Date post: | 05-Dec-2014 |
Category: |
Documents |
Upload: | vijaybijaj |
View: | 737 times |
Download: | 1 times |
© Nathan Soderborg, 2008. All rights reserved.
Implementing Structures and Processes for
Lean Six Sigmain Product Development
Dr. Nathan SoderborgDesign for Six Sigma Master Black Belt
North America Product Development
Ford Motor Company
WCBF Global Lean, Six Sigma and Business Improvement Summit
Orlando, Florida
October 15, 2008
2© Nathan Soderborg, 2008. All rights reserved.
Purpose & Outline
Discuss application of Lean principles to Six Sigma in Product Development (DMAIC and DFSS)
� Align projects with customer value
� Increase project throughput
� Root out waste in decision making & scoping
� Generate more profound and lasting benefits
Topics
� Avoid pitfalls in assigning project value
� Institute “pull” systems for filling the project pipeline
� Error Proof—prevent defects before they are created
� Scope wisely—segment large projects into sub-projects
3© Nathan Soderborg, 2008. All rights reserved.
Lean and Six Sigma
� Lean
The elimination of waste with the goal that all steps in a process add value from the customer’s perspective
� Six Sigma
� A statistics-based, data-driven, problem solving methodology
� DMAIC: Define, Measure, Analyze, Improve, Control,
focuses on finding and fixing existing defects in products and processes
� DFSS is a product development (PD) approach that complements DMAIC, focusing on
� Innovation to satisfy customers and improve profitability
� Discovering and preventing defects before they occur in
products or processes
4© Nathan Soderborg, 2008. All rights reserved.
Lean Thinking (Womack & Jones)
� Correctly specify value so you are providing what the
customer actually wants
� Identify the value stream for each product family and
remove the wasted steps that don't create value but do
create muda (waste)
� Make the remaining value-creating steps flow
continuously to drastically shorten throughput times
� Allow the customer to pull value from your rapid-
response value streams as needed (rather than pushing
products toward the customer on the basis of forecasts)
� Never relax until you reach perfection, which is the
delivery of pure value instantaneously with zero muda
5© Nathan Soderborg, 2008. All rights reserved.
Project Value
� Basic product of Six Sigma is a project
� Lean Principle: “Correctly specify value so [projects]
provide what the customer actually wants”
� What does the customer want? Typical answer:
� Defect free products
� Ability to perform the intended function or use
� Performance better than competition in attributes that matter
� Features that satisfy, even enthuse or excite
� Answer focused on value:
� Experience and characteristics that relieve cost from the
customer
� Performance and features for which the customer is willing to pay
6© Nathan Soderborg, 2008. All rights reserved.
Assigning Project Value
� A project is worth doing if
� It saves the end customer money
� Directly: reduced repair bills, service costs, etc.
� Indirectly: time & effort, product marketing/warranty costs that get passed back
� It leads to performance or features for which the customer will pay
(extra)
� Set a standard prioritization scheme and get on with it
� If possible, use existing accounting techniques for projects that cut
costs or will generate revenue
� For projects that prevent future costs
� Assign relative weights to different customer issues based on experience
� Work on issues with high weights and high likelihood of occurrence
� Avoid the waste of instituting complex project value calculations
and involving multiple layers of the finance department
Debating value does not add value
7© Nathan Soderborg, 2008. All rights reserved.
Pull Systems
� Lean Principle: “Allow customer to pull value
from rapid-response value streams as needed”
� “Pull” in lean production means to produce or process
an item only when the customer needs and requests it
� Lean manufacturers design their operations to respond
to the ever-changing requirements of customers
� Such operations avoid the traditional batch-and-queue
system many manufacturers must rely on
� Pull systems
� Should be convenient and easy to use
� React to needs—don't anticipate them
8© Nathan Soderborg, 2008. All rights reserved.
Example: Pull in a Manufacturing Process
� A light bulb is set up on a pole at an assembly line; when the light goes on, it is the signal (kanban) to
the producing station to wheel over a cart of
components to the assembly line
� A full cart is dropped off at the assembly line and an
empty cart is wheeled back to the producing station—the empty cart is the signal that authorizes
the producers to make more parts
� Finished components from the last operation at the
producing station are placed directly on the cart—if there is no cart there is no production
� The process eliminates double handling
� Parts are placed on the cart as produced
� Parts are taken off the cart during the first operation at the assembly line and put directly into an assembly
9© Nathan Soderborg, 2008. All rights reserved.
Filling the DMAIC Project Pipeline
For projects related to Defect Elimination,
pull comes from…
� Repair/Service Data
� In plant repairs (surrogate for customers)
� Warranty claims (sellers & customers)
� Survey Data
� Customer satisfaction
� Complaints: “Things Gone Wrong”
What are customers
complaining about
right now?
Where are customers
experiencing problems right now?
Prioritize on data from first few weeks or months of ownership—often correlates well to data from higher time in service
10© Nathan Soderborg, 2008. All rights reserved.
“When ____ arrived six months ago, he found ____'s quality
operations bogged down with drawn-out decision making and
interdepartmental finger-pointing.
“If there was a problem with an air conditioning system, for example, the
engineering department might suggest the plant didn't put oil in the unit,
while the plant might say the unit was improperly designed. Weeks of back-
and-forth e-mails would ensue, "and meanwhile, the customer is out there
saying, 'I'm hot,'‚" ____ said.
“____ scrapped the old system, created standard definitions of quality, and
established the view that customer satisfaction starts with a potential
customer's perception of a brand and continues though vehicle ownership and
repurchase.
“He launched dedicated interdepartmental teams to address problems in
minutes over conference tables, not weeks over e-mail.”
Detroit News, April 14, 2008
Recently In the News
11© Nathan Soderborg, 2008. All rights reserved.
Customer
brings vehicle
to dealer for
service
Customer Pull
Example
� “Every warranty claim received by a dealer is sent to
the plant where the vehicle is built and the issue is
‘mapped back’ to the work station where it might have
originated.”Ford’s Drive One campaign, answers for employees to FAQs related to Quality
Dealer fixes
issues and
records
information
Information is
sent to plant
immediately for
review
Plant groups
issues and
shares with
engineering
Issues with
highest
frequency are
next up projects
Project Portfolio
12© Nathan Soderborg, 2008. All rights reserved.
Reaction to Field Concerns is just the Beginning
� Even better than reacting quickly when (1) customers
discover failures, is to react quickly to (2) failures
discovered in development
� Find them before they escape
(ideally as soon after creation as feasible)
� Immediately mitigate their effect
(on customer and the organization)
� Even better than reacting quickly to failures discovered
in development is to (3) prevent creation of the defect
that leads to failure in the first place
� Because no process or person is perfect, we need to
plan for and address all three scenarios
13© Nathan Soderborg, 2008. All rights reserved.
Lessons from High Reliability Organizations
� “HROs” exhibit
� Preoccupation with failure
� Reluctance to simplify interpretations
� Sensitivity to operations
� Commitment to resilience
� Deference to expertise
� “…high reliability organizations are preoccupied with small,
emerging, early failures aka problems (failures in the sense of things
not working out exactly as expected). They see those small failures
as clues that the system is not as healthy as they thought it was.
Those early small failures are also easier to deal with than are full blown failures. And HROs spend a great deal of time and effort to
catch stuff while it is still small.”“Revisiting Mindfulness, Managing the Unexpected, and the Cerro Grande Staff Ride”http://www.myfirecommunity.net/documents/Santa_Fe_post_conference_reflections_kms.pdf
Karl Weick & Kathleen SutcliffeManaging the Unexpected: Assuring High Performance
In an Age of Complexity, John Wiley & Sons, 2001
14© Nathan Soderborg, 2008. All rights reserved.
Defect Creation
� Failures result from product or process defects (Defect: “an imperfection that impairs worth or utility;” for our
purposes, can include gaps to competition or company targets)
� Defects are not created at the moment the product is
manufactured or the process is instituted—they are created
much earlier, during development
� For example
� Wrong identification of requirements important to customer, wrong targets
� Incorrect assessment of environmental, usage, manufacturing conditions
� Selection of an inadequate or non-robust design concept
� Incorrect characterization of the system, e.g., use of an inaccurate model
� Improper optimization assumptions or methods
� Selection of inadequate materials, geometry, interfaces
� Mistakes in execution of design or build
� Using detection events that can’t detect defects, ETC.
15© Nathan Soderborg, 2008. All rights reserved.
Example
The I-35W Mississippi River bridge catastrophically failed during the
evening rush hour on August 1,
2007, collapsing to the river and riverbanks beneath. Thirteen people
were killed and approximately one hundred more were injured.
The defect that led to the collapse was not created
when the bridge was built.
When was it created?
16© Nathan Soderborg, 2008. All rights reserved.
“Although the Board's investigation is still on-
going and no determination of probable cause has
been reached, interim findings in the investigation
have revealed a safety issue that warrants
attention," said NTSB Chairman Mark V.
Rosenker. …
“This review discovered that the original design
process of the I-35W bridge led to a serious error in
sizing some of the gusset plates in the main truss.
“Undersized gusset plates were found at 8 of the
112 nodes (joints) on the main trusses of the bridge.
These 16 gusset plates (2 at each node) were
roughly half the thickness required and too thin to
provide the margin of safety expected in a properly
designed bridge.”
NTSB URGES BRIDGE
OWNERS TO PERFORM
LOAD CAPACITY
CALCULATIONS BEFORE
MODIFICATIONS; I-35W
INVESTIGATION
CONTINUES; Jan 15, 2008, http://www.ntsb.gov/Pressrel/2008/080115.html
http://minnesota.publicradio.org/display/
web/2008/06/03/astaneh_bridgecollapse
17© Nathan Soderborg, 2008. All rights reserved.
� Apparently the defect was created in two parts: first, in the analysis and decision making process during design; second,
with the later modification to the bridge
� Nevertheless, the defect could have been detected prior to the
collapse by engineering analysis and calculation
“The… bridge was brought down by too much weight
from construction materials and pavement added to the
roadway years before, a structural engineering expert
reports in a paper delivered today… Hassan Astaneh's
research concludes MnDOT and the consulting firm it
hired could have prevented the collapse.”…
“In January the National Transportation Safety Board reported the bridge's
gusset plate at node U10 was undersized. Last August, Astaneh's research
found the same thing. But Astaneh says despite this design flaw, the plate
would have held up if it weren't for the extra weight.”--“Study: Heavy construction materials, added pavement brought down 35W bridge,” by Sea Stachura,
Minnesota Public Radio, June 4, 2008, http://minnesota.publicradio.org/display/web/2008/06/03/astaneh_bridgecollapse/
18© Nathan Soderborg, 2008. All rights reserved.
Waste Due to Defects
� A defect created at a certain point in the development
process cannot be fixed without revisiting that point in
the process
� Work completed after the defect is created is subject to
be re-done or at least re-evaluated: WASTE
� The further downstream the defect is
detected, the more re-work is required
to mitigate it
� When the customer detects the defect, there is not only
re-work needed to fix it, but potentially severe
implications for corporate image, customer satisfaction,
sales, profits, etc.
19© Nathan Soderborg, 2008. All rights reserved.
Filling the DFSS Project Pipeline
For projects related to Defect Detection
& Mitigation, pull comes from…
� Process Health Indicators
� Launch Concerns
� First time through and re-work metrics
� Number of part changes during development
� Results of Detection Events
� Checks against Design Standards
� Computer Simulations
� Physical Tests
Where are resources
being spent now?
Where are the
emerging problems?
(Things we know we know)
20© Nathan Soderborg, 2008. All rights reserved.
Filling the DFSS Project Pipeline
For projects related to Defect Prevention
& Business Opportunity, pull comes from…
� History
� Sustained high warranty costs
� Chronic gaps to competition in
survey results (e.g., JD Power)
� Launch concern trends
� Part change trends
� Design Review Discoveries
� Defect Anticipation (e.g., using structured tools
such as FMEA, FTA)
� Risk Analysis
What area or discipline
have we not yet mastered?
What concerns are
revealed after thorough
review?
(Things we know we don’t know)
(Things we don’t know we don’t know)
“Pull systems react to needs, don't anticipate them.”
If capability won’t meet demand, the defect already
exists. Identify it and react.
Identification
21© Nathan Soderborg, 2008. All rights reserved.
Example
Toyota’s GD3 and Design Review
Based on Failure Mode (DRBFM)
� Good Design
“A fundamental principle for reliability is to avoid changing the conditions of
a good design.”
� Good Discussion/Good Design Review
“We implemented Good Design Reviews in the development process as a creative “breakthrough”… This breakthrough
emphasizes discovery of undetected problems caused by intentional or incidental changes. Good Design Review is a
process of thoroughly discussing design plans to discover
undetected problems (Good Discussion), and of formulating countermeasures to solve those problems one by one.”
Hirokazu Shimizu, Toshiyuki
Imagawa, Hiroshi NoguchiReliability Problem Prevention Method for Automotive Components, SAE 2003-01-2877
22© Nathan Soderborg, 2008. All rights reserved.
Example
� Design Review Based on Failure Mode
� “Pay close attention to intentional and incidental changes in new development items.”
� “Promote discussions based on FMEA and FTA results.”
Reliability Problem Prevention Method for Automotive Components—Development of GD3 Activity
and DRBFM (Design Review Based on Failure Mode), JSAE 20037158, SAE 2003-01-2877
23© Nathan Soderborg, 2008. All rights reserved.
Sources of Project Pull: Summary
� Chronic Repair/Service Issues
� Chronic Survey Gaps
� Design Review Discoveries
� Defect Detection Events
� Process Health Indicators
� Current Survey Data
� Current Repair/Service Data
From the Lean perspective: Proactive is Reactive
Reacting to the right things is how to be proactive
DMAIC
DFSS
Proactive
Reactive
24© Nathan Soderborg, 2008. All rights reserved.
Product Development Error Proofing
� Projects react to the identification of defects
� The final step in any project must be implementing
permanent corrective actions, determining…
� How, in the future, we will operate differently and in a standardized way
� What structures (controls) have been put in place to ensure this will continue to happen
� Examples of Design Error Proofing (SOPs)
� Established design guidelines (don’t reinvent good design
practices)
� Better methods for detecting defects earlier (including
simulation and component testing)
� More robust design concepts and solutions
Never close a project without institutionalizing the learning
25© Nathan Soderborg, 2008. All rights reserved.
Final Observations on Project Scoping
� Proper scope reduces waste & re-work
� Originally, we planned DFSS projects to take a system
through the entire PD process (months to years)
� Projects needed adjustment with movement of key people
� Hard to maintain focus over many months
� Changes in direction at higher levels negate previous work
� Subdivide big projects
� Multiple small projects, each focused on eliminating a single defect or few related defects
� Projects for each phase of the development process (e.g. Definition, Characterization, Optimization, Verification)—as
we know different types of defects arise in each phase
� Let the most current circumstances “pull” scope
26© Nathan Soderborg, 2008. All rights reserved.
Benefits of Focus
� Situation: A black belt is requested to split time
between two “equally important” DFSS projects
� Scenario 1: The black belt works on them
simultaneously and completes both in 6 months
� Scenario 2: The black belt focuses entirely on one
project for 3 months and completes it, then works on
and completes the second project in 3 months
Is one scenario better?
� Scenario 1: neither project is finished for 6 months;
average time to complete projects=6 months
� Scenario 2: one project is finished in 3 months; one in
six months; average time to completion=4.5 months
27© Nathan Soderborg, 2008. All rights reserved.
Example Deliverable Scopes for DFSS Projects
Institute technical training
Improve detection
measurement
system/gage RR
Add/intensify noise
content in a test or
simulation to better excite
failure modes
Develop/improve a
transfer function (model)
to discover failure modes
analytically
Create a new detection
event or standard
Develop/improve a
customer-correlated
metric, target
Replace a test or
simulation event with a
design standard
Replace a hardware test
by a virtual simulation
Move a detection event
from a system to
subsystem or component
level
Improve design to reduce
severity of a failure mode
Institute updated
procedures and error-
proofing to prevent
mistakes
Institute generic robust
design guidelines
Institute product-specific
design improvements,
e.g., optimize for
robustness
Implement a new, robust
concept
Increase Detection Capability
Move Detection Capability Earlier
Make Designs More Reliable/Robust
28© Nathan Soderborg, 2008. All rights reserved.
REFERENCES
� Tim P. Davis, “Science, engineering, and statistics,” Applied Stochastic Models in Business and Industry, Vol. 22, Issue 5-6,
pp401-430, 2006.
� James M. Morgan and Jeffrey K. Liker, The Toyota Product
Development System: Integrating People, Process and Technology, Productivity Press, 2006.
� James P. Womack and Daniel T. Jones, Lean Thinking: Banish Waste and Create Wealth in Your Corporation, Simon and Schuster, 2003.
Collapsed bridge photo from Wikipedia: This image or file is a work of a United States Coast Guard service personnel or
employee, taken or made during the course of that person's official duties. As a work of the U.S. federal government, the image or file is in the public domain (17 U.S.C. § 101 and § 105).
Standing bridge photo from Wikipedia: Permission is granted to copy, distribute and/or modify this document under the terms of
the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation