Driven Material Requirements Planning

variance of promise dates from actual receipt dates. The caution here is thatmanyofthesedatesareoftendeterminedandmanagedinitiallythroughcriticalflawsintraditionalmaterialrequirementsplanning(MRP).Finally, thenumberofalternativesourcesforapartormaterialcanfactorintothesupplyvariabilityequationbecausetheneteffectofmoresourcesmightbemorereliablesupply.

Heuristically, companies can use the following segmentation for the threesimplestcategoriesofsupplyvariability:

High supply variability. This part or material has frequent supplydisruptions.

Medium supply variability. This part or material has occasionalsupplydisruptions.

Low supply variability. This part or material has reliable supply(eitherahighlyreliablesinglesourceormultiplealternativesourcesthatcanreactwithinthepurchasingleadtime).

Both forms of variability can also be mathematically calculated andexpressedthroughstandarddeviationforeachparticularpart.Partvariabilitycanbecomparedwiththecoefficientofvariationwithinagroupofparts.Coefficientofvariation is alsoused inChapter6 for thedistributionpositioningexample.This analysis can be useful in understanding the relative distribution ofvariability within the part population, but it still requires a team to set thespecificboundariesofhigh,medium,andlowwithinthedistribution.

As shown previously in Chapter 6, Figure 7-7 illustrates how buffers atdifferent stageswithinamanufacturingprocess canexperiencedifferent levelsand types of variability depending on their relationships with each other.Arrowedlinesthatmovefromlefttorightrepresentsupplyvariability.Arrowedlines thatmove from right to left representdemandvariability.Comingoutofthebufferineitherdirection,theyaresmootherthanwhentheyenter,implyinglessvariabilitythatispassedalong.

Purchased parts tend to be influenced almost exclusively by supplyvariability.Oneexceptionisinpuremake-to-order(MTO)orengineer-to-order(ETO) environments, where there are no buffers at the subcomponent,intermediate component,or end item level.Apuremake-to-order environmentwouldindicatethattheinventorypositioningfactorsdictatedbufferingonlyforsome purchased items. This is an example ofwhy companies cannot skip the

Page 136: DDMRP - Demand Driven Material Requirements Planning

inventorypositioningstepeveninMTOandETOcompanies.Itcandramaticallyalterwhichitemswillendupinwhichbufferprofiles.

FIGURE7-7Multiplebuffersanddifferentformsandlevelsofvariability

FIGURE7-8Thecombinationofleadtimeandvariabilitycategoriesforpurchasedparts

Figure7-8showsthecombinationofleadtimeandvariabilitycategoriesforpurchased parts from the previous example. This matrix results from theassignmentofpartstobothleadtimeandvariabilitycategories.

Manufactured parts can be subject to both supply and demand variabilitydependingonhowthepositioningmodelisformulated.Manufacturedpartsarelesssubjecttodemandvariabilityifthemanufactureditemfeedsanotherlevelofbuffered component or end item. These parts are less subject to supplyvariabilityiftheyconsumecriticalpartsthatarereplenishedstrategically.Thisisdue to the dampening nature of the buffer breakwalls on the end supply anddemandvariability.

However,inmanycasestherecanbeablendofdemandtypesexperiencedbyabufferedposition.Anexampleofthistypeofmanufacturedpartisonethatisusedinsubassembliesortheenditems(someofwhichmightbebuffered)butis also a service part (whichmight go directly to the customer). This type ofmanufacturedpartprobablywouldbesubjecttomoredemandvariabilitythanapart that fed only some buffered subassemblies or end items. Thus, it is

Page 137: DDMRP - Demand Driven Material Requirements Planning

imperative that companies carefully apply the positioning factors described inChapter6.

Figure7-9showsthecombinationofleadtimeandvariabilitycategoriesformanufacturedpartsfromthepreviousexample.

Distributed parts or SKUswill tend to be affected by one variability typedependingontheirrespectivelocationsintheinternalsupplychain.Distributedparts or SKUs at central buffers can be largely immune from large demandvariability if the downstream positions that they feed are sized and managedproperly.PartorSKUbuffersatdownstream locationswillbeaffectedalmostexclusively by demand variability because they are protected by the centralbufferonthesupplyside.SeeChapter6formoredetailoninventorypositioningindistributionnetworks.

FIGURE7-9Leadtimeandvariabilitycategoriesformanufacturedparts

FIGURE7-10Variabilitycategoryranges

Thevariabilitycategorywillthenbeusedtosupplya“variabilityfactor”topartswithinaprofile.Figure7-10showsthevariabilityfactorrangeswithineachcategory.Thevariabilityfactorisappliedtoanothercalculationthatestablishesabase level of safety corresponding to lead time. These calculations will beexplainedlaterinthischapter.

Parttype,leadtime,andvariabilitycategoryassignmentarethethreebasicparameters of buffer profiles. The lead time and variability factor ranges aremeant as a conservative guide for planners and buyers to follow. It should benoted that precision in the factor percentage determination is rarelyworth thetimedevotedtoit.Roughlyrightisbetterthanpreciselywrong.Thedifference

Page 138: DDMRP - Demand Driven Material Requirements Planning

between37.6percentand38.1percentwillhavelittleeffectonthebufferlevels.More importantly, the very nature of a buffer allows for quite a bit ofimprecisionandapproximationsince thebuffers shouldbeadaptiveover time.Butbufferprofilesareonlyhalfthenecessaryrequirementstocalculatebuffers.Nextwewillturnourattentiontotheindividualparttraitsthatwillprovidetheothernecessarycondition.

Basedon these three factors (part type, lead timecategory, andvariabilitycategory), there are 36 basic buffer profiles.Depending on themanufacturingenvironment,therecouldbeevenmorederivationsandpermutationsthanthis.Ifthere is a certain global attribute that makes sense by which parts should begrouped that is not related to variability, lead time, or part type, then anothertypeofbufferprofileshouldbeexploredandpossiblyadded.

FIGURE7-11Basicbufferprofilecombinations

Figure7-11summarizes the36differentbasicbufferprofilecombinations.Eachbufferprofilehasbeendesignatedwithacodebasedonitscombinationofattributes.Withinthatnamingcode,thefirstlettersignifiestheparttype:“P”forpurchased,“M”formanufactured,“D”fordistributed,and“I”forintermediatecomponents.Nextistheleadtimecategory:“S”forshort,“M”formedium,and“L” for long. The third letter represents the variability category: “L” for low,“M”formedium,and“H”forhigh.Forexample,adistributedpartwithmediumleadtimeandlowvariabilityiscodedas“DML.”Apurchasedpartwithalonglead time and high variability is in the buffer profile known as “PLH.” Laterexamples will use this naming schema with respect to parts’ buffer profileassignments.

Page 139: DDMRP - Demand Driven Material Requirements Planning

IndividualPartAttributes

Theindividualpartattributesarepropertiesornumericalvaluesthatarespecifictothepartitself.Manyofthesepropertiesorvalueswillbefoundorcalculatedfromthecurrentpartmasterinformation.InDDMRPtherearethreespecificpartattributes that will determine buffer levels for purchased, intermediate, andmanufacturedbuffered itemsandfourspecificpartattributes thatwill factor infordistributedbuffereditems.

PartAverageDailyUsage

Average daily usage is a calculated rate of use for each specific part. It is acornerstoneofthebufferequations.Significantchangestothepart’sADUwilloften yield significant impacts to the calculated buffer zones. There are fourimportant considerations in the ADU calculation for each part. Theseconsiderations may require planners, buyers, and distribution personnel toconsultwithotherareasoftheorganizationforvalidation.

Length-of-PeriodConsiderationAny average is only as relevant as the period over which the equation wasapplied. If the average is calculated with a shorter horizon, it will be moreresponsivethanalongerhorizon.Butifthehorizonistooshort,theADUwillbeoverreactive and may reproduce the bullwhip effect as the calculated buffersjump between extremes. This will be especially true with products thatexperience relatively large demand changes within shorter windows of time.Examples would be products that are frequently promoted. Figure 7-12illustratesthiswithaproductthatgoesthroughfourmajorpromotionalperiodswithinayear.Thebarsrepresenttheamountofweeklydemandexperiencedbytheproduct.Thefourmajorpromotionalperiodscoincidewiththechangingoftheseasonsandrepresentsignificantupliftsindemand.

Figure7-13illustratesthedifferencesinADUvaluesgeneratedbydifferentlength-of-periodcalculations.Threelengthsarecompared:52weeks,12weeks,and 1week.All periods are past looking. Later in the chapter, alternatives topast-lookingperiodsareexplored.TheADUvaluesareestablishedbytakingthetotalweeklybucketanddividingby7.Thisisassuminga365-daycalendar.Ofcourse, the type of calendar used to calculate the ADU value will make adifference.Thecalendar should reflect theworkingcalendarof the locationofthedecouplingpoint.

Page 140: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-12Afrequentlypromoteditem

FIGURE7-13DifferencesinADUvaluefromdifferentperiodconsiderations

Figure7-13 demonstrates a significant difference between a period lengthrolling52weeksinthepastversusonethatrollsfrom1weekinthepast.The52-weekrollhasverylittlefluctuationinthecalculatedADU,whilethe1-weekrollproducesdramaticswings.Manyenvironmentswouldsimplynotbeabletoreacttothemassivefluctuationsinthebuffersthatwouldresultfroma1-week-period calculation. In this regard, the length-of-period consideration furtherconnects theconceptofa stockbuffer to time.Astockbuffer ismeant tobuytime, especially when there are shared resources responsible for supportingmanydifferentbuffersinthatenvironment.InthisexampleADUmovesfromalow of under 16 to a high of over 44 when using a 12-week period. In this

Page 141: DDMRP - Demand Driven Material Requirements Planning

example, this is sufficientmovement to allow thebuffer to reasonably flexupanddownwhilenotcreatingabullwhipeffectontheenvironment.

Frequency-of-UpdateConsiderationPerforming a calculation and updating a part master with the result of thatcalculationaretwodifferenttasks.ItisimportantthatthereisnottoomuchtimebetweenpartADUupdates.The longer the time frame, thechoppier theADUwill become. This can create buffers that jump or fall in away that transfersvariabilitytootherpartsoftheorganization.Stepsshouldbetakentoupdatethepart’s ADU on a frequent basis. In most cases this would mean daily or nolonger thanweekly updates. Ideally, the part’s recordwould be automaticallyupdatedastheADUcalculationisperformed.UnliketraditionalMRPwheretheincreased frequency of MRP updates increases system nervousness, underDDMRPthemorefrequenttheupdatesofrelevantinformationsuchasthis,themore stable the environment. Infrequent updates (monthly or quarterly) cancauseamajorbullwhipeffectinaDDMRPenvironment.

Past,Forward,orBlendedConsiderationADUatanypointintimewillalsobeheavilyinfluencedbythelength-of-periodhorizon that theconsideration isapplied to—thepast, the future,orablendofboth.

Figure7-14showsthedifferenceforthepromotionalitemfromthepreviousexamplewitha12-weekperiodconsiderationusinganADUbasedon12weeksofpastsalesversusanADUusing12weeksofforecastedsales.TheleftYaxisisthedemandfortheproduct,whiletherightYaxisisthecalculatedADUforeachtypeofperiodconsiderationforanyparticularweek.

An ADU value calculated from a rolling past period can be problematicwhenthereisanexpectedupsurgeindemandthatisnotbeingproperlyreflectedinthecurrentorimpendingADU.WithregardtothepromotionalexamplefromFigure7-14,thisproblemcanbeseenlooking12weeksintothepast.Thereisaknownupsurgecoming inperiods6–9;yet inperiod1(thecurrentperiod) theADUisreducingdramaticallyoverthenexttwoperiodssothattheADUwillbeunder 16 while the forward-looking value is at 27 and has been trending up.Additionally, the forward-lookingADU appears to better anticipate themajordrop-offindemandinweek43.

Page 142: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-14PastandforwardADUvaluesagainstthedemandforthepromotionalproduct

Is forward lookingalwaysabetterway tocalculateADU?Remember thatthisscenarioassumesthattheforecastedrateofsaleswillactuallyoccur.Thisiswheresomedegreeofintuitionmustbebroughttobearwithintheenvironment.Howconfidentistheplanningteaminthenumbersbeingforecasted?Arethesenumberscomposedof firmorders,orare theymorewishful thinking? If thesenumbersarerelativelyfirm,thenitcouldbeappropriatethatthispartbeplacedon a forward-lookingADU. “Relatively firm”means that themajority of thisdemandisalreadyspokenforthroughamajorplannedpromotion.

Note that a forward-lookingADU is the incorporationof forecast into theDDMRP buffers but not into the DDMRP ordering mechanism. This isextremelyimportanttokeepinmindandwillbefurtherdiscussedinChapters8and9.

IfcalculatingADUusingaforecastseemstooaggressiveorriskygiventheenvironment,thereisafinaloption.Giventhefactthatthefartheroneforecastsintothefuture,thelessaccuratethedemandsignalis,thenmaybethesolutionisto limit forward-lookingADUcalculation toashorter future timerange that ismore accurate. But that amount of time will create an ADU calculation thatmightbetooshortandoverresponsive.Inthiscase,blendingthepastsaleswiththefutureexpectedsaleswouldmakesense.Figure7-15showstheADUvaluewithablendedapproachcontrastedwithpast-andforward-lookingversions.

Alsoanotheroptionwillbeconsidered inChapter8 todynamically adjusttheADUvalueusingademandadjustmentfactor(basedonhistoricalpatternsorknownevents)ratherthanaforward-lookingADU.

Page 143: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-15Past,forward,andblendedversionsofADU

ADUExceptionsIfasignificanteventoccursthathasaltereddemandprofilesdramaticallywithinarelevantrange,thenplannersshouldbecautiousaboutadjustingtheADU.Thisrapid change in demand should generate what is called an ADU alert. Theparameters of an ADU alert must be defined in terms of quantity and timeagainstthecalculatedADU.Forexample,ifapart’sADUchangesmorethanxpercentwithinaytimeframe,thenthealertisgenerated.

Thisrequiresahighandlowthresholdandahorizontobeestablished.Forexample, ifacompanysetsahigh thresholdof300percentover threedays, itwouldbeaskingforADUalertstobegeneratedwhendailydemandwastriplethedailynorm(ADU)forthreedaysinarow.Conversely,alowthresholdcouldbeestablishedof10percentforfivedays,meaninganADUalertwouldoccurwhen demand was 10 percent of the ADU for five days in a row. These aresimplyexamples;thelengthandthethresholdwillbeuniquetotheenvironment.

If thesignificanteventsareanomalousandwill revert tonormalbehavior,then the abnormal usage (or lack of usage) should be excluded in the ADUupdateequation.However, if thedramaticshift is indicativeofwhat thefuturemightlooklikeforthispart, thenthatinformationshouldbeincludedorat theleast factored in to some extent through a demand adjustment factor.DemandadjustmentfactorsarediscussedinChapter8.

EstablishingADUforItemswithNoHistoryA companymay have no history by which to generate an ADU for strategicitems.Thiscouldoccurinacompanythathasalegacysystemornosystemof

Page 144: DDMRP - Demand Driven Material Requirements Planning

recordatall.Ifacompanyhasamodernplanningsystemandthesedataarenotavailable, then the lack of availability would often be due to poorimplementation.

ThecollectionofactualusagedataisanecessaryconditionforanadaptiveDDMRP implementation. Data need to be collected and phased in withestimations. Planners may need to consult with other relevant personnel tocalculateanestimationofADUthatseemsrealisticforeachitem.Thenoverthelength-of-period consideration, the actual ADU can be blended with theestimatedADU.Attheendofthelength-of-periodconsideration,anADUthatrepresentsactualhistoryovertheentirelengthofperiodiscalculated.Figure7-16illustratesthisconceptwithaproductthathasa12-weekpast-lookingADUcalculation.AnestimatedADUof3,000wasusedtobuildthebuffer,andinthenext12weeksanADUof3,224.9emerges.Eachweektheactualusagereplacesaweekofestimatedusage.

PartLeadTime

Anothercriticalindividualpartinputintothebufferequationisthepart’suniquelead time asmeasured in discrete units of time (most often in days). For anymanufacturedorintermediateitem,thisleadtimeshouldbethedecoupledleadtimeofthepart.ThevalueofdecoupledleadtimeisdiscussedinChapter6.Forpurchasedparts, thepurchasingleadtimefromthepartmastershouldbeused.Fordistributedparts,thetransportationleadtimefromthesourcingunitorhubshould be used. In some cases, additional time may be added for necessarystagingorreceipt ifapplicable. If thequalityassuranceor incominginspectiontimeissignificant,thenataminimumthistimeshouldbeincluded.

Page 145: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-16ActualizingADUoverthelength-of-periodconsideration

PartMinimumOrderQuantity

Ordering policies (minimums,maximums, andmultiples) complicate planningandsupplyscenariosbutareafactoflifeforplanners.Manyoftheseorderingpolicies are based on valid data and sound assumptions;many are not. It is agiven that there will be parts and SKUs that do require minimum orderquantities.Minimumorderquantities(MOQs)canaffectbufferlevels,especiallywhentheyarelargeinrelationtotherateofuse.Thesearecalled“significant”MOQsandwillhaveadirectimpactonthesizingofthebufferthroughthegreenzone. The qualifying characteristics that make an MOQ significant will beexaminedlaterinthischapter.

PartLocation

Location is an attribute that is unique to distributed part types chosen forstrategic replenishment assuming more than one forward distribution point.Refer to the distribution example from Chapter 6 for a description of theconsiderations for distribution. For each part that is distributed through thenetwork, a separateADU and lead timewill exist for each distribution point.Figure7-17showstheADUvaluesfromChapter6fordistributedpart123.

Page 146: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-17Distributedpart123ADUandleadtimevaluesbylocation

Each physical distribution point including the hub has a distinct averagedailyusageand lead time.Whilewarehouses1and2have thesamevalue forleadtime,itmustbeunderstoodthatthosenumbershavenorelationshiptoeachother; they are separate and distinct transportation times that can changeindependentofeachother.

Figure7-18summarizes thecombinationof the factorsdiscussedso far inthis chapter thatwill combine to create each strategicdecouplingpointbuffer.As previously discussed, individual part attributes will combinewith the leadtime and variability factors from a buffer profile assignment to create theseuniquebufferlevels.

CalculatingReplenishedPartBufferLevelsandZones

Replenished part buffers are composed of three color-coded zones: green,yellow, and red. Each zone has a specific purpose and will vary in size andproportiondependingonthecombinationofthebufferprofileandtheindividualpart traits discussed previously in this chapter. It is important to note that thebuffer is not simply divided into equal thirds.Understanding the purpose andcalculation of each zone is crucial to understanding how DDMRP buffersproducetheirresultsaswellashowtheycomparewithotherstockmanagementtechniques.

Figure 7-19 displays the individual and buffer profile attributes of anexamplepartthatwillbeusedtolearnhowtocalculatebuffers.ThisparthasthebufferprofileofMML,meaning it isamanufacturedpartwithamediumleadtimeandalowvariabilitysetting.ThisisshownintherowinFigure7-19called“Buffer Profile.” Additionally, in the “Buffer Profile” row, the lead time andvariability category has been assigned a factor. This factor is noted in theparenthesesafter the lead timeandvariabilitydesignator, respectively.For this

Page 147: DDMRP - Demand Driven Material Requirements Planning

example, the lead time factor is 0.5, and the variability factor is 0.33. Thesefactors fall within the ranges for each subcategory discussed earlier in thechapter.

FIGURE7-18Bufferprofilefactors

FIGURE7-19Examplepartbufferprofileandindividualattributes

TheGreenZone

As discussed earlier in this chapter, the green zone is the heart of the supplyordergenerationprocessembeddedinthebufferandwilldeterminetheaverageorderfrequencyandtypicalordersizeforthepartchosenforthisexample.Thegreen zone is determined by one of three factors.Whichever factor yields thegreatestnumberdeterminesthesizeofthegreenzone.Eachfactorneedstobecalculatedforourexampleinordertodeterminewhichwillqualifyasthegreenzonevalue.

Option1:AnImposedorDesiredMinimumOrderCycleAnordercycleissimplythenumberofexpecteddaysbetweenorders.Itcanbeanimposedfactorthroughtheuseofaproductschedulingwheelorbeadesiredaverage number of days between orders.Eitherway the equation is the same.TheequationforcalculatingthegreenzonebasedonordercycleissimplyADU×desiredor imposedordercycledays.For thepart inFigure7-19, thiswould

Page 148: DDMRP - Demand Driven Material Requirements Planning

yield70.

Option2:UsingaLeadTimeFactorAgreenzonecanbecalculatedusingtheleadtimefactordiscussedpreviouslyinthischapter.Thisleadtimefactorisexpressedasapercentageofusagewithinafullleadtimeofthepart.ThatpercentagewillfallwithinarangecorrespondingtotheleadtimecategoryofthepartillustratedinFigure7-19.Theformulaforproducing the green zone value using this technique is decoupled lead time×ADU×leadtimefactor.ForthepartinFigure7-19,thisoptionwouldyield60.

Option3:MinimumOrderQuantity(ifApplicable)Ifthegreenzoneisaboutsupplyordergenerationandfrequencyandtheparthasaminimumorderquantity,thentheminimumorderquantitycanberelevantindetermining thegreenzone. Inshort, thegreenzoneshouldneverbe less thantheminimumorder quantity. If theminimumorder quantity yields the largestvalueasagreenzone,thenthatminimumorderquantityisdeemedsignificant.Theminimumorderquantitymustbecomparedagainstthedesiredordercyclequantityvalueandthevaluecreatedbyusingtheleadtimefactor.ForthepartinFigure7-19,thisoptionwouldyield50.

Tosummarizetheoptions:

Calculatingthegreenzoneusingtheordercycleyieldsagreenzoneof70[ADU(10perday)×ordercycle(7days)].

Calculatingthegreenzoneusingtheleadtimefactoryieldsagreenzoneof60[ADU(10perday)×DLT(12days)×leadtimefactor(0.5)].

Theminimumorderquantityis50.

Thelargestvalueisactuallythedesiredordercyclevalueof70.

TheYellowZone

The yellow zone is the heart of the inventory coverage in the buffer. It is theeasiestandmoststraightforwardzonetocalculateinabuffer.Theyellowzoneisalways calculated as ADU multiplied by the decoupled lead time. In ourexampleinFigure7-19,theyellowzoneofthispartwouldbesizedat120pieces[ADU(10perday)×DLT(12days)].

Page 149: DDMRP - Demand Driven Material Requirements Planning

TheRedZone

The red zone is the embedded safety in the buffer. The higher the variabilityassociatedwiththepartorSKU,thelargertheredzonewillbe.Calculatingtheredzonerequiresthreesequentialequations:

1.Establish the“redbase.”Theredbase isestablishedbymultiplyingtheleadtimefactorbytheaveragedailyusagebytheleadtime.Thislead time factor corresponds to the same ranges used for the greenzone calculation but can have a different numerical value. In ourexample in Figure 7-19, the part falls in the medium lead timecategory. For simplicity in the example, we will use the samepercentage lead timefactor (50percent,or0.5) thatwasused in thegreenzonecalculation.Thusforthisexampletheredbasevalueis60units[ADU(10perday)×DLT(12days)×leadtimefactor(0.5)].

2.Establishthe“redsafety.”Theredsafetyiscalculatedasapercentageoftheredbase.Thepercentageusedisdeterminedbythevariabilityfactor.Liketheleadtimefactor,therearerangesofvariabilityfactorsdepending on whether a part experiences high, medium or lowvariability.Ourexamplepartfallsinthelowvariabilitycategorywitha variability factor of 33 percent (0.33). Our red safety value iscalculatedas20[redbaseof60×0.33(variabilityfactor)].

3.Calculatethetotalredzonebyaddingtheredbasetotheredsafety.Our example part will have a red zone of 80 [red base (60) + redsafety(20)].

Thismayseemlikeacomplicatedwaytocalculateasafetylevel.However,the safety level is related toboth lead timeandvariability throughvisible andindependent factors. The red base is the safety with respect to the ability torecover due to time. The red safety zone factors that number based on thevariability factor within that lead time period. Thus parts that have the samevariabilityfactorappliedbutareindifferentleadtimecategorieswillhavetotalredzonesthatareproportionatelydifferent.

Onceallzonesarecalculated,wecanaddthemtogethertogetatotalbuffer.Figure 7-20 summarizes the buffer calculations and subsequent zonecalculations.Thetopofthebufferiscalled“topofgreen”andisthesummationofallzones.Inthiscaseitis270[red(80)+yellow(120)+green(70)].

A few observations can be inferred at this point about the zones and the

Page 150: DDMRP - Demand Driven Material Requirements Planning

relationshipstheymayhavewitheachother:

The average order frequency can be calculated as the green zonedividedby theADU. In this example it is 7days.Thismeans, onaverage,withnovariability,thispartwillbeorderedevery7days.

The red zone divided byADU tells us just howmuch embeddedsafetyiswithinthebuffer.Theredzoneofthisbufferrepresents8daysofsafety[redzone(80)dividedby10(ADU)].

FIGURE7-20Examplepartbuffercalculationsummary

Theredbasewouldequalthegreenzoneiftheordercyclehadbeenlower than the green zone calculated using the same lead timefactor.

Therelationshipbetweenthegreenzoneandyellowzonecanalsotell us howmanyopen supplyorderswe can expect to see at anyone time. In our example, dividing the yellow zone (120) by thegreen zone (70)means that there is an average of around 2 opensupply orders at any one time. For longer lead time parts, thisestimateshowmany“ordersinthepipeline”canbeexpected.Thiscanprovideaquickwaytoanalyzewhetherthepartfitsthebufferprofileprovidedforitandwhetherthebuyerorplannerislaunchingsupplyorderswithinarelevanttimeframe.

Figure7-21showsasummaryofhowbuffersarecalculated.Notethatwhencalculating the buffers by hand, it is easiest to calculate the yellow zone first

Page 151: DDMRP - Demand Driven Material Requirements Planning

sincegreenandredbasecalculationsareapercentageofADU×DLT(yellowzone).Byestablishingtheyellowzonefirst,thegreenandredbaseequationsarequickertoperform.

Tofurtherpracticecalculatingbuffersaswellascontinuingtobuildthecaseforthevalueoffindingtherightinventorypositions,theexamplewithCompanyABCcontinues.

ContinuingwithCompanyABC

InChapter6anexamplecalledCompanyABCwas introduced todemonstratethe decoupling point positioning considerations. The various decoupling pointpositioningconsiderationswereusedtocreateamodelthatbetterprotectedtheenvironment from variability, leveraged inventory, and compressed totalinventory requirements in the face of certainmarket expectations.Thatmodelhad several iterations. Now the buffers for those different iterations arecalculatedinthefollowingexamples.

ThestartingsituationwithCompanyABChadthreefinisheditemsalreadychosen for strategic buffering due to their relatively short customer tolerancetimeinrelationtotheirdecoupledleadtimes.Figure7-22isthestartingsituationinChapter6withthethreeenditems(FPA,FPB,andFPC)circled.

FIGURE7-21Thebufferequationsummary

Page 152: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-22InitialdecouplingpointsforCompanyABC

Thebufferprofilesettingsandtheindividualpartattributesarerequiredforeachoftheseenditemsinordertocalculatetheirbuffers.Figure7-23containsthebufferprofileparameters thatCompanyABCisusing.Note thatCompanyABCcreatedaseparatebufferprofilecategoryforintermediatecomponentsthatimpactstheirrespectivevariabilityfactors.Thisisdonewiththerecognitionthatintermediate components tend to experience less variability since they areprotected by parent and/or component positions.These profile settingswill beappliedtothedifferentiterationsofpositioninginChapter6.

FIGURE7-23BufferprofilesettingsforCompanyABC

Iterations

Page 153: DDMRP - Demand Driven Material Requirements Planning

In the first iteration, the buffer sizes of FPA, FPB, and FPC will start theexample.Inordertocalculatethesebuffers,thebufferprofilesettingsinFigure7-23 are combinedwith each individual part’s characteristics. Figure7-24 hastherelevantpartdatatoperformthebuffersettingcalculation.Allparentitemshavebeensettomediumvariabilityinthisfirstiterationbecausetheyaresubjectto both demand variability and supply variability from a long, tightly coupledleadtimechain.

The buffer profile assignments are the sequence of part type, lead timecategory(withtheleadtimefactorinparentheses),andvariabilitycategory(withthevariabilityfactorinparentheses).Forexample,FPAhasaprofileMLM.Thismeans it is amanufactured part, has a long lead time, and is assigned to themediumvariabilitycategory.

Withthebufferprofilesettingsandthepartattributes,thebuffersandtheirrespectivezonescanbecalculated foreachend item.Figure7-25 contains thecalculationsforeachbuffer.Thelargerboldednumberrepresentseachzonesizeforeachpart.

For each part the yellow zone is calculated by multiplying the part’sdecoupledleadtimebythepart’saveragedailyusage.Forexample,FPAhasaDLTof20andanADUof250,yieldingayellowzoneof5,000.

Page 154: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-24FPA,FPB,andFPCpartattributesatCompanyABC

FIGURE7-25BuffercalculationsforFPA,FPB,andFPC

Each part’s green zone is determined by comparing three numbers andtakingthelargest.Inallthreepartsinthisfirstiteration,thegreenzoneissizedastheleadtimefactormultipliedbytheADUmultipliedbytheDLT.

Eachpart’sredzoneisarrivedatthroughtheredzonesequentialequation.First,thebaseisestablishedbyapplyingaleadtimefactor.Theproductofthatequationisthenmultipliedbythevariabilityfactor.Theredzonebaseisaddedto the red zone safety toget the total red zone setting.AsFigure7-25 shows,FPA’sbaseis1,250afteraleadtimefactorof0.25isappliedtofullusagewithinthe DLT. That 1,250 is then multiplied by the variability factor of 0.5. Thismakesthesafetyportion625.Next1,250and625areaddedtogethertoyieldatotal redzoneof1,875.Note:Thesame lead timefactor for thegreenand redzonesisbeingusedforthisexample.Thisisnotrequiredbutisdonesohereforsimplicity.

Figure 7-26 is the graphical depiction of the end item buffers in the firstiterationoftheCompanyABCexample.Thegreenzoneforeachpart isat thetopofthestack,theyellowzoneisinthemiddle,andtheredisat thebottom.Thenextiterationinvolvesdecoupling201inordertoreducethedecoupledlead

Page 155: DDMRP - Demand Driven Material Requirements Planning

timeforeachparentparticularlyFPA.Compressingeachofthedecoupledleadtimesforeachenditemshoulddefinitelylowertherespectivebufferlevels,butitwillalsorequirebuildinganadditionalbufferfor201.

Figure 7-27 shows the impact on FPA buffers by decoupling at 201.Decoupling201hasamajorimpactonsomeofthecriticalinputstothebufferequation.These changes are highlighted in the shadedboxes (“BufferProfile”and“DLTAfterDecoupling”).Decoupling201has compressed thedecoupledleadtimeforFPAtosevendays.ThatmovesFPAintoadifferentbufferprofilefor two reasons.First, the shorter lead timemoves thepart from the long leadtime profile (with a lead time factor of 0.25) to themedium lead time profile(withaleadtimefactorof0.4).Second,bydecouplingat201theenditemsaresubject tomuch less supply variability from that leg of the product structure.ThishasresultedinFPAmovingfromthemediumvariabilitycategory(withavariabilityfactorof0.5)tothelowvariabilitycategory(withavariabilityfactorof0.25).

FIGURE7-26EnditembuffersizesforCompanyABC

Page 156: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-27FPAbuffercomparison(firsttoseconditeration)

The impact on the buffer zones of FPA is significant. Figure 7-27 alsoillustrates a side-byside comparison of buffer sizing, with the first iterationlabeled“FPA1”andtheseconditerationlabeled“FPA2.”Thetopofgreenlevelforthefirstiterationis8,125,whilethetopofgreenlevelforFPA2is3,375.

Similar to the impactonFPA,decoupling201alsohasmajor implicationsforFPB.TheFPBdecoupledleadtimehasbeencompressedfrom23daysto9days.ForthesamereasonsFPBalsomovestoamediumleadtimeprofile(withaleadtimefactorof0.4)andalowvariabilitycategory(withavariabilityfactorof 0.25). Figure 7-28 depicts the buffer levels between the first and seconditerationforFPB.

FPCalsoexperiencesalargecompressionofdecoupledleadtime(from23to8days).ThesameshiftinbufferprofileoccurswithregardtoFPC.Figure7-29showsthecomparisonofthefirstandseconditerations.

Obviously,decouplingwith201reducesthelevelofaverageworkingcapitalcontained in theend itembuffers.Butbyhowmuch?Toanswer thisquestionthere are two requirements. First, the additional required buffer levels incomponents tosupport thesecompressed lead timesmustbecalculated. In thiscasethebufferrequiredfor201mustbecalculated.Second,theaverageon-hand

Page 157: DDMRP - Demand Driven Material Requirements Planning

levelsinallbuffersinvolvedinthecompression(componentandenditem)mustbe calculated and compared against the noncompressed scenario. Simplycomparingthetopofgreenlevelsprovidesadistortedinventoryimpactpicture.

FIGURE7-28FPBbuffercomparison(firsttoseconditeration)

Page 158: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-29FPCbuffercomparison(firsttoseconditeration)

Tomeet the first requirement, the buffer levels from the Company ABCexample can be calculated. The second requirement necessitates anunderstandingofhowDDMRPbuffersplanandreceivesupplyorders.Thusthesecond requirement will be deferred until Chapter 9 when Company ABC isrevisited, including the calculation of the full working capital impact in theexample.

Next thedecouplingpoint buffer for 201 is calculated.Figure7-30 showsthe profile assignment and part attributes thatwill determine 201’s buffer andzone levels.For thisexample,averagedailyusage is thesumof theend items(FPA,FPB,andFPC).If201wasusedinotherparentitemsorwassoldonitsown (e.g., a servicepart), thatdemandwouldhave tobe incorporated into theADU. Part 201 is assigned to themanufactured intermediate component, longleadtime(leadtimefactorof0.25)andmediumvariability(variabilityfactorof0.5)profile.

The 201 buffer appears substantial due to a relatively large amount ofdemandanditslongdecoupledleadtime.Thetopofgreenisat20,070.Atfacevalue thevastmajorityof the inventorysavings in theparentsappears tohavesimplyshiftedtothiscomponent.Thatistruefromaquantityperspectiveonly.

Page 159: DDMRP - Demand Driven Material Requirements Planning

Keepinmindthatenditemshavemoredirectmaterialdollarscontainedinthem,ataminimum.Furthermore,asdiscussedpreviously,itwillbedistortivetofocusonthetopofgreennumbers.Whatisneededtomakeafairfinancialcomparisonbetweeniterationsiswhat theaverageworkingcapital levelswillbeunder thedifferentscenarios.ThisisdiscussedinChapter9.

FIGURE7-30The201bufferprofile,partattributes,andcalculatedbufferlevels

ThenextiterationoftheCompanyABCexampleinvolvesbufferingthe203componentinordertopotentiallyeliminatetheneedtoholdFPAstock.FPA’scustomer tolerance time is three days, and buffering both 201 and 203wouldallowFPAtobemadewithinoneday, thusmovingFPAtoassemble toorderstatus. Part 203 is an intermediate component with its purchased itemsdecoupled,allowingittobeinthelowvariabilitycategory(variabilityfactorof0.2).Itsdecoupledleadtimeofsixdaysstillplacesit inthemediumleadtimecategory.Figure7-31showstheprofileassignmentandpartattributes thatwilldetermine203’sbufferandzonelevels.Figure7-31alsodepicts thecompletedbuffercalculationsfor203.

ThefinaliterationoftheCompanyABCexamplefromChapter6 involvedcompressing the decoupled lead time for 201. The purchased part 401P wasselected to be decoupled. By stocking 401P, the decoupled lead time forcomponent 201 is reduced from 19 days to 9 days. However, 9 days stillqualifies as a long lead timemanufactured part for CompanyABC. The leadtime category and lead time factor (0.25) remain the same. The variabilitycategory has been reduced from medium (variability factor 0.5) to low(variability factor 0.2) because external supplier variability has beenmitigatedby the401Pbufferposition. It isworthnoting thatafter401P isbuffered,201becomesthefirstbufferintheCompanyABCexampleinwhichthegreenzone

Page 160: DDMRP - Demand Driven Material Requirements Planning

qualifiesasthedesiredordercycle.

FIGURE7-31Component203’sbufferprofile,partattributes,andcalculatedbufferlevels

One final buffer calculation is required to complete the Company ABCexamplefromChapter6.Part401Pwasdecoupled inorder tocompress201’slead time. Next is the calculation of the 401P buffer. Figure 7-33 shows theprofileassignmentandpartattributesthatwilldetermine401P’sbufferandzonelevels.Figure7-33alsodepictsthecompletedbuffercalculationsfor401P.

Figure7-34 summarizes the buffer levels for all buffers for each exampleiteration at CompanyABC that were impacted by the positioning example inChapter 6. The buffers at 302P, 402P, 403P, 410P, 411P, and 404P are notdisplayed,as therearenosignificant implicationsfor thosebuffers.Somethingthat can be noted is that the positioning example tends to create a shift ofinventorytothelowerlevelsoftheproductstructure.

Page 161: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-32FPCbuffercomparison(beforeandafterbuffering401P)

FIGURE7-33Thebufferat401P

Page 162: DDMRP - Demand Driven Material Requirements Planning

FIGURE7-34AllbuffersfromtheCompanyABCexample

InChapter9 theCompanyABCexample is revisited inorder tocover theworking capital impactwith regard to averageon-hand levels for thedifferentiterations.

CalculatingReplenishedOverrideBuffers

Replenished override parts still utilize the three-zone systemof green, yellow,and red. The calculations of these zones, however, are overridden with amodifiedequationorareessentiallyuserdefineddependingon the limitations.These limitations would be beyond the limitations already discussed such asminimum order quantity or order cycle. An example might be spaces in avendingmachine;thereisafinitenumberofslotsineachmachinetobeutilized.Another example might be cash or space limitations where a company iscontractually bound. The zones still serve the same purpose but due to theimposedlimitationsmayhavelimitedeffectivenessintheirrespectivecapacities.Appendix C, which describes how DDMRP can be applied to the retailenvironment,providesanexampleofbufferoverridelogicinordertodealwithcertainrestrictionsorlimitations.

CalculatingMin-MaxBuffers

Min-max parts utilize only two zones—green and red. These zones are

Page 163: DDMRP - Demand Driven Material Requirements Planning

calculatedandadjustedinthesamewayasreplenishedparts;thereissimplynoyellow zone. Figure 7-35 shows the previous example part with a min-maxdesignation. The same green and red zone calculations occur, and the yellowzoneisdisregarded.

FIGURE7-35Examplepartwithmin-maxdesignation

Summary

The calculations involved in setting buffer levels are based on quite simpleequations.Mostpeopleabovetheageof12shouldbeabletoperformthemwithrelativeease.The secretof sizingbuffershas less todowith formalequationsand more to do with the considerations and organization of applying thoseequations.Thisrequiresaprocessofparametersettingandmaintenanceatbothaglobal (buffer profile) and individual part level. This processwillmost likelyimpactanorganization’sSalesandOperationsPlanning(S&OP)processesandthetoolsanddatabasesusedin thatenvironment.Theimpact toS&OPwillbediscussed in depth in Chapter 13 through the introduction and description ofDemandDrivenS&OP.

Page 164: DDMRP - Demand Driven Material Requirements Planning

CHAPTER8

BufferAdjustments

Chapter 7 discussed the considerations in calculating the initial levels of abuffer.Since today’s supplychainsare incrediblydynamic, thesebuffersmustadjustandadapt tochangingconditions.Byunderstanding theequations tosetthebufferzones,thenthefactorsthatcanchangeapart’sbufferoverthecourseoftimearealsounderstood.Thesechangescancomefrompartattributechangesorbufferprofilechanges.

RecalculatedAdjustments

Recalculated adjustments are automated adjustments to buffer levels based onchangestoindividualpartattributesorbufferprofileadjustments.

As discussed in Chapter 7, there are three critical factors for all bufferedparts thatdirectly impact thebuffer equations:ADU, lead time, andminimumorder quantity. ADU and lead time tend to have the most dramatic impactbecausetheyareinvolvedinallthreezonedeterminations.Theminimumorderquantity is only involved ingreen zonedetermination.Themost dynamicpartattributeistheADU,asitisconsistentlybeingrecalculatedandupdated.

As an example of each input change, Figure 8-1 shows the buffer inputsover a six-month time frame.Note that every input is static exceptADU.Thepart’s decoupled lead time (DLT) and buffer profile inputs [lead time factors(LTF)andvariabilityfactors(VF)]donotchange.Thispartseemstoexperiencesignificantgrowthoverthissix-monthtimeframeastheADUmovesfrom10onJanuary1to53byJune15.Atthispoint,part1234doesnothaveaminimumorderquantityorordercycleassignedto it.Thegreenzoneiscalculatedusingtheleadtimefactorof0.5(amediumleadtimepart).

Figure8-2showsthebufferzonesadjustingoverthecourseofthesix-monthperiod. As the ADU quintuples over that period, the buffer adjusts up

Page 165: DDMRP - Demand Driven Material Requirements Planning

accordingly.TherateofthisadjustmentdirectlycorrespondstotheriseinADU;all othervariables remain static.TheADU is representedby the solid line. ItsvaluecorrespondstotherighthandY-axiswhilethebufferzonevaluesrelatetothelefthandy-axis.

FIGURE8-1Part1234data

FIGURE8-2Part1234bufferadjustmentoversixmonths

In aDDMRPsystem,ADUwill alwaysbe changing since it is frequentlyrecalculated.Within some timeperiods the levelof changecouldbe relativelysmall, but it is changing nonetheless.What happenswhenmore static but not

Page 166: DDMRP - Demand Driven Material Requirements Planning

less significant part attributes change? For example, the part’s DLT directlyimpactsallzonesifthereisnominimumorderquantitypresent.Thisisthecasesofarwithpart1234.Figure8-3showsthechangeinthepart’sDLTonMarch15.

The shaded boxes represent the notable changes. First, theDLT has beencompressed from 10 to 5 days. If the part is purchased, this could occur bysourcing froma different vendor or a current vendor agreeing to a better leadtimenow that the company isno longer constantly rescheduling. If thepart ismanufactured, this compression could be due to additional capacity, processimprovement, reduction of work in progress, or additional decoupling (asillustratedwith theCompanyABCexample inChapters6and7).This shorterleadtimehasmovedthepartfromamediumleadtimefactorof0.5toashortleadtimefactorof0.7.Thishasadirect impactonthebuffer.Whilethegreenandyellowzonesshrinkduetothisleadtimecompression,theredzoneactuallyinflates due to the higher LTF. Figure8-4 illustrates the impact on the bufferzonesofcompressingtheleadtimefrom10to5daysonMarch15butwiththerestoftheinputsstayingthesame.

FIGURE8-3Leadtimechangeforpart1234

Finally,thepart1234exampleisusedtoillustratetheimpactofasignificantMOQonthebuffer.Inthiscaseaminimumorderquantitywasimposedonpart1234 on April 15. If this is a purchased part, this could occur due to a new

Page 167: DDMRP - Demand Driven Material Requirements Planning

agreementwiththesupplier(maybeinexchangefortheshorterleadtime).Ifthisis a manufactured part, this could occur in an attempt to save setups on acapacityconstrainedresource.Theleadtimecompressionfrom10to5daysonMarch15hasbeenmaintained.Figure8-5illustratestheimpositionoftheMOQ.Theshadedboxeshighlighttherelevantfactorsinthischange.AnMOQof400has become the green zone value. The column labeled “Green (LTF)” is thegreenzonevaluecalculatedusingtheleadtimefactor.

Figure8-6illustratestheimpactofimposingtheMOQonpart1234onApril15. There is an immediate jump in the green zone. Thiswill have significantimpacts on the amount of working capital contained in the buffer. WorkingcapitalimplicationswillbefurtherexploredinChapter9.

FIGURE8-4Part1234leadtimecompression

Page 168: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-5Part1234withaminimumorderquantityof400

FIGURE8-6Part1234withaminimumorderquantity

A buffer profile change causes a recalculation of the buffers to all partsassigned to that profile. The part 1234 example has already demonstrated theimpactofmovingapart toadifferentprofile(frommediumleadtimetoshortleadtime)andapplyingthedifferentfactorsassociatedwiththatprofile.Fromaglobalperspective,iftherearechangestothefactorswithinabufferprofile,thenobviouslyallpartscontainedinthatprofilewillbeaffectedsimultaneously.Forexample,iftheplanningteamdecidestouse0.6insteadof0.5fortheleadtime

Page 169: DDMRP - Demand Driven Material Requirements Planning

factorformediumleadtimeparts,thenallpartswithamediumleadtimewillbeaffectedunlesssubcategoriesarecreatedformediumleadtimeparts.

PlannedAdjustmentFactors

Buffers also can be manipulated through planned adjustments. Plannedadjustments are basedon certain strategic, historical, andbusiness intelligencefactors.Theseplannedadjustmentsaremanipulationsofthebufferequationthataffect inventory positions by raising or lowering buffer levels and theircorresponding zones at certain points in time. Thesemanipulations tend to beconfined to demand input manipulations, zonal manipulations, or lead timemanipulations.

DemandAdjustmentFactor

Thedemandadjustmentfactor(DAF)isamanipulationoftheADUinputwithina specific time period. This manipulation occurs by adjusting the ADU to ahistoricallyprovenorplannedpositionbasedonanapprovedbusinesscaseorasareactiontorapidchangesindemandwithinshortperiodsoftime.

Demand adjustment factors should not be indiscriminately used. Do notunderestimatethepowerandflexibilityofaproperlymanagedDDMRPsystem.The buffers are robust. They are designed to absorb variability. The moreresponsive the plant or supplying resource is, the more robust the buffers’performancewillbeforhighervariability.Thehigherthevariabilityfactorbuiltinto the buffers, the more robust the buffers’ performance will be for highervariability, albeit with the penalty of additional inventory. The longer thehorizon to see spikes, the more robust the buffers’ performance will be forhighervariability.

Withthatsaid,thebuffersareonlydesignedtoabsorbvariabilitytoacertainextent. It can often be the case that variability up or down can threaten theeffectivenessofthebuffertoprotectthatdecouplingpoint.Demandadjustmentfactorsshouldbeemployedwhenvariabilitythreatenstooverwhelmthebuffers.In this case the adjustment should be up. When variability will cause largeamountsofprolongedexcessinventory,theadjustmentshouldbedown.

Howdodemandadjustmentfactorswork?Figure8-7depictsasamplepartcalled ABC. The buffer-level zones appear to be relatively stable over asignificantperiodoftime.Theselevels,however,caneasilybemanipulatedbychangingtheADUvaluefeedingthebuffer-levelequationatcertainpoints.

Figure8-8illustratespartABCwithaDAFimplementedfromweek13to

Page 170: DDMRP - Demand Driven Material Requirements Planning

week25.ThetablebelowthegraphshowstheoriginalADU,theDAFappliedtoeach period, and the adjusted ADU value. For example, in week 18 anadjustment factor of 1.8 has been applied to the original ADU of 34.68 toproduceanadjustedADUof62.43.Perhapsthisisdonewiththeknowledgeofamajor promotionor seasonality.Anydemand adjustment above1produces aninflationaryeffectonthebuffer.

FIGURE8-7PartABCwithnodemandadjustment

FIGURE8-8PartABCwithdemandadjustmentfactor

Inweeks22–25afactorlessthan1isapplied.ThisdeflatestheADUvaluefeedingthebuffer-levelequations,causingatroughtoappearforasmallperiodoftime.Perhapsthiswasdonewiththethinkingthatthemarketwillbesaturatedand demand will drop off temporarily. This is a “de-promotion” effect. Thiscouldalsobeaseasonaleffect.Theapplicationofademandadjustment factordoesnotchange the selectioncriteria forgreenzonesizing.Thegreenzone is

Page 171: DDMRP - Demand Driven Material Requirements Planning

stilldeterminedby the largervalueof theMOQ,ordercycle,orapplicationoftheleadtimeaftertheadjustedADUhasbeenapplied.

Demandadjustmentsareusedforcommonsituationssuchasrapidchangesin demand, product transition, and seasonality. These planned adjustmentsseparatewhatisknown,thepatternofthedemand,fromwhatisunknown,thelevelofactualdemandthatwillbeexperienced.

RapidBufferAdjustmentA demand adjustment factor can be employed to immediately raise or lowerbufferlevelsifthereisarapidchangeindemandthatisindicativeofacomingtrend. This rapid change would be indicated by an ADU alert (discussed inChapter7).

AnADUalertdoesnotdirectlygenerateademandadjustment factor.Thealert should raisequestions that canbeconsideredat theSalesandOperationsPlanningmeeting.Why is the item experiencing significantly heavier demandwithin the recent past?Didweadd a significantnewcustomeroropen anewterritory?Didwerunapromotionthatmarketingforgottotelloperationsaboutit? Is therean impendingperception in themarket fueling thedemand thatweshould capitalize on? Conversely, why is the item experiencing significantlylighterdemandwithintherecentpast?Wasthereanaturaldisasterthatcausedamajor supply chain disruption? Is there negative feedback on social mediaaffecting the product? The answers to these questions could lead to theapplicationof a demand adjustment factor for a periodof timeuntil theADUcalculationnormalizestothenewlevelofdemand.

ProductIntroduction,Deletion,andTransitionDemand adjustment factors will often be used with product introductions,deletions,andvariousformsoftransitions.

ProductIntroduction.Whenintroducinganewproductthatwillbestrategicallybuffered,acompanyhastoestablishabufferposition.Thedemandadjustmentfactors can be usedwith regard to a new-product launch.The productwill beoffered to the market in week 4. The sales and marketing plan for this newproduct calls for the product to be selling 2,000 units per day 12weeks fromnow.That2,000willbecomethebaselinefortheapplicationoftheDAF.Insteadofspendingthecashandcapacitytobringthebuffertofullsizeimmediately,aDAFwillbeappliedstartinginweek2andcontinuetoweek11.Figure8-9 is

Page 172: DDMRP - Demand Driven Material Requirements Planning

thedemandadjustmentschemaforthisnewitem.This planned adjustmentwill ramp up the plannedADUover a period of

time,thuscreatingabufferthatalsogrowsovertime.Additionally,abufferwillbepresentonthe launchdate inweek4 tohandle initialdemand.ThedemandadjustmentfactorcanbeadjustedbasedontherealperformanceagainstthatplanifanADUalertistriggered.Figure8-10depictsthebufferzonesrampingup.

FIGURE8-9Ramp-updemandadjustmentschema

FIGURE8-10Bufferramp-upforthenewproduct

Page 173: DDMRP - Demand Driven Material Requirements Planning

ProductDeletion. Next is the application of the demand adjustment factor inbringingabufferdowninanticipationoftheproductnolongerbeingofferedtothemarket.Figure8-11depictstheramp-downforthisexamplepart.Theramp-downusesaDAFbelow1thatdecreasesoveratimeperioduntilitreacheszero.InthisexampletheapplicationoftheDAFbeginsinweek5andhasazerovaluebyweek9whentheproductisplannedtobediscontinued.

TheADUisrampeddowntocreateagraduallydiminishingbufferlevelandzonedefinition.Figure8-12 illustrates the impact the rampschemahason theproduct’sbuffer.

FIGURE8-11Ramp-downdemandadjustmentschema

Page 174: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-12Bufferramp-downforthediscontinuedproduct

ProductTransition.Twotypesoftransitionmayrequiretheuseofthedemandadjustmentfactor.Oneinvolvesthereplacementofanolderversionofaproductwithanewerversion.InmostcasestheADUoftheolderproductcanbeusedastheADUforthenewproduct.Thenewproductisareplacementandwillmostlikelyhavethesamecustomerbase.Anyuptickindemandduetonewfeatures,forexample,maybeable tobeactualized throughtheADUcalculationperiodoncethenewproductisintroduced.WhiletheADUforbothwillbethesame,becarefultonot“double-buffer.”Additionally,theriskofobsoletestockshouldbeminimized.

In order to manage this transition, both a ramp-up and a ramp-down areemployed simultaneously. The new product will be ramped up, while the oldproductwill be rampeddown. In both cases the effectivity date of both itemswill be used as a culmination point for the ramp-up and ramp-down schemas,which means that the buffers will overlap. Figure 8-13 shows the demandadjustmentschemaforbothproductsovera12-weektransitionperiod.Thenewversionwillbeofferedforsaleonweek9,andtheoldversionwillnolongerbeavailableonweek12.Thusthereisafour-weekoverlapwherebothareoffered(asofttransition).

Figure8-14showsthebufferimplicationsforbothversionsoftheproduct,theoldrampingdownwhilethenewisrampingup.Byweek11thenewproductis at full buffer strength in time for its effectivity date. The adjusted ADU(AADU) is shown for both the new (N) product version and old (O) productversionaswellas thebufferzonevalues.Thenewproduct is ina stacked-barformat,whiletheoldisinastacked-lineformatforcontrast.

Page 175: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-13DemandAdjustmentSchemaover12weeks

FIGURE8-14Demandadjustmentforthereplacementproduct

Acompanythatutilizesthisstrategydoesnotcommitresourcesearlierthannecessary.Thekeyisthatthisexamplehasplannedtheramp-upcurvetobeat

Page 176: DDMRP - Demand Driven Material Requirements Planning

100 percent ofADUby the time the old part is no longer activewhile at thesametimehavinganamountofnewinventoryinadvanceofthatcutoverdate.Thisstrategywillminimizeoreliminateobsoleteinventorywhileallowingforaseamlesstransitiontoanewpartfromthemarket’sperspective.Thiswillreduceoreliminatetherisksofmissedsalesduetoshortagesthattendtooccurthroughpoorlymanagedtransitionsorvariabilityofdemandduringthattransition.

Another form of product transition involves a known and dramaticimpendingshiftinitsdemandupordown.Thisrequiresrampingupordownaspreviously described. Figure 8-15 illustrates an example of an expected andsignificant demand uptick. The row titled “Week” represents the present andfuture inweekly buckets.Week 1 is the currentweek. The row titled “Sales”presentstheactualsalesforweek1andtheprojectedsalesforweeks2–17.Therowtitled“ADU(1Week)”istheADUvaluewithineachweeklybucket,whiletherowtitled“ADU(12Week)”istheADUcalculatedbasedona12-weekpastperiod.Therowtitled“ADUDifferential”showsthequotientbetweentheADUwithintheweekandtheADUbasedonthelast12weeks.

In week 6 (five weeks from the current time period) a significant andprolongedupsurgeispredictedtooccur.Demandsurgesfromapredicted80perweek to 400 and then to a steady 800 per week for the foreseeable future.Perhapsanewstrategiccustomerhassignedacontractcallingforlargeweeklyvolumes. The ADU differential goes from 1 to 4.8 in the span of only twoweeks. Figure 8-16 shows the difference in projected weekly demand(“Demand”),projectedADUwithineachweek(“ADU1Week”),andtheADUbasedonthepast12weeks(“ADU12Week”).Itwilltakeuntilweek17fortheADUtonormalizetothehigherlevelofdemand(theADUdifferentialto1.0).

FIGURE8-15Alargeexpectedupsurgeindemand

Page 177: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-16UnderstatedADUforaprolongedperiod

Thismeansthatthebufferswillbedramaticallyunderstatedforaprolongedperiodoftime.Thenewagreementwillnotbeabletobesupportedundertheseconditions. Figure 8-17 shows the projected lag in the buffer adjusting to theexpectedupsurge.

FIGURE8-17Laggingbufferadjustment

Inthiscaseademandadjustmentfactorcanbeusedtoflexthebufferupinadvance of the impending uptick. The planner will attempt to employ factorsstarting inweek2 thatwill rampthebuffersup to theexpectedrequired level.Byweek7,demandadjustmentfactorshavebroughttheADUdifferentialto1.Oncetheramphasbeenaccomplished,theplannerwillthencutoverthebuffercalculationtotheprojectedADUof114.3andreducetheDAFbackto1.This

Page 178: DDMRP - Demand Driven Material Requirements Planning

cutoveractuallyhappensinweek8.Figure8-18shows thedemandadjustmentfactorsemployedfromweeks2to7inordertoaccomplishtheramp-up.

ThenewADUwillthenbeactualizedoverthehorizon.Isthisforecasting?Yes,but it is tied to a strategic decision and a specific customer agreement.Figure 8-19 shows the buffer reaction to the demand adjustment factor. Theassumption is that thenecessary resourceswillbeavailable toaccomplish thisramp-up.Thebufferpatternprovidesavisibleandrealisticpictureofwhatmustbeaccomplishedtosupportthisnewstrategicdecision.

Notethattheseexamplesareshowingthevalueofbeingabletosimulatetheimpactonbuffersbasedonchangesinpartandprofileattributes.

SeasonalityAnother application of demand adjustment factors occurs with regard toseasonality.Many companies have products with seasonal uplifts and troughsthatmayposechallengestoDDMRPbuffersifnotproperlyaddressed.Tacklingseasonality will involve both ramp-up and ramp-down adjustments. This wasdemonstrated in Figure 8-8. But several interactive considerations should betakenintoaccountwhenconsideringwhenandtowhatextenttoapplydemandadjustmentfactorstocompensateforseasonalityofstrategicallybuffereditems:

Consideration 1. Severity of the seasonality (length andsignificance). The first consideration is the known length andseverity of the seasonal swing. The higher the change and theshorterthewindow,themoreseveretheseasonality.

Consideration2.LengthoftheADUcalculationperiod.Thelengthofthe ADU calculation period must be known and considered inrelation to the severity of the season. The longer the ADUcalculation period and themore severe the season, the higher thelikelihoodthattheADUwillbedramaticallyunderstatedduringtheinitialseasonalperiod.

Consideration3.Past, forward,orblendedADU. Inmanycases, aforward-looking ADU negates the need for a planned adjustmentfactor, as it anticipates the seasonal demand change. A blendedADUapproachmaybereactiveenoughdependingontheseverityoftheseasonandthelengthofthecalculationperiod.Inmostcasesapast-looking ADUwill leave the buffers vulnerable to significantseasonality, and demand adjustment factors will be needed to

Page 179: DDMRP - Demand Driven Material Requirements Planning

compensate.Consideration4.Leadtimesofcriticalcomponents.Longleadtimeitems (including parent items with long lead time components)become a factor in determining when demand adjustment factorsshouldbeapplied.Employingademandadjustment factor too latewill mean that despite the buffers being properly sized for thedemand change, they will be subject to a lag in supply that isequivalent to the supply lead time. For example, if a demandadjustment factor begins to be applied 6weeks in advance of theseasonal period, but the lead time of the item or lower-levelcomponents is 12 weeks, then the buffer will be essentiallyundersuppliedwhilewaiting for the long lead time item to arrive.Thiscouldresultinshortagesthroughouttheseasonalperiodaswellasoveragescomingoutof thatperiod.Thus theapplicationof thedemand adjustment factor must take lead time into account. InDDMRPthisiscalleda“supplyoffset.”Evenifaforward-lookingADUisemployedandtheleadtimeislongerthantheADUperiodcalculation,asupplyoffsetmayneedtobeconsidered.

FIGURE8-18Requireddemandadjustmentfactors

Page 180: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-19Thebufferwithdemandadjustmentfactorsapplied

If aparent itemhasabuffered long lead timecomponent, thenthecumulativeleadtime(asopposedtothedecoupledleadtime)oftheparentshouldbeusedfortheoffset—unlessthecomponenthasasufficient buffer position (typically a shared component) to absorbthe seasonal demand of the parent item without considering thecomponentleadtime.

Consideration 5. Resource capacity. Finally, determining whendemand adjustment factors should be employed may require theconsideration of resource capacity. If the seasonal severity orseasonalramp-upschemaoutstripsthesupplyingcapacity,thentheapplicationofthedemandadjustmentfactormustbepulledforwardintime.

ChickenTruffleSoupSeasonalExample.These considerations are illustratedin a fictitious company called SoupCo. SoupCo makes a type of soup calledChicken Truffle Soup. It is a specialty, low-volume product that has highlyseasonaldemand.Salesofthesoupareabout16,000casesperyear,implyingayear-roundaverageofaround307casesperweek.Butover74percent (11,800of 15,875 cases) of the yearly Chicken Truffle sales occur within a 12-weekwindow. Figure 8-20 shows the seasonal demand profile of Chicken TruffleSoup.

Page 181: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-20ChickenTruffleSoupdemandprofile

SoupCo uses a 12-week past-looking ADU calculation. Figure 8-21illustratestheADUcalculationusingapast-looking12-weekhorizon.Thepast-lookingADUresultsinasignificantlagthatleavesthebufferunpreparedfortheseverityofthedemanduplift.Thismeansstockouts.Additionally,itwillleavealarge amount of excess stock past the peak demand period. This potentiallycouldmeanstockthatwouldgopasttheexpirationdate.Thebuffersimplywillnot have the product available in time for the season. Thenwhen it becomesavailable,thelargepartoftheseasonwillbedone.

The lagof apast-lookingADU is further exacerbatedbya long lead timeassociated with the crucial component of the soup—truffles. To make andpackage Chicken Truffle Soup only takes a week, but the lead time for thetruffles is six weeks. Even if the truffles are buffered, that buffer will bedramaticallyunderstatedbasedonalaggingADU.ThismeansthatSoupComustoffset thedemandadjustment factorbyat least sixweeks inorder to ensure asmooth flow of ingredient supply to support the season. Since the firstsignificantupliftoccursatweek17,thisimpliesthattheDAFwillneedtobegintobeappliedbyatleastweek11.

Page 182: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-21ChickenTruffleSoupADUusingthepast12-weekhorizon

Finally, SoupCo has a packaging line capacity constraint to consider.Chicken Truffle Soup is packed on a dedicated specialty line that has limitedcapacity. This line can only pack 800 cases per week when operating at fullcapacity. For a period of nine weeks (weeks 20–28) the market will beconsumingmorethanthepackaginglinecanpack.Over thatnine-weekperiodthepackaging line canonly pack7,200 cases, andyet demand calls for 8,600cases.That1,400-casedifferencerepresentsnearlytwoweeksofpackaginglinecapacity.Figure8-22showsthecapacityofthepackaginglineindicatedbythedottedlinecorrespondingtotheleftYaxis.

This is also a consideration in determining when to start the DAFapplication.Itshouldbeatleastweek11inordertooffsetthetruffleleadtime.Fromweek11toweek20(theweekinwhichdemandoutstripscapacity),thereisatotaldemandof2,370cases.Withinthatsametimeframe,availablecapacityis5,600cases.Thedifference is3,230casesofavailablecapacity.Thatmeansthat the packaging line has the capability to absorb the capacity shortage ofweeks20–28withinweeks8–19ifnecessary.Analternativecouldbetoapplysmaller demand adjustment factors prior to week 11 in order to create a lesssteepseasonalramp-up.

PromotionalCampaignsStillanotherapplicationfortheemploymentofdemandadjustmentfactorsistocompensateforlargepromotionalcampaigns.Thesepromotionalcampaignsareplannedtoinducedemandsurges.Theycanoftenlastonetothreemonthsandcan be a significant marketing investment. Large promotional campaigns that

Page 183: DDMRP - Demand Driven Material Requirements Planning

create largeamountsofdemanduplift only to result in stockouts arenormallyviewed as disasters. Thus it is crucial that companies adjust their buffers inadvance of these planned campaigns. The same factors are considered inpromotional campaigns as for seasonality. Since a company typically hasexperiencewiththeselargecampaigns,thereisanexpectationoftheamountofsurgeexpected.

FIGURE8-22Consideringcapacityofthepackagingline

ApplyingDAFtoComponentsIfademandadjustmentistobeappliedtoaparentitem,itmaybenecessarytoapplythisfactor to thecomponentbuffers toensureaproperlysuppliedparentbuffer.Ifthecomponentisuniquetotheparent,thenthesamefactorappliedtotheparentcanbeappliedtothecomponent.Ifthecomponentisalongleadtimeitem,thenits leadtimeshouldbeconsideredindeterminingwhentoapplytheDAF(supplyoffset).

If the component is a shared item, then the demand adjustments (or lackthereof)ofallparentsmustbeconsideredindeterminingtheamountbywhichtoadjust thecomponent’sbuffer.Figure8-23showsanexamplewith twoparentitems (FPA and FPB) that have multiple shared components (ICB, PPB, andPPA).Thenumericalstripinthemiddleofthegraphiccorrespondstothelevelintheproductstructureinwhichthepartoccurs.Ifthesetwoparentitemshavedifferent DAFs applied to them, how does it affect the DAF that should beappliedtothecomponents?

ArelevantfactorindetermininghowtheDAFwillchangeatthecomponentlevel is to account for the parent-to-component ratio of each part across all

Page 184: DDMRP - Demand Driven Material Requirements Planning

productstructures.InFigure8-23thatratioisdisplayedintheblackboxtotherightofeachpartname.Forexample,ittakestwoICAsandfourPPCstomakeoneSAA.WhenaDAFisappliedtoaparentitem’sADU,anadjustedADUiscreated.ThisadjustedADUisthenmultipliedbythecomponentratiotogetthecomponent’sadjustedADU.InFigure8-24theratioforeachcomponentinthestructure is depicted in the columns “FPA Ratio” and “FPB Ratio.” PPA isinvolvedinbothstructuresbuthasadifferentratiowithitsdirectparentineachrespectiveproductstructure.

Figure8-24 shows the adjustedADUs for each component item and howthatconvertstoaDAFforeachcomponent.InthisexampleFPAhasaDAFof2(200percent),whereasFPBhasnoDAFapplied(aDAFof1).Ineverycaseofacommoncomponent, thereisablendingthatoccursbasedontheDAFofeachparentandtheratiorelativetoeachproductstructure.ICB’sADUhasadjustedto800,whichisaDAFof1.3ofitscurrentADU.

FIGURE8-23ProductstructureforFPBandFPA

Page 185: DDMRP - Demand Driven Material Requirements Planning

FIGURE8-24TheDAFeffectonFBAandFPBcomponents

ZoneAdjustmentFactor

Anotherway to adjust part buffers is through zonalmanipulations.This couldapplytotheindividualpartoragroupofpartsaffectedinthesamewaybytheadjustment(notnecessarilypartsinthesamebufferprofile).Azoneadjustmentfactor can be applied to any of the three zones of a buffer. The zones of thebufferservedifferentpurposes,sothezoneadjustmentfactorshouldapplytotheappropriatezonebasedontherationalefortheadjustment.

GreenZoneAdjustmentAsdiscussedinChapter7,thegreenzonedeterminesordersizeandfrequency.Thuswhenordersizeandfrequencyneedtobeadjusted,thegreenzonecanbemanipulatedupordown.Adjustmentsupoftenoccurwhenasupplyingresourcewithsignificantsetupissuesencountersacapacityconstraint.Thegreenzoneofcertain itemscanbe raised inorder tocreate less frequentand largerorders tosavethecapacitylostinadditionalsetups.Note,thisisnotraisingthegreenzoneinordertolowerunitcost;itisadeliberatestrategybasedonprotectingtheflowof relevant information and materials through a capacity constraint. Anycapacity saved at that point directly translates to more total output for thesystem,andthatprotectsallbuffersandcustomersfedbythatresource.

Downward adjustments in the green zone might occur when there is

Page 186: DDMRP - Demand Driven Material Requirements Planning

sufficient excess capacity, andmarket responsiveness is the key. This can berequired in some seasonal markets in which a supplying resource, aftercompleting the planned buffer buildup, must carefully watch market demandacrossproductsandquicklyrespondtoreplenishingitemsthatareuniqueforthisseason’shighmovers.ThismeansthatsomeproductsmayhavedifferentMOQsdependingonthechangingcircumstancesinrelationtosupplyingcapacity.

IfthepackaginglineinSoupCowasnotdedicatedtoonlyChickenTruffleSoup,thisstrategycouldbeusedtoaugmenttheuseofaDAF.Forexample,ifSoupCo has a whole line of seasonal truffle soups, larger batches of the fastmovers could be run in the seasonal buildup in order to prevent erosion ofpackaginglinecapacityduetosetups.Withintheseason,smallerbatchescouldthen be employed across the different varieties to respond to this particularseason’spull.

YellowZoneAdjustmentThe yellow zone is the heart of demand coverage in the buffer. As noted inChapter7, the yellow zone is calculated asADU× decoupled lead time. It isassuminga rateofdemandwithina responsewindow. In that regardayellowzoneadjustmentcouldbe triggeredbyknownorplannedevents thatdealwitheitherofthosetwocomponents.Typically,ayellowzoneadjustmentoccursasaresponse to a planned short-term promotional event or a planned or knownsupplydisruption.

A short-term promotional event is one that does not last long enough toemployademandadjustmentfactorschema.Itappearswithinashortwindowoftimewhereasignificantrateofdemandchangeoccursbutthendemandpatternsquickly turn to normal. Thiswindow of time is usuallywithin the part’s leadtime.A factor canbe applied to size theyellowzone to the expecteddemandwithinthattimeframe.ApartwithanADUof1,000andaleadtimeof7dayswouldhaveayellowzoneof7,000.Ifashort-termpromotionaleventlastingaweekwasexpectedtotriplesalesfortheweek,afactorof3couldbeappliedtotheyellowzonetosizeitfrom7,000to21,000.

Another scenario for a yellow zone adjustment occurs when there is aknownorplannedinterruptioninsupply.Maybeacriticalsupplierislocatedinaregionthathasanextendedholiday,orperhapsthereisaplannedupgradetoafacility. These events translate to the source being unable to respond for awindowoftime.Forexample, if theSoupCopackinglinewasscheduledtobedownforoneweek,ayellowzoneadjustmentfactorof2woulddoublethesize

Page 187: DDMRP - Demand Driven Material Requirements Planning

of theyellowzone to cover the twoweeksofdemand.Thebufferwould thenreturntonormal.

RedZoneAdjustmentThe red zone is the embedded safety in the buffer. The red zone should beadjustedwhen there is a knownor plannedbut temporary change in volatilitythat does not warrant moving the part(s) to another buffer profile. This istypically associated with a temporary change or transition with regard topurchasing,producing,ordistributinganitemorgroupofitems.Perhapsanewresourceormaterialisbeingbroughtonlinethatcouldcausemoredisruptionsintheshorttermthanwhatthebufferiscurrentlysizedfor.

LeadTimeAdjustmentFactor

A lead time adjustment factor could apply to an individual part or a group ofparts affected in the same way by what is prompting the adjustment (notnecessarilypartsinthesamebufferprofile).Theuseofaleadtimeadjustmentfactorcoincideswithaplannedorknownexpansionoftheleadtimesofagroupof items. For example, if amajor transportation route to awarehousewill betemporarily disrupted due to construction, a lead time factor may need to beappliedtoproductsatthewarehousethatcomethroughthatrouteoritsalternate.Ifthepreviousleadtimewastwodaysbutthenewleadtimeisestimatedtobethree days, on average, then a lead time adjustment factor of 1.5 should beemployedforthedurationoftheconstructionproject.Analternativewouldbetoamendthepartmastertoreflecttheleadtimechangedependingontheexpecteddurationforthisnewleadtime.

Summary

The third component ofDDMRP is strategic buffer adjustments.Recalculatedadjustmentsallowthebufferstorecalculatetheirrespectivelevelsbasedonkeyattribute changes of which first and foremost is average daily usage. Otheradjustments tobuffer levels canbe related toknownorplannedevents.Thesearecalledplannedadjustments.Withregardtotheseevents,thebufferswillbeeither significantly overstated or understated if not adjusted. There are manytypesofplannedadjustmentstrategiestobeemployeddependingonthespecificnatureoftheplannedorknownevent.

Page 188: DDMRP - Demand Driven Material Requirements Planning

CHAPTER9

DemandDrivenPlanning

ThefourthcomponentofDemandDrivenMaterialRequirementsPlanningisaproven and intuitive method of supply order generation—demand drivenplanning. In addition to lead time compression and variability dampening, thebuffersplacedatthedecouplingpointsaretheheartofsupplyordergenerationforDemandDrivenMRP.Theybecomea focal point for creating, promoting,protecting,anddeterminingrelevantinformationandmaterials.Theyalsocreatetheopportunityforamoreelegantandvisiblewaytogeneratesupplyorders.Itstarts with the consideration of what really is relevant information from ademandperspective.

TheShifttoActualDemand

Protectingtheflowofrelevantinformationandmaterialsisthekeytoprotectingandpromotingreturnoninvestmentperformance.Fromaplanningperspective,therightmaterialswillnotbeavailablewithouttherightinformation.Yetmostconventional planning starts with information that is suspect. The directconnectionbetweenforecasteddemandandsupplyordergenerationmeansthatsupplyordersarebeinggeneratedwithsignalsthatareknowntobewrong.Thisties up cash, capacity, space, and timeby dealingwith the resulting irrelevantmaterialsandforcesadditionaleffortstoattempttogettherelevantmaterials.

Themost relevantdemandsignal isasalesorder. It isaknownandstateddesire from a known customer to buy. Yet using this most accurate piece ofdemandinformationhaseludedplanningprofessionalsfordecades.Asdescribedin Chapter 4, decoupling opens the door for its effective use. This highlyaccurate form of demandwill provide the demand input into a daily planningequation for each buffered position. This equation is called the net flowequation.

Page 189: DDMRP - Demand Driven Material Requirements Planning

TheNetFlowEquation

Previously referred to as the “available stock equation” in DDMRP literaturebefore 2016, the net flow equation provides the supply order generationrecommendation signal (timing and quantity) for buffer replenishment. It is akey and unique aspect of DDMRP and should be performed daily on alldecoupled positions. The net flow equation is elegant and intuitive whileencompassing all the ranges of planning that most experienced planningprofessionalsareconcernedaboutwithregardtosupplyordergeneration.

Thenetflowequationissimple:

On-hand+on-order–qualifiedsalesorderdemand=netflowposition

Figure 9-1 illustrates the components of the net flow equation, and thefollowinglistexploresthecomponentsonebyone:

On-hand.Thequantityof stockphysicallyavailable. InFigure9-1thisisrepresentedbythelargeboxlabeled“On-Hand”inthemiddleof the graphic. The smaller boxes represent the actual quantity ofunitsonhandandavailableforuse.

On-order. The quantity of stock that has been ordered but notreceived. InFigure9-1 this is representedby the truckon the left-handsideofthegraphicthatisheadingtowardtheon-handposition.The smaller boxes in the trailer represent the amount of units onorder. This could be a single incoming order or several incomingorders. The on-order quantity is the total quantity that has beenorderedbutnotreceived,irrespectiveoftiming.

Qualified sales order demand. The sum of sales orders past due,sales orders due today, and qualified spikes. In Figure 9-1 this isrepresentedbytheordersthatarehighlightedin“Today”and“Day3.”Thereisnopastdueamountrepresentedinthisfigure.Therearetwo sales orders due today and three sales orders that havecombined to create a qualified spike. These two days of qualifieddemand are added together to get the total amount of qualifieddemandfortoday’scomputationofthenetflowequation.

Figure9-2givesanexampleofthenetflowequation.Thereare6,412unitscurrently available in stock (on-hand). There are 2,468 that are inbound (on-

Page 190: DDMRP - Demand Driven Material Requirements Planning

order).Thereare312unitstobeshippedouttodayand807unitsduetoshiponday3.Today’snet flowequation for thispart is6,412 (on-hand)+2,468 (on-order)–1,119(qualifiedsalesorderdemand)=7,761(netflowposition).

FIGURE9-1Illustratingtheelementsofthenetflowequation

FIGURE9-2Anexampleofthenetflowequation

The net flow equation answers all the question of any robust planningequation:

WhatdoIhave?Theon-handvalue.Whatiscomingtome?Theon-ordervalue.WhatdemanddoIneedtofulfillimmediately?Salesorderspastdueandduetoday.

Whatfuturedemandisrelevant?Qualifiedfuturespikes.

At this point a significant question still needs to be answered about thedemandcomponentofthenetflowequation.Salesordersdueinthepastarethesummationof thequantityof stock representedbysalesorderswehaveyet tofulfillbutwereduebeforetoday.Salesordersduetodayarethetotalquantityofstock represented by the sales orders scheduled to be fulfilled today. Futurequalifieddemandconsiders twoconditions. InFigure9-2 theday3 totalorderquantityqualifiesasaspikeamount(807). It isnot thenumberofsalesorders

Page 191: DDMRP - Demand Driven Material Requirements Planning

thatistriggeringthespikebutthetotalquantityofstockthatthosesalesordersrequirewithinaday.

QualifyingOrderSpikes

Twoconditionsmustbemetinordertoqualifyasanorderspike.InDDMRP,anorderspikeisaqualifyingquantityofknowncumulativedailydemandwithinaqualifyingtimewindowthatthreatenstheintegrityofthebuffer.Thatmeansthequalifying level (order spike threshold)and thequalifying timewindow(orderspikehorizon)mustbedefined.

Condition1:TheOrderSpikeThresholdThe order spike threshold is a level that qualifies a spike in a particularenvironment.Thesalesordersforthesamepartnumberforeachdayaretotaledand compared against this threshold. If the sum is greater than the threshold,thentheentireamount(notjusttheamountabovethethreshold)isincorporatedintotheavailableflowequationasaqualifiedspike.AnorderspikethresholdisdepictedinFigure9-3bythehorizontaldottedline.Salesordersforeachdayarerepresentedby theboxesabove thecolumn.Forexample,day6has foursalesordersduethatday.Thecumulativequantityofthosesalesordersisrepresentedbythecolumnbeneaththesalesorders.Thethreesalesordersduetwodaysfromtoday (day3) representenoughcumulativedemand toqualify as a spike.Thisalsooccursonday9,wheretwosalesorderscombinetoformaqualifiedspike.

Page 192: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-3Anorderspikethreshold

Theentirecumulativeamountofthespikeisqualifiedandincorporatedintothe equation.This typically prompts thequestionofwhy thewhole amount istaken andnot just theportion above the threshold.The entire spike amount istakentoguardagainstsuccessiveorclusteredspikescreatinganoverwhelmingsurgethatwouldcompromisebufferintegrityiftheentireamountwerenottakeninto account. Taking the whole spike represents the most conservative risk-averseapproach.

Settingthethresholdlevelinvolvesdeterminingalevelofdailydemandthatjeopardizes the integrity of the buffer. Three alternatives will be explored indeterminingthethresholdlevel.Thefirsttwosetthethresholdinrelationtotheredzoneofthebuffer.Sincetheredzoneistheembeddedsafetyinthebuffer,thenspikescanbeseenrelativetotheirpotentialconsumptionofthatsafety.InearlyDDMRPimplementationsadefaultheuristicspikethresholdof50percentoftheredzonewasutilized.

An alternativemethod relates to the border of the redbase and red safetyportionof thebuffer for finished items.Theredsafetyportionof the redzonedirectlyrelatestothevariabilityofthepartposition.Ifthefinisheditemisinthehigh variability category, then that part is subject to frequent spikes.Thus theorderspikethresholdcouldbesetfortheamountequivalenttotheredzonebasevalue.Anyvalueabovethatbaseamountpenetratesintothesafetyportionoftheredzone.

Still another way to determine the order spike threshold has to do withqualifyinganorderspikeinrelationtotheADUofaspecificpart.Thisisoftenthe most intuitive way for planners and buyers. It can also be historicallyvalidated.

Page 193: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-4DifferentwaystosettheorderspikethresholdforFPA

Figure9-4 shows the different order spike threshold options for part FPAfrom the CompanyABC example fromChapters 6 and 7. The total red zonevalue (red base + red safety) after compressionwas 875. The version labeled“OST(50%)”istheorderspikethreshold(OST)setat50percentofthetotalredzone (437.5). The version labeled “OST (RS)” is theOST set at the red basevalue(700).Theversionlabeled“OST(ADU)”istheOSTsetatthreedaysofaveragedailyusage[3×ADU(250)=750].

These alternatives are options for planning teams to evaluate how theywouldliketodetermineorderspikeamountsintheirspecificenvironment.Eachoption will most likely yield a different number of qualified spikes over anextendedperiod of time. In this example, themost qualified spikeswill occurwiththeOST50percentmethod.ThusinmostcasestheOST50percentmethodistypicallythemostconservativemethodoforderspikethresholddetermination.Thiswilltendtoyieldslightlyhigherlevelsofoverallinventorybutbestprotectservicelevels.Inreality,thedifferenceineffectivenessbetweenthemethodsisnegligibleiftheDDMRPmodelissensiblyconstructed,managed,andadapted.

No matter what method is used for OST determination, the OST shouldalwaysbewithinthetotalredzonevalue.Theredzoneisthesafetyembeddedinthebuffer.HavinganOSTvaluegreaterthanthattotalsafetyopensthedoorforinsufficientcoveragefromthebuffer.

Condition2:TheOrderSpikeHorizonThe secondconditionofqualifyinganorder spike is to set aqualifying future

Page 194: DDMRP - Demand Driven Material Requirements Planning

timewindowcalledanorderspikehorizon(OSH).Thishorizonisawindowoftime within which cumulative daily demand can qualify as a spike if thatcumulative daily value is above the OST. If the cumulative daily demand isoutsidethatwindow(fartherintothefuture),thenthatcumulativedailyamountwillnotbequalifiedinthenetflowequation.

TheOSHshouldbesettoatleastonedecoupledleadtimeinthefutureforthe buffered part. This allows enough time for the part’s buffer to properlycompensate for the spike. If the order spike horizon is beyond the sales ordervisibility horizon, it simply means that qualified spikes will routinely appearwithintheorderspikehorizon.Inthiscasethepart’sbufferprofileshouldbesettoahighervariabilitycategoryinordertocompensateforthislatequalification.

Figure9-5combinestheuseofanorderspikethresholdandanorderspikehorizon.TheOSH is representedby theboxwithhatched linesoverlaying thedaily sales order quantity positions for a period of seven days (today plus sixdaysinthefuture).Inthiscasethedecoupledleadtimeofthepart issixdays.ThustheOSHistoday+DLTofthepart.Whenthehorizonisimposed,theday9cumulativedemandisnotqualifiedin thenetflowequation.Itwill taketwomoredaysbeforethatdemandqualifiesasaspike.

Yetthisbegetsanadditionalquestion.Ifwearesoconcernedwithrelevantdemand signals and sales orders represent themost accurate form of demandsignal, then why not take all known demand into account in the net flowequation?

Allknownsalesordersarenotincludedinthenetflowequationbecausethatdemandisessentiallyalreadyaccountedforinthebuffer.Ifthedailysalesorderdemand is under the threshold, those orders represent normal or averagedemand.Thebufferswerebuiltusingthisaveragerateofuse.Thuswhatisduetoday,whatwasdueinthepast,andwhatqualifiesasaspikearereallyallthatisrelevantfromademandperspective.

Figure9-6depictswhatanorderspikequalificationscreencouldlooklikeina DDMRP-compliant information system. The part number (“EXAMPLE”),orderspikehorizon(6days),orderspikethreshold(100),andtoday’sdate(6/21)aredisplayedat the topof the screen.The informationdisplayed in the screencorrespondstotheexampleinFigure9-5.Figure9-5hasbeenaddedbeneaththesamplescreeninordertoshowhowthetwoareconnected.Inthesamplescreen,futuredaysaredisplayedinandescendingmanner.Thesalesorderswithintheorderspikehorizonhaveshadedboxes.

Page 195: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-5Theorderspikethresholdandhorizon

Page 196: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-6Sampleorderspikequalificationscreen

There are two sales orders (1234 and 1235) totaling 75 units that are duetoday.Thespiketwodaysfromtoday(6/23)iscomposedofthreesalesorders(1237, 1238, and 1239) totaling 180 units. That means that today’s totalqualifieddemandforthenetflowequationis255(0pastdue+75duetoday+180futurespikedemandonday3).Thereisanotherspikeon6/29madeupofsales orders 1250 and 1251 totaling 140 units but is outside the horizon andlabeledassuch.

An additional horizon and threshold consideration for finished items thathave large and known dependent demand orders against them in the remote

Page 197: DDMRP - Demand Driven Material Requirements Planning

futurecanoccurwithlargeplannedpromotionaleventsinwhichcustomershaveagreedtotakeasignificantamountofstockwithinashortwindow.Theeventisplanned, agreements have beenmade, and the demand is real. In this case anadditionalOSH andOST should be formulated.Wewill call themOSH2 andOST2.Theyneedtobeusedincombinationwitheachother.

TheOSH2canbeincreasedtothecumulativeleadtimeoftheparentpartinordertoaccountforthespikeagainstbufferedcomponentpositionsthatwouldbeunabletoabsorbtheupliftwithinthedecoupledleadtimeoftheparent.TheOST2willalsoincreaseandbeappliedtotheOSH2periodonly.

Figure9-7depicts thecombineduseofa close-inorder spikehorizonandthreshold(OSH1andOST1)andanOSH2andOST2.Theclose-inorderspikethreshold (OST1) is the solid line terminating at day 7 with a solid circle(OSH1).Thecumulativeleadtimeofthisfinisheditemis30days,indicatedbytheopencircleat theendof thedottedOST2line.The long-rangeorderspikethreshold(OST2)is500units.BeyondtheOSH1salesorder,demandbeginstobe spotty.Most customers have simply not ordered yet. But the known largeorders(1,000each)tosupporttheplannedpromotionarevisibleondays28and29.Theseordersaresignificant,and ifallowed toonlybequalifiedwithin theOSH1, they would overwhelm the finished buffer and the component bufferssupportingit.

FIGURE9-7Short-andlong-rangeorderspikethresholdsandhorizons

Page 198: DDMRP - Demand Driven Material Requirements Planning

Inthiscaseboththedemandsonday28andday29wouldbequalifiedasdemand in thenet flowequation.Thisamountof2,000wouldbeadded to the180(salesorders1237,1238,and1239)withintheOSH1andtheamountof75due today (sales orders 1234 and 1235) to create a total qualified demand of2,255.

Thisalsomeans thatanypromotionsor large-ordernegotiationsshouldbefinalizedinadvanceoftheOSH2horizon.Withoutthiscommitment,thebufferswillalwaysbeatrisktosignificantorders.

Onefurthernoteonspikequalificationinvolvesthecontinuedqualificationof spikes. Spikes continue to be qualified each additional time the net flowequationisperformed(typicallyonceaday).Figure9-8continues theexamplethatwehavebeenusing for spikequalificationon6/22 (day2), oneday laterthan the original date of the example. Today’s qualified demand within theOSH1is40(salesorder1236)plusthequalifiedspikeof180(salesorders1237,1238,and1239)foratotalof220.Thenextday(6/23)willresultinanewspikequalificationof140(salesorders1250and1251).Thatspikewillbecombinedwith the 180 due that day (sales orders 1237, 1238, and 1239) for a totalqualifieddemandof320withintheOSH1.

Cumulativedailyspikescontinuetobequalifiedeachdayfromthetimeinwhich they were first qualified in order to balance against any on-orderquantities that have been generated as a result of the spike inclusion. Todisregardthespikeafterqualifyingitonitsfirstdaywoulddistortthenetflowposition and make the buffer appear to be oversupplied. This will bedemonstrated later in this chapterwith the simulationof thenet flowequationoveraseriesofdays.

Page 199: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-8Qualifyingdemandon6/22withintheOSH1

SupplyOrderGenerationBasedonNetFlowPosition

Eachtimethenetflowequationiscomputedforeachbuffer,ityieldsanumberthatiscalledthenetflowpositionwithineachbuffer.Thenetflowpositionwilldictate whether a supply order will be recommended against the bufferedposition.Ifthepositionisbelowthetopofyellow(TOY),thenasupplyorderisrecommendedforaquantitythatisthedifferencebetweenthenetflowpositionandthetopofgreen(TOG).Figure9-9illustratestheTOG,TOY,andtopofred(TOR)positionswithinthe401PbufferfromtheCompanyABCexamplefromChapters7and8.

TOGisthesummationofallthezonesofabuffer.Inour401Pexamplethisis10,888.TOYis thesummationof theyellowandredzones,whichis8,938.TOR is the full quantity of the red zone, which is 2,438.When the net flowpositionisatorbelow8,938,asupplyorderwillbegeneratedtorestorethenetflowpositiontoTOG.Thismeansthatthetypicalorderwillbeatleastthefullamountofthegreenzone(1,950)orgreater.

Considerthefollowinginformationfor401P:

On-hand=2,652On-order=6,233

Qualifiedsalesorderdemand=712

Page 200: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-9TOG,TOY,andTORvaluesfor401P

Withthisinformation,thenthenetflowpositionfor401Ponthisparticulardayis8,173.ThismeansthenetflowpositionisbelowTOY.Figure9-10showsthe401Pnetflowposition.

TheTOGvalue of the 401P buffer is 10,888.This creates a supply orderrecommendation of 2,715. Once this supply order is approved, the net flowpositionchangesto10,888.Theorderrecommendationhasnowbecomeanon-orderquantity.Thisincreasesthetotalon-orderquantityto8,948(2,715newon-orderquantity+6,233oldon-orderquantity).Thenetflowequationafterorderacceptance is 2,652 (on-hand) + 8,948 (on-order) – 712 (qualified sales orderdemand).

Onacceptance,theorderisassignedarequestdateonedecoupledleadtimeintothefuture.Component401Pisapurchasedpart.Assuch,itsdecoupledleadtimeisthesameasitspurchasingleadtime.Theorderwouldbeassignedaduedate10daysfromthedateoftheapproval.Thetypeofcalendarabusinessusesmustbeaccountedforincalculatingtheduedate.

Additionally,thenetflowpositionistypicallydisplayedasapercentageofTOGandwiththezonecolorthatthepositionfallswithin.Intheprevious401Pexample,thenetflowpositionwouldbedisplayedas75.1percentandbecodedyellow. Figure 9-11 illustrates what a basic DDMRP planning screen should

Page 201: DDMRP - Demand Driven Material Requirements Planning

display,includingthebasicelementsofthenetflowequationandsupplyordergeneration logic as well as the lead time and due date for any recommendedorders.Thefigureusesthedatafor401PinFigure9-10.

FIGURE9-10Netflowpositionof401P

FIGURE9-11DDMRPsampleplanningscreenwith401Pdata

Note that today’sdate is July15.Anorder recommendation for2,715hasbeen made with a request date of July 25. This is 10 days from today. Thecolumn titled “Planning Priority” is the net flow position expressed as apercentageof the topofgreen.Theboxhas ayellow shade, denoting that thepercentage is inside the yellow zone. There are two expressions of bufferpriority.Oneisdiscreteandexpressedasapercentage.Theotherisgeneralandisexpressedasacolor.Thusforeachpart,aplannerorbuyercanquicklygeta

Page 202: DDMRP - Demand Driven Material Requirements Planning

sense of the part’s status relative to its planned buffer position.While this isquite powerful on a part-by-part basis, the true power of this view can reallyonlybeappreciatedwhenseenwithmultipleparts.

Figure 9-12 is the same planning screen but now with multiple bufferedparts.Thesequence isdeterminedby theplanningprioritycolumnpercentage.Thelowerthepercentage,thehighertheplanningpriority.Thissequencingnowgives a general and discrete sense of relative priority across multiple partscallingforresupply.Thisiscrucialwhenlimitationsorconstraintsarepresentinan environment. When dollars, time, space, and resource capacity are at apremium,itisextremelyadvantageoustobeabletoquicklyfocusonthehighestpriority.

With this view we can quickly see that there is a buffered position inimmediatetrouble.Part406Phasanetflowpositioncoloredredandisat19.8percentofthetopofgreen.Asupplyorderfor2,606isrecommendedandshouldbeimmediatelyapproved.Inthiscasewecanalsoseethat401Phasthelowestrelativepriorityamongallpartscallingforsupplyordergeneration.Wealsoseeapart(404P)thathasnorecommendedsupplybecauseitsnetflowpositionisinthegreenzoneat97.6percentofthetopofgreen.Inmostcasesthispartwouldbe filtered out of the planner or buyer view because no planning action isrequired.

This relative priority distinction is a crucial differentiator between theconventionalMRP planning alerts and actionmessages and the highly visibleandfocusedDDMRPapproach.ConventionalMRPisabinarysystem.YouareeitherOKornotOKwithregardtoeachpart.Thereislittlesenseofhowpartscomparewitheachother—youneedtoeitheractornotact.UndertheDDMRPapproach, planners and buyers can quickly judge the relative priority withoutmassiveamountsofadditionalanalysisanddataqueries.

FIGURE9-12DDMRPplanningscreenwithmultiplebuffereditems

Page 203: DDMRP - Demand Driven Material Requirements Planning

Sometimes therecommendedorderquantitycouldexceed the topofgreendue to order multiples. In these cases, supply order generation rules can beconstructedtoincorporatetheordermultiplebutminimizegoingoverthetopofgreen (OTOG). An example would be to limit the inclusion of an additionalordermultipletosituationsinwhichtheOTOGnetflowpositionwouldbelessthanthequantityunderthetopofgreenwithouttheinclusion.

Oncetoday’sordershavebeenfulfilledandanyon-orderhasbeenreceived,on-hand will be adjusted accordingly. This will create an ending on-handinventorythatwillbeusedfortomorrow’snetflowequation.Additionally,ifon-order is received andconverted toon-hand, then theon-orderquantitywill beadjusteddownfor tomorrow’snet flowequation. Ifon-order is receivedbut isunderaqualityorinspectionhold, itshouldstillbetreatedason-order,asit istrulynotavailableason-hand.

Thefollowingsimulationdemonstrateshowastrategicallybufferedpositionbehavesoveraseriesofdays.

SimulatingDDMRPSupplyOrderGeneration

Nowthatthenetflowequationanditsroleinsupplyordergenerationhavebeendiscussed, a simulation can be run to demonstrate the daily use of a net flowequationagainstabufferedposition.Figure9-13containstheinformationforapartcalled“Example.”Thisparthasanaveragedailyuseof10andadecoupledlead time of 7 days and is in the MMM buffer profile category—it is amanufactured item,withmedium lead time andmedium variability associatedwith it.The lead time factor is0.5, and thevariability factor is0.5.Theorderspike threshold has been set to 50 percent of the red zone (26) over an 8-dayorderspikehorizon.Thismeans thatanorderspike isdefinedas2.6 times theaverage daily usage or greaterwithin any future daily bucketwithin the orderspikehorizon.

Page 204: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-13Simulatedreplenishedpartinformation

Asimulationwillberuninwhichtheactivitiesagainstthisbufferedpositionfor21dayswillbedescribed.

Figure 9-14 demonstrates the starting situation for our example part. Thebufferisinthemiddleofthegraphic.Salesorderdemandisseenoveran8-dayhorizon to the right of the buffer (labeled “Demand Side”). This horizoncorrespondstotheorderspikehorizon(DLT+1day).Theorderspikethresholdof26 isdepictedby thedotted linehalfwayup the redzoneof thebuffer andextendingover theorderspikehorizon, terminating inasmallcircleonday8.Todayisday1,indicatedbythenumber1onthearrowtotherightofthebuffer.Thearrowispointingtowardthebufferbecauseitrepresentsdemandquantitiesin daily buckets coming at the buffer. On day 1we see a demand of 10 dueindicated by the light gray bar.All bars are intended to be to the scale of thebuffer level. On day 5 we see actual known demand for 5. As time movesforward1day,thenumber1willbereplacedby2onday2.

Ontheleft-handsideofthebuffer,weseeanotherarrowflowingfromlefttoright towardthebuffer(labeled“SupplySide”).Thisrepresents theinboundsupply to the buffer in daily buckets (represented by dark bars). The value ofeachsupplyorderisspecifiedonthebar.Thevaluesonthearrowarenegativebecausethearrowisindicatinghowmanydayslefttothesupplyorderreceipt.There is anorder for35pieces thatwillbe receivedbyendofday tomorrow.Tomorrowallsupplybarswillshiftrightby1day.Thesupplyfor35willthenbeat–1onthesupplyside.

Additionally, below the buffer and the supply and demand sides is theplanning screen for our example item. In this case the net flow position has

Page 205: DDMRP - Demand Driven Material Requirements Planning

prompted a supply order recommendation. Today’s starting on-hand is 65,indicatedbythedottedlineinsidethebufferintheloweryellowzone.Therearetwoopensupplyorderstotaling72units(ordersof37and35).Oneorderfor35isduetomorrow,whiletheotherisdue6daysfromnow.Thereisalsoqualifieddemand in at least two daily buckets. There are 10 due today and there is aqualifiedspikeon the8th for30.There isnopastdueamount.Totalqualifieddemandis40[10+30].Thenetflowpositionis97,indicatedbyasolidlineinthe upper half of the yellow zone. This net flow position yields a yellowplanningpriorityof61.8percent[97(netflowposition)/157(topofgreen)].

FIGURE9-14Simulationday1

Asupplyorder for60 is recommendedandwill be approved.The requestdatefortheorderis7daysfromtodayonthe8th.Thatnewlycreatedorderwillshowupinday2(tomorrow)asopensupply.Finally,theordersduetodaywillbeshippedout.Therearenoreceiptsoccurringtoday,sotoday’sendingon-handinventorywillbe55units[65(startingon-hand)–10(unitsduetoday)].

Day 2 is represented by Figure 9-15. The starting on-hand quantity is 55(yesterday’sendingon-handquantity).Herewecanseethatthedemandsidehasshiftedby1daytotheleft.Thedemandarrownowrangesfromthe2ndtothe9th,andtheorderspikehorizonspanstothe9th.Onthesupplysideweseetheorder for 60 that was created yesterday. There are now three orders of opensupply totaling 132 units. A supply of 35 will be delivered today, while ademandof18willbefulfilled.Thatmeansthatendingon-handinventoryforthedaywill riseby17 (35–18)up to72.Thenet flowposition is at139and issolidly in the green zone (88.5 percent of top of green). Thus there is no

Page 206: DDMRP - Demand Driven Material Requirements Planning

additionalrecommendedquantity.The spikedueon the8th continues tobequalified aspart of thenet flow

equationbecauseitbalancesagainsttheopensupply.Ifthespikewereremovedfrom thenet flowequation, thenet flowpositionwould riseby30 to169andshowover the topofgreen.Thismay lead to the impression that thebuffer isoversuppliedwheninfactitisnot.

Figure9-16showsthecontinuationofthesimulationonday3.Supplyhasshiftedtotherightby1day.Theorderfor35wasreceivedonday2andisnolongervisible.On-ordernowtotalsat97(60+37).Demandhasshifted to theleftby1day.Qualifieddemand isat47 [17 (due today)+30 (qualified spikedueonthe8th)].Startingon-handisat72.Today’snetflowpositionisat122(72+97–47).Thispositionat122isayellowplanningpriority(77.7percentoftopofgreen).Anewsupplyorderisrecommendedfor35[157(topofgreen)-122(netflowposition)].Thisnewsupplyorderwillhavearequestdateof the10th(7daysfromtoday).Endingon-handwilldropby17,asnosupplyreceiptswilloccurand17willbesentoutthedoor.

FIGURE9-15Simulationday2

Page 207: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-16Simulationday3

Figure9-17showsthecontinuationofthesimulationonday4.Startingon-handinventoryisat55.Theexistingsupplyordershaveshiftedtotheright,andanewsupplyorderfor35isnowvisible.Totalon-orderis132.Thedemandsidenowrangesfromthe4thtothe11th.Thereare6unitsduetoshiptoday,andthespike continues to be qualified. Total qualified demand is 36. The net flowpositionisgreenat151(96.2percentoftopofgreen).Thereisnosupplyorderrecommended. On-hand will drop by 6 to 49, as no incoming supply will bereceived.

Figure9-18showsthecontinuationofthesimulationonday5.Startingon-handinventoryisnowat49.Supplyordershaveshiftedtotherightby1dayandstill total 132. The demand side now ranges from the 5th to the 12th. Thequalified demand total is 35 [5 (due today) + 30 (spike)]. Today’s net flowpositionisgreenat146(93percentoftopofgreen).Thereisnorecommendedsupplyorder.

Accordingtothenetflowequation,thebufferisproperlyplanned,andyettheon-handis in theredzone. Is thisaproblem?This is thefirst timewecanactually see thedifferencebetween theplanningandexecutionperspectivesofDDMRP.ThebufferwasbuiltusinganADUvalueof10.Thatmeansthatwitha current on-hand of 49 there are nearly 5 days of average coverage stillcontainedinthebuffer.Thenext3daysarelightondemand,andthereisopensupplyscheduledtobereceivedintheshortterm.

Page 208: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-17Simulationday4

FIGURE9-18Simulationday5

On-handhaspenetrated into the redzone,but thisviewof the redzone isbuiltforplanning.Ifthenetflowpositionhaddippedintotheredzone,thenthatisabigproblem.On-handdippingintotheredisexpected.Theredzoneispartofthebufferandisintendedtobeused.Ifon-handdipstoofarintothebuffer,atsomepointitbecomesaproblem,butonday5thisisnotthecase.Ithasonlyslightly penetrated the red zone, has light upcoming demand, and has asignificant amount of open supply due in the near term. Planners at this time

Page 209: DDMRP - Demand Driven Material Requirements Planning

shouldnotexpediteexistingsupplyorlaunchaneworder,asthisissimplynotrequired.ExecutionurgencyiscoveredindepthinChapter10.

Theon-handsituationwillerodefurther,as5areduetodayandnoreceiptswilloccur.Endingon-handwilldropto44.

Figure9-19showsthecontinuationofthesimulationonday6.Startingon-handisat44.Thatnumberstillrepresentsover4daysofaveragecoverage,anda supply order is due to be received today. There is no recommended supplyorder, as the net flow position is green at 137 (87.3 percent of top of green).Endingon-handis72,as37arereceivedandonly9areshippedout.

Figure9-20showsthecontinuationofthesimulationonday7.Theon-handposition of the buffer is now 72. On-order has dropped to 95, and qualifieddemandisat40.Thereisnorecommendedsupplyorder,asthenetflowpositionisgreenat127(80.9percentoftopofgreen).Endingon-handwillbeat62,as10areshippedoutandnonearereceived.

Figure9-21showsthecontinuationofthesimulationonday8.Startingon-handis62.On-orderis95,andqualifieddemandisonly30(thespikehasfinallycome due, and there are no additional spikeswithin the order spike horizon).Thereisnoadditionalsupplyorderrecommendation,asthenetflowpositionisgreenat127(80.9percentoftopofgreen).Endingon-handwillactuallyriseto92despite the fulfillmentof the spike.This isdue to the receiptof the supplyorderof60.

Figure9-22showsthecontinuationofthesimulationonday9.Startingon-handisat92,andon-orderconsistsofonly35.Onthedemandsidealargedrop-inorderof30hasbeenacceptedfordeliverytomorrow.This30isontopofthe6thatwasalreadyordered.Thiscreatesaqualifiedspike.Thedemandquantitydueon the10thwill nowbe included in thenet flowequation.Thus the totalqualifieddemandis41.Thenetflowpositiononday9is86[92(on-hand)+35(on-order)–41 (qualifieddemand)].Thisactuallyplaces thenet flowpositionbelowtheon-handposition.Asupplyorderfor71witharequestdateofthe16thisrecommended.Endingon-handinventorywillbe87,as5areshippedoutandnonearereceived.

Page 210: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-19SimulationDay6

FIGURE9-20Simulationday7

Page 211: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-21Simulationday8

FIGURE9-22Simulationday9

Page 212: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-23Simulationday10

Figure9-23showsthecontinuationofthesimulationonday10.Startingon-hand is at 87.On-order now sits at 106 (35ofwhichwill be received today).Qualified demand sits at 36 (including the drop-in order approved yesterday).Thereisnosupplyorderrecommendation,asthenetflowpositionisatthetopofgreen(157).Endingon-handwillbe86,as35arereceivedbut36areshippedout.

Figure9-24showsthecontinuationofthesimulationonday11.Thestartingon-handposition is at86.Despitehavinghada largedrop-inorder, thebufferseems to be in excellent shape.On-order is 71, and qualified demand is at 9.Thereisnorecommendedsupplyorder,asthenetflowpositionisgreenat148(94.3percentof topofgreen).Ending inventory is77,as therearenoreceiptsand9arefulfilled.

Figure9-25showsthecontinuationofthesimulationonday12.Thereisnorecommended supply order, as the net flow position is green at 138 [77 (on-hand)+71(on-order)–10(qualifieddemand)].Endingon-handwillbe67.

Page 213: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-24Simulationday11

FIGURE9-25Simulationday12

Figure 9-26 shows the continuation of the simulation on day 13. The netflowposition [67 (on-hand)+71 (on-order)–20 (qualifieddemand)]yields arecommended supplyorder for39with a requestdateof the20th.Endingon-handwillbe47.

Figure9-27showsthecontinuationofthesimulationonday14.Startingon-handis47.Thenewsupplyorderfor39isnowvisibleandcombineswith theexistingorderof71foratotalon-orderamountof110.Qualifieddemandisonly

Page 214: DDMRP - Demand Driven Material Requirements Planning

at6.Thereisnorecommendedsupplyorder,asthenetflowpositionisgreenat151(96.2percentoftopofgreen).Endingon-handwillbe41.

FIGURE9-26Simulationday13

FIGURE9-27Simulationday14

Figure9-28showsthecontinuationofthesimulationonday15.Startingon-hand is41; it’s in the redbutnotdeepenough tocausemuchconcernsincealargesupplyorderwillbereceivedtomorrow.Thenetflowpositionisgreenat140.Thereisnosupplyorderrecommendation.

Page 215: DDMRP - Demand Driven Material Requirements Planning

Figure9-29 shows the continuation of the simulation on day 16.On-handhaseroded tonearlyhalf the redzone. If therewasno impending supply, thismightbecauseformajorconcern.Stillthesituationdictatesthattheplanneratleast check on the incoming order status. Upon being assured that it will bedeliveredtoday,noadditionalactionshouldberequired.Theseverityoftheon-handsituationandtheleveloftheresponsearedictatedbytheamountofcurrentandnear-term futureon-handpenetration into the redzone.Onceagain this isthe execution component of DDMRP and is explored in depth in the nextchapter.

FIGURE9-28Simulationday15

Page 216: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-29Simulationday16

The net flow position on day 16 is green at 130 (82.8 percent of top ofgreen).Thereisnosupplyorderrecommendation.Endinginventorywillbe91,as71arereceivedand10areshippedout.Thisisalmosttooeasy—whatcouldpossiblygowrong?

Figure9-30showsthecontinuationofthesimulationonday17.Startingon-handisonly51.Whathappened?Thesupplyorderfor71wasreceived,butoninspection40werefoundtohavequalityissues.Those40wereplacedonholdandmovedback toon-order status.Thisbringson-order to79.The itemswillneedtobereworkedandwillbeunavailableforseveraldays.The“On-Order”columnshowsayellowshading to indicate that theon-orderquantitycontainsitems on hold. The net flow position is yellow at 112 (71.3 percent of top ofgreen). A supply order for 45 with a request date of day 24 has beenrecommended.Endingon-handwillbe33,as18areshippedoutandnonearereceived.

Figure9-31 shows the continuation of the simulation on day 18. The on-handsituationiscauseforconcern.Theon-holdopensupplyisstilldaysawayfrom being made available. On-hand still covers the next 3 days of knowndemand with a supply order due on the 20th. No expedite is ordered oradditional supply is launched at this time. There is no supply orderrecommendation,asthenetflowpositionisgreenat145(92.4percentoftopofgreen).Endingon-handinventorywillbe21(12shippedoutandnonereceived).

Figure9-32showsthecontinuationofthesimulationonday19.Thestartingon-handsituationhasnowerodedtolessthanhalftheredzone.Theopensupplyis still one day away from being received, and the reworked parts have beenprogressingslowerthanexpected.Atthispointtheplannerdecidesthattheon-hand status dictates action. The planner requests an expedite of the closest-insupply order for 39 pieces. The supply order for 39 now has an exclamationmarkoverit,indicatingthatitisonexpeditestatus.Thisorderismovedupontheschedule,andovertimeisapplied tobring theorder inat theendof today.Endingon-handwillincreaseby31,as8areshippedoutand39arereceived.

Page 217: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-30Simulationday17

FIGURE9-31Simulationday18

Page 218: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-32Simulationday19

Figure9-33showsthecontinuationofthesimulationonday20.Thereisnorecommended supply quantity, as the net flow position is green at 131 (83.4percentoftopofgreen).Endingon-handwillbe46.

Figure9-34showsthecontinuationofthesimulationonday21.Startingon-hand isat86!Whathappened?Theamountonqualityholdwasmoved toon-handstatus.On-orderisat45,on-handisat86,andqualifieddemandisonly5.Thereisnorecommendedsupplyorder,asthenetflowpositionisgreenat126(80.3percentoftopofgreen).Endingon-handwillbe81.

Page 219: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-33Simulationday20

FIGURE9-34Simulationday21

Whatcanbelearnedfromthissimulation?First,thatthebuffersarerobust.Ifproperlymanaged,thesebufferscanhandlemanyformsofvariabilitysuchasupticks in demand, supply problems, and drop-in orders, all of which werepresent in the simulation. Figure9-35 shows a run chart for net flow and on-handpositionsoverthesimulated21days.Eachtimethenetflowpositiondropsintotheyellow,itisimmediatelyrestoredtogreenbylaunchingasupplyorder.Thisiscrucialtokeepingthebufferproperlysuppliedbymaintainingtheflowofrecommended amounts. Lags and lapses in launching these supply orderswillonlypresentadditionalvariabilitytothebufferandcancompromisetheabilitytomaintaintheintegrityofthedecouplingpoint.

Thisbufferperformedwellconsideringthechallengesthatwerethrownatit.Considerthatdemandforthisperiodwasover20percenthigherthanwhatthebufferwasbuiltfor.Overthese21days,demandaveraged12.4,whilethebufferwas built for an average daily usage of 10. This buffer should definitely beadjustingup,asthisupliftindemandbeginstoraisethecalculatedADU.

This simulation alsobegins to showuswhatwecan expect fromon-handperformance over time with regard to DDMRP buffers. Figure 9-36 is adistributionchartshowingthefrequencyofon-handvalueoverthesimulated21days.The×axisisinbinvaluesof10.Overthecourseofthesimulation,thereweretwoon-handoccurrencesbetween21and30;thusweseeabarwithadata

Page 220: DDMRP - Demand Driven Material Requirements Planning

labelof2inthe30positiononthe×axis.Weseeasingleuniformdistribution(asopposedtoabimodaldistribution)ranginginvaluefrom21to92.On12ofthe21days,theon-handvaluewasabovethetopoftheredzone.Averageon-handoverthe21-dayperiodwasat59.24.

FIGURE9-35Reviewingthesimulationresults

FIGURE9-36On-handpositionfrequencyover21days

CalculatingAverageOn-HandInventory

Page 221: DDMRP - Demand Driven Material Requirements Planning

Understanding how the different components of the net flow equation worktogether is crucial to grasp how to calculate the average on-hand inventoryposition.When the net flow equation yields a quantity in the yellow zone, asupplyorderwillberecommendedforaquantitytobringthenetflowpositionup to the top of the green zone. The green zone defines the average orderfrequencywhenthequantityisdividedbytheaveragedailyusage.

Figure9-37 depicts the buffer levels for a sample part to demonstrate theaverageon-handcalculation.Withagreenzoneof60andanaveragedailyusageof10,theaverageorderfrequencywouldbe6days.Theaverageorderquantitywould be 60. That means in a perfectly average world, 60 would be orderedevery6days.Eachsupplyorderwouldbedue6daysapart.Iftheparthasaleadtimeof18days,thenwewouldexpecttoseethreeopensupplyordersatanyonetime.

Figure9-38shows the starting situation forourexample.Startingon-handinventoryis110.Thenetflowpositionisat the topofyellow(280).Asupplyorder isbeing recommended for60due in18days (day19 if today isday1).Opensupplyshows180units,andyetwearelaunchinganeworder.Thatmeansthatanorderwillbereceivedat theendof today.Thatorderwas launched18days ago. Tomorrow therewill still be three supply orders in the “On-Order”columnincludingtheneworder.

Figure9-39 represents theorderhistoryof thispart, assuming theaverageorder size and frequency of the last 18 days. A supply order is due today(representedinthe“–1”column).Twoadditionalsupplyordersareexpected—onefor60duein7daysandanotherfor60duein13days.

FIGURE9-37Asamplepart

Page 222: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-38Today’snetflowpositionfortheon-handcalculationexample

Tomorrowthenewlygeneratedsupplyorderwillbevisibleat–18.Figure9-40showstheorderhistoryonday2.Asupplyorderwasreceivedyesterday.Thenextsupplyorderisdue5daysfromday2onday7.

Returningtooursituationtoday,on-handwillbeconsumedthroughouttheday,bringingtheendingon-handto100.Theorderfor60willbereceivedattheend of today and will be reflected in the beginning on-hand inventory fortomorrow(160).Eachdayuntil thenewsupplyorderisreceived,on-handwilldrainby10units.Ifon-handinventorystartsat160onday2,itwillendday2at150.Onday3on-handwilldropto140.Onday4itwilldropto130.Onday5itwilldropto120.Onday6itwilldropto110.Onday7on-handwilldropto100andthenberestoredto160onday8,asthesupplyorderwasreceivedfor60attheendofday7.

Figure9-41depictsthisrecurringbehavior,assumingaveragedemand,ordersize,andfrequency.Theon-handpositionrangesbetweenahighof160andalow100.Day1showsthedailydemanddrainfrom110to100.Asupplyorderisreceived,pushingon-handupto160.Demandthendrainsofffor6consecutivedaysbackdownto100beforethenextsupplyorderisreceivedthatday.Attheendofday19,thesupplyorderisreceivedthatwascreatedonday1.

FIGURE9-39Supplyorderhistory

FIGURE9-40Orderhistoryonday2

Page 223: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-41On-handpositionassumingaveragedemand,ordersize,andorderfrequency

This exercise gives insight into some interesting characteristics aboutDDMRPbufferswithregardtopredictinginventoryandopensupplyranges.

AverageOn-HandRange

With this examplea rangeof expectedon-hand is established.This rangewillbecomethefoundationformakingpredictionsaboutbufferbehavioraswellasmeasuring the performance of a particular replenished part over a past timeframe.InDDMRPthisrangeiscalledtheaverageon-handrange.Theaverageon-handrangecorrespondsdirectly to twozonesof thebuffer: the redand thegreen.Thelowerlimitoftherangeisthetopofred.Theupperendofthelimitisthetopofredvalue+thegreenvalue.Inouron-handexample,thatrangeis100(lower)and160(upper).

InthebeginningofChapter7,wediscussedthevalueofinventoryandtheneedtobeabletofindarangethatkeepsusbalancedbetweenthenegativesofhavingtoomuchandhavingtoolittle.Figure9-42isaduplicateofFigure7-1,showingtheTaguchilossfunctionofinventoryvalueandtheconceptualoptimalrange.Thereisalossofvalueaswemovetowardtheextremesoftoolittleortoomuchandoutsideofanoptimalrange.Wenowhaveawaytocalculatewhattheoptimalon-handrangeis.

Toshowhowthisworks,thepartwillbeusedthatwassimulatedearlierinthischapterfor21days.Itstopofredzonevaluewas52.Theyellowzonewas

Page 224: DDMRP - Demand Driven Material Requirements Planning

70,and thegreenzonewas35.Theoptimalon-hand range is thevalueof thegreenzone.Oneithersideofthegreenarethewarningranges.Totheleftisthelow-levelwarning.Thatvalueis thevalueof theredzone(52).Totheright isthe high-level warning and is the remaining amount of the yellow zone (35)whenthegreenzoneissubtracted.Thelossfunctionreachestoolittle(pointA)atzeroon-handandtoomuch(pointB)at122.The122isthetopoftheyellowzone.ThisisgraphicallyrepresentedinFigure9-43.

FIGURE9-42FormerFigure7-1showingtheoptimalzoneandTaguchiinventorylossfunction

FIGURE9-43Inventorylossfunctionforthesimulatedpart’svalues

Whenconsideringpastperformance, thisgreen rangecan tellushowwellthebufferhasperformed.Toomanyseverepenetrationsbelowtherangecantellusthatthebuffer’ssafetywasfrequentlyrequired.Toomanyon-handpositionsabove the green range will signal that the buffer might be oversupplied oroverstated relative to theactual requirement.Figure9-44 showsa run chart ofthe21-daysimulationwithregardtotheon-handpositionontheon-handrange.The run chart displays on-hand values and is converted to a Taguchi lossfunctionviewcolorscheme.Thegreenisstillthegreenzoneof35.Ithasbeen

Page 225: DDMRP - Demand Driven Material Requirements Planning

stacked on top of the red zone. On the upper side of the green zone is theremaining amount of the yellow zone of 35 (green zone -yellow zone). If on-handisatzeroorlessthanzero,thatisinthe“TooLittle”range.Ifon-handisoverthetopofyellowvalue(122),thatisinthe“TooMuch”range.

Aswecansee,thebuffer’son-handpositionspentalmosttheentirefirsthalfof thesimulationwithintheoptimalon-handrange.Drop-inordersandqualityholds created deep penetrations into the low warning range. The on-handpositionneverwentabovetheoptimalrange.

JudgingpastperformanceisdiscussedinmoredepthinChapter12.

FIGURE9-44Runchartagainstoptimalandwarningranges

AverageOn-HandTarget

Now that the average range is established, then an average amount can becalculated.InDDMRPthisiscalledtheaverageon-handtarget.Itisinmiddleoftherange.Theequationtocalculatetheaverageon-handtargetisthetopofredvalue+greenzonevalue/2.Inouraverageon-handexamplepart,thetopofredis100,andhalf thegreenzonevalueis30.Thustheaverageon-handtarget is130.Figure9-45nowhastheaverageon-handtargetillustratedasthesolidlinethroughthemiddleoftheaverageinventoryrange.

This equation is mathematically similar to an inventory calculation madeunder theconventionalapproach,whereon-hand issupposed toaverageout tosafetystock+one-halftheminimumorderquantity.Thisconventionalequation,

Page 226: DDMRP - Demand Driven Material Requirements Planning

however,rarelyproducesareasonableapproximationoverlongerperiodsduetoMRP’ssusceptibility tonervousness,variabilityofdemand, inventory inflationtypically caused by safety stock ordering, and shortages created by supplycontinuityvariability.

Understandinghowtocalculate theaverageon-handtargetwillnowallowthe CompanyABC example fromChapters 7 and 8 to be finished.With thisequation theworkingcapital implicationsof thebufferscanbe judged.This iscoveredlaterinthischapter.

AverageOpenSupplyOrders

Theyellowzonedividedbythegreenzoneistheaveragenumberofopensupplyorders.Inthisexamplethereshouldbethreeopensupplyordersoneachday.Ifthegreenzonewas90insteadof60,thenthereshouldbeanaverageoftwoopensupplyordersatanyonetime.

FIGURE9-45Averageon-handtargetof130

This makes mathematical and intuitive sense when you consider that thepurposeoftheyellowzoneistoprovidetheheartofthedemandcoveragewithinonefull lead time.Theyellowzone is thepipelineorconveyorbeltportionofthebuffer.Itextendsthebufferbackintothesupplychain.Aslongasordersareapprovedwhenthenetflowequationcallsforresupply,thatpipelinewillremainintact.

Ofcourse,theseequationsarebuilttoaverages.Withinshorterwindowsof

Page 227: DDMRP - Demand Driven Material Requirements Planning

time, they will not be accurate due to variability. With the 21-day planningsimulation, there was plenty of variability within the simulation window. Forexample,theaverageon-handtargetwouldbecalculatedas69.5[52(topofredvalue)+35(greenzonevalue)/2].Yetattheendof21days,theaverageon-handamountwas 59.24.Thiswas, of course, due to heavier demand than expectedand supplydelays.Averages are,however,valid touseover longerperiodsoftime—the strategic relevant range.Assuming themodel is set andmaintainedproperly,thereisareasonablepredictionabouton-handlevelstoberealizedbythatmodeloverthatlongerrange.

DecoupledExplosion

WhenconsideringdecouplingandtheDDMRPsupplyordergenerationprocess,anobviousimpactemerges.Whenasupplyorderisgeneratedatahigherlevel,decouplingstopstheexplosionofthebillofmaterialatdecouplingpointsplacedatlowerlevels.Theexplosioncanandwillbestoppedbecausethatdecouplingpointisbuffered.Thenetflowequationisthenindependentlycalculatedatthatpoint.Supplyordergenerationonlycontinuesifthatposition’snetflowequationcallsforresupply.Underthatcondition,theexplosionthenbeginsagainrelativetothatpart’srespectivecomponents.

FIGURE9-46ExplosiondifferencesbetweenMRPandDDMRP

Figure9-46illustratestheexplosiondifferencesbetweenconventionalMRPand DDMRP. MRP uses a full “requirements explosion.” APICS defines arequirementsexplosionas:

The process of calculating demand for components of a parent item

Page 228: DDMRP - Demand Driven Material Requirements Planning

requirementsbythecomponentusagequantityspecifiedinthebillofmaterial.(p.149)

Thiscalculationcontinues through thebillofmaterialover theentireplanninghorizon.Thustheword“explosion”relatestodependentrelationshipsbetweenaparentitemanditscomponentforaprojectedperiodoftime.

The explosion on the left of Figure 9-46 represents conventional MRP,whereanydemandatthehigherlevelistypicallydrivenallthewaythroughtothepurchasedlevelandprojectedintothefuture.TherearesomeexceptionstothisruleinMRP,buttheyaresimplythat—exceptions.Forexample,MRPwillstoptheexplosionwithinaparticular leg if there issufficienton-handstocktocoverthedemandrequirementinthattimeperiod.Thisoccurrence,however,ismorehappenstanceratherthananyrealplan.It tendstohappenonlyifthereisresidual on-hand inventory remaining due to changes in schedule or ordermultipledifferences.Remember,MRPfundamentallywantstonetallpositionsto zero projected available balance (for positionswith safety stock, the safetystocklevelisthenewzerolevel).

DDMRPutilizesa“decoupledexplosion”depictedontheright-handsideofFigure9-46.AdecoupledexplosionisacriticaldistinguishingcharacteristicofaDDMRP system. The term itself is an oxymoron. It means “independentdependence.”Yet that is exactlywhat is occurring. This concept is crucial inpreventingnervousnessbecausemostchangesathigh-levelparentswillnotbebigenoughtopassthroughthebuffersatthedecouplingpoints.Thismeansthatflowis largelyprotectedagainst thesystemnervousness that is transferredandamplified in conventional MRP. This is especially true for decoupling pointsplacedatcommoncomponents(acommonstrategy)asthesmoothingbenefitofaggregationagainstabiggercalculatedbufferlevelisrealized.

Can conventional MRP systems decouple explosions? There are at leastthreetacticsthatshouldbeexploredwithregardtothisobjective,butallcomewithaprice.

1. Safety stock could be added at desired decoupling points in asufficientquantitytoguaranteeon-handquantities.However,thiswillcreate expedite requirements ason-handbalancesdropbelowsafetystock.Thisessentiallymeansthereisnorealdecoupling.Thesignalistransferredandamplifiedthroughthebillofmaterial.

2.Astopexplosionordercanbeused.ManyconventionalMRPsystemsallow the deployment of a “stop explosion flag” or an “externally

Page 229: DDMRP - Demand Driven Material Requirements Planning

planned” setting for specifically designated parts. In this case theexplosioncandefinitelybestoppedatthatposition,buttherearetwosubsequentchallenges.First,thebufferlevelsmustbeconstructedatthatposition,andsecond,thenetflowequationmustbeemployedtorestarttheexplosion.ConventionalMRPsystemscan’tautomaticallyrestart the explosion, and they most certainly won’t restart theexplosionaccordingtoanetflowpositionbasedonaDDMRPbufferconstruct.Thismeans that itwill fall toplanners tomanually restarttheexplosionforeachdecoupledintermediateandpurchasedposition.ThishasbeenthescenarioformanyearlyDDMRPimplementations.

3.Amultilevelmasterproductionscheduleprocesscanbeused.Inthiscaseanexplosioncanbestoppedandthenrestarted,butatwhatcost?Setting up and effectively maintaining multiple master productionschedules and their connections will prove extremely difficult formostplanningteams.Furthermore,thenetflowequationmuststillbeperformedagainst theDDMRPbuffer levelsbetween thesedifferentmaster production schedules, and there is no mechanism inconventionalMRPsystemstoeffectivelydothat.

YetMRPandDDMRParenotalwaysdifferent.WithincertainparametersMRPandDDMRPareidentical.Figure9-47depictsanarea in theexample inwhichMRPandDDMRPbehaveexactlythesameway.Thisareaishighlightedwithin theboxonbothexplosions.Thus theconceptofadecoupledexplosionillustrateshowDDMRPsuccessfullycombinesboththedependenceofMRPandthe independence of strategic decoupling required for an effective solution fortoday’s volatile and uncertain environments. There is independence at thedecoupling points, but between decoupling points there is dependence. ThatdependencebetweendecouplingpointsisnodifferentfromconventionalMRP.That is why there is still MRP in DDMRP. A decoupled explosion is acornerstoneoftheplanningmechanisminaDDMRPsystemandallowstheflowofrelevantinformationandmaterialstobepromotedandprotected.

SodoesthismeanthatwithDDMRP,everythingwillgoaccordingtoplan?Willthemanufacturingleadtimesalwaysbeaperfectreflectionofhowlongitwilltaketomovematerialthroughadecoupledleadtimechain?Ofcoursenot,but that iswhy there is a buffer, in particular the red zone portionwithin thebuffer.ThisbringstolightanotherveryimportantdistinctionbetweenMRPandDDMRP. InMRP,because there isnodecoupling,everythingmustgoexactly

Page 230: DDMRP - Demand Driven Material Requirements Planning

accordingtoplaninordertogetthepredictedresult.InDDMRPtheassumptionis that almost nothing will go exactly according to plan. DDMRP is roughlyright,whereMRPispreciselywrong.

FIGURE9-47MRPandDDMRPexplosionsimilarities

Figure9-48depictsaproductstructureforaproductcalledFPD.Iftoday’sFPDnetflowequationcallsforaresupply(belowtopofyellow)of25pieces,asupplyorder for25willbegenerated fordelivery insevendays (the lengthofthedecoupledleadtime).ThiscreatesademandforanFPDparentorderreleasesixdaysfromtoday(thedecoupledleadtimeminusthemanufacturingleadtimeofFPD).Thecomponentrequirementsinvolvedinthisparentorderdemandareforcomponents208and210.Component208isbuffered,andsoitisassumedtobeavailable,andnofurtherexplosionhappensdownthatleg.Theshadedrowsin the figure represent the buffered components (208, 410P, 412P) and definewherethedemandexplosionwillstop.

Figure 9-48 also depicts the decoupled explosion that drives through thenonbuffereditems;however,asshowninthetablebelowtheproductstructure,210isunbuffered.Itsmanufacturedleadtimeistwodays.Anorderreleasewillbe created four days from today (FPD decoupled lead time minus 310manufacturing lead time). This will then generate an immediate releaserequirementforcomponent310.

InsummaryFPDrequires208and210onday6.Part210requires310onday 4. Part 310 requires 410P and 412P on day 1. An order source is alsoavailable showing what specific component(s) drove the requirement. This iscalledsingle-levelpegginginMRPandisstillverymuchapplicableinDDMRPbetweendecouplingpoints.

Page 231: DDMRP - Demand Driven Material Requirements Planning

Therewill alwaysbe an immediate release at the terminusof theparent’sdecoupled lead time chain unless there is residual on-hand inventory that issufficienttostoptheexplosion.Thatimmediatereleaserequirementisbasedonthe assumption that the stock is available at the decoupling point and thedecoupledleadtimeforthehigh-levelparentresponsiblefortheoriginaldemandwillbehonored.

FIGURE9-48Adecoupledexplosionexample

To illustrate howan explosionwould stopwithin the decoupled lead timechain,wewillallowforresidualon-handinventorywithintheexample.Figure9-49illustratesasituationinwhichFPDgeneratesthesamedemandrequirementfor 25 units. The demand is passed through 210 (evident in the “Net ItemRequirement”column),buttheexplosionstopsat310becausethereisresidualon-hand inventory. Thus no demand allocation is passed through to 410P and412P.

What about components being protected against large parent demandspikes? If parent item spikes are not qualified in advance, this could be a bigissueforlowerlevelsandfurtherreinforcestheneedfororderspikequalificationattheenditemlevel.IntheFPDexample,theorderspikehorizonwouldbeatleastsevendays.Ifthespikegeneratesaresupplyrequirement,then310wouldgetanimmediaterequirementforrelease.

Page 232: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-49Adecoupledexplosionexamplewithresidualon-handinventory

Will 410P and 412P have enough stock on hand to cover the immediaterelease? Of course, that depends on the size of the spike. At some point nostockingsolutionwillbesufficienttocoverenormousparentspikes.Butif410Pand412Parecommoncomponents,thentheirrespectivebuffersarebuiltuponamuchlargeramountofconsolidateddemandproducingarelativelylargerbuffer.This is evident in theCompanyABCexamplewith regard to component201.Thisoftenmeansthatalargespikefromasingleparentperspectiveisnotalargespikefromacomponentperspective.Additionally,FPDdoeshaveanamountofsafetyifthespikeoutstripstheon-handlevelsof410Por412P.Thepropensityand size of the spikes on FPD determine how that red zone safety level wascalculatedinthefirstplace..

Figure9-50illustratesanFPDdemandof125,qualifyingasaspikeagainstthebuffered208and412Ppositions.Inthecaseof208,thedemandisnotduefor sixdaysbuthasqualifiedasa spikeagainst the208order spike threshold.That means it will be included in the net flow equation for 208 until it issatisfied. Will it generate a requirement for resupply against the 208 buffer?Perhaps,but thatnet flowequation isdeterminedseparately.That is theentirepointofthedecoupledexplosion.

Figure9-50 also shows a net requirement of 75 being passed through the310positionto410Pand412P.The50on-handquantityfor310isnettedfromthe demand of 125. For 412P an immediate release of 75 is required, and thedemand represents a spike. Will it generate a supply? As with 208 it couldgeneratea supplyorder. It isdefinitelyaqualifieddemand in412P’snet flowequationfortoday.

Let’saddanadditional layerofcomplexity to theexample to illustrate thesimilaritiesofMRPandDDMRPwithinthedecoupledleadtimechain.Figure9-51 shows another parent component generating demand for the commoncomponents 210, 310, 410P, and 412P. In this example FPD is calling for aresupplyof 25, andFPE is calling for a resupplyof 60.Thismeans thegross

Page 233: DDMRP - Demand Driven Material Requirements Planning

requirement for 210 is 85.This is seen in the table under the “Order Source”columnshowingdemandfrombothsources.Component310has50inresidualon-handandsoonlypassesanettedrequirementfor35to410Pand412Peach.

AllofthisservestoillustratethatconventionalMRPandDDMRPgiventhesame inputsessentiallybehave the samewith regard tononstocked items.Thedecouplingpointsdefinetheboundariesofthatcommonbehavior.

FIGURE9-50Anorderspikefor208

FIGURE9-51Grossrequirementfor210

HybridModelSupplyOrderGeneration

ThehybriddistributionmodelwasintroducedinChapter6.Itinvolvedamixedmodelofhub-and-spokeconfigurationforslower-movingitemsandmake-and-ship on fast-moving items. Figure 9-52 is the hybrid model from Chapter 6.

Page 234: DDMRP - Demand Driven Material Requirements Planning

Products1,2,and3havebuffersatthespokeonly,whileproducts4and5areinthehub-and-spokeconfiguration.

Generating supply orders for the slow-mover parts in the hub-and-spokeconfigurationisastraightforwardapplicationofthenetflowequationatboththespokes and the hub. But how to plan and generate supply orders for the fastmoverswhen there isnoclear signalprovidedbya centralbuffer?Rememberthatthesefastmoversareshippedtotheforwarddistributionlocationsfromthemanufacturing plant upon completion. In most cases there will be signals forresupply on fast movers from the spokes at various times and for variousquantities. The average order size for a specific forward location is by farsmallerthantheminimumrunsizeatthesourcingunit.Howshouldthesourcingunitdeterminewhenistherighttimetoruneachfastmover?

Figure 9-53 shows the hybrid configuration for product 1 across alllocations.Thereisnocentralbuffer,onlyspokebuffers.Thezonevaluesoftheproduct1bufferateach locationaregiven inFigure9-53.Notice thefar rightcolumnunder“Total.”Inthiscasethegreenzonesacrossalllocationshavebeensizedtototalthesourcingunitminimumorderquantityof10,000.Eachlocationhasadifferentleadtimethatiscalculatedbytheplantleadtime(sevendays)+transportationtime.Locations1and4arethreedaysawayfromtheplant,whilelocations2and3areoneandtwodaysawayrespectively.

FIGURE9-52ThehybridmodelfromChapter6

Page 235: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-53Bufferlevelsacrossalllocationsforproduct1

The average daily usage for each location is displayed.The total networkaveragedailyusageisthesummationofalllocations’usage.Eachlocationhasadifferent number of days’ safety contained in its red zone (location redzone/locationADU).Location1has7.5days.Location2has6days.Location3hasjustunder7days.Location4has7.5days.Intheaggregatethetotalsafetyacrossthenetworkis7days[12,244(totalofredzones)/1,750(totalADU)].

Supply order generation for a fast-moving product must consider theaggregatenetflowpositionacrossanetwork.Figure9-54displaysthesituationfor product 1 on an example day. The column labeled “NFP” is the net flowpositionateachlocation.The“Priority”columnisthepercentageofthenetflowpositiontothetopofgreen.Itiscolor-codedbasedonitspositionrelativetothetop of yellow.Below the TOY value yields a yellow coding; above the TOYvalueyieldsagreencoding.The“Order”columnistheamountneededtorestoreeachlocation’sbuffertothetopofitsgreenzone.Thetotalamounttorestoreallbufferstothetopofgreenis10,851.ThisisgreaterthanthesourcingunitMOQof10,000.Asupplyorderisgenerateddespitetwolocations’netflowpositionbeinggreen.

Page 236: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-54Product1priority(example1)

FIGURE9-55Product1priority(example2)

Figure9-55showsanotherexample inwhicha supplyorderwouldnotbegenerated. Despite location 2 having a yellow planning priority, the totalaggregatedemandagainstthesourcingunitislessthantheMOQ.

[This example demonstrates the use of net flow position across a specificgroupinordertodealwithaspecificlimitation.Inthiscasethegroupwasthedistributionlocationsforaspecificproduct,andthelimitationwasthesourcingunit MOQ. This example opens the door to a crucial concept with regard toDDMRPsupplyordergenerationinrelationtolimitationsassociatedwiththosegroups.Thisconceptiscalledprioritizedshare.

PrioritizedShareSupplyOrderConsideration

Prioritizedshareisasupplyordergenerationschemaappliedtoagroupbasedonthe relative net flow positions within that group and the imposition of aconstraintorthreshold.Thetypesofgroupcanincludewithinalocation,acrosslocations,bysupplier,andbyproductstructureorsharedcomponent.Therearethreemainusesofprioritizedshare:discountoptimization,freightoptimization,andcoverageoptimization.Eachusewillbeexplored.

DiscountOptimization

Suppliers frequently offer discounts for meeting certain conditions. Thesediscountsoftenrepresentasignificanteconomicadvantagetothebuyingentityif properly managed. Examples might include free freight or a percentagediscountforordersthatmeetagiventhreshold.Thisthresholdmightbethatthe

Page 237: DDMRP - Demand Driven Material Requirements Planning

totaldollaramountoftheordermustreachaspecifiedminimumorthatthetotalordermustfillatruck..

One example: Let’s say a company buys electronic components from asupplier called SuperTech. SuperTech has a policy that itwill pay the freightcostsifanorderfillsatruck.Theorderdoesnothavetobeforjustoneitem.Amixedloadcanbeordered,andas longas theentire truck is filled,SuperTechwill pay the freight. This is a real economic benefit to the customer, but thecustomerneeds to takeadvantageof it inawaythatpenalizes it the leastwithregardtoinventory.

There are 30 total pallet positions in a large trailer. The minimum orderquantityandmultipleforeachpartisonepallet.Thiscustomerbuys12differentcomponents from SuperTech. Many items have green zones sized to theirrespectiveminimumorderquantities(onepallet).Figure9-56showsthebuffersof allSuperTechparts.Eachpart’s net flowposition is indicatedby theblackdiamond symbol. There are five parts with net flow positions in yellow. Arethesefivepartsenoughtofillatruck?

Figure9-57 is theplanningscreen forSuperTechparts.Thescreen is firstsortedbypriority.Priorityisdeterminedfirstbycolorandthenbypercentageofnet flow penetration in relation to the top of green. The “Order” column issimplyhowmuchquantityisrequiredtomovethenetflowpositiontothetopofgreen.Thefivepartscallingforresupplyareevidentat thetopoftheplanningscreen(parts10,12,11,6,and9).

Twocolumnshavebeenaddedtotheplanningscreentohelpdeterminehowto take advantage of the SuperTech free freight offer. “Qty per Pallet” is thenumberofeachpartperpallet.“PalletsReq”isthenumberofpalletsneededtorestorethenetflowpositiontothetopofgreen(“Order”/”QtyperPallet”).Thepartscallingforresupplyonlytotal28pallets,but30palletsarerequiredforfreefreight. How can the customer best prioritize the share of the additional twospacesinthetrailer?

One option would be to order an extra two pallets of the part numberscalling for resupply. This would put at least one or two of those parts in anOTOGnet flowposition.Anotheroptionwouldbe to take thehighest-prioritygreen parts. Parts 1 and 2 have relatively significant green zone penetrations.Theseareparts that arenot calling for resupplybuthave thedeepestnet flowpenetrationintotheirbuffersrelativetoallothergreenparts(8,3,4,5,and7).IfthegreenzoneisdeterminedbyMOQ,thenthiswillcauseanOTOGnetflowpositionaswell.Thatisthecaseforbothparts1and2.Ifthecustomermustgo

Page 238: DDMRP - Demand Driven Material Requirements Planning

over the top of green,what is the best decision tominimize excess inventoryliability?

FIGURE9-56PurchasedSuperTechpartnumbersandtheirbufferpositions

FIGURE9-57PlanningscreenforSuperTechparts

Aneasywaytodeterminetheminimalovertopofgreenliabilityofaddingtwoadditionalpalletstotheorderistocomparethenetflowpercentagewithoneadditionalpalletforallpartnumbers.Figure9-58showstheimpactofordering

Page 239: DDMRP - Demand Driven Material Requirements Planning

anadditionalpallet thanisrequiredforeachspecificpart.Thecolumn“Pallets+1”showsanadditionalpalletforall items.Thecolumn“NFP(+1Pallet)” isthenetflowpositionwiththeadditionalpalletquantity.

The column “Priority (+ 1 Pallet)” is the net flow position after theadditional pallet quantity has been figured in divided by the top of the greenzoneforeachpart.Whenover100percent,anyprioritycolumnshoulddisplayOTOG and be shaded a light blue. The column “OTOG%” is the percentageamount over the top of green.This can then be easily sorted to show the twoitemswiththesmallestOTOG%impact.Figure9-59showsthepartssortedbytheleastamountofOTOGimpactoneachoftheirrespectivebuffers.

Thisanalysisshowsthatorderinganadditionalpalletofpart6andorderingonepalletofpart1willresultintheleastamountofovertopofgreenliability.Thus the prioritized share approach provides a simple and quick way to takeadvantage of the free freight offer while minimizing the impact on excessinventory.

Thesametypeofprioritizedshareanalysiscanbeperformedforminimumspendthresholds.Let’susethesameexamplebutwithanalternativeSuperTechpolicy. In this case SuperTech offers free freight for orders above $19,000.Figure9-60 shows the dollar value per pallet of each SuperTech part and thetotalspendfor the itemscallingforresupply.Onepallet is theminimumorderquantityandmultipleforeachitem.Thefiveitemscallingforresupplyonlytotal$18,025ofthe$19,000spendingthreshold.Prioritizedsharewillbeusedtohelpdetermine the best way to spend the remaining amount tomeet the thresholdwhileminimizinginventoryliability.

Page 240: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-58OTOGpercentagewithadditionalpalletforeachpart

Page 241: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-59PartssortedbyOTOG%

FIGURE9-60Totaldollaramountofpartscallingforresupply

TheplannerwillneedtofindthebestcombinationofadditionalpalletorderstomeettheminimumspendthresholdwiththeleastamountofadditionalspendandtheleastamountofOTOG%liability.Figure9-61isthesameasFigure9-59butwiththeadditionalcolumnofpriceperpallet(“$perPallet”).Thiscolumnwill be used in combination with the “OTOG%” column to select the bestcombination of parts for additional pallet ordering to satisfy the aboveconditions.

In this case, the best combination of minimizing additional pallets andOTOG%andmeeting the spend thresholdwith the least amount of additionalspend is to order additional pallets of parts 1 and 12.Ordering one additionalpalletofeachtotals$1,000ofadditionalspend.Thiswillbringthetotalorderto$19,025, only $25 over the required minimum spend. Additionally, thisconfigurationallowsfortheleastamountofOTOG%tocapturethefreefreight.

Page 242: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-61PartssortedbyOTOG%withdollarperpalletperpart

FreightOptimization

Prioritizedsharecanbeusedtooptimizefreightspendfromhubtospokeorhubto hub (multihub configuration), assuming that full truckloads will result inbetteroverallfreightrates.JustasinthecaseofthefirstSuperTechexample,thereceivinglocation’s(spoke)netflowpositionsonallreplenisheditemscomingfrom the supplying location (hub) will be analyzed from a prioritized shareperspective,andafulltruckloadwillbebuilttominimizeOTOG%liability.

Inthiscasemixedpalletsmightbeapossibility,allowingformultiplegreenitems to be simultaneously brought to top of green status. Depending on thenumberofgreenitemsandtheseverityoftheirpenetration,thiscouldeliminateanyoverthetopofgreenendingsituations.

CoverageOptimization

The prioritized share schema can be used to allocate scarce quantities ofinventory to spoke buffers in order to produce the best aggregate networkcoverage.Thiswasthefirstknownapplicationoftheprioritizedshareschema.

Page 243: DDMRP - Demand Driven Material Requirements Planning

When there isnot enough inventory to cover all of apart’sdemandacross alllocations calling for resupply or there is a desire for a matched run out of adiscontinuedproduct, prioritized sharewill step across the locations allocatinginventorytothetopsofthelowestzonesfirst.Firsttopofred,thentopofyellowandthentopofgreen.

Thisavoidsonelocationthathasadeeppenetrationtotakealltheinventory,restoring its net flow position to healthy while others remain relatively introuble.Inthisapplicationtheprioritizedshareschemaislookingtobalancealllocationswithaboutthesamelevelofprotectionfromanetflowperspective.

FIGURE9-62Productprioritizedshareexample

Figure 9-62 shows how scarce supply at a hub is allocated against fourlocationscallingforresupplyusingtheprioritizedshare.Product5hasdemandsforresupplyacrossthenetworkincludingarednetflowpositioninlocation1.Atotal of 21,347 units are required to restore all locations to a full net flowposition, but there are only 15,000 units available at the hub. If location 1 isallowed to restore itsnet flowposition to100percent, itwill leaveonly3,750remainingfortherestofthelocations.Thiswillresultinanunbalancednetworkfromanetflowperspective.Supplyorderstoalllocationswillbemodifiedusingprioritizedsharetoallocatethe15,000availabletoachievearelativelyequalnetflowpositioninalllocations.

First,inventoryisallocatedtobringlocation1’snetflowpositiontothetopof red. This requires 750 pieces [4,875 (TOR) – 4,125 (NFP)]. This leaveslocation 1 with the deepest yellow penetration. It has priority for additional

Page 244: DDMRP - Demand Driven Material Requirements Planning

inventoryshare.Now14,250pieces remainavailable.Location1hasayellowzoneof6,500.Location3(thenextdeepestpenetrationinyellow)requires1,905torestoreittothetopofyellow.Location2requires1,808torestoreittothetopofyellow.Location4 requires384 to restore it to the topofyellow.The totalrequirementtorestoreallpositionstothetopofyellowis10,597.Ifthisdemandwerelessthanthetotalavailable,prioritizedsharewouldlooktobalancethenetflowpositionsinyellowacrossthelocations.

Sincethereisenoughavailableinventorytorestorealllocationstothetopofyellowbut not enough to restore all to the top of green, prioritized sharewillneedtobalancealllocationsinthegreenzone.Thereare3,653piecesremainingin available inventory against a total remaining demand of 10,000. Theprioritized share schema will balance out each position’s priority percentage.Figure9-63showsastep-by-stepapplicationofprioritizedsharetothisexampleandthetotalallocatedtoeachlocation.

At the end of this sequence, all product 5’s net flow positions across alllocationsarebalancedtoaprioritypositionof83.5percent.Figure9-64showstheplanningscreenwiththeupdatednetflowpositions.

FIGURE9-63Prioritizedsharesequence

Page 245: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-64Updatedplanningscreen

Min-MaxSupplyOrderGeneration

There are three types of part classifications in DDMRP. To this point thischapter has been exclusively dedicated to replenished parts.Replenished partsare the majority of buffered parts in most DDMRP implementations. Theapplicationofthenetflowequationisexactlythesameforreplenishedoverrideparts.That leavesonlythesupplyordergenerationconsiderationsformin-maxparts.Min-maxpartsarenonstrategicbufferedpositions.Thepartsassigned tomin-max status tend to be low variability and readily available, but they stillrepresentdecouplingpoints.

Thedailyapplicationofthenetflowequationappliestomin-maxpositions.Using themin-maxexample fromChapter7, the use of the net flow equationwill bedemonstratedwith amin-maxposition. InFigure9-65 the starting on-handquantityis40.Thereare70on-orderunitsand32inqualifieddemand.Thenetflowpositionis78(40+70–32).Asupplyorderwillbegeneratedfor72units.

CompletingtheCompanyABCExample

TheexamplethatwasstartedinChapter7cannowbecompleted.Younowhavethecapabilitytocalculatetheaverageon-handquantityforeachbuffer.Thiswillallow us to judge the merits of different decoupling position options. Thiscomparisonismadefromtwoperspectives:quantityandcash.Theaverageon-hand targetequationyieldsaquantity.When thatquantity ismultipliedby thedirectmaterialcostoftheitem,theresultisacashvalueassignedtothetarget.

Page 246: DDMRP - Demand Driven Material Requirements Planning

Notethatonlythedirectmaterialcostisconsidered.Thisisthebestwaytomakea trueworkingcapitalcomparisonwithoutworryingaboutoverheadandlabor allocations. These types of allocations can be distortive. Directmaterialcostcannotbedistorted.Ifthoseallocationsareconsidered,thepicturebecomesevenmoreskewedinfavorofthedecouplingpointsplacedlowerintheproductstructures;thosepartsdon’tgetthat“valueadded”thathigher-levelcomponentsget and consequently look “cheaper” to hold. However, there is no realdifferencefromatrueworkingcapitalperspective.

FIGURE9-65Netflowpositionformin-maxexample

Establishing the direct material cost of all items starts with knowing thedirectmaterialcostsofallpurchaseditems.Parentitemdirectmaterialcostisthesumofthecomponentdirectmaterialcosts.Figure9-66showsthethreeproductstructureswiththedirectmaterialcostsforallpurchaseditemsindollarsbeloweachitem.

Withthematerialcostsofthepurchaseditemsidentified,itisthenpossibleto calculate the directmaterial cost for all parent items. Figure 9-67 shows asummaryofthedirectmaterialcostsforallitemsinallproductstructuresforthisexample. Each parent’s direct material cost is the sum of its immediatecomponent direct material cost. For example, the direct material cost of 202($235) is the sum of the directmaterial cost of 306 ($125) and 305P ($110).These directmaterial costs are combinedwith the calculated average on-handtarget equation in order to show the working capital implications for each

Page 247: DDMRP - Demand Driven Material Requirements Planning

decouplingiteration.TheCompanyABCexamplebeganwithallenditems(FPA,FPB,andFPC)

being buffered. Figure 9-68 shows the buffer values in the starting situationbefore the decoupling point selection example. Each end item’s estimatedaverageon-handdollarvalueiscalculatedinthelowerright-handcornerofthetable.Thedirectmaterialcostisthetotalvalueofthecomponentsineachitem.Theaverageon-handtargetforeachitemisestablishedbyaddingthetotalredzonevalue tohalf thegreenzonevalue (theaverageon-hand target equation).Theaverageinventorydollarsforeachpositionisestablishedbymultiplyingtheaverageon-handtargetby thedirectmaterialcostof the item.The total forallestimatedparentsatthebeginningoftheexampleis$3,260,723.

FIGURE9-66Productstructureswithpurchasedpartmaterialcosts

Page 248: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-67Alldirectmaterialcosts

The intermediate component 201 is then selected for decoupling. Thiscompresses the end item buffer levels dramatically but does require aninvestmentinabufferfor201.Figures9-69,9-70,and9-71showthereductioninaverageon-handtargetvalueandaverageinventorydollarsforeachenditempositionwhen 201 is buffered (shaded boxes). Figure9-72 shows the averageon-handtargetandaveragedollarinvestmentforthe201buffer.

Theinvestmentrequiredforthe201buffermustbenettedagainstthetotalreductioninparentiteminventory.Figure9-73summarizestheinventorypictureto thispoint.Bybuffering201, a totalof$281,163averageon-handdollars isexpectedtobereleasedfromthesystem.

Inthenextpositioningiteration,thecomponent203wasbuffered,allowingfortheFPAenditembuffertobecompletelyeliminated.Figure9-74showstherequiredinvestmentfortheestablishmentofthe203buffer.

Page 249: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-68Startingaverageon-handinventoryposition

Page 250: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-69FPAon-handanddollarcompression

Page 251: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-70FPBon-handanddollarcompression

Page 252: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-71FPCon-handanddollarcompression

FIGURE9-72Component201averageon-handtargetandinventorydollars

Page 253: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-73Inventorysummaryafterthefirstdecouplingiteration

FIGURE9-74Component203averageon-handtargetandinventorydollars

FIGURE9-75Inventorysummaryaftertheseconddecouplingiteration

Aftertheseconddecouplingiteration,twocomponentbuffers(201and203)havebeenadded,andoneenditembuffer(FPA)hasbeeneliminated.Figure9-75isaninventorysummaryaftertheseconditeration,showingatotalestimatedinventoryreductionof$537,413.

The final decoupling iteration for Company ABC establishes a bufferedposition for the purchased component 401P. This allows for a dramaticcompressionofthedecoupledleadtimeof201.Figure9-76showstheon-handtargetandinventorydollarcompressionfor201.

However, inorder to accomplish the201compression, an investment in abufferat401Pisrequired.Figure9-77showstheon-handtargetandinventory

Page 254: DDMRP - Demand Driven Material Requirements Planning

dollarinvestmentfor401P.Finally, this example started in Chapter 7 is complete. Figure 9-78

summarizes the inventorydollar compression throughall iterations, showing atotalestimatedreductioninon-handinventoryinvestmentof$1,325,828.

FIGURE9-76Component201on-handanddollarcompression

FIGURE9-77Component401Paverageon-handtargetandinventorydollars

Page 255: DDMRP - Demand Driven Material Requirements Planning

FIGURE9-78Inventorysummaryafterthethirddecouplingiteration

Summary

TheplanningcomponentofDemandDrivenMaterialRequirementsPlanningisasimple, intuitive,andhighlyvisiblewaytogeneratesupplyorders.Itsuseofqualified sales orders means that all components of the supply generationequation(on-order,on-hand,andqualifiedsalesorders)areknownandcontainrelativelylittlevariability.Thiscombinedwiththedecouplingpointpositioningandbuffersmeansthatnervousness,supplycontinuityvariability,andthebull-whipeffectaremitigated.Thenetflowequationandnetflowpositionallowforquick, intuitive, and informative views across groups of items, giving a realsenseofrelativepriorityandhowtobesthandlethatrelativepriority.

Page 256: DDMRP - Demand Driven Material Requirements Planning

CHAPTER10

DemandDrivenExecution

In DDMRP, a careful distinction is made between planning and execution.“Planning”istheprocessofgeneratingsupplyorderrequirementsusingthenetflowequationandtheelementsofdecoupledexplosionexplainedinthepreviouschapter. Planning ends once the recommendation has been approved andbecomesanopensupplyorder(purchaseorder,manufacturingorder,ortransferorder).

InDDMRP,“execution” is themanagementofopensupplyordersagainstrelevantcriteria.Thesecriteriaaredefinedintwobasiccategoriesnecessarytoprotectandpromoteflow:bufferstatusandsynchronization.Figure10-1depictsthefourbasicDDMRPexecutionalertsinthesetwocategories.

Bufferstatusalertsaredesignedtoshowthecurrentandprojectedstatusofthedecouplingpointpositions (independentpoints)across theDemandDrivenOperatingModel. These alerts use the current and projected on-hand positionrather than the net flow position. If there is no on-hand, then the decouplingpoint isnotdecoupledandwillmost likelypassonvariability.Thiscanoccurdespiteagreennetflowposition.Thenetflowequationistoplanthatpositiononly.On-hand tellsus if theposition is capableofperforming itspurpose—tomaintaindecoupling.Thetwobufferstatusalertswillbediscussedindepthlaterinthischapter.

Synchronization alerts are designed to highlight problems with regard todependencies.DependenciesstillexistinDDMRP(seethesectionondecoupledexplosion in Chapter 9). These dependencies are about known demandrequirementsversusprojectedsupplyavailability.Whilethebuffersmitigatethetransference of variability up and down the chain, synchronization is stillimportant in DDMRP between decoupling points and particularly between adecouplingpoint and thecustomer.Thebetter thevisibility to synchronization

Page 257: DDMRP - Demand Driven Material Requirements Planning

problems, the less variability is transferred to and between buffers and to thecustomer.Twoformsofsynchronizationalertswillbediscussedindepthlaterinthischapter.

BufferStatusAlerts

In order to really understand how andwhy buffer status alertswork, two keyperception changes must occur. The first change requires challenging howconventionalprioritymanagementtypicallyworks.Thesecondchangeisashiftin thebuffercolordisplaybetweenplanning (using thenet flowequation)andexecution(focusingoncurrentandprojectedon-hand).

FIGURE10-1DDMRPbasicexecutionalerts

ChallengingPrioritybyDueDate

To understand the power behind DDMRP execution, first the problem withconventional planning systems when it comes to execution needs to beunderstood. The “P” in MRP stands for “planning.” Material RequirementsPlanning inherently is a planning system and not an execution system.Conventional MRP systems lack real-time visibility to relative prioritiesassociatedwithpurchaseorders(POs),transferorders(TOs),andmanufacturingorders (MOs) throughout the internal manufacturing operation and across thesupplychain.

Without this effective priority approach, conventional tools force supplychains (i.e., suppliers,manufacturing, fulfillment, and customers) to employ arudimentaryandarbitraryprioritysystemcalledprioritybyduedate.Duedatesdrive everything when it comes to execution. Common practice is that if

Page 258: DDMRP - Demand Driven Material Requirements Planning

suppliers are late, it counts against them in their performance report.When itcomes to MRP, all plans are developed assuming that dates are maintained.Everything must go according to plan, or the outcome (final due date) is injeopardybecausethereisnoslackintheplannedschedule.Ifamanufacturerisconsistently not able to meet customer due dates, then there are negativebusiness implications.These include lostopportunitiesand increasedexpedite-relatedexpenses.Companiesareacutelyawareoftheimportanceofhittingduedates, especially in this current hypercompetitive market. This ripplesthroughoutanorganization,reinforcingtheneedtomeasureandactaccordingtoprioritybyduedate.

Thusprioritybyduedatebecomesthedefaultmethodtomaintainexpectedcustomerservicelevels.Whenthingsareclosertobeingdue,theybecomemoreimportant.Iftheyarepastdue,theyareveryimportant.Iftheyarepastpastdue,then they are even more important. Teams of expediters are employed todetermine how important things really are. Schedules are constantly changed(causing even more ripples), and overtime and expensive expedites areemployed inanattempt tosolve theproblem.Duedate isall thatmost supplychainpersonnelknow,andduedateisallthatmostsupplychainpersonnelhaveattheirdisposal.

What if thismodeofoperationismoreapartof theproblemrather thanasolution? This book has emphasized from the beginning the importance ofprotectingandpromotingtheflowofrelevantinformationandmaterials.Whatiftheprevailingprioritybyduedatemodeofoperationisactuallymoredistortivethanrelevant?

Ask any buyer a very simple question: “Would you rather have yoursuppliersbeontimeorneverstockyouout?”Thisquestionusuallyelicitsaveryinteresting reaction.Most respondersblurt out, “On time, of course!”but thenpause with a quizzical look on their face as they think about it harder andconsiderotherpossibilities.Intuitively,theybegintoconsiderthattheremaybeamisalignment betweenwhat companies use as ametric andwhat effect thatmetricmighthave.Suppliers canbe consistently100percenton time, andyetthecompanystillhasshortages.ThisisespeciallytruewhenduedatesgeneratedoutofconventionalMRPsystemsarebasedpurelyontheassumptionsandtheplanatthetimeoftheiractualcreation.Yetweknowthatthoseassumptionsareextremely short-lived, as conventional MRP is highly subject to nervousness(demand signal distortion and change) and supply continuity variability (delayaccumulation). The “real” requirements are changing constantly, but the due

Page 259: DDMRP - Demand Driven Material Requirements Planning

datesstaythesameastheywereonrelease.Evenifasupplierhitstheduedates,there is still a significant chance that there could be shortages or excessinventorybasedon thechanges thathaveoccurredbetween the time theorderwascreatedandthetimeinwhichitwasreceived.

Thusprioritybyduedaterarelyconveystherealday-to-dayinventoryandmaterials priorities. Priorities are not static. They change as variability andvolatilityoccurwithintheactivelifespanofPOsandMOs—thetimefromwhenthey are opened until they are closed. This life span is called the “executionhorizon.”Customerschangetheirorders.Qualitychallengesoccur.Therecanbeweather-or customs-relatedobstacles.Engineeringchangeshappen.Suppliers’capacity and reliability can fluctuate temporarily. The longer the executionhorizonis,themorevolatilearepriorities.Thismeansthatthecompanyismoresusceptible to adverse material synchronization issues and shortages. Thisincreased variability and volatility guarantees that despite our best attempts atplanning, reality will deviate from the plan. Conventional MRP requires thateverythinggoaccordingtoplanfortheduedatestobetherelevantdrivingforcebehindpriority.Moresophisticatedschedulingsystems likeadvancedplanningand schedule or optimization (APS or APO) only make this variability andvolatilityworsewithitsmorefrequentrescheduling.

Additionally, other challenges are associated with driving priority by duedate. Request and promise dates change frequently due to nervousness andsupply continuity variability. These changes often create confusion anddisagreementbetweensuppliersandcustomersaboutthe“real”dates.Supplierscould view their on-time performance as high because they delivered to theirpromisedate,whereascustomersseeitmuchdifferentlyfromtheirviewoftherequestdate.ItisoftenthecasethatMRPsystemswillevenrequestthingswithduedatesinthepast!

Aligning a supplier’s schedule with a customer’s real priorities underconventionalMRPapproaches is a huge challenge.Amanufacturing companycanhaveseveralopenPOstoasupplierallwiththesameduedate.Figure10-2isanexampleofsuchacase.Notethattherearethreeordersalldueonthesameday(POs821158,831145,and831162).

Page 260: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-2DeterminingPOprioritybyduedate

Theseorderscouldhavenothingtodowitheachotherfromthecustomer’sperspective.Fromasupplier’sperspective,however,theymightallconvergeatabottleneckresource.Ifthesupplierrealizesthatitsimplycannotfulfillalltheseordersontherequiredduedates,howcanitdecidewhichisthemostimportant?Optionscan includecalling thebuyerorchoosing theorderbasedonwhat thesupplierperceives tobe itsbestuseof its timeorperhapsby the lowestordernumber.

Ifthedecisionistocallthebuyer,canthebuyerquicklyconveythecorrectpriority?InmostcaseswithconventionalMRPtools,theanswerissimply,“I’llhave to get back to you.” Determining the correct priority will require anadditionalamountofdataanalysisandpotentialcommunicationwithplanning.Additionally,whateveranswerisderivedfromthisconversationwillmostlikelychange as timemoves forward and theMRPdeckof cards gets shuffledonceagain.ThissituationisexacerbatedwhenanAPOorAPSsystemisinplace.

If thesupplier insteaddoeswhat itperceives tobe thebestuseof its timeandcapacity,thefactthatitmightpicktherightpriorityforthecustomerwouldbe completely coincidental. Picking priority based on order number seemsarbitraryatbest.Boththeseoptionsrisksuppliercapacitybeingoutofalignmentwithactualcustomerneed.Isthereasimpleandintuitivewaytofixthis?

Figure10-3isthesamegraphicasFigure10-2butwithoneadditionalfieldaddedthatmakesimmediateprioritydeterminationrelativelyeasy.Thisview,ifprovided to the buyer, would allow for a definitive and quick response.Furthermore, if thisviewwereprovided to thesupplierona recurringbasis, itwouldnegatetheneedforthephonecallinthefirstplace.Additionally,itwouldreducetheriskthatthesupplierwouldmakeadecisioncountertotheinterestofthecustomeroroneofanarbitrarynature.

Page 261: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-3Bufferstatusincluded

Justasininplanning,thereisageneralreferencethatiscolor-basedwithadeterministic percentage. Out of this comes two critical forms of relevantprioritynotavailableinconventionalMRP.Firstisasenseofhowasinglepart’sstatus is relative to its own established buffer level necessary to maintaindecoupling point integrity. Second, it allows a sense of how a part’s status isrelative tootherparts’ statuses.This iscrucial for theexample inFigure10-3,where the supplier needs to quickly understand which order is the mostimportant. Figure 10-3 clearly shows that despite order number sequence andduedate,thehighestpriorityorderisPO831145.

The priority by due date problem does not just affect the traditionalcustomer-supplierrelationship;ithassignificantimplicationsforamanufactureraswell.ThemanufacturingfloorandmanufacturingorderprioritydeterminationisusedtofurtherextendtheprioritybyduedateversusprioritybybufferstatuscomparisoninthisnextexampleinFigure10-4.Thesearemake-to-stock(MTS)manufacturingorderswithdifferentduedatestochoosefrom.

Iftoday’sdateis5/11,thismanufacturingresourcehastodeterminewhichisthemostimportanttorunnext.Inatypicalenvironmenttherearethreeordersthatwouldcontendforthehighestpriority:MOs831145,821158,and831162.TheotherMOsaretypicallydisregarded.Shouldthemanufacturersimplyworkthesequencegiventoit?Thatistheschedule;yetdowehaveenoughrelevantinformationtomakeagooddecisiontobestpromoteandprotectflow?

Figure10-5 now gives themanufacturing floor amuch clearer picture onpriority. In thiscaseoperatingaccording toprioritybyduedatewould lead tothewrong sequence froma flowperspective.Theorder thathas the latestduedate (MO645181with a date of 5/14) has the buffer that is in real jeopardy.Ultimately,prioritybybufferstatus isaboutaligningefforts tobestprotect theDDMRP model. The DDMRP model was built under careful and strategicconsideration (see Chapters 6, 12, and 13). Aligning efforts to the strategic

Page 262: DDMRP - Demand Driven Material Requirements Planning

operatingmodelofabusinessinrealtimeisthepinnacleofeffectiveexecution.In fact, ifbufferstatusesaredisplayed, that raises twoquestions:Whydisplayduedatesatall?Whatadditionalvaluedotheybringversustheriskofconfusionanddistortion?

Seeingcolorandbufferpercentagesandrelatingthattopriorityisintuitiveformostpeople;redisdanger,yellowiscaution,andgreentypicallymeansOK.How these colors and percentages are determined are key to effective andcollaborativeexecution.

FIGURE10-4Dispatchlistfor5/11

FIGURE10-5Manufacturingschedulewithbufferstatus

PlanningVersusExecutionDisplay

Thebufferstatuscanprovideaquickand intuitiveway toalignefforts tobestprotecttheDDMRPmodel.Bufferstatusalertsdonotusethenetflowequation;theyutilizeon-handvaluesonly.Thisseparatestheactivitiesdedicatedtosupplyordergenerationfromtheactivitiesdedicatedtoopensupplyordermanagement.Thus buffer status alerts represent a different perspective than DDMRPplanning.

AsdiscussedinChapter9, theon-hand target range iscalculatedas topofred(TOR)toTOR+thegreenzonevalue.Theon-handpositionshouldfluctuatebetweenthosevalues.Iftheon-handisconsistentlyaboveTOR,thenwedon’t

Page 263: DDMRP - Demand Driven Material Requirements Planning

have any real concerns. The buffer position is functioning as planned; it hasadequateon-handtoensuretheintegrityofthedecouplingpoint.ItisOKfroman execution perspective, and “OK”means green. Thuswhen considering thebuffer statusalerts for themanagementofopensupplyordersagainstbufferedpositionswhenon-handisaboveTOR,thebufferstatusalertwilldisplaygreen.

Yellowand red aredeterminedby the severityof theon-hand situation inrelationtothetotalredzonevalue.Thisshouldmakeconceptualsensewhenyouconsiderthenatureoftheredzone.Theredzoneistheembeddedsafetyinthebuffer—thecushionagainstvariability.Whenon-handdipsintothatsafetyzone,acautionaryflagshouldgoup;itshouldbecoloredyellow.Ifon-handcontinuestoerodeintothesafetyzone,thatcautionaryflagbecomesmoreurgent;itshouldchangefromyellowtored.

Thisrationalegiveswhatisneededtodeterminecolorandpercentage.First,there is apoint atwhichyellowshould turn to red.This is called theon-handalertlevel.Theon-handalertlevelisthepointthatescalatespriorityfromyellowtored.Themostcommonwaytodeterminetheon-handalertlevelistosetitto50percentof the redzone.Second, thediscretepercentagevalue is calculatedbasedupon theon-hand level as apercentageof theembedded safety (the redzone).Thelowerthepercentage,thelesssafetyremains.

Asanexample,takeapartwithaTORof50andon-handalertlevelof25(50percentoftheredvalue).Onday1anon-handvalueof72wouldproduceanexecutionbufferstatusdisplayingagreenvalueof144percent.Thecolorgreenisassignedduetotheon-handquantitybeingabovetheTORvalue.Onday2anon-handalertvalueof26wouldproduceanexecutionbufferstatusdisplayingayellow value of 52 percent.Yellow is displayed because the on-hand value isbelowtheTORvalueofthebufferbutabovetheon-handalertvalue.Onday3anon-handalertvalueof20wouldproduceanexecutionbufferstatusdisplayinga red value of 40 percent. Red is displayed because the on-hand quantity isbelowtheon-handalertvalue.Figure10-6depictsthebufferstatusalertforthisexamplepartoneachsuccessiveday.

FIGURE10-6Bufferstatusalertexample

Page 264: DDMRP - Demand Driven Material Requirements Planning

Thisshiftincolorassignmentcanbedifficulttoimmediatelygraspbecauseit means that an on-hand value that would represent a yellow designation inplanning is actually green from an execution perspective. The key tocomprehendingthischangeissimplytounderstandthattheequationanddisplayforplanningare separateanddistinct fromexecution.Planningutilizes thenetflow equation (ofwhich on-hand is only one component) in order to generateorders. The execution view with buffer status alerts is designed to providerelevantinformationonwhatiscriticalwithregardtomanagingordersthathavealready been created.Which dowe need to expedite?Which can afford to belate?

Figure10-7representstheplanningsimulationfromChapter9butdisplayedfrom an execution view. This chart only shows the on-hand value against thebufferstatusalertzonesoverthecourseofthe21days.Thisgivesaclearpictureabout what a planner would be thinking with regard to on-hand status (asopposedtosupplyordergeneration)onanyparticulardayoverthe21days.

The TOR value of this part was 52, andwe are setting the on-hand alertlevelto50percentofthered,whichis26.Figure10-8representsaday-by-daybufferstatusforthesimulatedpart.AsindicatedinbothFigures10-7and10-8,thereisonlyonedayinwhichtheon-handleveldropsbelowtheon-handalertlevel.Duringthesimulation,thatdropactuallypromptedtheplannertoexpediteanexistingorder.Thisisthepurposeofbufferstatusalerts—toquicklypointoutwheresafetyhasbeensignificantlyeroded.

FIGURE10-7Planningsimulationwithexecutioncolorscheme

Page 265: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-8Day-to-daybufferstatusforasimulatedpart

Bufferstatusalertsfocusontheon-handposition.Thecurrenton-handalertfocusesontoday’son-handbufferpositionagainstthedefinedexecutionbufferstatus definition. The projected on-hand alert focuses on a projected on-handlevelinthenearfuturetypicallyuptooneleadtimeinthefuture.

CurrentOn-HandAlert

The current on-hand alert is designed to show personnel what replenishedpositionsarecurrentlyintroublefromanon-handperspectiveonly.Forplanningand purchasing personnel, these alerts aremeant to identify partswhere opensupplymayneedtobeimmediatelyexpedited.Formanufacturingpersonnel,thecurrenton-handalertprovidesrelevantinformationaboutwhichmanufacturingordersshouldtakeprecedence.

Page 266: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-9Samplemanufacturedparts’currenton-handalertscreenforCompanyABC

Figure10-9 iswhat a currenton-handalert screenmight look like forourprevious example for Company ABC. All on-hand alert levels are set to 50percentoftheredzonevalue.Initssimplestform,thealertscreenonlyneedstoprovidea listofpartswithprioritysortedfromhighest to lowest.Theon-handalert screen is displaying for manufactured items only. This screen would beutilized by a planner. A separate on-hand alert screen would most likely beutilized by purchasing personnel for purchased items. Figure 10-10 is thepurchasingon-handalertscreenforCompanyABC.

Partsingreencanbeexcludedfromthelist,astheyarenotencroachingintothesafetyzone.Ofcourse,afterbeingalertedtoapartintrouble,mostpersonnelwould need quick access to additional information that may be involved inactions related toattempting toprotect thebufferposition—mostnotablyopensupply orders against the buffered position that are candidates for potentialexpedite.

Figure10-11depictsadrill-downviewfromthemanufacturedpartscurrenton-handalertscreen.ItdisplaysopensupplyordersagainsttheFPBposition.Ineachcasearequestdate(theduedategeneratedusingdecoupledleadtime)andpromise date (the due date that the scheduler has confirmed) are displayed.Additionallyweseethat today’sdateisMay8andthetopofredandon-handvaluesaredisplayedthatgenerateanon-handalertstatusof20.7percent.Thereare three openmanufacturing orders supplying FPB (MOs 15781, 15852, and15999).Twoof theseorders arepastdue.One is severelypastdue relative toFPB’sADU(100).Thatwouldmost likelyexplaintheheavilyerodedon-handposition.

Page 267: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-10Samplecurrenton-handalertscreenforpurchaseditemsatCompanyABC

FIGURE10-11OrderactivityforFPB

Theplanner’simmediateconcernmustbeforMO15781.Acallismadetothefinalqualitycontrolstationforinformationontheorder.Theplanneristoldthattheorderhaspassedinspectionandtheentirequantitywillbemovedtoon-handwithinthehour.Withanadditional363units,theon-handlevelwillmoveto475(112+363).Anon-handvalueof475willmovetheon-handalertfromredat20.7percenttoyellowat88percent.Figure10-12isthecurrenton-handalertdisplayafterMO15781hasbeencompletedandtheFPBon-handpositionhasbeenadjusted.Atthispointtheplannerfeelscomfortablewiththepositionandtakesnofurtheraction.

Figure10-13depictsadrill-downviewfromthepurchasedpartscurrenton-hand alert screen. It displays open supply orders against the 401P position.Today’s date isMay8.There are three open orders against the 401Pposition(POs24366,24413,and24587).Thebuyer’smostimmediateconcernshouldbeforPO24366,whichshowsthattheorderisintransitandthesupplierhasgivenatrackingnumberforit.Whencheckingwiththelogisticscompany,thebuyeristold that the order will arrive late in the day, possibly after hours, due toinclement weather. The buyer immediately makes arrangements for receiving

Page 268: DDMRP - Demand Driven Material Requirements Planning

personneltostaylatetomeetthetruck.Fordistributionnetworks,thecurrenton-handalertworksthesamewayas

describedinthepreviousexampleswiththeexceptionthattheon-handstatusforeach part number or SKU must be location specific as well. Figure 10-14displays the current on-hand alert for a distribution network. There are twolocations(RiversideandDallas)thathaveseverelyerodedon-handpositionsforproduct FPT. Additionally, the Chicago location for FPT is near the on-handalertthreshold.

FIGURE10-12Theupdatedcurrenton-handalertscreen

FIGURE10-13Orderactivityfor401P

Anexpandedversionofthecurrenton-handalertdisplaywouldgiveaclearpictureofbufferintegrityacrossthenetworkforallproducts.Figure10-15isanexampleofsuchaview.Therearefourlocations:Riverside,Dallas,Atlanta,andChicago.Thenetworkdistributesfiveproducts(FPD,FPE,FPG,FPK,andFPT)from these locations. In this view, an additional column can be added thatsummarizestheoverallnetworkhealthforeachproduct.Thisisrepresentedbythe“Network”column.Howthiscolumnispopulatedcanbeconfiguredbasedontheindividualnetwork(all locationsmaynothaveequalimpactorstrategicimportance). In this case a very simple schema creates a color coding fornetworkhealth;one red locationequalsayellowdesignation; twoormore red

Page 269: DDMRP - Demand Driven Material Requirements Planning

locationsequalareddesignation.Thisnetwork-widecurrenton-handdisplayincombinationwithprioritized

share logic can help personnel decidewhich locations are in themost troublewhentherearelimitedamountsofsupplyavailable.Thiscanalsopointtobetteroptions to help positions in trouble. For example, for product FPT there islimitedstockatthehub.Thehubmustserviceitsowncustomersaswellasthespoke locations.Oneoption tohelp theDallas locationwouldbe tocross-shipinventory from the Atlanta spoke. It is a less than ideal solution, but it mayprovidethemosteffectiveandexpedientwaytoprotectsalesinallmarkets.

FIGURE10-14Currenton-handalertdisplayforadistributionnetwork

FIGURE10-15Networkon-handstatus

ProjectedOn-HandAlert

Theotheron-handfocusedbufferstatusalertistheprojectedon-handalert.Theprojectedon-handalertcalculatesredzonepenetrationinthenearfutureinordertowarn supply chain personnel about impending buffer integrity problems. Inorder todothis, thealertrequires timingandquantity informationfrombotha

Page 270: DDMRP - Demand Driven Material Requirements Planning

demandandsupplyperspective.Likethenetflowequation,itutilizeselementsofdemand,supply,andon-hand.Unlikethenetflowequation,itisnotdesignedto recommend new supply orders but to point out positions that may needexistingsupplyordersexpeditedinordertomaintainthebuffer’spurpose.

The projected on-hand alert takes today’s on-hand inventory and projectson-hand status for each future day based on the average daily usage or thequantityandtimingofknowndemandallocations,dependingonwhichislarger,andthequantityandtimingofexpectedsupplyorderreceipts.Inthiscase,duedatesarerelevantforbothdemandallocationsandsupplyorderreceiptsbutonlytomakeaprojectionofwhatbufferstatusmaylooklike.Itisworthmentioningthat a projection is exactly that—a projection. It is the best guess about theneartermfutureusingthebestinformationavailable:knowndemandallocations,calculatedrateofdemand,andknownopensupplyandcurrenton-handlevels.

To demonstrate how the projected on-hand alert works, the 21-daysimulation of DDMRP supply order generation is revisited from the previouschapter. The format of the simulation provides an excellent way to articulatehow this particular alertworks. Figure10-16 is the simulated environment onday1ofthesimulation.

Onday1theon-handinventoryis65units.Theprojectedon-handalertwillstartwith thatvalueanduseknowndemandallocationsandanticipatedsupplyorder receipts to project on-hand levels for each day through one lead time(sevendays)ofthepartbeginningwithday2;thisalsocorrespondstotheorderspike horizon used in the simulation. Figure 10-17 is a projection of on-handlevels against thebuffer status alert color zones.Theon-handalert level is 50percentoftheredzone.

Page 271: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-16Day1

FIGURE10-17Projectedon-handalertstatusbyday

The projected buffer status is generated by the projected starting on-handpositionagainstatopofred,whichis52.Inthiscasethereappearstobelittlecauseforconcernifthingsgoevenremotelyclosetoplan.

Whatiftheknowndemandallocationsarenotavailableorareonlyvisiblein the very nearterm future? In these cases, the projected on-hand alert canproject on-hand levels using ADU. Figure 10-18 depicts this simulated partusingonlyADUas thedemandvalue foreach futureday.Withoutaccountingforknowndemandallocations,especiallyqualifiedspikes,thisformofthealertmayoverestimatefutureon-handpositions.

AmoreconservativeapproachwouldbetoconsiderbothADUandknown

Page 272: DDMRP - Demand Driven Material Requirements Planning

demandallocationsandtakethegreaterofthetwoonanyparticularday.Figure10-19 depicts projected on-hand alert status for the simulated part using thehighest value each day, either known demand allocations or the average dailyusage value. The shaded boxes in the “Average Daily Usage” and “SalesOrders”rowsrepresentthequalifyinghighestdemandinputforthatday.

FIGURE10-18ProjectedbufferstatusalertwithADU

FIGURE10-19Projectedon-handalertusinghighestdailydemandinput

Whenthereislesschangeinanenvironment,theprojectedon-handalertisless impactful. In more stable environments there is less need for advancewarning. But what about environments where dates are constantly moving?Suppliers push back delivery dates.Manufacturing is frequently shuffling theschedule,movingjobsforwardandbackward.Inthesecasestheprojectedbufferstatusalertcanbeextremelyvaluable.

Figure10-20isasampleprojectedon-handalertscreenforCompanyABCpurchased parts. The projected on-hand alert should alert to the first on-hand

Page 273: DDMRP - Demand Driven Material Requirements Planning

alert-levelpenetrationwithinoneleadtimeintothefuture.Ifapartisprojectedtobe stockedout, adark redcolorwillbedisplayed. InDDMRP thedark redcolor is more severe than red. A negative percentage means that the part isprojected to be stocked out with demand. When a part is stocked out withdemand, flow is definitely impeded. These potential situations are crucial forvisibility.

FIGURE10-20Sampleprojectedon-handalertscreenformanufactureditemsatCompanyABC

There are three parts on the alert screen, and one requires immediateattention.Part401PisprojectingtobestockedoutwithdemandtwodaysfromnowonMay5.Inordertounderstandwhyitisprojectedtobestockedoutwillrequireadrill-downonthe401Porderactivity.Atime-sequencedsummaryofplannedopensupplyordersisneededwithknowndemandallocationsinordertodeterminepotentialcorrectiveactions.

Figure10-21isanorderactivitysummaryfor401P.Demandallocationsandopensupplyordersarelistedinthecolumn“Order#.”Since401Pisapurchaseditem, anymanufacturing order represents a demand allocation,while purchaseordersrepresenttheopensupplyorders.Eachorderhasaquantityassignedtoit.The orders are sequenced based on the “Date” column. For manufacturingorders,thedaterepresentstheparentorderreleasedate.Thisiswhentheorderwillrequirethestatedquantityof401P.Thatquantityissubtractedfromtheon-hand position on that future date. For purchase orders, the date represents theanticipated delivery of the order based on the best information available. Thequantityofthoseordersisaddedtotheprojectedon-handpositionatthosedates.The“ProjectedOn-Hand”columndisplays theprojectedon-handbalanceaftertheorderquantitieshavebeensubtractedoradded.

The current on-hand quantity is 2,108. There is one manufacturing order(MO 2574) set to release on May 5 for a quantity of 2,213. This demandallocation against the401Pposition exceeds theprojectedon-handbalanceon

Page 274: DDMRP - Demand Driven Material Requirements Planning

May5by105units.Thenextsupplyorderisnotexpecteduntilthedayafterthemanufacturing order’s scheduled release.Additionally, another open supply isdue the very next day. Apparently this issue was caused by a machinebreakdown at the supplier that delayed PO 61325 by three days. An updatedsupplierpromisedateofMay6hasimmediatelygeneratedaprojectedon-handalert.

FIGURE10-21Orderactivitysummaryscreenfor401P

What can the buyer do? The supplier cannot get the order in sooner thanMay 6, and the manufacturing orders are not under the buyer’s control. Thebuyer shouldconsultwith themanufacturingplanner about thebestoption forthecompany.Planningwillalsoseethisproblembutfromtheperspectiveofadifferenttypeofexecutionalert—amaterialsynchronizationalert.Thisexamplewillbecontinuedlaterinthe“MaterialSynchronizationAlerts”section.

In most environments, paying attention to projected on-hand alerts andacting appropriately will reduce the number of current on-hand alerts. It willnevereliminatetheon-handalerts,asvariabilityintheshorttermwillstillplayarolewithon-handlevels(qualityholds,breakdowns,spoilage,etc.).

Amixed-mode operation has bothmake-to-stock (replenished) orders andmake-to-order (nonbufferedanddirect tocustomer)ordersmoving through thesame resource base. In conventional operations most operations defer to the“real”customerorder:themake-to-orderitemanditsduedate.YetinDDMRP,thisdefaultmodeofoperationrequiresreexamination.

Itiscrucialtoprotectcustomercommitments,butitisworthareminderthataDemandDrivenOperatingModel is configured to best protect total flow.Amake-to-orderorderrepresentsonecustomer,whileastrategicstockbuffercanrepresenthundredsofcustomers.Thustheactualpriorityconsiderationbetween

Page 275: DDMRP - Demand Driven Material Requirements Planning

stockbuffersintroubleandamake-to-orderorderwithafirmcustomerduedatecan really only be defined at the timewhen the conflict occurs. Choices willhave to be made, but those choices should be impacted by the relevantinformationcontainedinthebufferstatusalerts.

SynchronizationAlerts

Asdiscussedin thepreviouschapter therewillstillbemanydependenciesandsynchronizationpoints that arenotdecoupled inDDMRP.Whenvisibilitycanbe promoted to keep these points from experiencing as little disruption aspossible, then there will be less supply continuity variability passed to thedecouplingpointsandcustomers.Whenlessvariabilityispassedtodecouplingpoints, there is less working capital required to buffer those positions.Whenthereislessvariabilitypassedtothecustomers,salesaretypicallyprotectedandpromoted.

DDMRP uses two types of execution alerts to promote visibility andmanagementtosynchronizationpoints:materialsynchronizationalertsandleadtimealerts.

MaterialSynchronizationAlerts

Materialsynchronizationalertsdisplaysupplyshortfallsagainstknowndemandallocations.Inordertodemonstrateamaterialsynchronizationalert,thepreviousprojected on-hand alert example (401P) is continued but from a differentperspective.Thatprojectionhascreatedapotentialsynchronizationproblemforanorderrequiring401P(MO2574).Theordercallsforaquantityof2,213,andthe projected on-hand balance is expected to be short by 105 pieces.Manufacturing orders are under the control of planning, not purchasing. Thusplanningpersonnelneedtoseewhentheordersandbufferpositionsundertheircontrol are in potential trouble. Remember, more visibility to relevantinformationdirectly impacts theamountofvariabilityexperienced, transferred,andamplifiedinanenvironment.

Figure10-22providesanadditional reference for thisexample. Itdisplaysthe product structure of FPA. The boxed-in area is the relevant part of theproduct for this example. The intermediate component 201 is the nextdecouplingpointbeingfedby401Pand301.Part201isbuffered.

Part 301 is a nonbuffered item. Thus a projected stockout at 401Pwouldcreateasynchronizationproblemforthedemandallocationsdriventhroughthe301 position from supply orders generated at 201 (refer to the decoupled

Page 276: DDMRP - Demand Driven Material Requirements Planning

explosion section in Chapter 9 for additional information about decoupledexplosionbehavior).Supplyisinsufficienttocovertheentiredemandallocation.It is projected to be short by 105 pieces. When this occurs, a materialsynchronization alert will be issued for the manufacturing order(s) affected.Figure10-23isanexampleofamaterialsynchronizationalertforMO2574.

The material synchronization alert needs to show planners the relevantfactors that trigger that alert as well as the factors involved in a potentialreconciliationoftheissue.Thealertdisplaystheordernumber(MO2574)thatisimpacted,thepartnumberthattheorderisagainst(301),thereleasedateoftheorder(May5),andthequantityoftheorder(2,213).The“Shortage”columnisdisplaying the shortage toMO 2574 (–105) and the part number creating theshortage(401P).The“ParentBufferStatus”columnisprovidingareferencetothestateofthenext-higherbufferedpositionintheproductstructure(201)thatMO2574issupplying.Inthiscaseitisdisplayingthecurrentbufferstatusalertof201.Thebufferappearstobeinrelativelygoodshape.

FIGURE10-22Productstructureforexample

FIGURE10-23MaterialsynchronizationalertforMO2574

Itshouldbenotedthatbufferstatusalertsareavailabletotheplannerforallmake item buffers including 201. This material synchronization alert should

Page 277: DDMRP - Demand Driven Material Requirements Planning

impacttheprojectedon-handof201(less105thanplanned)onthedatethatMO2574 is expected to be received at that position. This amount is insufficient,however, togenerateaprojectedon-handalertagainst the201position.Figure10-24showstherelationshipsbetweentheexecutionalertsinthisexampleandwhoisalertedbythem.Thebuyerseesthecurrentandprojectedon-handalertsfor 401P, while the planner sees alerts for manufactured items—the materialsynchronizationalertsfor301andcurrentandprojectedon-handalertsfor201.

Intheprojectedon-handalertexample,thebuyer’shandsareessentiallytiedwith regard to401P; the supplier simplycannotget the410Psupplyorders insooner. Yet there are options open to the planner to deal with the materialsynchronizationalertforMO2574.

Option1.Movetheorderreleasedatebacktocoincidewiththenextplanned401Psupply.Since theparentbufferstatus is inrelativelygoodshape,ashortdelayofthesupplyordershouldhaveminimalimpactonthebuffer;thereisplentyofsafetyremaining.Thisoptionallows the planner to protect the scarcity in the 401P position if401Pmighthaveadditionalneedsdefinedintheshortterm.

Option2.Reducetheorderby105pieces.Thiswillgetsupplytothe201 position as planned from a timing perspective. This optionmightbeselectedifthe201positioniscurrentlyorprojectedtobeintroubleinthenearterm.

FIGURE10-24Executionalertrelationshipsandtheirpersonnelassignments

Page 278: DDMRP - Demand Driven Material Requirements Planning

Option3.Movetheorderreleasedatebacktocoincidewiththenextplanned401PsupplybutthenexpediteorderMO2574.Thiswouldservetoprotectbothbufferedpositions(401Pand201)ifnecessarybut would require additional manufacturing efforts or potentialdisruptionfromtheexpedite.

All theseoptionswould remove thematerial synchronization alert and theprojectedstockoutalert(aprojectedon-handalertwouldstillbepresentbutjustnotassevere).

In DDMRP a material synchronization alert will occur for three primaryreasons:

Insufficient supply. Material synchronization alerts are triggeredwhenlevelsofsupplyarelessthanrequired.Thisisparticularlytruewhenavarianceoccursonnonbuffereditems.Thiscanhappendueto quality or significant yield loss issues. Insufficient supply canalsobetriggeredbyperiodsofsignificantlyheavydemand.

Latesupply.Whenapromisedateonasupplyorderispushedlaterintime,itmaytriggermaterialsynchronizationalerts.Thiscanalsocreateacurrentorprojectednegativeon-handpositioninthecaseofabufferedpart.Thiswoulddrivecurrentoron-handalerts for thebuffered item and material synchronization alerts for demandallocationsthatareshorted.

Figure10-25isaplannedrelationshipbetweenamanufacturingorder (demand allocation) and a supply order. The supply order isrequiredtohavefullallocationforthemanufacturingorderonMay5.

Figure10-26showsasupplyorderthathasmovedlaterintime.It is now expected to be received on May 7. This creates asynchronizationissueforthemanufacturingorderonMay5.

If the promise dates of nonbuffered parts are pushed later intime, then material synchronization alerts typically follow anydemandallocationsagainst thosenonbufferedpositions.Thepart isnonbufferedandisplannedtonettozero.Thusanydelaywillcreateasynchronizationproblembecausethereisnobuffer.

Page 279: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-25Aplannedrelationshipbetweenasupplyorderandmanufacturingorder

FIGURE10-26Asupplyorderexpectedlaterthanplanned

EarlierStartCommitment.Ifthestartdateofademandrequirementis changed to an earlier date, material synchronization alerts mayfollow.Thisisparticularlycommonwhenthedemandrequirementis dependent on nonbuffered components. Since the component isnonbuffered, there is a dependent relationship based on when thedemand was originally required and the timing created for thesupplyorderofthecomponent.

Figure10-27illustratesthemanufacturingorderstartdatebeingmovedearliertoMay3.ThiscausesasynchronizationproblemwiththesupplyorderthatissettoarriveonMay5.

LeadTimeAlert

A lead time alert is an execution alert for strategic nonbuffered items. Thesestrategic nonbuffered items may not come in sufficient volume to justify

Page 280: DDMRP - Demand Driven Material Requirements Planning

stocking but typically create major synchronization issues when they arerequired.Theycould come fromaproblematic supplieror from a problematicgeographicregionorbesubjecttohandlingortransportationdifficulties.Theseitemsareplannedinthesamemannerasanynonbuffereditem;yettheyarepaidspecialattentionintheexecutionhorizon.

FIGURE10-27Amanufacturingordermovedearlier

Thebasicideasbehindtheleadtimealertareasfollows:

1. It is beneficial to know about synchronization issues before asynchronization issue occurs. This allows adjustments andcontingenciestobeplannedandexecutedinordertohavebetterflowperformance.

2.Supplierperformancecanbeinfluencedbyhoworganizedaparticularcustomer iswith regard to the customer’s orders. If a customer cancommunicateinawaythathelpsprioritizeasupplier’sactions,thereis typically an appreciation and attention paid to that customer’sorders.

3.Itcanbebeneficialtoestablishaclear“papertrail”showingwhatledto any synchronization issues or costs to avoid them. This is ofparticularimportanceinindustriesthatusefinancialpenaltiesforlatedeliveries. Establishing a clear chain of events that led tosynchronization issuesmay provide recourse for cost recoverywithproblematicsuppliers.

Page 281: DDMRP - Demand Driven Material Requirements Planning

Leadtimealertsareusedtopromptpersonnel tocheckuponthestatusofcritical non-stocked parts before those parts become a synchronization issue.Leadtimemanaged(LTM)supplyordersaretracked,andatadefinedpointinthepart’sleadtime,personnelarepromptedforfollow-up.Typically,thispointistwo-thirdsofthewaythroughtheleadtimeofthepart,asmeasuredfromthepromisedate.Thefinalthird,calledtheleadtimealerthorizon,isthendividedinto threezonesof typicallyequalproportions(green,yellow,andred).Figure10-28illustratestheleadtimealertconcept.

The lead timemanagedparthasa lead timeofninedays.Typically,LTMpartshavemuchlonger lead timeitems(30+days).Apartwitha lead timeofnine dayswas chosen as an example in order tomake the dates easier to seegraphically.Asupplyorderfor thatpart is launchedonMay1andisdueninedays lateronMay10.The lead timealerthorizon is the last threedaysof theorder’sleadtime.Whentheorderisthreedaysawayfrombeingdue(onday7),theorderwillenterthegreenzoneoftheleadtimealerthorizon.Thisisthefirstleadtimealertforthispart.Onday8itenterstheyellowzoneoftheleadtimealerthorizon.Onday9itisintheredzoneofthealerthorizon.

FIGURE10-28Aleadtimemanagedpartandleadtimealertzones

Thezonecolorsof the lead timealerthorizonsaresimplyacountdowntotheexpectedarrivaldateoftheorder.Theprogressionfromgreentoyellowtored is onlymeant to connote how close the order is to that date. Red simplymeansthatitisdueinthenearterm.Iftheorderdoesnotarriveontheexpecteddate,thenitprogressestoadarkredcolor,meaningitislate.Iftheordergoeslate, itwillmost likely create amaterial synchronization alert for the demandallocationthatcausedittobeorderedinthefirstplace.

Theentryofanorderintotheleadtimealerthorizonshouldpromptabuyer

Page 282: DDMRP - Demand Driven Material Requirements Planning

orplannertofollowupanddocumenttheorderstatus.Anyinformationderivedfromthatfollow-upcouldbevaluableinpredictingandmitigatingtheimpactsofsynchronization issues and, aswell, could be valuable in documenting a clearpicture about why any synchronization issues would occur. Follow-up anddocumentationshouldoccuras theorderprogresses intoeachzoneof the leadtimealerthorizon.

Note that the lead time alert horizon is not an insertionof additional timeintothepart’sleadtime—itoverlapsthelastportionoftheleadtime(typicallythelastthird).

Figure10-29depictsasimpleversionofaleadtimealertscreen.The“LeadTime” column is present in order to understand the size of the lead time alerthorizonandeachzone.Theleadtimealerthorizonhasbeensettoone-thirdofthepart’sleadtime.PartPPLhasa90-dayleadtime,givingitaleadtimealertzoneof30days.Eachzone is10days insize.The lead timealerthorizon forPPFis20days.Itsgreenzonevalueisset to6,yellowto7,andredto7.TheleadtimealerthorizonforFPZis15days.Green,yellow,andredareallsetto5days.

In the lead timealertscreen inFigure10-29, thedate isMay5.Therearethreeorderswithleadtimealerts.Eachorderisinadifferentzonewithinitsownrespectiveleadtimealerthorizon.PO112032isintheredzoneofitsleadtimealerthorizonwith7daysremaininguntilexpectedreceipt.Itstotalredzoneis10days.Ithasbeenintheredzoneforthe3precedingdays.

The“Current?”columninFigure10-29 tells theplannerorbuyerwhethertherehasbeenanupdatednoteontheorderforthezonethattheorderisin.Thecolorgreenand“YES” indicate thatanotehasbeenenteredagainst thiszone.Thusthereshouldbeaminimumofthreenotesineachorderrecordastheorderprogresses through its lead timealert horizon.BothPO113562andMO5741have yet to have their zone status update recorded. They are colored red anddesignatedwitha“NO.”

FIGURE10-29Asimpleversionofaleadtimealertdisplayforbuyersandplanners

Page 283: DDMRP - Demand Driven Material Requirements Planning

FIGURE10-30Updatedleadtimealertscreen

If date revisions occur, this can adjust the lead time alert status. Forexample, if the zone status follow-up for PO 113562 resulted in the buyerfinding out about a significant delay, then the new promise date should beimmediatelyenteredintotheorderrecord.Ifthesupplierisnowsayingthatdueto equipment failure the orderwill be delayedby five days, the newdue datemust be adjusted toMay 22. Figure 10-30 shows the updated lead time alertscreen.

PO113562nowhasaduedateofMay22,placingitinthegreenzoneofitsleadtimealerthorizon.Itszonestatusiscurrent,anditwillgothroughatleasttwo more zone status updates (yellow and red). Furthermore, it has beendocumented why the date has changed. This date revision has most likelycreatedamaterialsynchronizationalert(s)and/orprojectedon-handalerts.

The lead time alert horizon does not have to be set as a percentage of apart’sindividualleadtime.Agloballeadtimealerthorizoncouldbeappliedtoallleadtimemanagedpartsorgroupsofpartswithsimilarleadtimeproperties.

Synchronization alerts by themselves can provide excellent visibility forenvironmentsinwhichthereis limitedopportunitytoemploydecouplingpointbufferssuchasengineer-to-orderandextensivemake-to-order.

Summary

This chapter has explained four basic concepts and alerts for supply ordermanagement in DDMRP environments. These alerts and the concepts behindthemaredesignedtocreatehighlyvisibleandcollaborativeexecutionacrosstheDemand Driven Operating Model. There are nearly limitless permutations oftheseconceptsbasedonthecircumstancesofindividualenvironments.Allthesepermutations focus on current and projected decoupling point buffer integrity,ordersynchronization,andthepromotionandprotectionoftheflowofrelevantinformationandmaterials.

Page 284: DDMRP - Demand Driven Material Requirements Planning

CHAPTER11

DDMRP Impacts on the OperationalEnvironment

SofarthisbookhasdescribedhowDDMRPworkswithregardtosupplyordergeneration and management. This chapter is about how DDMRP affects thelarger operational environment around it, particularly vertically integratedmanufacturers or larger manufacturing environments that have complexschedulingandshopfloorexecutionneeds.

DDMRPStrategicBufferCriteria

AppreciatingtheuniquecapabilitiesofDDMRPbuffersandtheimpacttheycanhave throughout a larger manufacturing environment requires a revisit of thebasicelementsofthebuffers.Thepreviousfivechaptershaveservedtodescribethepositioning,sizing,andoperationalaspectsof thesetypesofbuffers.Thesechaptershavedefinedthenecessaryconditionsforastockbuffertobeastrategicbufferandeffectivelyminimizeormitigatethebullwhipeffect.Thesenecessaryconditionscanbesummarizedintosixcriticaltestsforastockbuffertobecalleda strategicDDMRP replenishmentbuffer.Withoutmeeting these conditions, abuffer is not DDMRP compliant, will not sufficiently dampen the bullwhipeffect,andwilloftenforceexpensiveperformance-erosivecompromisesintoanenvironment.

TheDecouplingTest

As previouslymentioned, decoupling is about creating independence betweentwolinkedprocesses,events,orareas.Astockbufferthatpassesthisteststopsthe transference and amplification of variability (demand signal distortion andsupplycontinuityvariability)upanddownthechainaswellasbreaksthelead

Page 285: DDMRP - Demand Driven Material Requirements Planning

timedependencyequationforbuffersizingandsupplyordergeneration.

TheBidirectionalBenefitTest

Well-manageddecouplingpointbuffersprovidebenefitsforbothconsumersandsuppliers of the position. DDMRP buffers clearly pass this test. Processes orareasthatconsumestockfromthepositiongetinstantavailabilityand/orshortertimes, as the buffered positions are intended to always have stock available.Processesorareasthatsupplythebufferedpositionsgetaconsolidateddemandsignalthatcorrespondstoactualneedandorderpatterns.

TheOrderIndependenceTest

Order independencedistinguishes inventory that isassigned toaplannedstockpositionfromworkinprocess.Work-in-processinventoryisalreadycommittedto a particular order—it is unavailable to any other orders (withoutmaking adiversionbetweenorders).Atrueplannedstockpositionholdsinventoryforanypotential requirement from a consumer, whether it be different customers ordifferentparentitems.Thisisofparticularimportanceinenvironmentsthathavesharedresourcesandsharedcomponents.

This is essentially the difference between a make-to-stock system and amake-to-order system. Both systems have inventory in them beyond thepurchasedmaterial level, but thenatureor flexibilityof the inventory ismuchdifferent.ThismeansthataDDMRPsystemisprimarilyamake-to-stocksystemwithenhancedrulesetsabouthowthestrategicstockpositionsareplaced,sized,andmanaged.

ThePrimaryPlanningMechanismTest

Strategically decoupled stocking points create shorter and independentlymanaged planning and execution horizons. All buffer planning (supply ordergeneration) elements (inputs and outputs) are performed at the buffer itself.Inputs include thebuffer parameters and settings and all relevant demand, on-hand,andopensupplyinformation,whereasoutputswouldsimplybeasupplyorderwithaquantitytorestorethepositionandrequestdateforreceipt.

TheRelativePriorityTest

Both for planning (supply order generation) and execution (supply ordermanagement), a buffered position must provide a sense of priority for the

Page 286: DDMRP - Demand Driven Material Requirements Planning

position itself aswell as anunderstandingofhow thatpriority relates tootherbufferedpositions. InDDMRP, this isprovided throughageneral referenceofcolorandadiscretereferenceofpercentage.Therelativeprioritytestisaboutabufferedposition’sabilitytoconveyrelevantinformationthatallowsoperationspersonneltoquicklydeterminewhichitemsrequirethemostattentionorscarcematerialsorcapacity.

TheDynamicAdjustmentTest

Anysystemthat isexpectedtoenduremustbeabletoadapt tothecomplexityand volatility experienced in today’s supply chains. In DDMRP, the strategicstockingpoints are theprimary shock absorption system in thepromotion andprotection of flow in a Demand Driven Operating Model. To that extent theshockabsorptioncapabilitymustbeable toadapt(raiseor lower itsprotectionlevel)aschangeoccursintheenvironment.

NotethatinDDMRPthereisoneexceptiontothistest:replenishedoverrideparts.Thesepartsorpositions,due tospecific imposedrestrictions,havestaticlevels.

DDMRPVersusSafetyStockandOrderPoint

Now that we have the criteria by which to judge a DDMRP strategicreplenishmentbuffer,somecomparisonscanbemadeagainstmoreconventionalforms of stock buffering and order generation. Demand Driven MRP hasfrequentlybeencomparedwithconventionalsafetystocksystemsororderpointsystems. How do these types of conventional mechanisms stack up to theDDMRPbuffercriteria?

SafetyStockandtheBufferCriteria

TheAPICSDictionarydefinessafetystockas:

1)Ingeneral,aquantityofstockplannedtobeininventorytoprotectagainstfluctuations in demand or supply. 2) In the context of master productionscheduling, the additional inventory and capacity planned as protectionagainstforecasterrorsandshort-termchangesinthebacklog.Over-planningcanbeusedtocreatesafetystock.(p.154)

While thereare twodefinitions, the firstdefinition is incrediblybroadandlacksanyrealsubstance.Theauthorsacknowledgethatlimitingthedefinitionofsafety stock to the first definition would mean that DDMRP buffers would

Page 287: DDMRP - Demand Driven Material Requirements Planning

qualify as safety stock positions with clearly defined rules about placement,sizing,andmanagement.

Theseconddefinition,however,definestheconventionalapproachtosafetystock and characterizes what most people refer to when making comparisonsbetweenDDMRP and safety stock.Whilemany safety stock approaches havesubtle (and often proprietary) differences, they are all built around the samebasicprinciples;definingastockleveltocoveragainstvariability(demandandsupply)overadefinedlengthoftime—typicallyaplanninghorizon.Thelongerthe planning horizon, the greater the rate of forecast error and the higher thesafety stock. Safety stock supplements supply orders that are generated by ademand allocation explosion using the projected available balance; these aretypicallyforecastedorplannedordersthatareprimarilydrivenfromaforecast.Thus the conventional safety stock approach is a planned supplementaryinventoryposition toguardagainstvariation. If forecast error ishigh, then thesafetystocksupplementarypositioncanbecomequiteanextraordinaryfinancialcommitmenttostatisticallycoverthaterror.

Howdoes theconventionalapproach tosafetystock relate to theDDMRPstrategicbuffercriteria?

Thedecouplingtest—FAIL.Safetystockfailsthistest.Theplanninghorizonsandleadtimeequationsarenotdefinedorimpactedbytheplacementofsafetystockpositions.

The bidirectional benefit test—FAIL. Safety stock fails this test.Safety stock is designed to protect in only one direction. Safetystock is designed toprotect supply continuity. It is focusedpurelyon protecting the consumption side of the position. But just howmuchprotectiondoesitreallyprovide?

Itcaneasilybearguedthatsafetystockactuallyexacerbatesoneelementofthebullwhipeffect—demandsignaldistortion.Thesafetystocklevelbecomesthe“newzero”thatMRPsystemsattempttonetagainst.Thismeans thatwhen theavailablebalance isprojected tobe below the safety stock level, an order is immediately generatedandcodedasanexpedite.Thusanypenetrationintothesafetylevelcreates additional orders with a high degree of urgency. Thisattribute can create massive amounts of confusion, distortion, andnoise for supplying resources to the safety stock position,particularlyasthepicturechangesfromMRPruntoMRPrun.Thus

Page 288: DDMRP - Demand Driven Material Requirements Planning

the use of safety stock can be an amplifier to demand signaldistortion.

This is contrary to DDMRP thinking, where current andprojected on-hand penetration into the safety level (red zone) isexpectedtoberoutine.InDDMRP,asthepenetrationbecomesmoresevere,thefocusisonexpeditingexistingsupplyorders,notcreatingadditionalsupplyorders.

Additionally, conventional safety stock mechanisms are mostcommonly implementedonlyat thepurchasedandend itemlevels.Thislimitstheirdampeningeffectonsupplycontinuityvariabilitytoonly those levels, opening the door for nervousness and delayaccumulation to be experienced to a larger degree between thoselevels.

Theorder independence test—PASS.Safety stockpasses this test.Safetystockinventoryisnotpredisposedtoanyparticularorder.Itisavailabletoanysourceofconsumption.

Theprimaryplanningmechanismtest—FAIL.Safetystockfailsthistest.While theconventionalsafetystockmechanismdoesgeneratesupply orders, those orders are only supplementary. The vastmajorityofplanningactivity(supplyordergeneration)forpositionsthataresafetystockedareforecastedorplannedordersandhavenorelationshiptothesafetystockbufferitself.

Therelativeprioritytest—FAIL.Safetystockfailsthistest.Whiletheconventional safety stock mechanism will inform that you havepositions under safety stock level, there is no ability to prioritizeagainst them. Every penetration is deemed urgent and requiresimmediate attention. When everything is a priority, there is nopriority. It is impossible without significant additional effort andanalysis to determine just how urgent each situation is and howmuchmoreurgentonesituationisoveranother.

The dynamic adjustment test—PASS/FAIL. In the authors’experiences most safety stock mechanisms are user defined andstatic.Thereareglaringexceptions,particularly in thefast-movingconsumergoodssegmentwhere thesafetystockequationsare tiedtochangesinarollingforecast,forecasterrorrate,andthedaysofsupplyintendedforsafetystock.

Page 289: DDMRP - Demand Driven Material Requirements Planning

ConventionalsafetystockmechanismssimplydonotscorewellagainsttheDDMRP buffer compliance criteria, the criteria that have been formulated tobetterprotectandpromoteflowinthemorecomplexandvolatilecircumstancesofthetwenty-firstcentury.

OrderPointandtheBufferCriteria

Let’s now turn our attention to the comparison of DDMRP buffers to orderpoint.Similar towhatwasdonewithsafetystock,orderpoint systemswillbecomparedwiththeDDMRPbuffercompliancecriteria.

TheAPICSDictionarydefinesorderpointas:

Asetinventorylevelwhere,ifthetotalstockonhandplusonorderfallstoorbelow that point, action is taken to replenish the stock. The order point isnormally calculatedas forecastedusageduring the replenishment lead timeplussafetystock.(p.117)

AtfacevaluethisdefinitionseemsmuchmorecompatiblewiththeDDMRPruleset:

The decoupling test—PASS. Order point passes this test. Inconventional order point, the planning horizons and lead timeequations are defined by the placement of order point positions.MRP stops the explosion at order point positions because thosepositionsaredesignedtohaveon-handinventoryavailable.ThisisconsistentwithaDDMRPapproach.

Thebidirectionalbenefit test—FAIL.Orderpoint fails this test.Asper the APICS definition, order point incorporates a conventionalsafety stockmechanism.That incorporationmakesorderpoint failthis test for the reasons previously described (exacerbating thedemandsignaldistortionelementofthebullwhipeffect).

Theorder independence test—PASS.Order point passes this test.Theinventoryattheorderpointpositionisavailabletoallpotentialconsumers.ThisisconsistentwithaDDMRPapproach.

The primary planningmechanism test—PASS.Order point passesthistest.Allsupplyordersaregeneratedattheorderpointpositionincludingtheproblematicsafetystockorders.

Therelativeprioritytest—FAIL.Orderpointfailsthistest.Generated

Page 290: DDMRP - Demand Driven Material Requirements Planning

supplyordersaresimplyquantitiesanddaterequirementswithoutaway tocompareandcontrast them.Additionally, the incorporationofasafetystockmechanismfurthercomplicatestheprioritypicture.

Thedynamicadjustmenttest—FAIL.Orderpointfailsthistest.Mostorderpointsystemsareuserdefinedandstatic.

According to theDDMRP buffer compliance criteria, order point systemsappeartobecloserinnaturetoDDMRPsystems—adistantcousinsotospeak.Yetweshouldbecarefulininferringthatthecloserrelationshipimpliessimilarresults.

Evenwhen order point passes a criterion, there are significant differencesworthexploring.Thesedifferenceswillcreateamajordisparityinperformancebetween the two. This is evident when exploring the primary planningmechanismtest.Whileorderpointpassestheplanningmechanismtest,therearemajordifferencesinthenatureoftheorderingequation(aboveandbeyondtheinclusion of the safety stock trigger) that should be understood. Order pointsystemsuseadifferentequationtodeterminewhetherabufferrequiresresupply.

Inorderpointtheequationislimitedtoon-handpluson-order.Thereisnodemandelementintheequation.Nopastduesalesorders.Nosalesordersduetoday.Nospikequalification.Thismeansthatdemandinorderpointsystemsisconsumption-based and purely historical. It is only recognized after it hasoccurred.Thishappensintheformofanon-handadjustmentinthenextday’sorderpointequation.

Yet the demand elements in the net flow equation ofDDMRP are highlyrelevant pieces of information. When visibility to relevant information isobscured,weknowthatthereisadirectrelationshiptotheamountofvariabilityexperienced,and that is thecasewithorderpoint systems.Theyaresubject tomore demand variability because they disregard highly relevant and accuratesales order information. The susceptibility to variability comes at the price ofhighertotalinventoryrequirementstocovertoprotectthebufferposition.

DDMRPImpactsonScheduling

With these six DDMRP buffer characteristics in mind, we will now turn ourattention to howDDMRPcan affect schedules at both thehighermaster levelandresourceschedulingarea.

DDMRPandMasterProductionSchedulingAssumptions

Page 291: DDMRP - Demand Driven Material Requirements Planning

Within the conventional planning schema, theoutput of themaster productionschedule(MPS)andMaterialRequirementsPlanningisimportanttounderstand.

TheAPICSDictionarydefinesmasterproductionscheduleas:

A set of planning numbers that drives material requirements planning. Itrepresents what the company plans to produce expressed in specificconfigurations,quantitiesanddates.(p.101)

Figure 11-1 displays the conventional relationship between the MPS andMRP as described inChapter3. The output of all this activity relies on threecrucialassumptionsforsuccess.Theseassumptionsmustbeexaminedatdepthinlightofwhatwenowknowaboutconventionalplanninglimitations.

Assumption1.Demandsignalsareknownandaccurate. Assumption 2. Lead times for supply order release, receipt, andsynchronizationarerealistic.

Assumption3.Materialandcapacityareavailableonthespecifieddates.

Essentiallyanyscheduleatanylevel inoperationsmakesthesethreebasicassumptions.Whyreleaseascheduleif thesebasicassumptionsarenotbehindit?Therelativevalidityofthesethreebasicassumptions,however,combinestodeterminejusthowrealisticascheduleis.Whenthescheduleismorerealistic,itismore likely tobemaintained.Conversely, the less realistic theschedule, themore likely that schedule will be disrupted. Schedule disruptions lead toperformanceerosionandcostlycompensationatalllevels.

Page 292: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-1TheconventionalMPS-MRPschema

How valid are these assumptions in the conventional MPS and MRPapproach? The answer to this question will determine just how realistic theoutputoftheconventionalapproachisonaroutinebasis.

Assumption 1. Demand signals are known and accurate. Tyingorder generation and scheduling directly to forecast means thatactual demand will vary from the forecasted demand used togenerate the schedule.The longer the planning horizon, the largerthe variance between forecasted and actual. As these variancesoccur, the demand signals change with every MRP run, creatingmassivenumbersofadjustmentsthroughouttheenvironment.

Assumption #2. Lead times for supply order release, receipt, andsynchronizationarerealistic.Withnodecoupling,delaysfrequentlyaccumulate,dramaticallyaffectingwhenorderscanbereleasedwithfull allocation. Safety stock is rarely placed at the intermediatelevelsinordertoprovideevenpartialsupplyvariabilitydampening.Positions are netted to zero, leaving no margin for error. Witheverythingcoupledtogetherandnomarginforerror,thescheduleismuchmorecomplexandfragile.

Furthermore,withnoexecutioncapabilitybuiltintoMRP,thereis no ability to see how those potential delays will affect theenvironmentinthenearterm.Thereislittletonovisibilityaboutaproblem until the problem has already been encountered, andwithoutdecouplingthereareguaranteedtobemoreproblems.Underthese circumstances synchronization (and flow) quickly breaksdown.

Assumption3.Materialandcapacityareavailableonthespecifieddates. When synchronization breaks down, material and capacityarefrequentlynotavailableasplanned.Materialarriveslateorgetsdiverted to cover shortages elsewhere. Capacity is frequently notavailableduetoscheduleslidesanddeviations.

Simplystated,theoutputoftheconventionalMPSandMRPprocessisjustnotvery realistic to startwith, and itquicklygetsworse inexecution, runningrampant through operations. As stated in previous chapters this conventional

Page 293: DDMRP - Demand Driven Material Requirements Planning

approachwill work only if everything goes exactly according to plan. Is thisnewstoplanners?Ofcoursenot;thatiswhycompromisesandworkaroundsarepresentinalmosteverycompanyusingthisprocess.Isthisnewstooperationalpersonnel?Ofcoursenot;thatiswhytheyblameplanningforsomanyoftheirproblems.

How does the DDMRP approach stand up to the same set of basicassumptions? The assumptions don’t change; the operational approach does.Willthatoperationalapproachproducearealisticoutput?

Assumption1.Demandsignalsareknownandaccurate.Theuseofqualified sales orders means that demand signals are much morerelevant,accurate,andtimely.

Assumption 2. Lead times for supply order release, receipt, andsynchronizationare realistic.Theuseofdecouplingpoints createsshorter independently planned andmanaged horizons. Thatmeansthat less variability (demand and supply) is passed through thesystem. This results in synchronization dates that are at the sametimemorerealisticthroughoutthesystemyetlessimportantduetothecushionat thebuffers.Additionally, theuseofdecoupled leadtimeforbuffersizingmeansthatthebuffersarerealisticallysizedatthedecouplingpoints toabsorbdemandandsupplyvariabilityandmaintaintheintegrityofthedecouplingeffect.

Assumption3.Materialandcapacityareavailableonthespecifieddates.DDMRPplansstrategicpositionstoalwaysbeavailable.Atdecoupling points, DDMRP is designed to never net to zero.DDMRP screams with multiple alerts at planning and operationspersonnelifthosepositionsareanywherenearzerofornetflowandon-hand. The DDMRP buffers represent points of stored capacityandmaterials.Ifthesepositionsaremaintainedwithon-handalwaysavailable, then the stored capacity andmaterials at thosepositionsare always available. This means that under DDMRP, thisassumption isvalidmostof the timefor thestrategicpoints in themodel.

Additionally,theuseofDDMRP’sexecutioncomponentsbringsdegreesofvisibilitytoopensupplyordersthatmustbeexpeditedtomaintain stockbuffer integrityand tomeetcritical synchronizationneeds.

Page 294: DDMRP - Demand Driven Material Requirements Planning

DDMRPmovesanenvironment from theconventional statementof“whatwecanandwillbuild”toastatementof“thecapabilitytobuildwhatwecanandwill sell.” Indeed, the whole notion of a conventional master productionscheduletiedtoasupplyordergenerationcalculator(MRP)evaporatesunderaDemand Driven Operating Model. In its place we find a “master settings”componentofDemandDrivenSalesandOperationsPlanning(DDS&OP)tiedtothe supply order generation calculator of DDMRP. Figure 11-2 depicts themastersettingsfeedtoDDMRPfromDDS&OP.

There are three primary inputs provided by the master settings: bufferprofiles, partdemanddata, andpartprofile assignment.Bufferprofiles are thegroupings and settings for replenished parts (part type, variability, and leadtime). Part demand data have two elements: planned adjustment factors (ifapplicable)andaveragedailyusage.Plannedadjustmentfactorsarefactorstobeapplied to the ADU of parts or groups of parts. Average daily usage is theaverage rate of use for each replenished part (past, forward, or blended). Partprofile assignment is the assignment of each replenished part to a particularbufferprofile.

Theseinputsarecombinedwiththeinventoryrecordfile(on-handandopensupplyrecords),theproductstructurefile(billofmaterial),andexternalordersfor components (sales orders) to perform the net flow equation and generateexecutionalertreportsforeachpart.Supplyordersareconveyedtoscheduling(manufacturing orders) and execution (purchase orders and stock transferorders).Alertsareconveyedtoappropriatepersonnel.

Simplystated, theoutputof theDDMRPapproachproducesamuchmorerealisticstarttotheapplicationofanorganization’sresourcestothefulfillmentof market demand. Supply orders are generated based on a model built topromote and protect flow to fulfill known and accurate demand signals. ButDDMRP’s impactonoperationsdoesnotend therewithvalid, prioritized, andsynchronizedsupplyorders.Inmanyorganizationsitwilldramaticallyinfluenceandimprovetheabilitytoeffectivelyscheduleresourcesandmanageworkflowforthebesttotalprotectionandpromotionofflow.

Page 295: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-2DDMRPmastersettingsfeed

DDMRPShopFloorSchedulingImplications

With regard tomanufacturing orders, the impact ofDDMRPon scheduling isnot limited to the simple generation of a supply order signal. Conventionally,manufacturing order requirements are generated, and it is up to scheduling towork it out. This presents a huge challengewhen resources andmaterials aresharedand there isno relative senseofpriority—yousimplyhave todo it all,andMRPassumesthatyoucan,soyoudothebestyoucan.UndertheDDMRPapproach,however,schedulingtheapplicationofmaterialsandcapacitycanbemuchclearerandevensimplerduetotwoparticularimpacts.

Impact1.DecoupledSchedulesSofartheDDMRPdecouplingeffecthasbeenexploredattheproductstructurelevelinordertogenerateandmanagesupplyorders.Supplyordersexistat theproduct structure (discrete part number) level. Yet with regard to schedulingmanufactured items, there is an additional level that must be understood: therouting level. Routing was defined in Chapter 6 (“Strategic InventoryPositioning”).

Simplyput, theroutingdetailsthewayapartnumberatthelowerlevelofthe product structure is converted to a part number at a higher level in the

Page 296: DDMRP - Demand Driven Material Requirements Planning

productstructure.Figure11-3depictsthedifferencebetweenaproductstructureviewandaproductstructureviewwithroutingsequencesbetweenpartnumbers.Withroutingsidentified,wecanseehowapartisconvertedintoahigher-levelpart.Thesquareboxesarethepartnumbers,andtheroundedshadedboxesareidentifiedactivitiesandtheirrelativesequence.

In this instance, converting a purchased item (PPA) into an intermediatecomponent (ICA) requires three activities in a particular sequence: cut, form,andgrind.Mostroutingswouldincludetimeperpieceaswellasotherpiecesofrelevant information for the resources performing the action or scheduling theactivity. In thiscase, that information is immaterial to illustrating theDDMRPimpactonscheduling.Theconversionoftheintermediatecomponent(ICA)intothe finished item (FPZ) requires three additional sequenced steps: prep, paint,andlabel.

Whendecouplingpointsareplacedattheproductstructurelevel,theyalsoserve to decouple at the routing level. The routing level is the level atwhichresource scheduling is created. Having fewer dependencies to scheduledramaticallysimplifiestheresourceschedulingprocessbecausetherearefewerthingsthathavetofittogether.

In Figure 11-4, scheduling two groups of three processes independentlyshould be easier than scheduling six processes together.There are simply lessdependencies.Asimpleanalogy:Isiteasiertoscheduletwomeetingswiththreebusy people or one meeting with six busy people? And that is anoversimplification,becauseinresourceschedulingweneedtofindtimeslotsina particular sequence. The chances that open time slots can be found toaccommodate three resources in the right sequence are routinely better thanroutinelyfindingsixopentimeslotsintherightsequence.

Page 297: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-3RoutinginformationinFPZproductstructure

Page 298: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-4Coupledversusdecoupledscheduling

Figure11-5depictsamuchlargernetworkofdependencies,onewithmanydifferent integration points and a significant amount of concurrent activity—ascheduling puzzle across shared resources. The rounded shaded boxes areresourceassignments.Therectangularboxeswithnoshadingarepartnumbersin the bill ofmaterial. Shared resources can be identified using the numericalassignment.Forexample,thesameresource(031)isrequiredontheroutingforcomponentS12andcomponentS21.

FIGURE11-5Aschedulingpuzzle

Page 299: DDMRP - Demand Driven Material Requirements Planning

Figure11-6depictsthesameenvironmentbutwithadecoupledschedulingapproach.Decoupling points have been placed at part numbers in the productstructure.Onelargeschedulingpuzzlehasbeenreplacedwithfoursmallerandeasier-to-schedule portions defined by the decoupling point placements atcomponentsS21,ICD,andICS.

Perhapsabetterdepictionofthedecoupledscheduleapproachistolaythemside by side, as they are independent of each other due to decoupling pointplacement.Figure11-7depictsthisview.

FIGURE11-6Adecoupledschedulingapproach

FIGURE11-7Decoupledschedulinglegs

Page 300: DDMRP - Demand Driven Material Requirements Planning

Yes, there are shared resources between these scheduling legs, but thedecoupling point buffers will help build a deconflicted schedule across thoseresourceschedulesbasedonnetflowpriority.Thiswillbediscussedinthenextsection(“VisibleandPriority-BasedSchedulingSequencing”).

Thestabilityofthescheduleisalsoimpactedbydecoupling.Theonethingthat is constant is change. Run rates can vary, quality issues can occur, andresources can go down from time to time.These events impact schedules.Asdiscussedpreviously,anychangeinahighlydependentsystemcreatesaripplethat affects everything that is dependent. This was previously referred to asnervousnessandwasillustratedattheMRPlevel.Changesindateorquantityinonesupplyorderrippledouttoaffectdependentsupplyorders(changesindateandquantityorcancellation),evencreatingadditionalsupplyorders.

Nervousness also exists at the routing level. This is called schedulingnervousness.Changesinscheduledactivityordelaysatcertainoperationsrippleout to affect the schedule.Advanced planning and scheduling systems perfectthisdamagingeffect.Small changes in schedulesare transmitted to the restofthe system, making the schedule a constantly changing work in process. Themore frequent the changes, the more chaotic the schedule. If the schedule isconstantly changing, then there is no schedule. The more dependenciesscheduledtogether,themorechangetherewillbe.

In a decoupled schedule, the changes that occur in one scheduling leg areisolated from theother legs.Theexception iswhen there are shared resourcesamongthelegs.Underthiscircumstance,DDMRPwillhelpmitigatetheimpactofthatvariabilitybyprovidingclearvisibilitytowhichschedulehaspriorityforthat particular resource. This will be discussed later in this chapter in the“DDMRPandWIPPriorityManagement”section.

Impact2.VisibleandPriority-BasedSchedulingSequencingAsdiscussedinChapter9,DDMRP’snetflowequationdeterminessupplyordergeneration.Assupplyordersformanufactureditemsarebeinggenerated,thenetflowpositionofeachitemcanbeusedtoestablishschedulingsequenceacrossasharedresourcebase.

Figure 11-8 represents a manufacturing plant with shared resources. Thisplant makes nine different finished items (FPA, FPB, FPC, FPD, FPE, FPF,FPG, FPH, FPI). All finished items go through the same assembly area. Thisplantalsohas18differentintermediatecomponents.Ninearefabricated(SAA,SAB, SAC, SAD, SAE, SAF, SAG, SAH, SAI), and the others aremachined

Page 301: DDMRP - Demand Driven Material Requirements Planning

(ICA, ICB, ICC, ICD, ICE, ICF, ICG, ICH, ICI).All finished itemsconsistofmachinedandfabricatedparts.

Figure 11-9 is the DDMRP planning screen for the finished items in theplant. Supply orders for three items are being recommended (FPB, FPE, andFPA).Theplanningprioritycreatestherelativeprioritybetweentheordersandestablishes the assembly scheduling sequence. FPB will be scheduled inassemblyfirst.

The same process is applied to the establishment of the machining andfabrication schedules. Figure 11-10 is the DDMRP planning screen forintermediate items. Two machined and two fabricated items are beingrecommended for replenishment. The planning priority will establish thescheduling sequence for both areas. ICBwill be scheduled inmachining first,andSAFwillbefirstonthescheduleforfabrication.

FIGURE11-8Asampleplantwithsharedresources

FIGURE11-9Netflowpriorityforfinisheditems

Page 302: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-10Netflowpriorityforintermediateitems

FiniteSchedulingwithDDMRP?

Anymanufactureditemstockbuffersshouldnotbethoughtofsimplyasunitsofstock.Theonlywaythatyoucanhavestockedpositionsofmanufactureditemsis tohave investedboth capacity andmaterials to create the stock.Thus thesebuffers should be thought of as storage tanks of capacity and materials.Furthermore,ifthesebuffersareplacedatstrategicpoints,thentheyarestrategicstorage tanks of capacity andmaterials.Additionally, these strategic tanks aredesignedtobemaintainedatcertainprescribedlevelsbasedonDDMRPbuffersizingcriteria.

WhatwehavelearnedaboutDDMRPbuffersuptothispointcaneasilybetranslatedintocapacityterms.Figure11-11relatesthebufferofamanufactureditemtothecapacityofaparticularresource.

Inthiscasethebufferofpart123isbeingrelatedtoassemblycapacity.Part123takes30minutesperpieceonaverageinassembly.Thatmeansthatatopofgreenof455piecesrepresentsatotalof13,650minutesofassemblytime.Theaverageon-handof155translatesto4,650minutesofassemblytime.Thisistheaverage amount of stored assembly capacity that is expected in the buffer. Aminimumorder size of 120 pieces corresponds to the green zone.Thismeansthateachorderforresupplywillrepresentatleast3,600minutesofcapacity.

This consideration brings up an interesting extension to or perhaps aredefinitionoftheterm“finitescheduling.”TheAPICSDictionarydefinesfiniteschedulingas:

Assigning no more work to a work center than the work center can beexpectedtoexecuteinagiventimeperiod.(p.63)

Page 303: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-11Translatingbufferquantitiestocapacityterms

This definition is focused solely on the scheduling of a resource.But canthis definition be expanded to include thewayDDMRP recommends supply?Thecriticalelementsofthedefinitiondistilldowntoschedulingcapacitytothedefinedlevelandnotbeyond.DDMRPmanufactureditembuffersarestrategicstoragetanksofcapacitythatarereplenisheduptoalevel(topofgreen)andnotbeyond.

TheDDMRP buffers are strategically placed; they are resupplied up to acertainpointandnotbeyond.Supplyorders for their replenishmentessentiallydirectactivityinthesystem.Thisisinterestingandsimilartoaspecifictypeoffinite scheduling known as “drum scheduling.” Drum scheduling was mostfamouslyintroducedtotheworldin1984intherevolutionarybook,TheGoal,by Eliyahu Goldratt and Jeff Cox. The APICS Dictionary defines drumschedulingas:

The detailed production schedule for a resource that sets the pace for theentiresystem.Thedrumschedulemustreconcilecustomerrequirementswiththesystem’sconstraint(s).(p.52)

Yet in DDMRP, less actual resource scheduling precision is requiredbecause the buffers allow for time and flexibilitywithin short ranges of time.They naturally allow for a deconflicted scheduling sequence tied to strategicbuffer status (planning and execution priority). Thus DDMRP manufactureditembufferscanbedefinedas“drums.”Theyarestrategicallyplaced.Theysetthepaceofthesystem.Theyaresupplieduptoadefinedpointandnotbeyond.For environments that have a fair amount of manufactured item buffers,

Page 304: DDMRP - Demand Driven Material Requirements Planning

DDMRP can be characterized as a simple, even elegant, finite schedulingalternative.

AdditionalSchedulingSequenceImpacts

Even when there are major scheduling complications, DDMRP can provideadditional insight into the sequence.For example, in foodprocessing, amajorcomplicationissomethingcalledallergensequence.

An allergen sequence defines the order in which food processors mustprocess foodswithcertainknownallergens.Forexample,all items thatdonotcontaineggsshouldberuninadvancedofallproductsthatdocontaineggs.Inthis case a sequence number can be provided to each item. For example,productswithouteggsreceivea1,whileproductswitheggsreceivea2.

ThissequenceassignmentcanthenbecombinedwiththeDDMRPplanningprioritytocreateasequencedschedulethatisbothsafetoprocessandaligns(asbestaspossible) toactualbuffer requirements.Figure11-12showsaDDMRPplanningqueuethatissortedbyallergensequence.Allitemsrecommendedforresupplyarenowassignedasequencebasedontheirallergenassignment.

Despitehavingthehighestplanningpriority,“NutSurprise”isthelastonthesortedallergensequencelist.Runningoutofsequencewouldrequiresignificantdelaysforrequiredcleaningactivities.Thiscoulderodecapacityandjeopardizeadditionalstockedpositions.Incircumstancessuchas this, itmightrequire theuseofhighervariabilityprofilesinordertocompensateforpotentialsequencingconflicts.

FIGURE11-12DDMRPresupplyorderssortedbyallergensequence

Of course, there will be environments and circumstances that dictateadditionalandmoredetailedschedulingcapability.Theseinclude:

1. When more detailed scheduling will result in better lead time andvariability control feeding into a stock position. Thiswill allow fortheminimizationofinventoryatthestockposition.

Page 305: DDMRP - Demand Driven Material Requirements Planning

2.Whenseriesofresourcesareneededtocreateafinalproductandthereis no ability to decouple before the finished level in the productstructure. In this case,more detailed scheduling effortsmight betterprotect customer commitments. This is particularly necessary whenthe resources are shared across many items. Config-ure-to-orderenvironmentsoftenqualifyunderthisreason.

3.Whentherearesharedresourcesinvolvedintheproductionofmake-to-stock and make-to-order items (mixed-mode operations). Moredetailedschedulingwilloftenbeneededtobetterdeconflictcapacityrequirementsbetweenmake-to-stockandmake-to-orderitems.

Whenthesecircumstancesarepresent,aDemandDrivenOperatingModelusescontrolpointscheduling.TheAPICSDictionarydefinescontrolpointsas:

Strategiclocationsinthelogicalproductstructureforaproductorfamilythatsimplifytheplanning,scheduling,andcontrolfunctions.Controlpointsincludegating operations, convergent points, divergent points, constraints, andshipping points. Detailed scheduling instructions are planned, implemented,andmonitoredattheselocations.(p.33)

Control points don’t decouple. They are places to transfer and amplifycontrolwithin agivenarea so the transferenceandamplificationofvariabilitywithinthatareacanbeminimized.Typicallytheyareplacedbetweendecouplingpoints or between decoupling points and customers. These environments andcircumstances in addition to control point determination and scheduling arethoroughly described in the bookDemandDriven Performance: Using SmartMetricsbyDebraSmithandChadSmith.

DDMRPandWIPPriorityManagement

After a schedule has been determined, DDMRP can further influencemanufacturing resources with regard to released work sequencing. By givingresources or thosemanaging the resources the visibility to the on-hand bufferstatus corresponding to the orders on their respective dispatch lists, they caneasily and quickly determine if the sequence should be altered based on thechanges occurring in the environment. Figure 11-13 is a scheduled sequencebasedonplanningpriorityfromthepreviousexampleinFigure11-9.

Page 306: DDMRP - Demand Driven Material Requirements Planning

FIGURE11-13Originalscheduledsequence

Should this schedule simply be frozen? Should manufacturing resourcesalwaysworkthesequenceintheschedule?Yes,iftherearenorelevantchangesbetween the time inwhich the sequencewas created and the activity is to beperformed.Iftherearerelevantchangesinthattimerange,thenvisibilitymustbepromotedinordertoconveyanychangesinpriorityandsequence.Figure11-14isanexampleoftheminimumamountofrelevantinformationthatresourcesorthosemanagingtheresourceswouldneedtoseetoshiftthesequenceofthereleased manufacturing orders. In this case, the released orders have beensequencednotintheordertheywerescheduledbutintheorderthatcorrespondstotheirbufferstatus.

FIGURE11-14Releasedmanufacturingorderssequencedbyon-handbufferstatus

FPA’son-handsituation isbecomingcritical.Asituationsuchas thismaydictate a change in the sequence in which these manufacturing orders areworked. At a minimum it should prompt discussions between appropriatepersonnel.

Summary

This chapter illustrated the impact DDMRP can have on the operationalresources that interact with it. It is the explanation and depiction of howDDMRPfitsintoaDemandDrivenOperatingModel—essentiallyprovidingtheheartbeatoftheactivityinthemodel.Tosummarize,Figure11-15isagraphical

Page 307: DDMRP - Demand Driven Material Requirements Planning

comparisonofconventionalMRP,Lean,andDDMRP.Thegraphic shows the independentplanninghorizonsbetweendecoupling

points. These independent horizons provide demand and supply variabilitymitigationwhileallowingforsynchronizationwithin thehorizon.Supplyordergeneration is illustrated by the dashed arrows pointing to the next lower leveldecouplingpoints.Theelementsofvisibleandcollaborativeexecutionarealsoillustrated, with the on-hand buffer reports pointing to each leg in the bill ofmaterial. This is meant to show that while executing against the initialreplenishment signal the statuses of the next buffered positions are visible toallowforcoursecorrectionifnecessary.

FIGURE11-15MRP,Lean,andDDMRP

Page 308: DDMRP - Demand Driven Material Requirements Planning

CHAPTER12

DDMRPMetricsandAnalytics

InthefirstsectionofChapter7,thetwocriticalprerequisitesareestablishedforinventory toactasanasset for the flowof relevant informationandmaterials:placement and sizing.Placement determines thepoints of independence in theplanning, scheduling, and execution process, and the sizing of the inventorypositionallowsthatindependencetobemaintained.

Withregardtoplacement,Chapter6discusseskeyfactorstoconsiderwhenapproaching the placement of decoupling points. This is a strategic design orblueprint that is not to be taken lightly. Once that design is completed andimplemented,howcanvisibilitybecreatedtomakesurethatthebufferlevelsatthose decoupling points are protecting and promoting the flow of relevantinformationandmaterialsforthebestreturnoninvestmentinourcomplexanddynamicenvironments?Aretherightsignalsbeingconveyedwithoutdistortionin a timely fashion? Are the right materials available when needed? Is theinventory in excess? Is the blueprint performing as designed? How can it bebetter?ThesearethekeyquestionsforthemetricsandanalyticsassociatedwithDDMRP.IfDDMRPdesignsarefundamentallydifferentfromconvention,thenithintsthatatleastsomenewmetricsandanalyticsmayberequiredtoanswerthesequestions.

MeasuringRelevantInformation(SignalIntegrity)

DDMRP planning is the process of generating supply order requirementsprimarily driven by the decoupling point net flow position. There is anassumptionthatasystemcannothavetherelevantmaterialswithoutfirsthavingthe relevant information (supply order signal) to act upon. Relevance isdeterminedbytimingandaccuracy.Ifthereisadelayinconveyingthesupplyordersignal,thentheDDMRPmodelisnotoperatingwithrelevantinformation

Page 309: DDMRP - Demand Driven Material Requirements Planning

and is less responsive. If thesupplyorder signal is inaccurate (toohighor toolow),theDDMRPmodelisnotoperatingwithrelevantinformationandiseitherlessresponsiveorwasteful.ThussupplyordersshouldbeconveyedassoonaspossibleandinproperquantityaspertheDDMRPmodel.

Figure12-1isanexampleofareportusedtomonitortheintegrityofsupplyorders for timingandaccuracy.Adate range isprovidedon the left-handsidecoveringa15-dayperiod(October1–15).Inthisexampleacompanymakesfourtypes of soup,which are all buffered:ChickenTruffle,ChickenNoodle,BeefStew,andMinestrone.Withineachproductcategorytwocolumnsareprovided:“DDMRPRecommendation”and“Actual.”

FIGURE12-1Trackingnetflowsignalintegrity

TheDDMRP recommendation is the quantity recommended to restore netflowtothetopofgreenonthatspecificdate.Aquantitywillonlyappearifthenet flow position is in the yellow or red zone. “Actual” is the quantity of theapproved supply order on that specific date line. For example, on October 3,Chicken Truffle Soup is calling for a supply order of 1,200 in the “DDMRPRecommendation” column. Its net flow position is in the yellow zone of thebufferandrequires1,200toberestoredtothetopofgreen.Asupplyorderwasapprovedonthesamedayforaquantityof1,800.

Three of these soups have problems with signal integrity. As describedabove,asupplyorder forChickenTrufflewasapprovedonOctober3.Supply

Page 310: DDMRP - Demand Driven Material Requirements Planning

order generation was late by one day and well above (1,800) the requiredquantity(1,321).Aninitialrequirementof1,200wasknownonOctober2;yetthere was no corresponding supply order. The next day the net flow positionerodedfurther,drivingalargerrecommendedamount.Anorderwasapproved,but that order far exceeded the recommended amount. It drives the net flowpositionoverthetopofgreen(OTOG)forthefollowingtwodays.Inthiscasetherearebothtimingandaccuracyissues.

Chicken Noodle exhibits problematic signal integrity from a timingperspective.OnOctober 6 therewas a recommended supply order quantity of2,400 but no approved supply order. This situation persists until the net flowpositionisdrivenallthewayintored,requiringaresupplyof6,200.Finally,onthatdayasupplyorderisapprovedforthequantityrecommended.Inthiscasethereareonlytimingissues.

Finally,BeefStewappears tobechronicallyoversupplied.OnOctober3asupplyorderwasapproved for1,200despite thenet flowpositionbeing inanOTOG position. This drives the net flow position even further aboveOTOG.ThenagainonOctober13anothersupplyorderisapprovedfor1,200despitenoorderrecommendationandanetflowpositioninthegreen.

Couldtherebegoodexplanationsbehindtheplanner’sbehaviorforeachofthethreepartswithsignalintegrity?Absolutely,butquestionsneedtobeaskedandreasonsneedtobedocumented.

Minestroneappearstobetheonlyitemthatisoperatinginaccordancewiththemodelandmethod.October3producesarecommendationfor1,900,andonthesamedayanorderisapprovedfor1,900.ThenetflowpositionshowsgreenthenextdayandcontinuesuntilOctober10,whenarecommendationof2,150isapprovedwithanewsupplyorderforthesameamount.

MeasuringDecouplingPointIntegrity

ThefirstsectionofChapter7alsoprovideduswithakeypieceofinformationthatwillbecomeanimportantfoundationforthebulkofthesenewmetricsandanalytics. Figure 12-2, repeated here from Chapter 7, shows the Taguchiinventory loss function. Inventory will be an asset to flow at the decouplingpointswhenitismaintainedbetweentoolittle(pointA)andtoomuch(pointB).Between these two points is an optimal range for inventory to bemaintained.Withinthisrangeitcanabsorbvariabilitybutisnotexcessiveinventory.

Thisbookdescribesawaytocalculatethisrangeaswellasthebehaviorthatshouldoccurovertimetomaintaintherangeandeventheelementsrequiredto

Page 311: DDMRP - Demand Driven Material Requirements Planning

make improvements.Chapter7 provides the sizing logic forDDMRP buffers.Chapter8 provides the logic for adjustments.Chapter9 provides the planninglogic for the buffers. Chapter 10 presents the rationale behindmonitoring thecurrentandprojectedabilitytomaintaintheintegrityofthedecouplingpoint.

Chapter 3 describes a typical effect in conventional planning called thebimodal effect. The bimodal distribution is primarily driven by two inherenttraits ofMRP: its hard-coded trait to net to zero at each part position and itsrequirement tomakeeverythingdependent in itsplanningequation.Netting tozeromeansthereisnoplannedbufferandeverythingisperfectinquantityandtiming.Treating everything as dependentmeans that change anywhere createschange everywhere. These traits create a constant oscillationwith regard to apart’spositionbetweentoolittleandtoomuch.

To know how problematic any inventory bimodal distribution really isrequires the optimal range for each part to be defined.Without this definitionanybimodalanalysis is limited towhetherapositionwasstockedoutor therewasasignificantamountofexcess.ThebuffersizinglogicofDDMRPprovidedinChapter7andtheplanninglogicinChapter9giveaclearrangeinwhichwecanjudgehowseverethebimodaleffectreallyis.

FIGURE12-2Theinventoryvalueloss(Taguchi)functionillustrated

Chapter 7 provides the zone calculations and total buffer size. Chapter 9illustratestheplanninglogictokeepthebufferproperlysuppliedand,basedonthat,thetargetedon-handinventoryrange.Thisrangeisdefinedasthetopofredto the top of red + green. It is this range that becomes the defined “optimalrange”inthelossfunctiondepiction.

Chapter9 also showshow todepict apart’spastperformanceagainst thisoptimalrange.Figure12-3isarepeatofFigure9-43.AsdescribedinChapter9,thetopoftheredzonevalueforthesimulatedpartis52,itsyellowzoneis70,and its green zone is 35.Theoptimal on-hand range is the value of the green

Page 312: DDMRP - Demand Driven Material Requirements Planning

zone.Warning ranges appear on either side of the green zone. The low-levelwarning range is on the left; that value is the value of the red zone (52).Thehigh-levelwarningrangeisontheright; thatvalueis theremainingamountoftheyellowzone(35)afterthegreenzoneissubtracted.Thelossfunctionreachestoolittle(pointA)atzeroon-handandtoomuch(pointB)at122.Thisvalueof122isthetopoftheplanningyellowzone.

Figure12-4,whichoriginallyappearedinChapter9asFigure9-44,showsthe run chart of the part over the 21-day simulation depicting the on-handpositionagainsttheoptimalandwarningranges.

FIGURE12-3OptimalrangedepictedfromChapter9planningsimulation

FIGURE12-4Simulatedparton-handperformanceover21days

Chapter 10 brings additional clarity to our range analysis with regard tojudging the severityof a lowon-handposition.The red zone is the embeddedsafetyinthebuffer.Thatsafetyisexpectedtobeused.Themorethaton-handerodes, however, the more severe the situation. The lower warning range isstratifiedintoyellowandredzones.Yellowisthepreliminarywarning,andred

Page 313: DDMRP - Demand Driven Material Requirements Planning

is amore severewarning of a potential stockout.A dark red color is used todepictstockouts.Figure12-5,originallyFigure10-7,showstherunchartforthesimulated part against the execution perspective ranges with the dark redstockoutrangeshown.

Combining the lessons inChapters9 and10,we now have a distinct andspecificdefinitionofhowtoanalyzebufferperformanceovertime.Figure12-6is the conceptual restatement of the loss function for analysis purposes. Thisrestated version provides tighter ranges to judge how buffers are performingovertime.

Eachzoneservestoprovidemorespecificinformationabouthowabufferisperforming against its planned performance—how often it moves out of thenominalrangeandtowhatseverity.Thiscanprovideanindicationofhowstableand reliable the position is over a period of time. Figure 12-7 depicts thesimulatedpartdatafromChapter9againsttheDDMRPanalyticscolorscheme.

FIGURE12-5Planningsimulationwithexecutioncolorscheme

FIGURE12-6TherestatedlossfunctionzonesforDDMRPanalytics

Page 314: DDMRP - Demand Driven Material Requirements Planning

FIGURE12-7Simulatedpartrunchartforanalyticspurposes

The zone labeled “DarkRed (ExcessiveStock)”would be triggeredwhentheon-handlevel isover the topof theplanninggreenzoneof thebuffer.Thezonelabeled“Red(High)”isthevalueoftheplanninggreenzoneofthebufferofthispart(35).Thezonelabeled“Yellow(High)”istheremainingamountoftheplanningyellowzonesubtractedfromtheoptimalon-handrange(70–35=35).Thezonelabeled“Yellow(Low)”istheupperhalfofthebuffer’splanningredzone (on-handalert threshold set to50percentof the redzone).Thezonelabeled“Red(Low)”isthelowerhalfofthebuffer’splanningredzone.Finallythezonelabeled“DarkRed(StockedOut)”isanon-handquantityofzero.

Overthe21-dayperiod,theon-handlevelwaswithintheoptimalrange12of the21days.Eight times it landed in the lowyellowzone.Onlyoncedid itlandinthelowredzone(Day19).Onceitlandedinthehighyellowzone(Day9). Run charts like this represent a distribution of occurrences or positionsagainstthelossfunctionscheme.Figure12-8depictstheon-handpositionsoverthe 21 days in a distribution curve against the loss function scheme for thesimulatedpart.

Withalargeenoughdataset,thedistributionofoccurrencescanbecharted,producing a bell curve against that scheme. Figure 12-9 is a conceptual bellcurveofoccurrencessetagainstthelossfunctionscheme.Onboththeleftandrightextremesofthecurve,therearedashedboxesstartingattheedgeoftheredzoneandmovingoutward.Theseboxesbecomethefocusofmetrics,analysis,andimprovementactivitiesregardingDDMRPbuffers.

These outlying positions specifically represent breakdowns and threats toflow. The larger the number of outlying occurrences in either direction, the

Page 315: DDMRP - Demand Driven Material Requirements Planning

largerthethreattotheflowofrelevantmaterialsrepresentedbyshortages(totheleft) or excess inventory (to the right). Both are targets for improvement andelimination.

Eliminating these outlying positions will improve flow under the currentbuffer definitions and even allow for themigration of parts to buffer profileswithlessvariability.Thiswillshrinkthedistancebetweentheextremepositions,encouragingatighterdistribution.Thistighterdistributionagainstsmallerbufferdefinitionsyieldsbetterflowwithlessaverageinventory.Thistranslatesdirectlytobetter returnon average capital.Once this concept is grasped, basic reportsemergetoidentifytheseoutlyingeventsandthepersonnelthatimpactthem.

FIGURE12-8Simulatedpart’son-handdistributioncurve

FIGURE12-9Outliersofthedistribution

OutlyingEventReports

Page 316: DDMRP - Demand Driven Material Requirements Planning

Outlyingeventreportsaredesignedtopromotevisibilitytotheoutlyingeventsthatbreakdownmaterialflowonbothsidesofthelossfunctiondistribution(thedashed-boxareasofFigure12-9).

Figure12-10isasampleofthistypeofreportforpurchaseditemsthataredisruptingflowonthe“toolittle”sideofthedistributioncurve.ThereportisforactivityoverthemonthofMay(thedaterangeMay1–31).Thereportshowsthepartnumberandthesupplierofthepart.Thecolumn“ParentItems”representsthenumberofparentsofwhichthispartisacomponent.Thecolumn“NumberofStockouts”liststhenumberoftimesoverthedaterangethattheparthasbeenstockedout.Thecolumn“TotalStockoutDays”showsthetotalnumberofdaysthat the part has been stocked out over the date range, as each stockoutoccurrencecouldlastformultipledays.Finally,thenumberoftimestheon-handpositionhaserodedpasttheon-handalertthresholdisrepresentedinthecolumn“NumberofOn-HandRed.”Anystockoutwouldalsobecountedasanon-handred penetration since a part would have to pass through the on-hand alertthresholdbeforebeingstockedout.

FIGURE12-10Purchaseditemsimpactingbufferintegrity

In this report PNWHeat Treat has two problematic parts: PPA and PPC.PPAhad3stockoutoccurrencesrepresentingatotalnumberof10stockoutdays.Thisendangers4separateparentitems.Thepart’son-handpositionhaserodedseverelyon5occasionsoverthedaterange,3ofwhichdeterioratedtostockout.PPChas 12 parent items, 3 stockout occurrences, and a total of 6 stock days.Onlyoneoftheon-handalertpenetrationsdidnotresultinastockout.ColumbiaFabricationhasonepart (PPZ) thatwas stockedout for22daysover thedaterange.

Thistypeofreportcanbegeneratedforanytypeofdecoupledpartposition(purchased, intermediate, finished, and distributed). Figure 12-11 depicts anoutlier report for disruptions to intermediate part buffers. In this case the

Page 317: DDMRP - Demand Driven Material Requirements Planning

supplyingresourceisidentified.ICTissuppliedbytheDriveControlCell,has8parent items, and has stocked out on 5 separate occasions for a total of 21stockoutdays.Additionally,ithaspenetratedtheon-handalertthresholdatotalof9times,5ofwhichresultedinastockout.

Figure12-12isanothersampleofanoutlierreport.Inthiscaseit issortedbysupplier.The“NumberofParts”columndisplayshowmanydifferentpartssuppliedbythissupplierarerepresentedinthereport.PNWHeatTreathas11itemsoverthedaterange(May1–31)thathaveeitherstockedoutorpassedtheon-hand alert threshold (shown in the “Number of On-Hand Red” column).PNWhasatotalof15stockoutoccurrencesagainstthe11parts.Thatmeansthatat leastoneof thosepartshasstockedoutmorethanonceover thedaterange.Those15occurrenceshaveresultedinatotalof29stockoutdays.These11partsalso have generated 19 on-hand alert penetrations (15 ofwhich progressed tostockout.)

FIGURE12-11Intermediatebuffereditemsreport

FIGURE12-12Suppliersimpactingbufferintegrity

These types of outlier reports give visibility to the level of variabilityexperiencedbybuffersandthesourcesofthatvariability.ThisvisibilitywillbeinvaluableindrivingimprovementactivitiesinaDDMRPsystem.

Turningourattentiontotheothersideofthelossfunctionspectrum,wewillfocus on samples of outlier reports that make excess inventory visible (the

Page 318: DDMRP - Demand Driven Material Requirements Planning

dashedboxontherightsideofthecurveinFigure12-9).Figure12-13isanexcessinventoryreportbyaplannerovera31-dayrange.

Itisdesignedtopointoutplanners(andthepartstheymanage)thatarehavingdifficultykeepingon-handinventoriesdown.Threeplannersarerepresented inthisreport.Thetotalnumberofpartsundereachplanner’scontrolisdisplayedinthe column labeled “Number of Parts.” Nick has 152 parts. Carmine has 113parts. Julia has 49 parts. The column labeled “Beginning Excessive On-Hand(DarkRed)” is populatedwith parts and their cumulative values in the excesson-hand zone.This zone is the zone labeled “DarkRed (ExcessiveStock)” inFigure12-7andthezonerepresentedonthefarright(“DarkRed”)inFigure12-9. The number in parentheses is the total number of parts in the zone in thebeginningof theperiodand their cumulativevalue justwithin that zone.Nickhad19partsintheexcessivezone.Thevalueofjusttheamountintheexcessivezoneofthose19partsis$121,633.

Thecolumnlabeled“EndExcessiveOn-Hand(DarkRed)”isthenumberofpartsandtheircumulativezonevalueattheendofthedaterange.Nickreducedthenumberofpartsintheexcessivezonefrom19to17andthetotaldollarvaluefrom$121,633to$112,361.Thereductionof$9,272isdisplayedinthecolumnlabeled“Excess InventoryChange.”Thecolumnlabeled“BeginningHighOn-Hand (Red)” represents the number of parts and their value within the zonerepresentedinFigure12-7labeled“Red(High)”andthezonelabeled“R”ontherightsideofthedistributioninFigure12-9.Nickbegantheperiodwith26itemsinthehighon-handzoneandendedwith12.Highon-handinventorydroppedby$97,939withintheperiod.

Thenumbersofparts and theirvalues in the “Excessive”columnsarenotcounted in the“HighOn-Hand”columns.Adding thenumbers togetherwouldproduce the totalamountofon-hand inventorypartsandvaluesofconcernforeachplanner.ForNick thatwouldbe29parts and$179,654at the endof theperiod.

Page 319: DDMRP - Demand Driven Material Requirements Planning

FIGURE12-13Plannerswithinflatedinventorypositions

FIGURE12-14Partswithinflatedinventorypositions

Figure 12-14 is an outlier report displaying parts with inflated inventorypositions and their respective change over a defined period (May 1–31). Thisreportincludesthespecificpartnumberandtheplannerassignedtoit.Thepart’sbeginning and ending excessive on-hand position and the change between thetwoisdisplayed.Thesameistrueforeachpart’shighon-handposition.

PartFPTiscontrolledbyNick.Overthedaterangethispart’sexcessiveon-handvalue has been eliminated ($1,256). Its highon-hand inventory has beenalmosteliminatedwithanendingvalueof$136.CarminehasmanagedtoreduceFPH’sexcessiveinventorybyasignificantamount($3,690.21).Sincetheexcessinventory has yet to be eliminated at the end of the period, the high on-handremainsunchanged.FPI,however,hasaddedinventoryovertheperiod.Itshighon-handrangewasfilled(from$356to$1,254),resultinginanoverflowtotheexcessiveinventoryrange($1,231).

There are countless derivations and alternatives to organize and displayoutlierdata.Theseare justa fewsamples thathavebeeneffective inDDMRPimplementations to date. Quantity, working capital, and even capacity can beincludedinthesereportsinordertohighlightvisibilitytooutliersandhowtheyrelatecriticalconstraintsandconcernstothedesiredDemandDrivenOperatingModel.

MeasuringVelocity

AnimportantmetricinanyDemandDrivenOperatingModelissystemvelocity.Ifthedecouplingpointsarestrategicinnature,thenmeasuringvelocityatthosepointswill provide visibility and insight into total systemvelocity. If velocitythroughdecouplingpointsislow,thenitwillindicatethattotalsystemvelocitymaybebreakingdown.Ifvelocitythroughdecouplingpointsisincreasing,then

Page 320: DDMRP - Demand Driven Material Requirements Planning

itindicatesthattotalsystemvelocityisincreasing.Buthowtomeasurevelocityatdecouplingpoints?

ADDMRPsystemusesameasurementcalled“order frequencyvariance.”Orderfrequencyvarianceisthedifferencebetweenplannedorderfrequencyandactual order frequency. Planned order frequency is calculated by dividing thegreenzoneofabufferbytheaveragedailyusage.Ifabuffereditemhasagreenzoneof200andanaveragedailyusageof40,thentheplannedorderfrequencyisfivedays.Actualorderfrequencyiscalculatedbydividingthetotalnumberofsupply orders generated against a position by the number of days in theconsideredperiod.Ifthesamepartwasordered7timesovera22-dayperiod,theactual order frequency is 3.14, greater than the planned order frequency. Thismetricismoreinsightfulwithalongerrange.

A“flow index” is anothermeasureof decouplingpoint velocity. It relatestheplannedor actualorder frequencyacross agroupofbuffered items. It is arelativeview.Figure12-15 isanexampleofa flow indexacross100bufferedpurchaseditems.Thisflowindexwascreatedbydividingtheparts’greenzonebytheirrespectiveaveragedailyusage.

Theflowindexspecificallyrangesfromanorderfrequencyof1to50.Partswith an average order frequency higher than 50 are lumped in the “More”column.Theheightofthebarscorrespondstothenumberofpartsthatshareanaverage order frequency within a one-day range. For example, there are fourparts thathaveanaverageorderfrequencyfrom1.01to2.Theparts to thefarrightareslow-movingparts.TheirgreenzonesarelargelyoutofproportionwithADU.Thisistypicallyduetorelativelylargeminimumorderquantities.

Page 321: DDMRP - Demand Driven Material Requirements Planning

FIGURE12-15Aflowindexforpurchaseditems

There are ninepartswith an averageorder frequencyhigher than45days(fouronday46andfive in the“More”column).Figure12-16 lists thesepartsand their respective average order frequency values. Part 182 has the mostinfrequent order frequency of just under 167 days.Each of these parts has itsgreenzonesetasitsMOQ.

These typesofparts typicallyposechallenges toplanningpersonnel.Theyrequiredisproportionately largeamountsof inventory.That inventory typicallyrepresents shared materials and capacity. Their infrequent order patterns canevencomplicateschedulingsequences.

FIGURE12-16Partswithloworderfrequency

There isaheightenedsensitivity to things thatmove tooslowly.Butwhataboutpartsthatmovetoofast?Cantherebesuchathing?Thingsthatmovetoofast can create unnecessary amounts of transactional activity and additionalsetups thatcouldresult incapacityerosion,especiallyatabottleneckresource.Figure12-17liststhepartswithanorderfrequencyhigherthan3.

When severely high-frequency parts are within a group that also hasseverelylow-frequencyparts,a“trade”canbemadethatseekstosmoothflowacrossthegroup.Smoothingflowacrossthegroupwillincreaseflexibility,thusminimizing total working capital commitments, protecting service levels, andminimizing expedite expenses. This trade prevents capacity erosion fromadditional setupscreatedby lowering theMOQsof infrequentlyorderedparts.Byraisinggreenzone limitsofhigh-frequencyparts, itcreatessetupspace forloweringinfrequentpartMOQs.Figure12-18shows thechanges tobothhigh-

Page 322: DDMRP - Demand Driven Material Requirements Planning

and low-frequency parts. High-frequency-part green zones have been set to athree-dayordercycle(ADU×3).Low-frequency-partMOQshavebeencut inhalf.

Figure12-19 is the updated flow indexwith the changed values for high-and low-frequency parts. The distribution of order frequencies has becometighter.

DrivingImprovementinDDMRP

By implementing these simple measures to signal integrity, decoupling pointintegrity, and operating velocity, DDMRP can transform an environment in arelativelyquick timeframeandprovideapath tocontinued improvementovertime.Figure12-20 shows a conceptual view of this progression through threestages.Stage1isthepartinitsinitialbimodaldistributionwithlargeroutlyingbehavior against the buffer criteria. Stage 2 represents initial DDMRP resultsstabilizing the part against its initial buffer criteriawith fewer outlying eventsanda singleuniformdistribution.Stage3 showsa tighteruniformdistributionagainst a compressed buffer definitionwith even smaller amounts of outlyingoccurrences. The arrow in Stage 3 represents the compression of the buffervalues.

FIGURE12-17Partswithhighorderfrequency

Page 323: DDMRP - Demand Driven Material Requirements Planning

FIGURE12-18Changestohigh-andlow-frequency-partaverageorderfrequencies

FIGURE12-19Updatedflowindexofpurchasedparts

Any sustained improvement approach in aDDMRP systemhas a primarydirective:constantlystrivetoreduceworkingcapitalcommitmentswithminimalexpediteexpensesandwithouterosionofservicelevels..Withthisdefinitioninmind,compressionshouldnotoccurunlessthemajorityofoutlyingoccurrencesareidentifiedandeffectivelyeliminated.

Page 324: DDMRP - Demand Driven Material Requirements Planning

FIGURE12-20Acompany’sprogressionthroughDDMRP

AsdescribedinChapter7,therearethreeprimaryfactorsthatcancombineto create the inventory commitment represented in the levels of a buffer:minimumorder quantity, lead time, and the variability factor. Partswith largevariabilityfactorsareanimmediatetargetforimprovementbytheeliminationofoutlyingevents.Theeliminationoftheoutlyingeventsmeansthatthestandarddeviation of the position is reduced. This results in amuch smaller spread invariability.Asmallervariabilityfactorwillresult inasmalleraverageon-handcommitment.Thevariability factordirectly impacts theplanningredzone,andtheredzonecomprisesasignificantportionoftheaverageon-handequation(redplushalfthegreenzone).

Page 325: DDMRP - Demand Driven Material Requirements Planning

Partswithminimumorderquantities thatqualifyas thegreenzonecanbeexploredforreduction.Thiswilldirectlyimpacttheplanninggreenzoneofthebuffer.ThelargertheMOQisinrelationtoaveragedailyusage,thebiggerthetargetforMOQreduction.AflowindexforpartswithMOQ-determinedgreenzonesisusedtoidentifythepartswiththemostpotentialimprovementforMOQreduction.

Parts in long lead time categories can be systematically explored for leadtime reduction. Lead time tends to have the single biggest impact on buffersizing,asitwillaffectallthreezones,ifthegreenzoneisafactorofleadtime.Ifthepartsarepurchased, thenalternativemethodsofsupplyshouldbeexploredwith existing or alternative suppliers. In some cases, it may be financiallybeneficial to pay a higher price per piece if the supplier is oneweek away asopposed to 90 days away. If the parts are manufactured items, then newdecouplingpointsorfastermethodsofproductioncanbeexplored.

Summary

Thischapterwasaboutcreatingvisibilitytothecriticalfactorsthatmaintainandimprovethereliability,stability,andvelocityoftheflowofrelevantinformationandmaterials—theveryfoundationofaDDMRPsystem.Aseriesofmeasures,concepts,andreportswerefeaturedtoensurethisend:

Signal integrity. Properly operating a DDMRP system requiresaccurate and timely supply order generation. A provenmethod tomakesignalintegrityvisiblewasintroduced.Thismethodmeasurescompliance toDDMRPsupplyorder recommendations in termsofbothtimingandquantity.

Decoupling point integrity. Properly operating a DDMRP systemrequires the ability to maintain and improve decoupling pointintegrity.This requires clear visibility to showhow thebuffer hasperformed against the control limits defined by the buffer sizinglogic. When the buffer has performed outside those limits, thevisibility to theoutlyingoccurrencesand their respectivecauses isvital tosystematicallyworkingtowardtheireliminationinordertobetterprotectandpromoteflow.

Decouplingpointvelocity.ProperlyoperatingaDDMRPsystemalsorequiresvisibilitytothevelocityoccurringatthestrategicbuffers.

Page 326: DDMRP - Demand Driven Material Requirements Planning

Focusingontheabovemeasureswillopenthedoorforbasicimprovementavenues for buffer compression through the reductionofMOQs, lead time, orvariability.BuffercompressionresultsinimprovedROI.

Page 327: DDMRP - Demand Driven Material Requirements Planning

CHAPTER13

TheDemandDrivenOrganization

Consider the broader implications that a flow-based strategy has for anorganization—theimpacton thenecessarystrategicandtacticalcomponents tocontrol, measure, adapt, and improve in the New Normal. The better thatorganizationsareatunderstandingandimplementingthesecomponents,thenthemoresuccessfulandsustainableaDDMRPimplementationwillbeandthemoresustainabletheresultingROI.

TheDemandDrivenAdaptiveSystem

This book describes the components and details of Demand DrivenMRP. InChapter5DemandDrivenMRPwasdescribedasacomponentoftheDemandDrivenOperatingModel.InChapter11the“mastersettings”ofDDMRPweredescribed as a component of a larger process knownasDemandDrivenSalesandOperationsPlanning(DDS&OP).ThusDDMRPisacomponentofalargerframeworkforthecomplexityandvolatilitythatorganizationsmustsuccessfullynavigate. This larger and emerging framework is a Demand Driven AdaptiveSystem(DDAS).ADemandDrivenAdaptiveSystemisdefinedas:

A management and operational system designed for complex and volatilemanufacturersandsupplychains.ADemandDrivenAdaptiveSystemusesaconstant system of feedback that connects the business strategy to thesettings and performance of a Demand Driven Operating Model through aDemand Driven Sales and Operations Planning Process (DDS&OP). ADemandDrivenAdaptiveSystemfocusesontheprotectionandpromotionofthe flow of relevant information andmaterials in both the strategic (annual,quarterly,andmonthly)andtactical(hourly,daily,andweekly)relevantrangesof decision making in order to optimize return on equity performance aschangeoccurs.

Page 328: DDMRP - Demand Driven Material Requirements Planning

Figure 13-1 is a depiction of the DDAS framework. This frameworkeffectivelymarries two important timehorizons: the tacticalandstrategic.Thevertical dashed line bisecting the diagram depicts the border of these twohorizons.Thesetwotimehorizonsareincrediblyimportanttounderstandwhenit comes to relevant information. In the tactical horizon (hourly, daily, andweekly), the informationanddecisions thatare relevantaredistinctlydifferentfrom the strategic horizon (annually, quarterly, and monthly). For example,DemandDrivenMRPaspartoftheDemandDrivenOperatingModelexistsinthe tactical relevant range; it is a supply order generation and managementmethod.

FIGURE13-1TheDemandDrivenAdaptiveSystemschema

Workingfromrighttoleft,strategicplanningtakesmarketintelligenceandthe desired business objectives to make strategic decisions. This strategicplanningprocess isaclosed-loopprocessandaswell is linked to thebusinessplan.ThestrategicprocessisaffectedbytheoutcomeofthefeedbackloopfromDDS&OP,whereROIopportunitiesandmarketopportunitiesareidentified.Theclosed loop for strategic planning has the goal tomaximize the system return

Page 329: DDMRP - Demand Driven Material Requirements Planning

accordingtotherelevantmodelfactorsofvolume,rate,andcapability.Whentheloopprocessiscomplete,theintendeddirectionispassedtothebusinessplanintermsoffinancialexpectation.

The business plan is also a closed-loop process taking into account thefinancialexpectationsfromthestrategicdirectionaswellasthedemanddrivenmodel projections. These projections include working capital, lead time,capacity, and other relevant projections. In addition, potential volumeopportunitiesarealsoidentifiedbasedonhowthecurrentmodelisperforming.Oncethisprocessiscomplete,thentheplanparametersarecommunicatedtotheDDS&OPprocess.All theseactivities to thispointare in thestrategicrelevantrangeandtypicallyareplannedinmonthly,quarterly,andannualtimebuckets.

DDS&OP is the lynchpin between the strategic relevant range and thetacticalrelevantrange.DDS&OPdevelopstheDemandDrivenOperatingModelbased in part on themodeling criteria described inChapters6, 7, and 8. Thisincludesdecouplingpointpositions,bufferprofiles,andpartplanningproperties.Thismodel is then run in conjunctionwith actual orders to generate and thenmanage all supply orders in the model. The DDMRP planning engine andexecutionalertsarethekeytothatgenerationandmanagement,respectively.

DemandDrivenVarianceAnalysisfeedsbacktoDDS&OPtodeterminethestability,reliability,andvelocityoftheDemandDrivenOperatingModeloverapast period. This is described in Chapter 12. The variance analysis highlightsareasforimprovementorchangesthatneedtobemadeintheoperatingmodel.Thefeedbackprovidesinformationonhowwellthecurrentmodelisperforming.The variance analysis also provides input to the DDS&OP process so thatDDS&OP can project business financial results based on the current modelcapabilityandperformance.

Each loop is both a closed loop and a tightly integrated chain to the nextloop to provide an aligned adaptive system. Effective deployment allows thecompany to achieve and sustain improved return on investment. TheDemandDriven Adaptive System is how companies will successfully compete in theNew Normal. Just as DDMRP provides the engine in the Demand DrivenOperatingModel,DDS&OPbecomes the critical linkage in aDemandDrivenAdaptiveSystem.

DemandDrivenSalesandOperationsPlanning

DDS&OP includes managing the model and establishing the parameters forcapability.Actualdemandrunsthesystem.AsopposedtotraditionalSalesand

Page 330: DDMRP - Demand Driven Material Requirements Planning

Operations Planning (S&OP), where the outcome is a master productionschedule—a statement of what can and will be built— DDS&OP defines acapabilitysothecompanycanprofitablybuildwhatcanandwillbesold.

Demand Driven Sales & Operations Planning is required as the linkagebetweenstrategyandtacticstodothenecessaryintegratedreconciliationandatthe same time take advantage of all the demand driven capabilities that haveemerged and can be exploited for new markets. The one thing that hasn’tchangedistheimportanceofthepeopleinvolvedtomakethisprocessasuccess.S&OPisaboutwheredecisionsaremadeandwheredecisionsshouldbemade.Thesecretforsuccessistheintegratedreconciliationprocess.

DemandDrivenSalesandOperationsPlanningisdefinedas:

AbidirectionalintegrationpointinaDemandDrivenAdaptiveSystembetweenthestrategic (annual, quarterly, andmonthly) and tactical (hourly, daily, andweekly)relevantrangesofdecisionmaking.DDS&OPsetskeyparametersofa Demand Driven Operating Model based on business strategy, marketintelligence, and key business objectives aligned with strategic informationand requirements.DDS&OPalso projects themodel performancebasedonthe strategic information and requirements and various model settings.Additionally, DDS&OP uses variance analysis based on past modelperformance(reliability,stability,andvelocity)toadaptthekeyparametersofaDemandDrivenOperatingModelandrecommendstrategicalterationstothemodelandprojecttheirrespectiveimpactonthebusiness.

Executed properly, the Demand Driven Sales and Operations Planningprocess directly links the strategic plans for the business with the review ofrelevant performance measures for operational execution and continuousimprovement. With simulation, DDS&OP develops into a bidirectionalintegratedbusinessmanagementprocess.DemandDrivenSalesandOperationsPlanningis the integratedbusinessplanningprocess thatprovidesmanagementwith the ability to strategically direct its businesses to achieve and sustaincompetitive advantage on a continuous basis. It does this by integratingcustomer-focused business plans for new and existing products with thecapability of flexible operations and by recognizing and exploiting the newcapabilitiesthatarenowpossiblefromtheDemandDrivenOperatingModel.

DDS&OPbecomesthebidirectionalintegrationpointbetweenstrategyandtactics in the Demand Driven Adaptive System schema. Figure 13-2 is thevertical representation of the DDS&OP schema. Plan parameters arecommunicatedtoDDS&OPfromthebusinessplanthatalignswiththefinancial

Page 331: DDMRP - Demand Driven Material Requirements Planning

expectations. The financial expectations come from the strategic planningprocessthatcombinesmarketintelligencewithbusinessobjectivestomakethestrategicdecisions.Theplanparametersare thensplit intothreemainplanninginputs—thedemandplan,capabilityplan,andperformancetargets.Thedemandplanissplitbetweenthecurrentportfolioandexpectednewactivities.

Thesethreeplanninginputscollectivelycreateanexpectationofwherethecompany is going in the future. They are used to define the model and partparameters that will define the Demand Driven Operating Model as the firstthreestepsofDDMRP.ThisisfurtherdescribedinthefiveDDS&OPstepslaterinthischapter.

FIGURE13-2DemandDrivenSalesandOperationsPlanningschema

Thestrategyandbusinessplanare typicallyperformedbyproduct linesormarketing product families with a horizon sufficient to secure or change thenecessaryresources.InDDS&OPthisdemandplanistranslatedtothestrategicbufferpositionsasdefinedintheDemandDrivenOperatingModeltocalculatetherequirementfortherelevantfactorstobeconsidered.Forexample,thiscouldbe space, capacity,working capital, orwhatever other limiting relevant factorthatexistsforthecompany.Thiscalculationprocessisfurtherdescribedlaterinthis chapter. These requirements are then considered for feasibility in theDDS&OPprocess.

Page 332: DDMRP - Demand Driven Material Requirements Planning

Themodelandpartparametersareessentiallythecapabilitycontrolsettingsof theDemandDrivenOperatingModel.Theseparameterswere referred to inChapter11asthemastersettingsofthemodel.Notethatthereisafeedbackloopfrom themodelandpartparameters to theplanparameterscalled theDemandDrivenModelProjection.This feedback loopprovidesa forward-lookingviewofhowtheDemandDrivenOperatingModelshouldperformgivenanychangesin the main planning inputs (demand plan, capability plan, and performancetargets). This feedback loop also includes visibility to leadership of currentmodelconfigurationandperformance.

ThecurrentmodelperformanceismonitoredthroughtheotherclosedloopcalledDemandDrivenVarianceAnalysis.Thisfeedbackloopprovidesrelevantinformation on how themodel has performedhistorically against targets.Thisvisibility then allows for adjustments to those parameters. Additionally, thisinformationisleveragedthroughthemetricsandanalyticsdescribedinChapter12 to focusonoutliers.When theoutliersareaddressed, then theoverall flowdispersion of the process is reduced, flow improves, and the capability of theoperational execution is improved. In addition, as the operational execution isimproved,newmarketopportunitiescanbeconsideredinthebusinessplanningand strategic planning processes. Improved operational execution is comparedwiththecompetition,andmarketscanbetargetedthatvaluethatdifferentiator.This can include improved customer service delivery, shorter delivery times,flexible configurations, etc. These operational improvements open marketopportunities to be leveraged by the business to achieve the strategic plan orprovideadditionalrevenueopportunities.

Thisredefinitionfromnewproducttonewactivitiesopensthescopeoutsideabusiness-as-usualsituationandincreasesappealtoamuchbroaderaudienceinthe business. This changing context has a dramatic effect on the view andunderstandingof thedemandandsupplyprocesssteps.The inherentflexibilityof thedemanddrivenMRP(DDMRP)methodology tosenseandadapt to realcustomer demands provides the capability of the overall strategic plan to beroughly right rather than precisely wrong through the traditionally usedinflexiblemasterproductionschedule.

DDS&OP is the lynchpin between the strategic relevant range and thetactical relevant range. Coming from the strategic relevant range, DD&SOPprovides model parameter management. Coming from the tactical relevantrange, DDS&OP provides the control, measures, adaptation, and projectionsnecessarytorunthebusiness.

Page 333: DDMRP - Demand Driven Material Requirements Planning

Theoverallprocessbeginswith thedevelopmentof a strategicmodel thataligns with the business plan. This defines the high-level direction for thecompany—markets,growth,products,etc.Thisstrategicplanning is setby thesenior-level executives of the company. These strategic plans are then furtherrefinedintospecificbusinessplanning,includingthefinancialandnonfinancialgoalsandobjectivesbybusinessunit.TheDemandDrivenSalesandOperationsPlanningprocessiswherethesehigh-levelplansareconnectedtotherealityofoperational capability. Note in Figure 13-1 that there are two very importantlinkages from the business plan parameters to themodel and part parameters.Operational capability and performance targets from the business plan whencombinedwiththedemandplanthendeterminetheintendedoperatingmodelforthecompany.

DDS&OP does not provide a master production schedule to tactical flowmanagement.Infact,whenusingDDMRPastheenginetotheDemandDrivenOperating Model, a master production schedule is not necessary. DDS&OPprovides targets in the form of average daily usage (ADU) aroundwhich thestrategic decoupling buffers are built. Supply order generation is based on thenetflowequation,whichthenreplenishesthesestrategicbuffersbasedonactualconsumption. The demand driven model is about capability rather than adefinitivemonthlyorweeklyrigidschedule.

DDS&OP addresses the volatile, uncertain, complex, and ambiguousbusinessworldwith an adaptive capability.The toppart ofFigure13-2 is thevision and the development of a common understanding of the goals andobjectives of the business. The transition point through DDS&OP is thecollaboration between the goals and objectives with the reality of operationalcapability. This approach distinguishes DDS&OP from conventional S&OPapproaches.

ConventionalSalesandOperationsPlanning

MRPwas codified in the early 1960s and enabled the concept of time-phasedplanning and dependent demand. MPS and capacity planning came next andprovided the ability to have doable schedules and stability. The result wasclosed-loopMRP. This level of integrated manufacturing planning led to theintroductionofS&OPinthe1980sinordertogetmanagementinvolvedtosteertheboat.

TheSalesandOperationsPlanningprocess,accordingtoRichardLingandAndyColdrick,istheprocessthatbringstogetheralltheplansforthebusiness

Page 334: DDMRP - Demand Driven Material Requirements Planning

(customers, sales, marketing, product development, manufacturing, sourcing,andfinancial) intoone integratedsetofplans.Thisprocess is iteratedmonthlyandreviewedbytheseniormanagementteamatanaggregatelevel.Theprocessmust reconcile all supply, demand, and new-product requirements at both thedetailandaggregatelevelsandtietothedesiredbusinessplan.S&OPprovidesthedefinitivestatementofhowthecompanyplanstocompeteinitsmarketfortheneartointermediatetermcoveringahorizonsufficienttoplanfornecessaryresources inaddition to supporting theannualbusinessplanningprocess.Notethe use of the word “process.” S&OP does not mean the meetings wheredecisionsaremade;S&OPisanongoingadaptiveprocesstoeffectivelymanagethebusiness.

Companies have traditionally focused more on supply and financialconsiderationsintheirplanningprocesses.Thewholeschoolofthoughtaroundsupply chainmanagement—and even thename “supply chainmanagement”—reinforces theparadigm that theplanningprocessbeginswith supplyand thenmust be reconciled to demand. This is a left-to-right process from supply todemand.However,demandandnew-productintroductionrepresentthesourceofmost change for a company today. As Charles Darwin said, “It is not thestrongestorthemostintelligentwhowillsurvivebutthosewhocanbestmanagechange.”TheS&OPprocessmustfocusfromrighttoleft—fromdemand—andthenenableasupplycapabilityfrommanufacturingthroughthesupplybasethatcan sense changes in customer demand and adapt planning and production inreal time. This was introduced by Ling and Coldrick in the 1990s as“breakthroughS&OP.”

Early attempts at integration often focused on the commercialization andintroductionstageoftheproductdevelopmentfunnel.Theaimsweretoensurepreparation for launch and phase-in/phase-out and to understand possiblecannibalization effects,motivatedbyhelpingproductionnot tobe caughtwithtoo little or toomuch inventorywhen introducing a new product. Progressiveorganizations,often thosedrivingveryaggressive innovationagendas, realizedthatconnectingonlythebackendoftheprocessstartingattheintroductionstagemissedsignificantopportunitiestomanagethecompleteinnovationfunnelinanintegratedway.

The scope alsowas broadened in another direction by thosewho saw theneed to manage new activities beyond the narrower definition of product.Althoughthelistisdifferentineverycompany,acommonthemeinopeningupthisstepbeyondjustproductisidentificationoftheactivitiesthat:

Page 335: DDMRP - Demand Driven Material Requirements Planning

Haveasignificantimpactondemandandsupply(volumeandvalue)andanyothersupportresources

Needtobemanagedacrosstheentirebusinesswithdecisionsdriventhroughastructuredreviewprocess

Requirevisibility andmanagement across a portfolioof activities,leading to better prioritization, resource allocation, and decisionslinkedtotheoveralldesiredbusinessstrategy

Planningactivityhas significantlychangedover the last10years.Aswithtraditional left-to-right S&OP in the 1980s, the notion of integrating new-productplanningwithsupplyanddemandplanningof theexistingportfolio inright-to-leftS&OPwassomethingofabreakthroughat the time,despitebeingcommonsense.Withtheincreasingfocusoninnovationandtheuseofstageandgate decision processes as well as innovation funnel management, theopportunity exists to integrate these emerging approaches and develop themsymbiotically with S&OP. The next evolution of S&OP is thus by necessitycalled Demand Driven Sales and Operations Planning. DDS&OP isaccomplishedthroughfiveintegratedsteps.

TheFiveStepsofDemandDrivenSalesandOperationsPlanning

Figure13-2 showshow the strategic relevant range is translated to the tacticalrelevant range and vice versa. That bidirectional linkage is DDS&OP and iscomposedoffivestepsinsequentialorder:

1.Strategicbusinessmanagementdirectionandreview2.Integratedreconciliation3.Managingtheportfolioandnewactivities4.Managingdemand5.Managingsupply

Underlying this five-step process is a committed management team withcommunicationclarity.

1.StrategicBusinessManagementDirectionandReview

The only source of sustainable competitive advantage is to exploit thecompany’suniqueoperationaladvantagethatprovidesvaluetothecustomerandprovidesROItothecompany.Inthetraditionalleft-to-rightS&OPapproach,the

Page 336: DDMRP - Demand Driven Material Requirements Planning

managementreviewistypicallythelaststep.InDDS&OP,thisseniorbusinessmanagementdiscussionmustbethefirststepinestablishingtheoverallintendeddirectionforthecompanyaswellasdefiningthekeymetricsthatwillbeusedtoevaluate performance. The senior businessmanagement review establishes thelevel of ownership of the process and specifies how to handle matrixmanagement acrossmultiple divisions,which can spanmultiple countries andcontinents.Thereviewisnotnecessarilyaonce-a-monthmeetingbutratherisaniterative process for seniormanagement to review performance and set futureobjectives.

DDS&OP is the linkage from the overall company strategy to theoperationalcapabilityandviceversa.DDS&OPandDDMRPare symbiotic innature. Each methodology benefits from the existence of the other. ThroughDDMRP,operationsdevelopsauniquecompetitiveadvantagewithshorterleadtimes,lowerinventory,andhighercustomerservicesthroughthefocusonflowandtheleveragingofstrategicdecouplingbuffers.Seniormanagementcanthenchoosetoexploitthesenewcapabilitiestoexpandthecurrentmarket,enternewmarkets,introducenewproducts,orengageinanycombinationofthesepossiblestrategies. This kind of scenario planning sets the stage for the integratedreconciliation of the different views that are now possible with this agilecapability.

ItisimportanttorealizethatDDS&OPisnotaboutamonthlymeetingbutratheradynamicongoingprocess.Theseniormanagementdirectionandreviewisanongoingprocesswhereseniormanagementhasitsfingeronthepulseofthebusiness with the capability to provide direction between the formal monthlymeetings.Ifnot,inadynamicenvironmentanecessarychangecouldslipbyforafullmonthifthisopportunityorchangeisnotedimmediatelyaftertheformalreviewmeeting.

2.IntegratedReconciliation

Integratedreconciliationhighlightstheimportanceoffinancialinvolvementandleadership early in the DDS&OP process. This changes the agenda from adiscussion about volume by product lines and simply reconciling demand andsupply to a true business discussion about capabilities and opportunities forcontributionmargin.Thisstepisaboutreconcilingdifferentviewsandscenarios.There is inherent value in discussing these different views, assumptions, andreasons for those scenarios. This increases the understanding of what thenumbersreallymean,whichfocusesattentionontheassumptionsunderpinning

Page 337: DDMRP - Demand Driven Material Requirements Planning

thenumbers,alongwithopportunities,vulnerabilities,andpossiblecapabilities.Theconversationisprincipallyaboutwhathaschangedbothinsideandoutsidethe organization since the last review and why. Without assumptions, theconversationissimple:Whatnumberschanged?

When the focus changes from just numbers to assumptions and differentscenarios, the need for marrying medium- and long-term forecasting withforesightandmarketintuitionbecomesevenmoreapparent.

Someofthequestionsinvolvedinintegratedreconciliationinclude:

Whatistheimpactofintegratingnewactivities,demandandsupply,onthebusiness(notjustthesupplynetwork)?

Whataretheemergingissuesandgaps?Whataretheopportunitiesandrisks?Whatscenariosarerelevanttomakebetterdecisionsinthefuture?Whatdecisionsshouldandcouldwemake,andwhichonesshouldbeescalatedtotheseniormanagementreview?

Whatoperationalcapabilitycanbeexploitedtoprovidevaluetothecustomer?

Volume and value information and assumption changes are required toanswer these questions.Understanding these questions leads to the imperativethat finance is an integral part of all five steps, whereas in many traditionalS&OPprocesses,financeisaddedonlyatthepre-S&OPmeetingandtheS&OPmeetingratherthanbeinganintegralpartoftheentireprocess.

The reconciliation step is not a meeting as such, but rather an iterativeinteractiveprocess runby a senior-level cross-functional team in thebusiness.This team highlights key issues and decisions required for the seniormanagement team’s attention and discussion. In fact, the reconciliation teamdetermines the agenda for senior businessmanagement review. Participants inintegratedreconciliationarethefutureexecutivesforthebusiness.Thisprocessis truly a key training ground for the next generation of presidents and vicepresidentsbecausethisteamhasaholisticunderstandingoftherelevantbusinessdrivers. This is fundamentally different from the pre-S&OP meeting intraditionalS&OP,wherethemainfocuswasonvolume,itsimpactonresources,andattemptstobalancecapacityandloadwithoutconsideringthefinancialandstrategicimpactofthosedecisionsonthebusiness.

Understandingintegratedreconciliationbroadensthescopeofnewactivities

Page 338: DDMRP - Demand Driven Material Requirements Planning

anddemandandsupplymanagement.Integratedreconciliationasaprocessleadsdirectly into the senior business management review, which focuses onunderstandingchangeincluding:

What is our current performance, and how does it compare withwhatweexpected?

Whatdecisionsarestilloutstanding?Whatdecisionshavebeenmadealreadyinintegratedreconciliation?Areweontrackwiththeintendedbusinessplan?Arewestillontrackwithourstrategicintent?

3.ManagingthePortfolioandNewActivities

Recognizing how different the future will be from the past and present isimportant in understanding the business issues that connect to the DDS&OPprocess. DDS&OP is all about managing change and its consequences to thecompany, its resources, and its people. Figure 13-3 depicts five differentportfolio models with their different emphases on DDS&OP. As models 1through 5 show, a company could have anywhere from a future demand plandevoid of new activity (model 1) to one with a high degree of new-productintroduction (model 5). A company operating under model 5 has the new-activitydevelopment impetuscomingfromproducts thatarenewto theworld.Thisisbydefinitionacompanyembracingastrategyofinnovation.

TheDDS&OPprocessinportfoliomodel1couldbemoretraditionalsincethere is no new activity in the next few years.Demand and supply balancingwouldbetheemphasisinmodelparametermanagement.Becausemedium-andlong-rangeforecastingforstandardproducts inmarkets thatarenotgrowingisrelatively straightforward, the ADU would most likely be calculated from ahistoricalbasis.Abusinessembracingthismodeltypicallywouldbefollowingastrategy of cost leadership. An example of this type of business would becommoditychemicals,wherethepriceissetbythemarket.

Page 339: DDMRP - Demand Driven Material Requirements Planning

FIGURE13-3Differentportfoliomodels

Inmodel2, there ismorenewactivity,but it is relatively straightforward,andthebusinessappearstohavelineargrowth.Newactivitywouldplayapart,but it would be a minor role. An example of this would be an industrialchemicalsorganizationwhosemainbusinessiscommoditiesbutprovidessomespecialty chemicals. This company may be acquiring small businesses toaugment thenew-to-uscategoryandachieving thatgrowthbyacquisition.Thestrategyhereisprimarilycostleadership,buttheresponseinspecialtychemicalscouldbedifferentiatedserviceand lead timebecauseof thehighermarginsontheseproducts.

The most challenging business model for traditional S&OP is portfoliomodel 5,where today’s portfoliowill not exist in four years’ time. These arebusinesseswithahighdegreeoftechnologicalchangeandrapiddevelopmentofnew products. There is a high degree of volatility and risk for the businessembracing thismodel.DDS&OPbringsattention to those risks so theycanbeassessedandplansputinplacetomanagethem.

Portfolio management, including new products, is a critical step in theDDS&OPprocess.Thisdirectlylinksthebusinessplantothestrategicplan.ThetraditionalS&OPmodelofdemandandsupplybalancingwouldappeartobeoflittlerelevancetoexecutivesinthesehighlyvolatileenvironments.Theabilitytosuccessfully manage uncertainty and a range of numbers in the integratedreconciliation step and the importance of simulation and its impact onprofitabilityhaveenormousconsequencesforthebusiness.ThisiswhyDDMRPis the logical methodology to effectively manage that variability. DDMRPprovides the ability to withstand the uncertainty while leveraging resources,

Page 340: DDMRP - Demand Driven Material Requirements Planning

materials,andoperationalcapabilitytobuildwhatcanreallybesoldratherthanwastingtimeandmaterialsonitemsthatarenotreallyrequired.

Measurements such as time to market and time to profit are immenselyimportant. Manufacturers of electronics, mobile phones, software, andcomputersare in thisportfoliomodel.Thestrategynormally followed in thesecompanies is product differentiation coupled with service differentiation.DDMRP provides the company the ability to sense what the actual demandreallyisandthenadaptthesupplynetworktoreacttothatrealdemand.

Many food and drink companies and fast-moving consumer goods andpharmaceuticalcompaniesareexamplesofportfoliomodels3and4.Typically,theywouldfollowproductorservicedifferentiationorcustomerrelationships.Ifyourbusinesshasaportfoliosimilartomodels3,4,and5,spendingtimeonlyimplementing thedemandandsupplyprocess,suchis thecasewith theS&OPtraditional model, provides little insight and value to the management of thecompany.

Understanding the business strategy is essential to understanding theemphasesonthewayDDS&OPwillwork.DDS&OPproductportfoliomodelsgohandinhandwithunderstandingthecompany’sstrategicmodels.Strategiesare about choices and trade-offs, and each business needs to understand andarticulate theprincipalstrategy it is following. It isnotunusual to find thatanorganizationmighthavedifferentbusinessunitsfollowingdifferentstrategiesindifferent areas around the world. DDS&OP must integrate these differentstrategies to provide a complete picture for the senior business managementreview. These are typically represented in product lines or families. Theseproduct lines or families are then converted to projected ADU in the nextsection.

4.ManagingDemand

Demand management, including the accountability for forecasting, hasdevelopedsignificantlyoverthepastfewdecades.IntheearlyyearsofS&OP,alotofeffortwentintoagreeingtoavolumeforecastemphasizingasinglesetofnumbers.Demand forecastingwasveryoftenpartof the supplychainor salesorganization,andforecastormasterscheduleaccuracywasseenastheprincipalmeasure rather thancustomer service.Someorganizations evenwent so far assaying, “You did not forecast this; therefore, we cannot make it!” obviouslyalienatingsalesandmarketing.Thethinkingthatsalesandmarketingformonehomogeneousorganizationwithasingleviewofthenumbersmissesthefactthat

Page 341: DDMRP - Demand Driven Material Requirements Planning

thesetwofunctionshavedifferentdriversandobjectives.By the mid-1990s, people realized the importance of including sales and

marketing in the forecastingprocess.At the same time, “Customer,Customer,Customer!” was fashionable, and sales became the focal point for forecastingandtheone-size-fits-allsolution,ignoringtheimportanceofmarketinginput.Inmanybusinessesfollowinga“customerrelationship”strategy,salesleadershipisappropriate, but in organizations with product and service differentiationstrategies, marketing is the principal driver of medium- to long-term demandprediction.

Givingsalessingleaccountabilityfortheforecastledtosomeorganizationsspendingtoomuchtimeanalyzingdetailedhistoryinafutileattempttogettheforecastaccurate insteadofbeingwith thecustomersgainingknowledgeaboutfuture trends. Against this background of trying to get the forecast accurate,therewasagrowingrealizationthatthereisinherentuncertaintythatdiffersbymarkets,channels,andsectors,aswellaswithdifferentproductsandcustomers.After years of complaining about forecast accuracy and trying to crank thehandlefasteronthesameolddetailedforecastingmachine,companiesbegantowakeup to forecasting forwhat it is—predicting the futurebasedon thepast!However, given the volatility today, the past bears little resemblance to thefuture.Bynomeansdoes this remove the responsibility for forecasting, but itdoes lead to newand innovativewaysofmaking amore informedprediction.Apartfromagreeingwithaforecastnumber,animportantpieceofknowledgeisto understand the likely range (high and low) and forecast confidence factor,together with the supporting assumptions. Providing numbers withoutdocumentedsupportingassumptionsisunhelpful.Insomecompanies,theruleisthataforecastnumbercannotbechangedunlessanassumptionisalsochanged.Forecasting is very important in the strategic relevant range of the DemandDrivenAdaptiveSystemschema.

Today,we understand that a robust demand plan over the necessary timehorizonispossibleonlybyreconcilingcross-functionalviews;volumeandvaluemustbeintegrated.Thisisaccomplishedinthestrategicrelevantrange.Finance,logistics, and supply chainmanagers are committed to this output. In general,salesinputbymajorcustomers(withinputfromaccountmanagers)andchannelsis important in theshort term, typically thefirst four tosixmonths.Marketingprovidesinformationbeyondfourmonthsbasedonmarketshare,businessgoals,andbrandorproducthealthandmarketinginvestment.Strategicmarketingandresearch and development in many cases have a role beyond 12 months,

Page 342: DDMRP - Demand Driven Material Requirements Planning

particularly in new activities. There must be reconciliation between foresight(i.e., strategicmarketing) and forecasting (i.e., neartermmarketing and sales).Theseareguidelinesonly to illustrate thecollaborativeapproachandwillvarydepending on the business. The responsibility of finance, supply chain, andlogistics is to ensure that thevolumeand financial forecast are reconciled andaligned.Thedemandplanisatanaggregatelevelinthestrategicrelevantrange,and the aggregatemarketing families are chosen, understood, and used by allfunctions.Simulationsattheaggregatelevelaremorehelpfulthantryingtodo“what-ifs?”atthestock-keepingunit.

TraditionalS&OPtendedtousemanufacturingfamilies. Indemanddrivenenvironments and product differentiation businesses today, typically theaggregateproductfamilyisthebrandorthebrandandtechnology.Whywouldonechooseanaggregategroupwithlittlerelevancetomarketingandsales?Thedifficultyisthenhowtotranslatetheforecastfrommarketingintoaninputthatcan be used by operations. This translation is done through the DDS&OPprocess by utilizing the DDMRP methodology with the strategic decouplingpositions.Theproductfamilyprojectionsaresupportedbytheoperatingmodelwith those strategic decoupling positions. Those projections require a time-phasedviewtoaccountforworkingcapital,space,andcapacityovertime.Whenanimbalanceoccurs,thentheADUcanbeadjustedwiththeplannedadjustmentfactor to level the load over the planning horizon.This is described in a latersection.

5.ManagingSupply

Supplymanagementalsohasbroadenedinitsscope.Traditionally,itappliedtojustmanufacturing,butnowitisextendedfrommanufacturingtoawiderviewof sourcing that encompasses other resources, including external ones. Inmultinational organizations, it has been extended to supply chain optimizationand risk management, making the best sourcing decisions from the scenarioplanning process. This has challenged the planning capability of manyorganizations, requiring planners who are capable of moving beyond theirtraditional role—that of management and execution in detail at single supplypoints—totheroleofoptimizingsupplynetworksbytestingdifferentscenariosandmakinginformedrecommendationsandhencethebestdecisions.

DDMRP allows the evaluation of different supply strategies through thecalculationandcomparisonofthenecessarybufferstoeachscenario,similartothe evaluationof different demand strategies.Thesedifferent supply strategies

Page 343: DDMRP - Demand Driven Material Requirements Planning

are not converted precisely to a master production schedule to drive supplyorders as in traditional S&OP. Leveraging the existence of the strategicdecouplingbuffers,thecompanynowhasthecapabilitytosensechangingactualcustomer demand and then adapt supply through the entire supply network inrealtimetogeneratethemaximumROI.

DemandDrivenSalesandOperationsPlanningProjections

AcriticalaspectofDemandDrivenSalesandOperationsPlanningistheabilityto project the Demand Driven Operating Model performance given specificcapabilitiesagainstaprojecteddemand.Theseperformanceprojectionstypicallyincludeworkingcapital,space,andcapacityimplications.TheDDMRPmethodsdescribedinthisbookmaketheseprojectionsrelativelyeasytoderivegiventheproperinputs.

Chapter7describedthecriticalinputstothebufferequation.Figure13-4isa repeatofFigure7-18 showing thekey elementsof thebuffer equation.Anyoneof theseelementscanbechangedandtheequationrerun inorder to judgethe impactof that change.For example, the implicationofmoving apart to adifferentprofilecaneasilybecalculatedbychangingtheleadtimeorvariabilityfactor. The implication of changing any of the part traits can also easily becalculated.

One of these part traits is average daily usage. Chapter 7 described thevarious considerations for determining the average daily usage in order tocalculatecurrentdecouplingpointbuffers.ThisisoneofthemastersettingsoftheDemandDrivenOperatingModel.

DDS&OP can also projectmore remote periods of time using a projectedADUinput.This inputwillprovideapoint-in-timepictureofwhat thebuffersfor a particular partwill look like.This picture can then be used to judge theworkingcapital,space,andcapacityimplicationsofdemandatthatlevel.

Theconnectionfromtheoperationalplanningusing theADUasdescribedabove, through the tacticalplanning to theDDS&OPprocess, is accomplishedthrough theuseofmedium-and long-termforecastsbyproduct families.Thesemedium-andlong-termforecastsarethentranslatedtothestrategicdecouplingbufferADUstocalculatethestrategicbuffersthatwouldbenecessarytosupportthat level of business. Once these buffers are calculated, then they can beconverted to working capital investment, space, or critical resource capacity.Thesebuffersare simulatedover theplanninghorizon toensure that sufficientresourcesexisttoexecutetheplan.Ifthereisinsufficientcapacityinfuturetime

Page 344: DDMRP - Demand Driven Material Requirements Planning

periods, then the planned adjustment factors can be used to level the load bybuildingthefast-movingproductaheadofthedemandsurge.

Asopposed to the traditionalplanningbyproduct family that is typical inS&OP,DDS&OP translates that forecast by product family into the projectedADU for strategic buffer items according to the Demand Driven OperatingModel,andhencethissupportsthefinancialexpectationfromthebusinessplanto calculate critical resources. Once these calculations are completed, theinformationcanbeagaindisplayedinmarketingproductfamiliestosupporttheDDS&OPreconciliationmeeting.

FIGURE13-4Bufferequationelements

Asanexample,consideracompanythatmakesfourproducts(itemsXYZ,ZYX,ABG,andGJK).Throughthedemandmanagementprocess,aprojectionhas been made for six months from now, as described in Figure 13-5. Thisprojected rate of demand is displayed in the “Projected ADU” column.Additionally,Figure13-5has thenecessarycomponentsof thebufferequation(leadtimeandbufferprofileattributes).

XYZhasacurrentADUof100.SixmonthsfromnowitsprojectedADUis150. Its decoupled lead time is five days. The parentheses in the “DecoupledLeadTime”columnrepresenttheleadtimecategoryandleadtimefactor.XYZisintheshortleadtimecategoryusinga75percentleadtimefactor(LTF).ThedesiredordercycleforXYZisthreedays.Itsminimumorderquantity(MOQ)is500. Finally, XYZ is in the medium variability category using a 50 percentvariabilityfactor.

Given these inputs, it is relatively simple to produce the projected bufferlevels using the projected ADU. Figure 13-6 shows current versus projectedbufferzonevalues forall four items.Additionally,Figure13-6alsoshows thecurrentversusprojected targetedon-hand inventoryposition (redzonevalue+one-half green zone value). The row titled “Green” displays the green zone

Page 345: DDMRP - Demand Driven Material Requirements Planning

quantityofthebufferaswellasthemethodofcalculation.

FIGURE13-5ExamplecompanyforDDS&OPprojections

FIGURE13-6Currentandprojectedbuffercalculations

WorkingCapital

As described in Chapters 7, 9, 11, and 12, DDRMP buffers are intended toalwayshavestocktomaintaintheirdecouplingprotection.Theaveragequantity(numberofunits)ofstockiscalculatedusingasimpleequation(redzonevalue+one-halfgreenzonevalue).Giventhisequation,anadditionalequationallowsus to convert the average quantity to an average amount of working capital.Figure 13-7 compares the current versus projected average working capitalrepresentedbytheaverageon-handquantityforeachofthefouritems.Ineachcase the average on-hand inventory levels (both current and projected) aremultiplied by the working capital per unit. This working capital per unitrepresents the direct material cost per unit only. This is consistent with theCompanyABCexampleusedinChapters6,7,and9.

To support the predicted future rates of use, an additional $108,160 inaverage working capital will be required. This is neither good nor bad. Thefeasibilityisforthebusinessleadershiptodecidegiventhecircumstancesofthebusiness. It simply means that given the current assumptions (same buffer

Page 346: DDMRP - Demand Driven Material Requirements Planning

profilesandsamepartattributes),thebufferswillneedtocontainmorecapitaltosupporttheincreasedbusinesslevel.

Thosecurrentassumptionscanallbechallengedoverthenextsixmonthsinordertochangetheprojection.Forexample,reducingpartleadtimes,reducingdirectmaterialcosts,orusingalowervariabilityprofilewouldallyielddifferentprojections. If there are real capital constraints, then these avenues can beexploredtoimprovethefeasibility.

Space

In a similar fashion the current and projected average targeted on-hand levelscan be converted to space requirements such as pallet positions. Figure 13-8compares the current versus projected pallet position requirements representedby the average on-hand quantity for each of the four items. In each case theaverageon-handinventorylevels(bothcurrentandprojected)aremultipliedbytheunitsperpallet.

In this case an additional column has been inserted called “ProjectedMaximum Pallet Positions.” This represents the number of pallet positionsrequiredwhenallbuffersareatthetopoftheiraverageon-handrange(redzonevalue + green zone value). This is done to give a sense of range in palletpositions that could be required even when the buffers are deemed to beoperatingwithintolerance.Thisrangemightbeimportantifacompanyhasrealwarehouse or storage limitations. Under the projected rates of demand, anadditional 84 pallet positions will be needed on average. As many as anadditional 133 positions could be needed if all items are at the top of theiraverageon-handrange.

Page 347: DDMRP - Demand Driven Material Requirements Planning

FIGURE13-7Workingcapitalcomparison(currentversusprojected)

FIGURE13-8Palletpositionrequirementscomparison(currentversusprojected)

Again, is this feasible?That is for thebusiness leadership todecidegiventhe circumstances of the business. It simply means that given the currentassumptions(samebufferprofilesandsamepartattributes),thebufferswillneedtoaddmorepalletpositionstosupporttheincreasedbusinesslevel.Thiscouldbeof strategic importance sinceadditional storage space, if required, couldbedifficult or costly to obtain in a short period of time. For example, buildingadditionalspacewillrequiredesign,permitting,andconstructiontime.Usingathird-party warehouse might require a complex logistics plan calling foradditional personnel and transportation. No matter what the specificcircumstances, it is important that the business gains this visibility well inadvanceofactualprojectedneed.

Capacity

AsignificantpartoftheDDS&OPprocessistodetermineifsufficientcapacityexiststosupporttheproposedfuture.UsingDDMRPastheoperationalplanningandexecutionmethodologyallowsa company to ensure that overall sufficientcapacityexistswithoutcommitting toamasterproductionschedule that forcesproductiontobuildwhatisontheschedule,ratherthanpossessingtheagilitytobuild what the customer desires to purchase. This is an excellent example ofwhereroughlyrightwilloutperformpreciselywrong.

Assumethattheresourcebeingconsideredisthelathedepartment.Currentlythelathesarenotaconstrainedresource,butthemanagementteamhasconcerns

Page 348: DDMRP - Demand Driven Material Requirements Planning

aboutthegrowthinbusinessrelativetotheoverallloadonthisdepartment.Thecompanyhas five lathes thatareavailable,eachfor1,200minutesperday(20hours)oftotalproductioncapacity;thatmeansthereare6,000minuteseachdayof total lathe capacity. For simplicity, all five lathes have the same processcapabilityandrate—theyareidenticalinnature.

FIGURE13-9Lathedepartmentcapacityload(currentversusprojected)

Forcalculationofrough-cutcapacityrequirements,thecurrentandprojectedADUsaremultipliedbytheminutesperunitonthelathe.Thisgivesusarough-cut lathe capacity requirement per day to support the current and projectedADUs. Figure 13-9 shows that there is sufficient current capacity. However,giventhecurrentdemandprojections,itcanbeseenthattheprojectedloadsixmonths from now far exceeds the current lathe capacity. This is detailed inFigure13-10.

Page 349: DDMRP - Demand Driven Material Requirements Planning

FIGURE13-10Compositionoflatheload(currentandprojected)

Now choicesmust bemade in themanagementmeeting.Changing bufferprofileswillhaveno impactsince thisanalysis isbasedondemandprojection,availablecapacity, anda specificpart attribute (minutesperuniton the lathe).One or more of the following things must be considered for manipulation toincreasecapacityordecreaseload.

LatheManipulationCurrentlythefivelatheswork20hoursperday.Anadditional1,200minutesperdaycouldbeaddedtothetotallathecapacityifalllatheswentto24hoursperday.Thatwouldbring totalcapacity to7,200minutes, stillbelow the required8,575minutesrequired.Additionally,itwouldrequiremoreoperatorsandleavenotimeforpreventivemaintenance.

Another consideration would be to add more lathes. If all lathes wereavailablefor1,200minutesperday,itwouldtakeatleastthreeadditionallathestomeet the load. This is not a trivial investment. Additionally, it would takemorelatheoperators.Insomepartsoftheworld,theavailabilityandleadtimetohire a qualified operator is sometimes more difficult to come by than the

Page 350: DDMRP - Demand Driven Material Requirements Planning

acquisitionofthemachine.

LoadManipulationOnecritical factor incalculating this load is the time it takesperunit foreachitem.TheitemrequiringtheheaviestloadperunitisXYZ.Ittakes30minutesper unit of lathe capacity. Sixmonths from now it is projected to require 75percent of available daily lathe capacity. One way to manipulate the loadrequirementistoattempttoreengineertheproductinawaythatrequiresmuchlesslathetime.

Anotherwaytodecreaselatheloadwouldbetooutsourceproductionofoneormore of these items.Butwhich items are the right candidates? In order toanswer thisquestion, thecompanywillneed tounderstand the financial returngeneratedbyeachitemthatgoesacrossthelathe.Abasictenetofmanagementaccounting is thatcompanies’profitsmaximizewhen thecompaniesmakeandsell the productswith the highest contributionmargin per unit of the scarcestresource. The scarcest resource six months from now is projected to be thelathes.Thuswewillneedtocalculatethecontributionmarginforeachproductinrelationtoitsimpactonlathingcapacity.Figure13-11showstherelativecashcontributionperlatheminuteforeachitem.PartXYZhasthelargestperminuteloadandthelowestrateofcashreturnonthelathe.Whatthismeansisthatwhenthe company is making XYZ, it is getting $5.83 in cash contribution versus$7.00whenmakingGJK.

In this contributionmargin calculation, we are considering only the priceminustrulyvariablecosts(inthiscasedirectmaterialcost).Thisisbecausetheonlytrulyvariablecostinthisenvironmentisthedirectmaterial;allothercostsareassumedtobefixedwithintheoperationalrelevantrange.Thevariablecostrepresentsarealcashoutlaydirectlyrelatedtoeachunitofeachparticularitem.Wemustunderstand therateofcashgenerationat the latheonly—thescarcestresource. Total labor and overhead are irrelevant and will only distort thepicture.Obviously, this is not the traditionalmargin as calculated in the ERPsystem item master, because standard cost considers nonrelevant costs in thedeterminationoffullyloadedcosts.Readersthatwishtoknowmoreaboutthisconcept are encouraged to read Demand Driven Performance: Using SmartMetricsbyDebraSmithandChadSmith.

Page 351: DDMRP - Demand Driven Material Requirements Planning

FIGURE13-11Contributionpermarginperminuteoflathetime

Whatdoesthismeanfromanoutsourcingperspective?Ifacompanyhasacapacityconstraintanditisgoingtooutsource,thenitshouldoutsourcetheitemthatproducestheleastreturnonthatcapacityconstraint.Itshouldkeeptheitemsthatproducethebestreturnonthatresourcein-house.Inthiscase,XYZshouldbe a candidate for outsourcing to bring the required lathe load down to 6,000minutes per day. That wouldmean outsourcing at least 86 pieces per day onaverage.

DemandManipulationIf the business is incapable ofmeeting all the demand, then demand could bemanipulateddownbyraisingthepriceofcertainitemsinordertomaximizetherateofreturn.Whichitemsshouldbechosenforapriceincrease?Theanswertothis also lies in examining the cash contribution from the lathe for each item.XYZ is projected to sell 150 per day in six months. It is the lowest cashcontributorintermsofthelathe.

Byraising theprice to$180perunitonXYZ, thecontributionmarginperlatheminutebecomesidenticaltoABG.Ifthisdoesnoterodedemandenough,thenXYZandABGshouldbeconsideredforadditionalpriceincreases.ABGisalow-volumeitem,andsoitsimpactintermsofrelievingtotallathecapacityislimited.

Figure 13-12 shows the price required for XYZ andABG to provide thesame contribution margin as the next-lowest product, ZYX. The XYZ pricewouldhavetomoveto$292perunit.ABGwouldhavetobepricedat$208perunit.AtthispointXYZ,ZYX,andABGcouldallbeconsideredforadditionalprice increases if theprojectedADUerosionwas insufficient for the availablelathe capacity. One thing is for certain; the company would be making atremendousamountmorereturnforthesamefixed-coststructure.

Page 352: DDMRP - Demand Driven Material Requirements Planning

ProjectedOrderFrequency

An additional factor to consider in managing supply is the current versusprojectedorderfrequency.Thiscouldberelevantwhenconsidering the impactonthenumberofsetupsandtheimpactoninboundlogistics.

FIGURE13-12Newcontributioncomparison

FIGURE13-13Currentversusprojectedorderfrequency

For simplicitywehave not dealtwith setups related to the lathe example.Yetlookingatprojectedorderfrequencywillgiveusagoodindicationofwhatthe impact will be. Figure 13-13 shows the current versus projected orderfrequencyforthefourproductsfromourexample.AscanbeseenintheFigure13-13,thefrequencyofproductionforGJKchangesdramatically—fromevery12.5daystoevery3.75days.Thisincreaseinthefrequencyofproductioncoulddramatically impact capacity if the setup time for GJK is significant on aconstrainedresource.

TheproductionofABGwillchangefromevery10daystoevery25daysonaverage. If the product has a shelf life, this is an issue that may need to beescalated to the management review. This product may be a candidate todiscontinue since the contribution margin per unit of constrained capacity isrelativelylowandtherewouldbeahighriskoftheproduct’sshelflifeexpiring.

In a distribution environment, order frequency will typically relate to the

Page 353: DDMRP - Demand Driven Material Requirements Planning

averagenumberofinboundreceipts.Themorethatitemsareordered,themoreshipments that tend tobe received.Thiscanputpressureon the receivingandinspectionoperations.Isthereenoughinspectionspace?Arethereenoughdockdoors?Arethereenoughpersonnel?

Summary

TheNewNormal has radically alteredwhat it takes to sustain and improve acompany’s competitive advantage. This alteration requires a new form ofstrategicandtacticalmanagement,onethatallowsacompanytosee,learn,andadapt its resources to the complexity and volatility of the New Normal. Thelegacy tactics inherent in the conventional planning and execution systems ascharacterizedbyMRPandMPSaresimplyinappropriateforthecircumstancesthat a company faces today. That inappropriateness translates directly to poorreturns on asset performance. These distort and confuse the picture andmakestrategicanalysisandpredictionextremelydifficult.Wearesimplystarvingforrelevantinformationinboththestrategicandtacticalrelevantranges.

Yet a newway has emerged: the DemandDrivenAdaptive System. Thisapproach effectively links through the DDS&OP process the strategic andtacticalrelevantranges,providingunprecedentedvisibilityandamechanismtoproducerelevantinformationforadaptationandprojection.Withmorerelevantinformation come more relevant materials and a better return on assetperformance.

The results of the demand driven approach speak for themselves. Typicalresultsinclude:

Servicelevelabove95percentInventoryreductionsof30to50percentExpedite-relatedexpensesdownsignificantlyoreliminated

Many case studies are available at www.demanddriveninstitute.com. SeeAppendixEforaDDS&OPimplementationchecklist.

ContributionofDickLing

Wewould like to recognize thecriticalcontributionofRichard (Dick)Ling tothischapter.DickistheoriginatorofS&OP.

Page 354: DDMRP - Demand Driven Material Requirements Planning

DickLinghasbeenhelpingcompanies largeandsmallwith theirbusinessplanning processes for over 40 years. He has found thatmost companies canimprovetheirbusinessplanningwithsomehelpandtherightfocus.Dickhasawell-deservedreputationasanexcellentcounselorandproblemsolver.

DickhasexperiencewithIBM,AristaInformationSystems,Xerox,andtheOliver Wight Companies. He is an educator, software developer, author,counselor, problem solver, and architect of improved business planningprocesses.Forthepast25yearssincehepioneeredthedevelopmentofS&OP,he has been educating companies about S&OP and helping them implementsuccessfully.

Dickviewscontributingtothischapterintwoways:asameanstohelplinkDDS&OPandDDMRP,atrulyexcitingalliance,andasatributetohislongtimefriendandcollaborator,AndyColdrick,whopassedawayinDecember2014.

Dick Ling and Andy Coldrick collaborated on S&OP for 25 years. Theyformedavery strongpartnership and specialized inpushing theboundaries ofS&OP.TheyhelpedbusinessesallovertheworldtomaximizeS&OP’spotentialto generate more cash and increase profits. Dick created S&OP, and he andAndyhavebeentwooftheleadingthinkersandconsultantsonitsevolutionandadvancement.

TheyledthethinkingonaligningtheS&OPprocesswiththestrategicintentofthebusinessandfutureportfolio.Beforethat,theywerethefirsttorecognizethat new-product activity and financial links to traditional S&OP were beingtreated as afterthoughts and were not being truly integrated. They pioneeredintegrationofthesetwopiecesandalsocreatedtheintegratedreconciliationstepto explode the single-number myth in vogue at the time. The importance ofunderstanding change, assumptionmanagement, and scenario planning with arange of views all reinforced management’s need for information that builtknowledge and know-how rather than data just supplying more and more

Page 355: DDMRP - Demand Driven Material Requirements Planning

numbers.

Page 356: DDMRP - Demand Driven Material Requirements Planning

CHAPTER14

ImplicationsforTechnology

Souder’s law states, “Repetition does not establish validity” Simply doingsomethingoverandoveragaindoesnotmakeittherightthingtodo;itsimplymeans it is the routine thing to do. The point of software is to enable andreinforce routine at scale and velocity. Yet if that routine is not appropriate,software becomes not an enabler of success but a generator of waste and aninhibitoroftheabilitytomanageassets.

Thuswehavecomefullcircle.JoeOrlickyhadverydescriptivewordsforprecomputer inventory management systems in his groundbreaking book,MaterialRequirementsPlanning,backin1975.Hisdescriptionisrelevantonceagain:

Traditional inventorymanagement approaches, in pre-computer days, couldobviously not go beyond the limits imposed by the information processingtoolsavailableatthetime.Becauseofthisalmostallofthoseapproachesandtechniquessufferedfromimperfection.Theysimplyrepresentedthebestthatcould be done under the circumstances. They acted as a crutch andincorporatedsummary,shortcutandapproximationmethods,oftenbasedontenuousorquiteunrealisticassumptions,sometimesforce-fittingconcepts torealitysoastopermittheuseofatechnique.(p.4)

Conventionalplanningsystemsareactingasacrutch.Theydoincorporatesummary, shortcut, and approximation methods based on tenuous or quiteunrealisticassumptions.Theirmandateduseforce-fitsconceptstorealitysoastopermittheuseofthetechniquesembeddedinthem.Ofcourse,thismeansthatanalternative way must be proposed, and that will have huge implications forformalplanningsystems.Arewereadyforthatchange?

Operations and Information Technology— Two Ships Diverging in the

Page 357: DDMRP - Demand Driven Material Requirements Planning

Night?

Itappearsthatwehavereachedthepointofdiminishingreturns.Wecanconfineourselves to and keep trying to optimize systems based on tenuous andunrealisticassumptions,orwecanseektobreakfromconventionforatruestepchange in performance.The emergenceof complex andvolatile supply chainshas created a fundamental gap betweenwhat the operations function needs inorder to stay competitive and innovative and what the current systems allowoperationstodo.Theproofofthisgapisfoundinthewidespreadproliferationofwork-aroundsbasedprimarilyonindividualized,error-prone,andnonscalablespreadsheets.Planningpersonnelactuallybelievetheyarethelesseroftheevils.Whatcanclosethatgap?

There are twomajor stumbling blocks to effectively closing the gap. TheprimaryobstaclehastodowiththedivergencebetweenITandoperationsthatistakingplace inmostorganizations.There is ahugeundercurrentof frustrationwithinmostsupplychain–centriccompanieswithregardtoplanningandcontrolsystems. Operations personnel are frustrated at the lack of solutions madeavailabletothemtocombatthevolatilityandcomplexitytheyareexperiencingin the environment (thus theneed forwork-arounds).Frommanyperspectivesthey see information technology closingdoors to innovation and improvementratherthanopeningthem.

Ontheotherhand,IThasbecomeextremelyfrustratedwithoperationsforworking around and outside the system. IT is concerned with things likesecurity, data, and transactional integrity. Working outside the system posesriskstoallthoseareas.Evenwhennewtechniquesdeliverpromisingreturns,thetechniques canbe easily culledby ITbecause theyhavenot beendone in thesystem. And when informed that the system is incapable of supporting thetechnique?Theresponseisakinto,“Thatcan’tbe;wehaveoneofthebestERPproductsonthemarket.Itsupportsallthebestpracticesoutthere.”

Thedrivebytopmanagement,guidedbyIT,toforceeveryonetoauniformuse of the existing technology—technology that the company spent a smallfortuneandtremendoustimeandefforton—isunderstandable.Itisproblematic,however,whenconsidering that the rulesbehind it are force-fitting techniquesbased on tenuous and unrealistic assumptions. Operations and distribution arebeingasked towork inamanner that supports thesoftware rather thanhavingthe softwarework in amanner that best supports theoperational flowand thestrategicmarketobjectivesofthecompany.

Decades ago ITgenerally reported to finance, and a logical case couldbe

Page 358: DDMRP - Demand Driven Material Requirements Planning

madebasedoncashflowandoperationalneeds.TheCFOwhounderstoodthecompany’soperationsandgrewupinthecompany’splantsimmediatelygraspedtheideasandtheneedforchange.IntheNewNormal,informationtechnologyiscore to the business, increasingly complex, and very political. IT now reportsdirectly to the CEO or COO, and inmany instances IT appears to bewhollydisconnected from the functions it provides service to and the strategicobjectivesofthebusiness.InmanycasesIThasanalltoocozyrelationshipwiththe ERP provider and implementation partner. Does IT even understand thenatureofthecompany’soperationsandtheglobalenvironmentitcompetesin?

Thismay appear to be a rant against IT. It isn’t. The authors are simplyrelayingwhattheyareseeinginsidelargemultinationalcorporations.ITprojectstake forever to negotiate and execute within the company, and both IT andoperations people typically walk away shaking their heads and disappointed.OperationsandITseemtobedivergingatanalarmingrate.WehaveobservedIT organizations that are completely incapable of understanding and orientingaroundthebusinessneedsandtheideaofsystemflow.Wehaveobservedthatoperations people have little knowledge of and regard for the needs andobjectivesoftheITorganization.Bothsidessimplydon’tknowwhattheydon’tknow.Thesolutionwill requireconvergencerather thancontinueddivergence.Wheretofindthatpointofconvergence?

Thesecondobstacleisontologicalinnature—whattypeofrealitydomajorsoftwareprovidersassumesupplychainsareattemptingtocontrolandmanage?In the authors’ experience working with dozens of different ERP systems,today’sERP,MRP,andDRPsystemsaresimplystuckintherulesfromdecadespastdominatedbylinearandcost-centricthinkingsimplybecauseithasalwaysbeenthatwayintheirhistory.

Do major software providers even remotely understand the nature of theproblem? At the time of this writing, the authors’ personal experiences andknowledgeof large softwareproviders’ innerworkings say theanswer is “no”or, at best, “not yet.” Will the tide turn? Fundamentally, it is very simple.Customers need to start demandingmore appropriate planning tools from bigsoftware,andthatrequirestheremovaloftheprimaryobstacletoimproveflow.

DDMRPSoftwareComplianceCriteria

Thisbookhasbeenwritteninordertoconveyablueprinttobringoperationsandinformation technology back into convergence with regard to supply chainplanningandexecution.Additionally,thebookismeanttoserveasablueprint

Page 359: DDMRP - Demand Driven Material Requirements Planning

to software providers by presenting specifications for coding more effectiveplanningandcontrolsystemsintheNewNormal.

Theauthorshaveestablishedthefollowingbasiccriteriaforsoftwaretobecompliant to the DDMRP method. The compliance criteria are intended toensurethatapieceofsoftwarehasenoughfeaturesandfunctionstoimplement,sustain,andevenimproveaDDMRPimplementation.Thecriteriawerewritteninsuchawaythattheyensurecompliancetothefundamentalprinciplesofthemethod but allow sufficient open space for competitive differences, creativity,andinnovation.

Component1:InventoryPositioning

The softwaremustbeable tocalculateand identify thedecoupledleadtimeformanufactureditems.

If the decoupled lead time calculation cannot be performed, thenDDMRPcompliancemust be limited to purchased and distributedpartsonly.

Component2:BufferProfiles

The software must be able to group parts into independentlymanagedfamilieswithvariablesettingsforzoneimpact.

The softwaremust be able to calculateDDMRPbuffers and zonevalues using a combination of buffer profile attributes and theindividualparttraitsofusage,leadtime,andordermultipleorordercycle.

Component3:DynamicBufferAdjustments

Thesoftwaremusthaveaprovisionfordynamicallyalteringbuffersforplannedoranticipatedevents.

Component4:DemandDrivenPlanning

ThesoftwaremustbeabletoperformtheDDMRPnetflowequationproperly, includingqualifying salesorderdemand (due today,pastdue,andqualifiedspikes).

Thesoftwaremustbeabletoproperlydisplaynetflowstatus(color,

Page 360: DDMRP - Demand Driven Material Requirements Planning

percentage, and quantity) for easy prioritization and supply ordergeneration.

All elements of the net flow equation should be visible or easilyaccessibleontheplannerworkbench.

Component5:HighlyVisibleandCollaborativeExecution

Thesoftwareshoulddisplayalertsbasedontheon-handbufferstatusfordecoupledpositions.

Of course, larger, more complex entities will need deeper and richerfeatures.ThislistrepresentsonlythebasicfeaturesthatanyDDMRP-compliantplanningandexecutionsystemshouldhave.

Page 361: DDMRP - Demand Driven Material Requirements Planning

PART3

Appendices

Page 362: DDMRP - Demand Driven Material Requirements Planning

APPENDIXA

AnMRPExample

Readers can watch a video version of this example in the video section ofhttp://demanddriven institute.com. The video is called The ConventionalPlanningPuzzle—JustHowCrazyDoesMRPMakeYourLife?”Thisexercisewill assume that all theMRP requirements and assumptions (as described inChapter3)are100percenttrue.

TheScenario

Thiscompanymakestwoenditems(productAandproductL).FigureA-1isthebillofmaterialforeachitem.AismadefromoneBandoneC.EachBismadefromoneD.LismadefromoneDandoneF.

FigureA-2describesthecharacteristicsofeachpart.Thelotsizeisapolicydecisionthatcontrolshowmanyofthatpartwillbeorderedatatime.APOS4isaperiodofsupplyforfourperiods.Thatmeans thatwhenAis required,MRPwilllookforwardandbuildavariablequantitysuchthatAshouldbebuiltonlyonce every four periods.Theother three parts (L,B,D) haveminimumorderquantities (MOQs) of 250, 100, and 1,000, respectively. These can be set bypolicy,byeconomicorderquantity,bysupplierrequirement,orforavarietyofotherreasons.AnMOQmeansthattheorderquantityisfixedbutthetimingwillvary.

Thelow-levelcodeisthelowestlevelinanybillofmaterialwherethatpartisstructured.EventhoughDisatlevel1forpartL,itisatlevel2forpartA,sothe lowest level where it is structured would be 2. The lead time is themanufacturingleadtimeforpartsA,L,andBandthepurchasingleadtimeforD.SafetystockhasbeensetonthepartlevelsforAandLandonthepurchasingmaterialforD.Thisisthemostcommonconfigurationforsafetystocksinceitisintendedtoprotectagainstvariabilityofsupplyordemand.

Page 363: DDMRP - Demand Driven Material Requirements Planning

SimulatingtheScenario

FigureA-3isthestartingsituationforpartA.PartAstartswith100onhandbuthas15allocatedalreadytoanotherorder.Alsointhefirsttimeperiod,thereare20scheduledtobereceivedonapreviousopenorder.Thatmeanstheavailableinventorywouldthenbe95=100(on-hand)–15(allocated)-10(requirement)+ 20 (scheduled receipt). In the next period the projected on-hand inventorywouldbe85=95(fromperiod1)–10(requirement).Inperiod3theprojectedavailablebalancefallsto70.Notethatinperiod5theprojectedon-handquantityis equal to the safety stock. There is no planned order generated until theprojectedavailablebalanceislessthanthesafetystock,whichhappensinperiod6. The projected on-hand quantity would be 5, which is less than the safetystock,soaplannedorderwouldlookoutthenextfourperiodsandincludethoserequirements. In this case the planned order release would be 45 and wouldoccurinperiod3toallowforthethreeweeksofmanufacturingleadtime.Period3willbetherequireddateforpartsBandC,andbothwillhavearequirementof45.

FIGUREA-1Examplebillsofmaterial

FIGUREA-2Thestaticdataforeachpart

Page 364: DDMRP - Demand Driven Material Requirements Planning

The planner would also receive an error message requesting that thescheduled receipt for period 1 be moved to period 5 since there is toomuchinventorywith the arrival in period 1. Unfortunately, the planner ignores thatmessageandallowsthosepartstoarrive.

Figure A-4 is the starting situation for item L. Similar to item A, theavailableinventoryfor itemLis280=50(on-hand)–20(requirement)+250(scheduledreceipt).Themasterschedulenotesthattherewillbetwoperiodsofspikedemandinperiods4and5.Thereissufficientinventoryforthefirstorder,but the second largeorderkicksoff aplannedorder receipt inperiod5.MRPwillplantoreceiveanorderinperiod5toreplenishthesafetystock.However,theorderreleaseshouldhavebeendonetwoweeksagosincetheleadtimeforthepartissixweeks.Theplannerwillreceivea“pastduefororderrelease”errormessage.The requirement of 250will flow to itemDand showpast due as arequirement.

FIGUREA-3TheinitialMRPplanningtableauforpartA

FIGUREA-4TheinitialMRPplanningtableauforpartL

FigureA-5isthestartingsituationforpartB.Bhasastartingbalanceof50

Page 365: DDMRP - Demand Driven Material Requirements Planning

andnorequirementsinperiod1.Thisparthasaservicepartdemandof10unitseveryotherweek.This ismarkedas independentdemand since thedemand iscomingdirectlyfromthemarket.ThedependentdemandiscomingfromitemAand is exactly what is required by A and exactly the timing required by A.However,aproblemwith itemBis that italsohas insufficient lead timetobemade—twoweeksinsteadoftherequiredfourweeks.Theplannerwillreceiveapastdueorderreleaseandexpediteerrormessagetoday.

The planner for part D was fine when he left yesterday. In fact, he wasprobablyfeelingprettygoodsincehehadtwicethesafetystockonhand.Whenthe planner arrives thismorning, as shown inFigureA-6, he is greeted by anorder past due for release and an expeditemessage.Being a good planner, hepicksupthephoneandcallsthesupplier.Thefirstquestionthesupplierasksishowmany are really needed. The planner responds that 250 units are neededdesperately.Thisistogetthe150requiredforthetwoordersplusreplenishthesafetystock.Thisplannerwentfromhavingtoomuchinventorytotoolittleinthe blink of an eye.He tries to solve the problem by offering the supplier anexpeditefee,airfreight,anincreaseinprice—anythingtogetthosepartsshippedin.However, thebest thesuppliercando is todeliver200 inperiod4and thebalanceinnormalleadtimeinperiod6.FigureA-7isthenewplanningtableauforDgiventheverybestthatthisplannercando.

FIGUREA-5ThefirstplanningtableauforpartB

Page 366: DDMRP - Demand Driven Material Requirements Planning

FIGUREA-6ThefirstplanningtableauforpartD

FIGUREA-7UpdatedtableauforpartD

Thisplannerwillcontinuetoreceivedailyerrormessagesrequestingthatthe200dueinperiod4beexpedited,inadditiontoreceivingamessagetoexpeditethe800fromperiod6.However, this is theverybest thatcanbedonefor thispart.Thetypicalreactionforaplanneristosuppresstheexpeditemessages.

Nowwhatabouttheotherplanners?Thisiswheretheplannersstarttobreaktherulestobringthingsintobalance.TheplannerforpartBisoblivioustotheproblemwithpartDsincehisplanner’sreportonlyprovidesinformationaboutpart B. There is no MRP error message about a component that will not beavailable.Theassumptionismadethatallpartswillbeavailableatthetimeofthe order release. This planner comes to work early this morning andimmediatelyreleasesanorderforBaftercheckingthatindeedthereissufficientstockofDforthatorder.TheBpartplannerdoesnotcheckthefullplanningforD since it does not affect his part B. He is anxious to get that order to theshopfloor since it is already a short lead time order.The order is supposed tohaveafour-weekleadtime.Theplannerreleasestheorderandissuesitwitha

Page 367: DDMRP - Demand Driven Material Requirements Planning

bigexpediteflagthatitisdueintwoweeks.FigureA-8istheupdatedviewofB.

NowitemD is in thesituationdepictedbyFigureA-9.Theplanner forDhasdoneeverythingthathecandobecausetheearliesttheordercanbereleasedistoday.NomatterhowhardtheplannerforpartDworks,hecannotreleaseanorderearlierthantoday.Hehasissuedthisorderwithonlyhalftheexpectedleadtime,whichwilllikelycauseanissueontheshopfloor.

Theplanner for itemLstoppedforanextracupofcoffeeonhisway intoworkthismorningandarrivesaftertheplannerforitemB.Uponhisarrival,heisgreetedwiththerealizationthatnotonlyistheorderforLpastdueforrelease;healsodoesnothavethesupplyofDheneedstobuildL.Soheviolatesthelotsizingruleandreleasesanorderfor100LsincethatisalltheDthatisavailable.Alsohe issues theorderwithabigredexpeditestickerandputsaduedateofweek5onit.HeisfeelingprettygoodaboutthissinceitgetsLbacktosafetystockinweek5.However,theMRPsystemwillnowplananadditionalorderforweek6torestorethesafetystock.Thisorderisalsopastdueforrelease,buttheplannercan’tdoanythingsincetherewillbenoDuntilweek4.FigureA-10 isthenewgridforL.

FIGUREA-8ThesecondplanningtableauforitemB

Page 368: DDMRP - Demand Driven Material Requirements Planning

FIGUREA-9TheupdateplanningtableauforitemD

FIGUREA-10NewplanningtableauforL

FigureA-11showstheimpactonDoftheLplanner’sdecisions.Nowthereis even a higher requirement that is past due, and the buyer tries to call thesupplieragaintoseeifthe200comingincanbeincreased.Theplannervowstoincreasethesafetystocknumberandtheleadtimesothatthiswillneverhappenagain!

Thenetresultofthissingledayisthat:

Disshortandcannotsupporttheparentorderschedule.Ddoesnotrecoverforfiveweeks.Inonedaythepartwentfromhavingdoublethesafetystocktoasevereshortage.

ExpeditefeeshavebeenpaidtotryandexpediteD.

Page 369: DDMRP - Demand Driven Material Requirements Planning

Leadtimeswereviolatedonthreeparts—D,B,andL.Lotsizeruleswereviolatedontwoparts—DandL.SafetystockswouldmostlikelybeincreasedonDandLandfurtherexacerbatethesituation

FIGUREA-11ImpactonpartD

Asanintellectualexercise,theinterestingthingisthatifonlylotsizingwereremoved as a rule, the companywould have sufficient inventory to fulfill allrequirements. See the part planning tableaus in FigureA-12with only the lotsize ruleschangedforallparts to lot-for-lot (LFL)planning.Thesafetystocksarestillkeptasdesigned.Onlyonepartfallsbelowthesafetystockcausinganexpedite—and that is just to replenish the safety stock,notbecause it is reallyneeded for production. There is sufficient inventory to cover the immediaterequirements. There are still some past due order releases, but they aresignificantlysmallerthanbefore.

However,intherealworld,typicallythereisn’ttheluxuryofbeingabletoorder everything only in the quantities that are needed.Usually some kind ofbatchingisemployedtosavesetupsortominimizecosts.

This short exercise explores why people use spreadsheets for planning!Meeting commitments means working around the system. The bigger thecompanyanddeeperthebillsofmaterial,theworsetheproblem.

Page 370: DDMRP - Demand Driven Material Requirements Planning

FIGUREA-12PartplanningtableausinanLFLscenario

Page 371: DDMRP - Demand Driven Material Requirements Planning

APPENDIXB

SimulatingDDMRPBuffers

ThisappendixwillrevealtheresultsofsimulationsdoneusingDDMRPtacticsin highly volatile and random environments. In order to demonstrate theresilience of DDMRP buffers, the simulations used random input fromindividualsatthe2014APICSInternationalConferenceandExpositioninNewOrleans. The results of the simulation were then displayed in front of a liveaudience.

AbouttheSimulation

Each participantwas handed a card and asked to provide 10 days of demandinput, which could be of any value from 0 to 500. There were no otherrestrictions placed on the input provided. The completed cards were simplysequenced as they were submitted by the visitors. The sequenced demandrecords were then entered into a DDMRP simulation tool. Figure B-1 is anexample of a completed card. In this one card you can see a high degree ofdemandvariability.

The simulation tool contained all relevant features for buffer sizing andsupply order generation as described in this book, including dynamicallyrecalculated buffers and order spike qualification as part of the net flowequation.Theorder spikehorizonwas set to only threedays in advance.Thislimited thebenefit that theorder spikequalificationwouldhaveon the resultsachievedbythebuffer.Theorderspikethresholdwassetat10percentoftheredzone. This will result in more order spikes being qualified than using a 50percentsetting.Thesimulationobeysleadtime,meaningthereisnoexpediting,yetalsonosupplydelays.

Each simulation featured two fictitious parts. The first part is called“Widget.” Figure B-2 displays the buffer settings for Widget. Widget is a

Page 372: DDMRP - Demand Driven Material Requirements Planning

purchased item. It is in the long lead timecategoryandusesa25percent leadtimefactor. Itwasplaced inahighvariabilitycategoryandgivenavariabilityfactorof120percent.

Thesecondpartiscalled“Gazoonk.”FigureB-3displaysthebuffersettingsforGazoonk.Gazoonk is amanufactured item. It is in themedium lead timecategory and uses a 40 percent lead time factor. It was placed in a highvariabilitycategoryandgivenavariabilityfactorof120percent.

To achieve representative simulation results, it is necessary to establish aperiodofpriorhistorytoenablethebuffertosettleintoanexpectedbehavior.Inthesimulationtoolwecreateafullyearofpriorhistorybasedon11monthsatanassumedstartingADU.TheopeningADUvaluewassetat30forbothpartsforthosefirst11months.

FIGUREB-1Demandinputcardexample

FIGUREB-2BufferdataforWidget

Page 373: DDMRP - Demand Driven Material Requirements Planning

FIGUREB-3BufferdataforGazoonk

The simulation uses a 90-day past-looking ADU calculation. The finalmonthofpriorhistoryispopulatedwiththevaluesprovidedinthefirstmonthoftheactualsimulationperiod.ThisallowedtheADUtonormalizetotheunknownADUlevelprovidedbytheparticipants.Assuch,inthemonthpriortothestartofthesimulation,theADUstartedclimbingfromtheinitialsettingof30toanopening value of 64.32 for the Widget and 81.44 for the Gazoonk. In bothsimulation trend graphs, youwill notice that supply orders arrive early in thesimulationdue to thepriorhistory.Thesewerebasedon the increasingbufferrequirements in the weeks leading up to the opening day of the simulationwindowasthenormalizationwasoccurring.

Figure B-4 shows the starting buffer levels for both parts given theirrespective ADU on the opening day of the simulation period. These ADUvalues, while in the process of normalizing, are still dramatically understatedgiventheirrespectiveaveragedemand-levelinputs.

WidgetSimulationResults

The simulation of the demand for Widget resulted in 100 percent customerserviceand6.43inventoryturnsforapartwitha90-dayleadtime.Thebuffersrapidly increased in size during the first fewmonths of the simulation, as thedemandprovidedby theparticipantswasmuchhigher than the startingADUsforeachpart.FigureB-5isasummaryofresultsfromthesimulation.

FigureB-6 is a trend graph showing buffer levels and zonal distributions(green, yellow, and red) and on-hand positions (line) over the course of thesimulation.Thetrendgraphdemonstrateshowitquicklyadjustedtothegreaterrateofdemandof151unitsperdayversus thestartingassumption.TheactualADU rate was 2.5 times the opening assumption and pressured the on-handposition,whichbottomedoutat1,369units.However,theincomingsupplyfromorders generated earlier in the simulation put the buffer back into a strongpositiontomaintainserviceforthebalanceoftheyear.

Page 374: DDMRP - Demand Driven Material Requirements Planning

FIGUREB-4StartingbufferlevelsforWidgetandGazoonk

FIGUREB-5Widgetresults

Page 375: DDMRP - Demand Driven Material Requirements Planning

FIGUREB-6Widgetsimulationtrendgraph

The Widget simulation was also run with no order spike qualification.FigureB-7summarizestheresultswithoutorderspikequalification.

Without thebenefitof three-days forwardvisibility tosalesorderdemand,thebufferstillachieved100percentcustomerservicefor theyear.Thismakessense since the order spike horizon is relatively small (3 days) in comparisonwith the part’s lead time (90 days); a small order spike qualificationwindowsimply has less value to extremely long lead time parts. Minimum on-handinventorydecreasedto980units—slightlylessthansevendaysofsupply.FigureB-8showsthetrendgraphwithoutorderspikequalification.

FIGUREB-7Widgetresultswithoutorderspikequalification

FIGUREB-8Widgetsimulationtrendgraph(noorderspikequalification)

Inbothscenarios20supplyordersweregeneratedduring theyearwithan

Page 376: DDMRP - Demand Driven Material Requirements Planning

averageordersizeofjustover2,900units.

GazoonkSimulationResults

Figure B-9 displays the results for the Gazoonk simulation. The simulationresulted in 100 percent customer service and inventory turns of 15.56. Theminimumon-handbalancewas380units and suggests that increased redzonesafety coverage would be appropriate to further reduce the risk of stockouts.Duringtheyear,48supplyordersweregenerated,supportingtherapidturnoverratefortheinventory.

FIGUREB-9Gazoonksimulationresults

FIGUREB-10Gazoonksimulationtrendgraph

FigureB-10 is theGazoonksimulation trendgraph.Thebuffersize flexedupanddown to agreaterdegree as the averagedailyusagedeviated from the

Page 377: DDMRP - Demand Driven Material Requirements Planning

meaninamore“seasonal-type”patternthanWidget.ThisfurtherillustrateshowDDMRP buffers are resilient to changing rates of demand, providing highservicewhilealsodrivingverypositiveinventoryturnover.

AsintheWidgetexample,orderspikequalificationwasturnedoffandthesimulation was run again. Figure B-11 shows the results of the Gazoonksimulation without order spike qualification. Under these conditions the partexperienced twodaysofstockoutandaservice levelof99.5percent.Thepartwasalsoorderedmorefrequently,asorderspikescouldnotbeconsolidatedintothenetflowequation.

Figure B-12 is the trend graph for Gazoonk with no order spikequalification.Onseveraloccasions theon-handpositionsgetdangerously low.In these situations we would expect to see expedite-related activity based oncurrenton-handorprojectedon-handalerts.

FIGUREB-11Gazoonksimulationresultswithoutorderspikequalification

Page 378: DDMRP - Demand Driven Material Requirements Planning

FIGUREB-12Gazoonksimulationtrendgraphwithoutorderspikequalification

While99.5percentcustomerservicewouldbeconsideredexcellentinmostcompanies,anadditionalsimulationwasrunthatadjustedthevariabilityfactorto 150 percent. This was an attempt to see what it would take to clear allstockoutsinthishighlyvariableenvironment.FigureB-13shows theresultsofthesimulationwithalargerredzone.

Serviceimprovedto100percent.Minimumon-handincreasedto235units,whileinventoryturnoverratedeclinedslightlyto14.78annualturns.Again,thisresultwasachievedwithoutany forwardvisibility tosalesorderdemand.Thisshowsus thatevenwitha limitedorder spikehorizon (3days) in relation toapart with a 21-day lead time, order spike visibility and qualification make adifference.Theclosertheorderspikehorizon,salesvisibilityhorizon,andpartlead time are to each other, the more powerful the effect of order spikequalificationonthebuffer’sperformance.

FIGUREB-13Gazoonksimulationresultswithlargerredzone

SimulatingtheImpactofMinimumOrderQuantities

AnadditionalsimulationonGazoonkwasperformedtodemonstratetheimpactofalargeminimumorderquantity(MOQ)onbufferperformance.GazoonkwasgivenanMOQof5,000.Thisrepresentsroughly33daysofconsumptionforapartwith a21-day lead time.Obviously thismeans that thegreen zoneof theGazoonkbufferwillbesizedtotheMOQ.FigureB-14summarizestheresultsofthesimulation.

While service was perfect, it came at a cost. Inventory turnover declinedroughly50percentto9.07.Supplyordersdeclinedfrom50to10.Averageorderfrequencywasover36days.Averageon-handbecame6,298,whichisroughly42daysofsupply.Minimumon-handincreasedto1,141asaresultofthelarger

Page 379: DDMRP - Demand Driven Material Requirements Planning

and less frequentordersize.FigureB-15shows theGazoonk trendgraphwiththelargeMOQ.

This is somewhat typical of experiences with regard to minimum orderquantities representingsubstantialmultiplesofusageover thepart’s lead time.For purchased items, this often represents an ineffective trade-off, as theresulting discount rarely justifies the impact the MOQ has on the flow ofmaterials.Thesamecanbesaid forminimumbatchsizes inproductionwhereefficiencymetrics cause large“artificialbatches” that impede flowand reducemanufacturingresponsiveness.

FIGUREB-14GazoonksimulationresultswithlargerMOQ

FIGUREB-15GazoonksimulationtrendgraphwithlargeMOQ

Summary

Thesimulationofbufferperformanceusingrandomdemandvaluesprovidedby

Page 380: DDMRP - Demand Driven Material Requirements Planning

random individualswas avery real and interesting test of theDemandDrivenMRP methodology. Other than the upper limit of 500, we had no idea whatdemandinputwe’dbegettingfromthestudyparticipants.Bothpartsthatweremodeled in the simulation achieved service levels of 100 percent while alsodrivingverysolidinventoryturnoverrates.It’scriticallyimportanttounderstandthatthisperformancewasachievedwithatmostthreedaysofforwardvisibilitytodemand.

We applied a high-variability safety threshold due to the unknown rate ofdemandthatdrovetheperfectservicelevelsachievedinthesimulation.Wealsoused the simulation to demonstrate how adjusting buffer parameters such asminimumorderquantityaffectsbufferperformance.

ThecoreconceptofDemandDrivenMRPbuffersisthattheyaredesignedto achieve constant material availability. The resilience of the buffers withregard to that objectivewasproved in the simulation examples.Supplyorderswere triggered based on actual sales and the penetration of the buffers. Highinventory turn rateswere achievedwithout the prevalent inventory distortionsseeninforecast-drivenmethodologies.DDMRPalsoprovidesuserswithaveryeasy-to-followsignalingsystemforplanningandsupplychainexecution.

Additional simulation scenarios and results are available atwww.demanddriventech.com.

AbouttheAuthor

Erik Bush has served as the CEO of Demand Driven Technologies since itsformationinthefallof2011.Underhisleadership,DDTechhasquicklymovedfrom a software start-up to a global provider of demand driven supply chainsolutions.DDTech’sglobalclientbasespanssixcontinentsandavastrangeofindustrysegments.

Page 381: DDMRP - Demand Driven Material Requirements Planning

PriortoformingDemandDrivenTechnologies,Mr.Bushspent31yearsinavarietyofrolesatIBM.HeservedastheVicePresidentofGlobalCapabilitiesleading the rapid growth and expansion of IBM’s network of GBS GlobalDeliveryCenters.Heledtheimplementationofsweepingchangesthatresultedin improved client service, rapid market share growth, and substantiallyenhancedcapabilities.Erikalsoservedas theVicePresidentofOperations forIBM’sGlobalBusinessServicesunitsinEuropeandtheAmericas.

Page 382: DDMRP - Demand Driven Material Requirements Planning

APPENDIXC

Applying DDMRP to the Apparel RetailEnvironment

TheNeedforaRetailApplicationofDDMRP

Retail shops are fundamental links of many supply chains. Worldwide retailsaleswereestimatedatUS$22trillionin2014.Wal-Martwasthefirstinthelistof the Fortune 500 companies in 2014 and 2015 andwill likelymaintain thatposition for several years. Most consumer goods, electronics, apparel, andfootwear, all sorts of hardware and home supplies, and many other productcategories are delivered to end consumers through retail stores. The actualdemandofaconsiderableamountofgoodsisdrivenbyretailsales.

But prevailing practices for retail materials planning based on the widelyused “forecast-push-and-promote” mode are far from achieving satisfactoryresults. Current reality proves that the retail industry throughout the worldsuffers from persistent high stockouts and lost sales coupled with seriousinventoryexcess.InDecember2015,thefollowingreportwaspublishedontheInternet,regardingtheperformanceofretailersintheUnitedStates:

According to a study by IHL Group, “out-of-stocks” accounted for $634.1billion in lost retailsales for theyearended in thespring—39percenthigherthanin2012.Likewise,overstockscontributed$471.9billioninlostrevenues,up 30 percent from three years prior. When a retailer has too muchmerchandise,itcutsintoitsmargins.1

Developing an effective material requirements planning and executiontechniqueattheretaillevelthatalleviatessuchsevereadverseresultsisacrucialneed.

DemandDrivenMRPwasdesignedtobeappliedmainlyinmanufacturingcompanies, including the supply chain portion that stretches from suppliers to

Page 383: DDMRP - Demand Driven Material Requirements Planning

distributioncenters.The retail levelwasnot considered in itsoriginalversion.Demand Driven MRP should be able to provide a world-class solution formaterialsplanningandexecutionattheretaillevel.However,someofitsregularrulesdonotconformwelltomaterialsmanagementinretailduetosomespecificcharacteristics of these environments with regard to high uncertainty of new-productsales,salescompositionandconcentration,buffersizing,andminimumdisplayquantities

ThisappendixprovidesadescriptionoftheapplicationofDDMRPtoretail,outliningthechallengesencounteredintheseenvironmentsandthewaytheycanbe overcome, proposing a general framework forwhat can be calledDemandDrivenRetailRequirementsPlanning.

ARetailApparelDDMRPExample

Thefirst implementationofDDMRPinLatinAmerica,whichturnedout tobethefirstDDMRPimplementationintheworldallthewaytotheretaillevel,wasled by the author and performed in Maquila Internacional de Confecciones(MIC),amanufacturerofchildren’sclothingbasedinMedellín,Colombia.Thiscompanydesigns children’s garments under licenses fromDisney,Mattel, andother similar brands and sells through its own chain of over 90 stores inColombia,Venezuela,DominicanRepublic,andCuraçao.

MICisalsoasupplierofsmall,medium,andlargestorechainsinColombia.The company designs the products, purchases yarn, and outsources fabricsweaving and finishing. In addition, the company owns the cutting and sewingproductionfacilitiesneededtomanufacturethefinalproducts.

What our team found in the initial analysis, when the project started inFebruary2013,exposedaprofoundintensificationoftheusualsymptomsofthetraditional push-and-promote model. This is the commonly applied model incompanies that wrongly focus on reducing unit costs as a way of increasingprofitsandthatshowaclearlackofsystemicmanagement.Thesewerethemainsymptomsthatwereobserved:

UnsatisfactoryservicelevelsConstantstockoutsinfinishedgoodsandrawmaterialsLongleadtimestomarketPurchasing,production,anddistributiondecisionsbasedonforecaststotheSKUlevelandintendedtominimizeunitcosts

Page 384: DDMRP - Demand Driven Material Requirements Planning

Excessiveamountsofwork-in-process(WIP)inventorythroughouttheplanttomaximizelocalproductionefficiencies

Significant excesses of inventory in the plantwarehouse of slow-movingproducts

Deliveries to stores every two weeks in order to reducetransportationcosts

Aclearsilomentalityacrossthecompany Chronic conflicts between sales and production and betweenproductionandpurchasing

Constant “scarcity sensation” in stores despite having excessinventories

Itisnotsurprisingthatunderthesepracticesandconditions,thecashflowofthecompanywasnegative.

Supply chains are a combination of inventory and flows that interact in asystemicanddynamicwaywithfeedbackloopsbetweenbothofthem.DDMRPdealsmainly with inventory. Production scheduling deals with flows, and themanner in which the scheduling is performed determines the production leadtimes.

Ourexperiencehasshownthat there isnopossiblewaythatacompanyinthe fashion industry (andmaybe inmany others industries) can achieve goodresultsiftherearelongleadtimesacrossthesupplychain.Ifacompanyhaslonglead times in an environment with high uncertainty and variability, it will bealmost impossible to become demand driven, properly sensing and quicklyadapting tomarket changes.MIC’s production lead timeswere as long as 45days.

We decided that the first task of the project was to substantially reduceproductionleadtimesusingLittle’slaw.Little’slawisthefundamentalequationof operations of which very few professionals working in manufacturing areaware.Intheauthor’sopinion,workinginindustrialoperationswithoutknowingandusingLittle’slawislikeaskingamechanicalengineertodohisorherworkwithoutknowingthatF=m*a.

The scheduling methodology of the Theory of Constraints, referred to asDrum-Buffer-Rope,and thekanbanmethoddevelopedatToyotaboth leverageLittle’slawtokeepWIPatalowandconstantlevelsotheleadtimeisshortandconstant.MaintainingconstantWIPrequires that thework that is released into

Page 385: DDMRP - Demand Driven Material Requirements Planning

thesystembeapproximatelyequaltotheamountleavingthesystem.Ourteamworkedwiththecompany’sschedulingteamtodevelopandusea

basicspreadsheetthatpermittedorderstobereleasedinarelativelysynchronousmanner,definingthesewingoperationasthemaincontrolpointofthesystem.Itwas not a world-class solution, but it worked well for this application. Theproductionleadtimewasreducedto15daysinamatterofafewweeks—a67percentleadtimereduction.

Atthispoint,webeganwiththeimplementationofDDMRP,leveragingitsfive components in prerequisite order. However, the lack of formalmethodologicalrulesformanagingMIC’sretailenvironmentmeantthatwehadto develop a working model that could conform to the challenges andparticularities of this domain. This construct was developed after spending asignificantamountofhourswiththeplanningteamofMIC, learningabout theexisting tools, policies, rules, expectations, and limitations and analyzing howthespeedofflowofrelevantinformationandmaterialscouldbeincreasedwhileoperatingbasedonactualmarketdemand.

To date (December 2015) this methodology has been fully or partiallyreplicatedinfourotherapparelproducersthatowntheirretailstores,obtainingsignificant and consistent results with regard to sales increases (up to 60percent), stockouts (less than 1 percent in high movers), and inventoryreductions (up to 40 percent),with other qualitative and very valuable effectssuchasimprovementsindecision-makingprocedures,workenvironment,clarityofpurpose,etc.

SpecialCharacteristicsandChallengesoftheApparelRetailEnvironment

There are two significant and specific characteristics of the apparel retailenvironmentthatcreateseriouschallengestomaterialsplanning:

1.Averyhighuncertaintyoftheactualsalesofnewproductslaunchedtothemarket

2. A very deep and extreme concentration of sales in a relative lownumberofSKUs

The first one arises from the common practice in the apparel industry offrequently launching new collections and styles following fashion trends. Theproblemisthatthereisanextremeuncertaintyinthedemandforanewproductonceitarrivesinthestores.Itistypicaltoseeinthesecompaniesthatdesigners

Page 386: DDMRP - Demand Driven Material Requirements Planning

and salespeople are constantly surprised by the low sales of a style that theyexpectedtobeamarketwinner,oronthecontrary,anotveryexcitingproductbecomingatruehighmover.

Our experience indicates that only10 to20percent of new styles that arelaunchedtothemarketwillhavesignificantsalesfigures.Theseproductswillbethe high movers. Also, around 40 to 60 percent of the new styles will havemarginal sales, if any, and will end up in an outlet point sold with largediscounts, up to 70 percent. These are the slow-moving items. The remainingstyleswillhavesomeintermediatesalesfigures.

The second special characteristic of the retail environment is that the fewhigh-movingstylesaccountforsome40to60percentofthetotalsalesinaretailpoint.Theremainingportionof totalsales isspreadout inthevastmajorityofmedium-andlow-movingstyles,eachofthemhavingaconsiderablylowersalesvolume,producingaveryhighconcentrationinthesalesdistributionasshowninFigureC-1.

Thisfiguredepictsrealaveragedailysalesandcumulativesales inaretailchain of one of our clients, with 42 stores of a well-known brand. The retailchaincarried761SKUs.The left-handverticalaxishas thesix-monthaveragedaily sales foreachSKU, sorted indescendingorder,with averagedaily salesshown on the solid curve. The right-hand vertical axis corresponds to thecumulativesalesoftheretailchainrepresentedbythedottedcurve.

Page 387: DDMRP - Demand Driven Material Requirements Planning

FIGUREC-1Averagedailysalesandcumulativesalesinarealapparelretailchain

Itwasasurpriseforthisclienttofindthatoutofthetotal761SKUs,around440ofthemhadzerosalesintheperiod.Around50SKUs(lessthan7percent)accountedfor60percentofthetotalsalesofthechain.Inthisexample,thesalesconcentrationisevendeeperthanwhatiscommonlyfound.

Thesefigureshavestrongrepercussionsforhowtheentiresupplychainofanapparel retail chain shouldbemanaged.Thisalso raises sensitivequestionsabout the enormous waste of money, materials, production capacity, labor,transportationcosts,etc.,causedbyslow-andnonmovingitemsthatwillhardlyberecovered.

They are even ofmore concernwhen other industrial sectors like food ormedicine are considered, where expiration dates of products require thatadditionalmoneybedevotedtodestroythem.Therearenolegaloutletsforfoodor medicine. The products are simply incinerated due to wrong practices insupply chain planning and execution, while literallymillions of people are indesperateneedoftheseveryitems.Ourexperienceworkingwithpharmaceuticalandfoodcompaniesconfirmsthatthisisthecase.

Page 388: DDMRP - Demand Driven Material Requirements Planning

As noted earlier, the two general characteristics—extreme uncertainty anddeepsalesconcentration—poserealchallengestomaterialplanningintheretailenvironment,andtheycallforamodelthatacknowledgesandfocusesonthesefundamentalfacts.Underthesecircumstances,maximizingthespeedofflowofrelevant information and materials while responding to actual demand is anabsolutemandate.

Also as mentioned previously, our experience shows that the standardoperatingmodel in this industry is theusual forecast-basedpush-and-promote,coupled with long delays in receiving the relevant information and withcommonly longer-than-30-day production lead times. This model, which weunderstand is the common denominator in this industry worldwide, worksfollowingthisgeneralpattern:

1.Thedesigndepartmentcompletesacollectionfollowingapreviouslydefinedportfoliostructureandcalendar.

2.Themarketingandsalesdepartmentsforecastexpectedsalesforeachstyle or SKU, for collections that will be released several monthsahead.

3. Operations performs a bill of material explosion for all SKUsaccordingtoexpectedsalesquantities.

4.Procurementplacespurchaseordersforalltherequiredrawmaterials,trims,andpackagingmaterial.

5. Production manufactures the entire collection against the definedforecast.

6.Productsarepushedtothestores,sometimesleavingnobufferattheplantlevel.Nobuffersareleftinrawmaterialsortrims.

7.Thecollectionislaunchedandarrivesatthestores.8. The market picks its high-moving items that will be sold out in amatterofweeks.

Page 389: DDMRP - Demand Driven Material Requirements Planning

FIGUREC-2Inventorylevelsinretailafewweeksafteranewcollectionislaunched

9.Storesstarthavingstockoutsoftheitemsthataccountforthemajorityof thesales,even ifasmallbufferwas leftat thecentral level.Thisinventory configuration isdepicted inFigureC-2. Items1, 2, and 3werethehighmovers.Items12–25turnedouttobeslowmovers.

10.Thereisnotanimmediateandsystematicfollow-upofsaleslevelsperSKU.Wehaveencounteredcompaniesthatmeasuresalesonemonthafterthecollectionislaunched.

11.Eveniftherewerenodelaysinobtainingandanalyzingsalesfigures,sincereplenishmentleadtimesinbothproductionandpurchasingareusually longer than 30 days, there is no possible way to replenishhigh-moving items (items that account for 50 to 60 percent of totalsales) when the market still wants them. There are significant lostsales.

12.Ontheotherhand,slowmoverswillnotbesold—sometimesnotevena single unit. These products consumed materials, working capital,production resources, and capacity and transportations costs. Theyalso occupy very valuable and constrained exhibition space in the

Page 390: DDMRP - Demand Driven Material Requirements Planning

storesthatcouldbeusedbyanotherpotentialhighmover.13.Aftersomemonths,theseslow-movingitems(around40to60percent

of the totalcollection)aresent tooutletstoresandaremarkeddownwithdeepdiscountsinordertorecoverthelargestpossibleamountoftheir total variable costs and related expenses. This also generates atremendous opportunity cost, not selling all products at full price inpremiumstores.

14. The excess inventory is significant, and a substantial amount ofworkingcapitalwascommittedforseveralmonths,creatinglongcashcycles.

15.Thenextcollectionisdesigned,andthecyclestartsalloveragain.

Theresultofthiscommonpatternisdevastating:highstock,lostsales,andhigh inventory excesses with extremely long cash cycles. This is where thefigurespresentedbytheIHLgroupreportfit.Itisa$1.0+trillion-dollarproblemintheUnitedStatesalone.

TheProposedModel

Atthispoint,itisabsolutelyclearthatagoodmodelforinventoryplanningandexecutionattheretaillevelmuststartbyacknowledgingandaddressingthehighuncertainty and deep sales concentration issues. In apparel retail (and in anyotherindustrythatisconstantlylaunchingnewshort-life-cycleproducts),itmustbe recognized that any new product that is delivered to stores is basically agamblingexercise.Therateofsuccessofsuchagambleinapparelisratherlow—between10and20percent.

Forecasterrorishuge.Asamatteroffact,someitemswillnotbesoldatallattheirintendedprices.

Sincenobodyreallyknowsorcouldknowinadvancewhichproductswillbe high movers or slow movers, a radically different model of operation isrequired.

TheproposedmodelisbasedontwofundamentalpropositionsofDDMRP,namely:

1.Thefirstlawofsupplychain:high-speedflowofrelevantinformationandmaterials.Demand sensingmust be doneonadailybasis, triggeringcontinuous replenishment signalsbackward in thesupplychain.There isanabsoluteneed for

Page 391: DDMRP - Demand Driven Material Requirements Planning

fastidentificationofhighmoversandthentheneedtoensuretheirfullavailability,replenishingtheminamatterofdays.

2. The first component of DDMRP: strategic inventorypositioning. The retail buffers are replenished by demandfromstrategicbufferspositionedinrawmaterials,intermediatecomponents,andfinishedproducts,carefullyanddynamicallysized according to the second and third components ofDDMRP.

Thebasicfeaturesofthemodelareconceptuallyquitesimple:

Produce and send minimum amounts to stores taking the leastpossiblerisk.

Sensetherealmarketdemandandidentifyhighmoversassoonaspossible.

Replenishthehighmoversfast.Guaranteefullavailability.

Ifanewstyleturnsouttobeaslowmover,itwillnotbereplenished.Thenonlyasmallamountwillneedtobecleared.Therisktakenwillbeonlyinthedisplayamounts.Therawmaterialsandtrimsbuffersmayeventuallybeusedforothernewstyles.

This way of operating allows a company to minimize stockouts of highmovers, thereforeachievingsignificantsales increasesandalsominimizingtheinvestmentandrelatedexpensesinproductsthatturnouttobeslowmoversandthatshouldnothavebeenproducedinthefirstplace(ifanyonehadtheaccuratecrystalballtodefinesuchbehaviorinadvance).

Inordertoimplementtheproposedconceptualmodel,radicalchangesinthetraditional way of operating must be made. The new pattern should be asfollows:

1.Thedesigndepartmentshouldcreateabufferofnewandunreleasedstylesaccordingtothecompanyproductportfolio.

2.Thisbuffermustbecreatedfollowingtherulesindicatedby“designfor manufacturing” techniques, without losing the brand-specificflavor.

3. Once a set of styles will be launched, the sales department must

Page 392: DDMRP - Demand Driven Material Requirements Planning

produce a forecast. This is absolutely unavoidable if there are longleadtimesforpurchasedmaterialsandlonger-than-15-dayproductionlead times. If lead times are lower than a week for purchased andproducedparts, the systemwill just adjust rapidly to actualdemandwithouttheneedforforecastingorlargebuffers.2

4.Aninitiallotofnewproducts,equaltothevisualdisplayofthestoresplus a buffer in the plant warehouse, is sized according to theproduction lead time and expected sales.Obviously, the shorter theproductionleadtime,thelowerthebufferswillbe.

5.Operationsperformsthematerialplanningexplosionforthislot,andlaunches purchase orders for the required components. If there arepurchased parts with long lead times, they should be positioned instrategicbuffers.

6.Manufacturingproducestheinitiallot,andthevisualdisplaysaresenttostores.

7.Actualdemandissensedfromday1,andsolditemsarereplenishedfromthefinishedproductsbuffer.

8.Withinafewdays,theplannersareabletoidentifythehighrunners.Our clients have reported that this trend is easily identified evenduringalongweekend.

9.An early replenishment production order is issued for highmovers.Sinceeithertherawmaterialsandtrimsareinabufferortheycanbeobtainedwith short lead times, this orderwill arrive on time to thecentral warehouse in order to continue replenishing high-movingitems,thusavoidingstockoutsandmaximizingpotentialsales.

10.Slow-movingitemsmustbeclearedinthestoreorsenttoanoutletassoon as possible, making room for new products. This should be aclearcompanypolicy.

11. The process is repeated by launching new products taken from thedesign buffer according to the amount of items thatwere cleared orsenttooutletsandaccordingtothecompanyportfoliostructure.Someofthenewproductswilleventuallybecomehighmovers,whichisthekey to the retailmodel: findasmanyhighmovers as possiblewhilelaunchingasmanyproductsaspossible inacoordinatedmannerandensuremaximumavailabilityofthesehighmovers.

This approach conforms well to how complex adaptive systems (CAS)

Page 393: DDMRP - Demand Driven Material Requirements Planning

shouldbemanaged.Thehighlyuncertainandemergentdemandpatternsshouldbecloselysensed,andtheentiresupplychainexecutionshouldquicklyadapttothem, in a self-organizing dynamic behavior. Also, in CAS the relevantinformation thatshouldbemonitored is in the tailsof thedistribution—that is,the high movers that account for a significant portion of sales and brandpositioningandtheslowmoversthatshouldberemovedfromthestoresassoonas possible, allowing for newproducts to be exhibited and eventually becomehighmovers.

AfewmonthsafteradoptingthisDemandDrivenRetailMaterialPlanningmodel,ourclientshave reported sales increasesofup to60percent, inventorydecreases of up to 45 percent, and therefore significant improvement in cashflowandreturnoninvestment.

RetailDDMRPBufferZoneConsiderations

Another major challenge that our team encountered while implementingDDMRPinMICwastherightsizingoftheretailbufferzones.

Thetypicalcharacteristicsoftheretailenvironmentforclothingare:

Averagedailysalesarelowerthan0.1unitinmorethan98percentof the SKUs (or its equivalent: sales frequency greater than 10days).

Replenishmentleadtimetostoresislowerthanthreedays.TypicallyitisonedaywhenDDMRPisimplemented.

Minimumdisplayinventoriesarerequired.Buffersshouldtakediscretevaluesof0,1,2,3,ormoreunits. There is extreme sensitivity of total inventory to the roundingpoliciesusedinthecalculationofthebufferzones.

Highersalesoccuronweekends.

These circumstances made it impossible to apply the conventionaltechniquessuggestedbyDDMRPforbuffersizingbecause the typicalaveragedailyusageandleadtimecombinationwouldyieldlevelsofzerounitsfortwoofthe threebufferzonesusing the regular roundingpolicies for integernumbers.Additionally, the minimum display quantities would not be respected, and itwouldnotbepossibletosizebuffersof1or2units,whicharethemostcommoninapparelretailenvironments.

Page 394: DDMRP - Demand Driven Material Requirements Planning

Numerically,ifthebufferzones’sizingformulasprescribedbyDDMRPareappliedtopartswithADUlessthan0.1andleadtimeequalto3days,theyellowzonewillbelessthanan0.3unitinmorethan98percentoftheSKUs.Theredzonewillalsobelessthanan0.5unit,andthegreenzonewouldbeequalto1unit (the regular MOQ for these environments). But buffer zones must haveintegervaluesforobviousreasons.

Roundingdowntozerowouldyieldabufferwithzerounitsintheredandyellowzonesand1unitinthegreenzone.Thereplenishmentmechanismofthenetflowpositionwouldnotworksincethereisnota“belowtopofgreen”point.Rounding to 1 unit would unnecessarily and significantly increase theaggregatedinventoryintheretailchain.

Another requirement of inventory management in retail is related to theminimumdisplayquantities.Forvisualmerchandisingreasons,theinventoryina retail storehas tohaveaminimumso thestorehasagoodappearance.Thisminimum amount should be respected in the buffer sizing. Given theseconstraints, a new technique and formula for sizing buffers in the retailenvironmentwasdeveloped.

Thelevelsofeachof thebufferzonesshouldbecalculatedbyconsideringthe average daily usage, lead time, and minimum display; using appropriaterounding rules; andmeeting the needs and specific policies of each companyregardingproductlaunchingandremoval.FigureC-3depictsanexampleofhowthebufferzonesaresizedaccordingtodifferentADUvaluesandagivensetoftheremainingvariables.

There are different rules and formulas that could work in differentenvironmentsandwithdifferentcompanypolicies,soeachspecificcaseshouldbeanalyzed.Themathisrathersimpleandshouldnotposeaproblemtoanyonewho would like to apply the rules and formulas to his or her specificenvironment if the mentioned conditions and requirements are carefullyobserved.

Page 395: DDMRP - Demand Driven Material Requirements Planning

FIGUREC-3DDMRPbuffersforretail—0,1,2,ormoreunits

RealizedResults

TheoperationsmanagerofMIC,alsoashareholderofthecompany,haspubliclyreportedthefollowingresultsachievedafter theimplementationofDDMRPintheretailstorechain3:

60percentincreaseinrevenuesforthe2013–2015period40percentdecreasedinventoryintheretailchain Longer product life cycles, requiring less renewal of the productportfolioandreducingthecomplexityinthesupplychain

Sales of high movers nine times higher during the high season(Christmas) of 2013, compared with those of the same period in2012

The elimination of a sense of scarcity in stores despite having alowerinventory

Decreaseinproductsshippedtooutletsandsoldwithhighdiscounts

Page 396: DDMRP - Demand Driven Material Requirements Planning

Radicalimprovementincashflowofthecompany Synchronization between the different functional areas of thecompany around the principle of increasing the speed of flow ofmaterialsandinformation

These results are consistent with those reported by other retailimplementationsofDemandDrivenMRPinothercountriesandotherindustries.Someofourothercurrentclientsreportsimilarresults.

In addition, it is worth mentioning that MIC applied the same DDMRPconceptsdevelopedfor theretailenvironment to implementavendor-managedinventorymodelin58storesofElÉxitoGroup,thelargestretailerinColombia.This implementation significantly contributed to the decision of this group togrant MIC the distinction as “Best Supplier of the Year” in its category for2014.4

ElÉxitomentionsonitswebsitethatthisawardwasgiventoMICbecause“thecompanydevelopedasupplychainmodelthathasallowedgrowthlevelsupto100%.”

In less than 18 months, MIC transformed from being in a deep crisis toachieving thisnotable recognition. Ithascontinued itsdemanddriven journey,and we are currently implementing a Sales and Operation Planning processleveragingthebenefitsoftheDemandDrivenAdaptiveSystemmodeltoensurethesustainabilityofthecompany’sresults.

AbouttheAuthor

DavidPovedaisacivilengineerfromtheAntioquiaSchoolofEngineering,withamaster’sinprojectmanagementfromtheUniversityofBritishColumbia.Heisthe founder and General Director of Flowing Consultoría, a Colombianconsulting company that specializes in transferring knowledge and supportingimplementationsof themostadvancedmethodologies inoperationsandsupplychain management with emphasis in DDMRP and Demand Driven AdaptiveSystems.HeisanEndorsedInstructorfortheCertifiedDemandDrivenPlannerProgram.

Page 397: DDMRP - Demand Driven Material Requirements Planning

Mr.PovedaservedasManagingDirectorofamidsizesteelwirecompanyinthe1990sandsince1997hasdedicatedhiseffortstoconsultingwithmorethan70 companies in South America. He has been an invited speaker to severalconferences in supply chain management in Latin America and the UnitedStates. He also has served as a member of the board of several industrialcompaniesbasedinMedellín,Colombia,since1993.

Page 398: DDMRP - Demand Driven Material Requirements Planning

APPENDIXD

DemandDrivenMRPDictionary

actively synchronized replenishment (ASR) The initial name given toDemandDrivenMaterialRequirementsPlanning(DDMRP).

ADUSeeaveragedailyusage.ADUalertAnalertindicatingasignificantchangeinADUwithinadefinedset

ofparameters(quantityandtime).ADUalerthorizonAdefinedshorterrollingrangewithinthebroaderrolling

horizonusedtotoalerttosignificantchangestoADU.ADUalertthresholdAdefinedlevelofchangeinADUthattriggersthealert

withintheADUalerthorizon.ADU-based recalculation A process of dynamically adjusting strategically

replenishedbuffersincorporatingarollinghorizon.artificialbatchAnybatchthatisnotafunctionofactualdemand.ASRSeeactivelysynchronizedreplenishment.averagedailyusage(ADU)Averageusageofapart,component,orgoodon

adailybasis.average inventory range This is the expected range inwhich the on-hand

inventoryvalueforaparticularbufferedpartshouldbeonanyparticularday.Therangeisdefinedbythetopoftheredplanningzoneofabufferuptoavalueofthetopofredzoneplustheentiregreenplanningzone.

averageon-handpositionTheredplanningzonevalueplushalf thegreenplanningzonevalueofabuffer.

blendedADUADUcalculatedbasedonacombinationofhistoricalusageandforecastedusage.

bufferpenetrationTheamountofremainingbuffer, typicallyexpressedasa

Page 399: DDMRP - Demand Driven Material Requirements Planning

percentage.buffer profile A globally managed group of parts with similar lead time,

variability,control,andordermanagementcharacteristics.bufferstatusalertsAlerts that show the current andprojected statusof the

decouplingpointpositionsacrossthenetworkofdependencies.bufferzoneAstratificationlayerwithinastockbuffer.Typically,bufferzones

arecolor-codedwithred,yellow,andgreenassignments.CDDLSeeCertifiedDemandDrivenLeader.CDDPSeeCertifiedDemandDrivenPlanner.CertifiedDemandDrivenLeader (CDDL)Aprofessional certificate from

the Demand Driven Institute and International Supply Chain EducationAllianceproclaimingthatapersonhassuccessfullytestedforproficiencyintheaspectsoftheDemandDrivenOperatingModel.

CertifiedDemandDrivenPlanner(CDDP)Aprofessionalcertificatefromthe Demand Driven Institute and International Supply Chain EducationAllianceproclaimingthatapersonhassuccessfullytestedforproficiencyintheDDMRPmethod.

controlpointsStrategiclocationsinthelogicalproductstructureforaproductor family that simplifies the planning, scheduling, and control functions(refertotheAPICSDictionary).

current on-hand alert An execution alert generated by current on-handpenetrationintotheredzoneofthebuffer.

customertolerancetimeTheamountoftimepotentialcustomersarewillingtowaitforthedeliveryofagoodoraservice.

DDASSeeDemandDrivenAdaptiveSystem.DDMRPSeeDemandDrivenMaterialRequirementsPlanning.DDOMSeeDemandDrivenOperatingModel.DDS&OPSeeDemandDrivenSalesandOperationsPlanning.decoupled explosion The cessation of a bill of material explosion at any

bufferedorstockedposition.decoupledleadtime(DLT)Aqualifiedcumulativeleadtimedefinedasthe

longestunprotectedorunbufferedsequenceinabillofmaterial.demand adjustment factor (DAF) A planned adjustment that is a

Page 400: DDMRP - Demand Driven Material Requirements Planning

manipulationoftheADUinputataspecifictimeperiod.DemandDrivenAdaptiveSystem(DDAS)Amanagementandoperational

systemdesignedforcomplexandvolatilemanufacturersandsupplychains.ADemandDrivenAdaptiveSystemusesaconstantsystemoffeedbackthatconnectsthebusinessstrategytothesettingsandperformanceofaDemandDriven Operating Model through a Demand Driven Sales and OperationsPlanning Process. A Demand Driven Adaptive System focuses on theprotectionandpromotionoftheflowofrelevantinformationandmaterialsinboththestrategic(annual,quarterly,andmonthly)andtactical(hourly,daily,andweekly)relevantrangesofdecisionmakinginorder tooptimizereturnonequityperformanceaschangeoccurs.

DemandDrivenMaterialRequirementsPlanning (DDMRP)Amethodtomodel,plan,andmanagesupplychainstoprotectandpromotetheflowofrelevant informationandmaterials.DDMRP is the supplyordergenerationandmanagementengineofaDemandDrivenOperatingModel.

Demand Driven Operating Model (DDOM) A supply order generation,operational scheduling, and execution model utilizing actual demand incombination with strategic decoupling and control points and stock, time,and capacity buffers in order to create a predictable and agile system thatpromotesandprotectstheflowofrelevantinformationandmaterialswithinthe tactical relevant operational range (hourly, daily, and weekly). ADemand Driven Operating Model’s key parameters are set through theDemandDriven Sales andOperations Planning process tomeet the statedbusiness and market objectives while minimizing working capital andexpedite-relatedexpenses.

Demand Driven Sales and Operations Planning (DDS&OP) Abidirectional integration point in a Demand Driven Adaptive Systembetween the strategic (annual,quarterly, andmonthly) and tactical (hourly,daily, andweekly) relevant ranges of decisionmaking.DDS&OP sets keyparameters of a Demand Driven Operating Model based on businessstrategy, market intelligence, and key business objectives (strategicinformation and requirements). DDS&OP also projects the modelperformancebasedonthestrategicinformationandrequirementsandvariousmodelsettings.Additionally,DDS&OPusesvarianceanalysisbasedonpastmodel performance (reliability, stability, and velocity) to adapt the keyparametersofaDemandDrivenOperatingModelandrecommendstrategic

Page 401: DDMRP - Demand Driven Material Requirements Planning

alterationstothemodelandprojecttheirrespectiveimpactonthebusiness.DLTSeedecoupledleadtime.dynamic buffers Buffer levels that are adjusted either automatically or

manuallybasedonchangestokeyparttraits.executionhorizonThelifecycleofanorderfromthetimetheorderiscreated

orreleasedtothetimeitisclosed.flowindexAverageorderfrequencycomparedacrossallparts.forwardADUADUcalculatedbasedonforecast.greenzoneThetoplayerofareplenished,replenishedoverride,andmin-max

buffer. If thenet flowposition is in thiszone, thennoadditional supply iscreated.

leadtimeadjustmentfactorAmultiplicativefactorappliedtoapart’sleadtime.

leadtimealertAnalertorwarninggeneratedbyanLTMpart.Analertwillbetriggeredwheneverthepartentersadifferentzoneinthebuffer.Greenisthefirstalerttobeencountered,followedbyyellowandthenred.

leadtimealertzoneThezoneassociatedwiththepercentageofleadtimethatprovides the definition for lead time alerts. The LTM alert zone has threesectionscolor-codedgreen,yellow,andred.

lead timemanaged (LTM) part A critical nonstocked part that will havespecialattentionpaidtoitoveritsexecutionhorizon.Typically,LTMpartsarecritical,longleadtimecomponentsthatdonothavesufficientvolumetojustifystocking.Aportionoftheleadtimeofthepart(typically33percent)will have a three-zoned warning applied to it. That portion is typicallydividedintothreeequalsections.

LTMpartSeeleadtimemanagedpart.marketpotentialleadtimeTheleadtimethatwillallowanincreaseinprice

orthecaptureofadditionalbusinessthrougheitherexistingornew-customerchannels.

material synchronization alert An alert against a demand allocation(s)whensupplyisprojectedtobeinsufficient tocoverthedemandat thetimethedemandissettooccur.

matrix bill of material A chart made up from the bills of material for anumberofproductsinthesameorsimilarfamilies.Itisarrangedinamatrix

Page 402: DDMRP - Demand Driven Material Requirements Planning

with components in columns and parents in rows (or vice versa) so thatrequirements for common components can be summarized conveniently(refertotheAPICSDictionary).

netflowequationAplanningcalculationtodeterminetheplanningstatusofabuffereditem.TheequationisOn-hand+on-order(alsoreferredtoasopensupply)–unfulfilledqualifiedactualdemand.Alsoknownasthe“availablestockequation.”

net flow position The position yielded by the net flow equation against apart’sbuffervalues.Alsoknownasthe“availablestockposition.”

nonbufferedpartsAllpartsthatarenotstocked.occurrence-basedrecalculationAmethod to adjust buffers based on the

numberandseverityofspecificoccurrencesinapredefinedfixedinterval.on-handalertlevelThepercentageoftheredzoneusedbybufferstatusalerts

inordertodetermineayelloworredexcecutioncolordesignation.orderspikehorizonAdefinedfuturetimeframeusedtoqualifyorderspikes

in combination with an order spike threshold. Typically, the order spikehorizonissettothepart’sdecoupledleadtime.

order spike threshold A defined amount used to qualify order spikes incombinationwithanorderspikehorizon.Typically,theorderspikethresholdwillbeexpressedasapercentageof the total redzone (orminvalue)of apart’sbuffer.

OTOGSeeoverthetopofgreen.Over the top of green (OTOG) A situation in which either the net flow

position or on-hand stock is over the top of the defined green zone,indicatinganexcessiveinventoryposition.

PAFSeeplannedadjustmentfactor.pastADUADUcalculatedbasedonhistoricalusage.planned adjustment factor (PAF) Buffer manipulations based on certain

strategic,historical,andbusinessintelligencefactors.planned adjustments Manipulations to the buffer equation that affect

inventory positions by raising or lowering buffer levels and theircorrespondingzonesatcertainpointsintime.Plannedadjustmentsareoftenbasedoncertainstrategic,historical,andbusinessintelligencefactors.

prioritizedshareAn allocation schema utilizing the net flow positions of a

Page 403: DDMRP - Demand Driven Material Requirements Planning

groupofpartsinordertoaccommodateaspecificlimitationorrequirement.projected on-hand alert An alert generated by a low projected on-hand

positionoverapart’sDLTbasedonon-hand,opensupply,andeitheractualdemandorADU.

qualified actual demand The demand portion of the net flow equationcomposed of qualified order spikes, past due demand, and demand duetoday.

qualifiedorderspikeAquantityofcombineddailyactualdemandwithintheorderspikehorizonandovertheorderspikethreshold.

ramp-down adjustment Manipulations to the buffer equation that affectinventorypositions,loweringbufferlevelsandtheircorrespondingzonesatcertain points in time. Ramp-down adjustments typically are used in partdeletion.

ramp-up adjustment Manipulations to the buffer equation that affectinventory positions, raising buffer levels and their corresponding zones atcertain points in time. Ramp-up adjustments typically are used for partintroduction.

red zone The lowest-level zone in a replenished, replenished override, andmin-max part buffer. The zone is color-coded red to connote a serioussituation.Theredzoneisthesumoftheredzonesafetyandredzonebase.

redzonebaseTheportionoftheredzonesizedbytheleadtimefactor.redzonesafetyTheportionoftheredzonesizedbythevariabilityfactor.relativepriorityTheprioritybetweenorders filteringby zone color (general

reference)andbufferpenetration(discretereference).replenished override part A strategically determined and positioned part

using a static (buffer zones are manually defined) three-zoned buffer forplanningandexecution.

replenished part A strategically determined and managed part using adynamic three-zoned buffer for planning and execution. Buffer zones arecalculatedusingbufferprofilesandspecificpartattributessuchasADUandDLT.

salesordervisibilityhorizonThetimeframeinwhichacompanytypicallybecomesawareofsalesordersoractualdependentdemand.

seasonality adjustment Manipulations to the buffer equation that affect

Page 404: DDMRP - Demand Driven Material Requirements Planning

inventorypositionsbyadjustingbufferstofollowseasonalpatterns.significantminimumorderquantity Aminimum order quantity that sets

thegreenzoneofabuffer.spike A comparatively large amount of cumulative daily actual demand that

qualifiesforinclusionintothenetflowequation.stockout(SO)Anitemthatisnotimmediatelyavailableinstock(refertothe

APICSDictionary).stockoutwithdemand(SOWD)An item that isnot immediately available in

stockandhasademandrequirement.Alsoknownasnegativeon-hand.stockout with demand alert A form of on-hand alert, triggered by a

strategicallystockeditemwithalackofinventoryonhandandthepresenceofademandrequirement.

strategic inventorypositioning The process of determiningwhere to putinventory that will best protect the system against various forms ofvariabilitytobestmeetmarketneeds,leverageworkingcapital,andmitigatethebullwhipeffect.

supplyoffsetAdjustingthetimingoftheapplicationofademandadjustmentfactortoaccountforlongleadtimecomponents.

synchronization alerts Alerts designed to highlight problems with regard todependencies.

thoughtwareTheanalysisandprocessemployedtodefinetherelevantfactorsand dependencies in an organization or system in order to constructappropriate business rules and operating strategies thatmaximize velocity,visibility, and equity. Within the DDRMP framework, thoughtware iscommonly referred to with regard to applying the inventory positioningfactors.

TOGSeetopofgreen.topofgreen(TOG)Thequantityof thetoplevelof thegreenzone.TOGis

calculatedbysummingthered,yellow,andgreenzonesofabuffer.topofred(TOR)Thequantityofthetopleveloftheredzone.topofyellow(TOY)Thequantityofthetopleveloftheyellowzone.TOYis

calculatedbysummingtheredandyellowzones.TORSeetopofred.

Page 405: DDMRP - Demand Driven Material Requirements Planning

TOYSeetopofyellow.yellowzoneThemiddlelayerofthebufferlevelcodedwithyellowtoconvey

asenseofwarning.Theyellowzoneistherebuildzoneforreplenishedandreplenishedoverridebuffers.

zone adjustment factor Adjusting part buffer zones by applying amultiplicativefactortothevalueofthezone.

Page 406: DDMRP - Demand Driven Material Requirements Planning

APPENDIXE

DDS&OPChecklist

A one-size-fits-alluniversal checklist forDDS&OP isnothelpfulbecauseofthe uniqueness of each company. DDS&OP provides a dynamic adaptableapproachforacompanytosetitsstrategyandthenrecognizeandrespondtorealdemand. However, the following checklist is a start to ascertain where yourcompany isat thecurrent time.Rather thanayesornoanswer, consideryourcompany’sprogressalongacontinuumoftheseconsiderations:

Is the purpose of DemandDriven Sales andOperations Planningunderstood?

DoestheDDS&OPprocesshaveanexecutiveprocesschampion?DoestheDDS&OPprocesshaveprocessstepowners? Are values and behaviors recognized as critical for a successfulDDS&OPprocess?

Istheemphasisonthefuture,understandingchangeanditsimpactonthesuccessofthebusiness?

Aretacticalissuesandproblemsdiscussedandquicklyresolved?Istherefinancialandvolumeintegration?Doperformancemanagementsystemsreinforceintegratedbehaviorandthedisciplineofexecution?

Is theDDS&OPprocesswelldocumented,withdocumentationforeachstepupdatedonaregularbasis?

Doesthebusinessmakeadistinctionbetweeninformationanddata?Isthereacommitmenttodataintegrity? Is there a new-product planning review that provides input to the

Page 407: DDMRP - Demand Driven Material Requirements Planning

managingdemandprocessstepandtheoverallDDS&OPprocess? Isthereamonthlyreviewofthefutureunconstraineddemandplanfor existing and new products based on inputs from marketing,sales,andfinance?

Is there aDDS&OPprocess at the supplypoint level that ensuresthatthereisavalidplantosupportordersandshipments?

Isthereaprocessthatinvolvesallthebusinessfunctionsinordertodevelop an integrated set of plans that reconcile the S&OPprojectiontothebusinessplan?

Isinformationpresentedtotheseniorbusinessmanagementteamonan exception basis with a focus on understanding and managingchanges?Aregraphicsusedextensivelytoimprovescheduling?

AsRalphWaldoEmersonsaidaboutlife,thesameistrueaboutDDS&OP—itisaboutthejourney,notthedestination.

Page 408: DDMRP - Demand Driven Material Requirements Planning

Endnotes

Chapter1

1.All theAPICSdefinitionsinthebookarefromthefourteentheditionoftheAPICSDictionary,(Blackstone,2013).

2.You can access the report at http://dupress.com/articles/success-or-struggle-roa-as-a-true-measure-of-business-performance/.

3. You can access the article at www.zerohedge.com/news/2013-02-12/how-rookie-excel-error-led-jpmorgan-misreport-its-var-years.

Chapter5

1.ThedefinitionisfromTheFreeDictionary,www.thefreedictionary.com.2. Liker, J. (2004). The Toyota Way: 14 Management Principles from theWorld’sGreatestManufacturer.NewYork:McGraw-Hill.

AppendixC

1. www.businesswire.com/news/home/20150506005233/en/Research-Report-Retailers-Lose-1.75-Trillion-Revenue.

2. It can be easily proved that purchasing materials from more expensivesuppliersbutwith short lead timesor shippingbyplane is averyprofitabledecision. This fact totally goes against the generalized but very unwisepractice of buying from cheaper producers in Asia or elsewhere that haveseveralweeks’orevenmonths’replenishmentleadtimes.

3. Gómez, J. (2015). “Reposición por Demanda,” Logismaster Congress,Medellín,Colombia.

4. “Ganadores de Proveedores de Éxito 2014,”www.grupoexito.com.co/es/proveedores/concurso-proveedores-de-exito/historia-de-ganadores?id=1232.

Page 409: DDMRP - Demand Driven Material Requirements Planning

References

APICS.(2013).APICSDictionary,14thed.Chicago:APICS.Goldratt,E.M.,&Cox,D.E. (1984).TheGoal:AProcessofOngoing Improvement.Great

Barrington,MA:NorthRiverPress.Liker, J. (2004). The Toyota Way:14 Management Principles from the World’s Greatest

Manufacturer.NewYork:McGraw-Hill.Noreen,B.G.(2013).ManagerialAccountingforManagers,3rded.NewYork:McGraw-Hill

Education.Orlicky,J.(1975).MaterialRequirementsPlanning:TheNewWayofLifeinProductionand

InventoryManagement.NewYork:McGraw-Hill.Plossl,G. (1992).Orlicky’sMaterial RequirementsPlanning, 2nd ed.NewYork:McGraw-

Hill.Ptak,C.,&Smith,C.(2011).Orlicky’sMaterialRequirementsPlanning,3rded.NewYork:

McGraw-Hill.Smith, D., & Smith, C. (2013).Demand Driven Performance: Using Smart Metrics. New

York:McGraw-Hill.

Page 410: DDMRP - Demand Driven Material Requirements Planning

Index

activelysynchronizedreplenishment(ASR)leadtime.Seedecoupledleadtime.actualdemand,23,133,149,237ADU.Seeaveragedailyusage.ADUalert,107ADUalerthorizon,107ADUalertthreshold,107ADUexceptions,109ADU-basedrecalculation,105advancedplanningandscheduling(APS/APO),209allergensequence,246analyticsview,255ASRleadtime.Seedecoupledleadtime.ASRLT.Seedecoupledleadtime.assetsvs.liabilities,93averagedailyusage(ADU),105averageinventoryrange,178averageon-handposition,176,181averageopensupplyorders,181

batches,33batchingpolicies,33bidirectionalbenefittest,231billofmaterial,28,45,60,182,239,250bimodaldistribution,9,253,265blendedADU,107BOMs,28breakthroughS&OP,273Brown,F.Donaldson,18buffer,39,93,96bufferlevels,111bufferpenetration,149bufferprofile,97,293

Page 411: DDMRP - Demand Driven Material Requirements Planning

buffersimulation,305bufferstatusalerts,207bufferzone,96bullwhipeffect,18,32,35businessplan,268

calculatingaveragedailyusage(ADU),109calculatingbuffers,111campaignscheduling,198capacity,142,238,246,283CDDL,52CDDP,51certifieddemanddrivenleader.SeeCDDL.certifieddemanddrivenplanner.SeeCDDP.challengingtraditionalS&OP,272coefficientofvariation,80,102commonusageofcomponents,58common-causevariation,30components,145conflictwithleanorpull,45continuousimprovement,263controlpoints,247conventionalplanning,22convergentpoint,62,247coverageoptimization,195criticaloperationprotection,60criticalpositioningfactors,57cumulativeleadtime,61,68currenton-handalert,214customertolerancetime,58

DDAS.Seedemanddrivenadaptivesystem.DDMRP,267DDMRPandretail,315DDOM.Seedemanddrivenoperatingmodel.DDS&OPchecklist,333decoupledexplosion,182decoupledleadtime,69decoupledschedule,240decoupling,34,37,41,45decouplingpointbuffers,39,231decouplingpointintegrity,253decouplingpoints,37,96delayaccumulation,30deletion,135

Page 412: DDMRP - Demand Driven Material Requirements Planning

demand,149,237,278demandadjustmentfactor,131demanddrivenadaptivesystem(DDAS),267demanddrivenmaterialrequirementsplanning(DDMRP),50,294demanddrivenmodelprojections,269,280demanddrivenoperatingmodel,53,268demanddrivenperformanceusingsmartmetrics,34,51,54,286demanddrivensalesandoperationsplanning,53,269demanddrivenvarianceanalysis,269demandmanipulation,286demandsignal,23demandsignaldistortion,32,37,57,209,231,234demandtrendsensing,133demanduncertainty,59demandvariability,80,101demand-drivenmanufacturing,51demand-drivenMRP,50,294demand-drivenplanning,149Deming,W.Edwards,17dependence,47dependentdemand,22discountoptimization,191dispatchingsystem.Seeprioritycontroldisposition.distributioninventory,103distributionnetworks,79,216distributionpositioning,79divergentpoints,62DLT.Seedecoupledleadtime.DRP(distributionrequirementsplanning),52drum,317drumschedule,246duedate,208dynamicadjustmenttest,232dynamicbuffers,294

Einstein,Albert,47enterpriseresourceplanning(ERP),291excessinventory,260execution,207,294explosionofrequirements,183

finiteschedule,245five-stepprocess(Ling/Coldrickmodel),274flatteningthebillofmaterial,28,35flow,16

Page 413: DDMRP - Demand Driven Material Requirements Planning

flowindex,49,262Ford,Henry,18forecastaccuracy,25,278forecasts,23forwardADU,107foursourcesofvariation,59freightoptimization,195frequencyupupdate,107goal,17,246Goldratt,Dr.Eliyahu,17,246greenzone,97,112greenzoneadjustment,146

highlyvisiblepriority,207hub,84hybridmodel,90,188hypersensitivity.Seesystemnervousness.

independence,47integratedreconciliation,275introductionadjustment,133inventory,93inventory,assetsvs.liabilities,93inventoryleverage,60inventorypositioning,71inventory,bimodal,9ISCEA,51itemtype,98

JIT,24

kanbans,45leadtime,99,109,237leadtimeadjustmentfactor,100,112,148leadtimealert,226leadtimealertzone,226leadtimecompression,96lead-time-managed(LTM)part,227lean,45,50,52lengthofperiod,105Ling,Richard(Dick),288Little’sLaw,16,317loadmanipulation,285location,110LTMpart,236

Page 414: DDMRP - Demand Driven Material Requirements Planning

managingtheportfolioandnewactivities,276manualadjustments,133manufacturingleadtime(MLT),60,67marketintelligence,268marketpotentialleadtime,58masterproductionschedule(MPS),236,272mastersettings,267,271,280materialrequirementsplanning.SeeMRP.materialsynchronizationalert,222matrixbillofmaterial,73metricsandanalytics,251minimumorderquantity(MOQ),110,112,266min-max(MM)parts,96,125,198MLT.Seemanufacturingleadtime.modelandpartparameters,270MRP,3,21,35,45,50,52,57,160,297MRPnervousness,27MRPsystemrequirements,297multi-hub,89multilevelmasterschedule,184multipleplants,79

nervousness.Seesystemnervousness.netflowequation,149netflowposition,158,190newnormal,12,18,27,292

occurrence-basedrecalculation,127Ohno,Taiichi,17oldS&OP,272on-handalertlevel,214on-handbalance,150on-orderstock,158openorders,150ordercycle,112orderfrequencyvariance,261orderindependencetest,232orderspikehorizon,152orderspikethreshold,151order-pointsystems,235Orlicky,Joseph,291Orlicky’sMRP3rdedition,51OSH.Seeorderspikehorizon.OSH1.Seeorderspikehorizon.OSH2.Seeorderspikehorizon,156

Page 415: DDMRP - Demand Driven Material Requirements Planning

OST.Seeorderspikethreshhold.OTOG.Seeovertipofgreen.outlyingeventsreport,257overtopofgreen(OTOG),161,194overflattening,45oversimplification,47

PAF.Seeplannedadjustmentfactor.partattributes,105pastADU,107plannedadjustmentfactor(PAF),131planning,207planninghorizon,26,49,82,183,233,279planningprojections,280planningvs.execution,49,212Plossl,George,15PLT.Seepurchasingleadtime.portfoliomanagement,276position,protect,pull,52positioning,293positioningfactors,57primaryplanningtest,232prioritizedshare,191prioritybyduedate,208prioritycontrol,243probabilitiesofsimultaneousavailability,71productdeletion,135productintroduction,133producttransition,136projectedbufferstatusalerts,218projectedon-handalert,218projectedorderfrequency,286purchasingleadtime(PLT),61

qualifiedactualdemand,150qualifiedorderspike,151

ramp-downadjustment,135ramp-upadjustment,133ramp-up/ramp-down,136rapidbufferadjustment,133recalculatedadjustments,127redzone,97,113redzoneadjustment,147redzonebase,113

Page 416: DDMRP - Demand Driven Material Requirements Planning

redzonesafety,114relativeprioritytest,232relativestandarddeviation.Seecoefficientofvariation.relevantinformation,18,23,251relevantmaterials,17,30,35,57,79,149,251,256,288relevantrange,267reorder-pointtechniques,235replenishedoverride(RO)parts,96,125replenishedpart,96replenishmentofstock,149requirementsexplosion,183retail,315returnonassets,5,16RMRP.Seeremanufacturingresourceplanning(RMRP).roughlyrightvs.preciselywrong,274routing,17,60,240runchart,175

S&OP,269safetystock,184,233salesandoperationsplanning.SeeS&OP.salesordervisibilityhorizon,58seasonalityadjustment,140shockabsorber,96shopfloorcontrolsystem,240signalintegrity,251significantminimumorderquantity,113SixSigma,52softwarecompliance,293Souder’sLaw,291spaceplanning,282spikethreshold,152stockout(SO),214stockoutwithdemand(SOWD),214stockoutwithdemandalert,214stopexplosion,184strategicbusinessmanagement,274strategichorizon,268strategicinventorypositioning,57,93strategicrelevantrange,268,274strategicallyreplenishedbuffers,93,231supply,279supplychains,293supplycontinuityvariability,33,55,57,181,209,222,231,234supplygeneration,96,158

Page 417: DDMRP - Demand Driven Material Requirements Planning

Supplyordergeneration,49supplyvariability,59supply-chainbullwhipeffect,59,102synchronizationalerts,222systemnervousness,27

tacticalhorizon,267Taguchifunction,10,95,255Taylor,Frederick,17technology,291TheoryofConstraints(TOC),52TOG.Seetopofgreen.topofgreen(TOG),115topofred(TOR),115topofyellow(TOY),115TOR.SeetopofredTOY.SeetopofyellowToyotaProductionSystem,17,50traditionalsalesandoperationsplanning,272transition,136

understated,139

variability,37,59,101variabilitycategory,104variabilityfactor,104variablerateofdemand,59varianceanalysis,269,271velocity,261

weeklybucket,28whereused,72wholesalenetwork,85WIPpriority,247work-arounds,7workingcapital,282work-in-process,247

yellowzone,97,113yellowzoneadjustment,147

zoneadjustmentfactor,146


Recommended