+ All Categories
Home > Technology > Understanding Reference Data with Aaron Zornes

Understanding Reference Data with Aaron Zornes

Date post: 14-Jan-2015
Category:
Upload: orchestra-networks
View: 2,863 times
Download: 5 times
Share this document with a friend
Description:
There’s growing recognition in the analyst community that reference data is a form of master data that requires its own governance. Locations, currency codes, financial accounts, and organizational hierarchies are so widely used in an organization that mismatches can result in: reconciliation issues, poor quality analytics or even transactional failures. While it’s easy to see how poor reference data management (RDM) can cause problems, many companies struggle with determining how to get started. Multiple questions arise: What’s the scope? How should one choose between RDM solutions? How do I compute ROI? To answer these questions and more, Orchestra Networks teamed up with Aaron Zornes, Chief Research Office of the MDM Institute and Godfather of MDM, for: Everything you ever wanted to know about Reference Data (but were afraid to ask). In this hour long webcast featuring Aaron Zornes (MDM Institute) and Conrad Chuang (Orchestra Networks) you will learn the: Characteristics of reference data, Key features of a reference data management (RDM) solution, Lessons learned RDM implementations, and more
Popular Tags:
33
Understanding Reference Data Management Aaron Zornes Chief Research Officer The MDM Institute Conrad Chuang Sr. Product Marketing Manager Orchestra Networks
Transcript
Page 1: Understanding Reference Data with Aaron Zornes

Understanding Reference Data Management

Aaron Zornes Chief Research Officer The MDM Institute Conrad Chuang Sr. Product Marketing Manager Orchestra Networks

Page 2: Understanding Reference Data with Aaron Zornes

Today’s Agenda

Part I: Reference Data Management Overview What is reference data? What is Reference Data Management (RDM)? Key requirements for RDM solutions Costs, savings & ROI scenarios

Part II: RDM Implementations

Q&A

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 3: Understanding Reference Data with Aaron Zornes

Founded in 2004 to focus on MDM business drivers & technology challenges

MDM Institute Advisory Council™ of 150 Global 5000 IT organizations with unlimited advice to key individuals, e.g. CTOs, CIOs, data architects

MDM Institute Business Council™ website access & email support to 35,000+ members

MDM Road Map & Milestones™ annual strategic planning assumptions

MDM Alert™ newsletter

MDM Market Pulse™ market research & multi-client studies

MDM Fast Track™ one-day public & onsite workshop rotating quarterly through major North American, European, & Asia-Pacific metro areas

MDM & Data Governance Summit™ annual conferences in London, NYC, San Francisco, Shanghai, Singapore, Sydney, Tokyo & Toronto

© 2012 The MDM Institute www.The-MDM-Institute.com

About the MDM Institute

“Independent, Authoritative, & Relevant”

About Aaron Zornes Most quoted industry analyst authority on topics of MDM, RDM & MDG

Founder & Chief Research Officer of the MDM Institute Founder & conference chairman for MDM & Data Governance Summits series

Founded & ran META Group’s largest research practice for 14 years M.S. in Management Information Systems from University of Arizona

Page 4: Understanding Reference Data with Aaron Zornes

What is Reference Data?

Reference data =“coded, semantically stable, relatively static data sets shared by multiple constituencies”

(people, systems, & other master data domains)

Customers

Product

Industry

Sales Person

Geo

Cost / Revenue

Acct

Business Unit ID

© 2012 The MDM Institute www.The-MDM-Institute.com

In the logical view, private & public forms of reference data connect domains & application; consistent values

(& semantics) required for multi-domain views & hierarchies

Page 5: Understanding Reference Data with Aaron Zornes

Errors in reference data will ripple outwards affecting quality of master data in each domain, which in turn affects quality in all dependent transactional systems

RDM needed in both operational & analytical MDM use cases where capability often used to provide attributes, hierarchies & KPIs

© 2012 The MDM Institute www.The-MDM-Institute.com

Why Reference Data? Why Now?

Central role of reference data means RDM becoming “starting point” for many organizations planning MDM & MDG

Systemic Failure

Inconsistent Reporting

Transaction Failure

Regulatory Non-compliance

Page 6: Understanding Reference Data with Aaron Zornes

RDM Prologue

In addition to MDM functionality, RDM systems also manage complex mappings btw different reference data representations & different data domains across enterprise

Governance of RDM is vital— manual or custom RDM often lacks change management, audit controls & granular security/permissions

Because reference data is used to drive key business processes & application logic, errors in reference data can have major negative & multiplicative business impact

© 2012 The MDM Institute www.the-MDM-Institute.com

Just as businesses no longer build own CRM, ERP, &MDM systems, so too are organizations beginning to acquire

commercial RDM, which can be easily tailored or configured & have full ongoing support of major software vendor

Page 7: Understanding Reference Data with Aaron Zornes

Reference Data Categories

Multi-Domain RDM Use Cases

Real-Time / Transactional RDM Use Cases

Public (External)

Countries & Subdivisions (FIPS10) Currencies (ISO 4217) Time Zones (ISO 8601)

Industry Classification (NAICS, ISIC) Security Prices

SWIFT BIC Codes (Payments) ICD-9/10 Codes (Healthcare)

ACORD/ISO Codes (Insurance)

Private (Internal)

Legal Entities Chart of Accounts

Organizations Employees

(i.e., much of HR & Finance Data)

Reference data required for transaction processing

© 2012 The MDM Institute www.The-MDM-Institute.com

Semi-Private? (Shared)

Customized Public Reference Standards (e.g. customized D&B)

Shared Private Data (Finance)

Page 8: Understanding Reference Data with Aaron Zornes

Why Manage Reference Data Independently? (“Hub of hubs”? Federated vs. Centralized?)

Customers

Product

Industry (NAICS, ISIC)

Sales Person

Geo (ISO3166,

FIPS)

Cost / Revenue

Acct

Business Unit ID

© 2012 The MDM Institute www.The-MDM-Institute.com

Geo

Geo

In the logical model, reference data connects domains & applications; in implementations local copies exist for each consumer; challenges include: governance, synchronizing,

versioning, & custom hierarchies/internationalizations

ERP

Finance HR BI/Analytics

Geo Geo Geo

Geo

Page 9: Understanding Reference Data with Aaron Zornes

Critique of Current Approaches for Multi-Domain Reference Data

RDM Solution Drawback Recommendation

Custom-built, manual solutions

Heavy TCO burden Avoid unless reference data demands are truly unique

Spreadsheets Difficult to govern, secure, version, & audit; no modeling, poor hierarchy management

Distribute data in spreadsheets; govern data in RDM solution

Repurpose hierarchy management solution (MSFT MDS, ORCL DRM)

Poor cross-domain support, no classification mapping, few enterprise integration options

Seek out multi-domain RDM solution with hierarchy management

Customize existing domain-specific MDM (Customer or Product)

Rudimentary data modeling, lifecycle mgmt capabilities, & governance features (esp. authoring & workflow)

Use multi-domain RDM solution to maintain connections & govern/update into CDI & PIM via data services

ERP / Enterprise Application

Limited governance, versioning, distribution; also reference data customized use in app may have limited appeal in other systems

Master in external platform. RDM can be used to govern baseline set, versions and adaptations

Real-time / industry-specific RDM

Premium priced R/T RDM solutions do not represent good economic sense

Leverage R/T RDM solutions for R/T use cases (trading, claims processing, payments)

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 10: Understanding Reference Data with Aaron Zornes

“Top 10” RDM Technical Evaluation Criteria

1. Administration of diverse reference data types 2. Ability to map reference data 3. Management of reference data sets 4. Architecture/performance 5. Hierarchy management over

reference data sets 6. Connectivity/integration 7. Import & export 8. Versioning support 9. Security & access control 10. E2E lifecycle management

© 2012 The MDM Institute www.The-MDM-Institute.com

Coming to market are RDM solutions characterized by multiple, diverse levels of integration w/ market-dominant MDM hubs as well as repackagings of existing mid-market

MDM solutions – HOW TO EVALUATE?

Page 11: Understanding Reference Data with Aaron Zornes

“Administration of Diverse Reference Data Types”

© 2012 The MDM Institute www.The-MDM-Institute.com

From R. Thompson,/Credit Suisse, “Multidomain Enterprise Reference Data,” 7th Annual MDM & Data Governance Summit New York 2012

Private Ref Data

Public Ref Data

RDM solution should support a w ide mix of data structures from name:value pairs to hierarchies (see criteria #5).

RDM Top 10 Eval Criteria #1

Page 12: Understanding Reference Data with Aaron Zornes

FINANCE LOCATIONS

“Ability to Map Reference Data” – pt. 1 (cross-domain mapping)

Issuing Country (ISO3166)

Name ISO 4217 Code

USA US Dollar USD CHN Yuan Renminbi CNY JPN Japanese Yen JPY

Official Currency

ISO3166

USD ASM USD IOT USD ECU USD SLV USD GUM USD HTI USD MHL USD FSM USD MNP USD PLW USD PAN USD PRI USD TLS USD TCA USD USA USD VIR CNY CHN JPY JPN

ISO 3166 Code

Name

USA United States of America CHN People’s Republic of China JPN Japan ASM American Samoa IOT British India Ocean Terr. ECU Ecuador SLV El Salvador GUM Guam HTI Haiti MHL Marshall Islands FSM Micronesia MNP Northern Mariana Islands PLW Palau PAN Panama PRI Puerto Rico TLS East Timor TCA Turks and Caicos Islands VIR Virgin Islands

RDM solutions need to preserve values & mappings between reference data sets – both in domain and across domains.

RDM Top 10 Eval Criteria #2

© 2012 The MDM Institute www.The-MDM-Institute.com

LOC

ATI

ON

& F

INA

NC

E

Page 13: Understanding Reference Data with Aaron Zornes

2012 VERSION 2007 VERSION

“Ability to Map Reference Data” – pt. 2 (temporal referential integrity)

2012 NAICS

Description

311224 Soybean and Other Oilseed Processing

221114 Solar Electric Power Generation

221115 Wind Electric Power Generation

221116 Geothermal Electric Power Generation

221117 Biomass Electric Power Generation

221118 Other Electric Power Generation

2007 NAICS

Description

311222 Soybean Processing

311223 Other Oilseed Processing

221119 Other Electric Power Generation - solar electric power generation

RDM solution needs to maintain links between versions, creating a migration path between versions of reference data. “Crosswalks” are

important for understanding how something changed.

MERGE

SPLIT

RDM Top 10 Eval Criteria #2

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 14: Understanding Reference Data with Aaron Zornes

RACI Tasks User

R Update sales hierarchies Rogers

R Change industry classifications Romanova

A Approve hierarchies and effective dates Stark

A Approve industry classifications Banner

A Approve merge into effective dated

Fury

“Mgmt of Reference Data Sets” (Governance workflows)

© 2012 The MDM Institute www.The-MDM-Institute.com

RDM Top 10 Eval Criteria #3

An RDM solution needs to support governance workflows; includes defining: responsible & accountable parties (including systems),

permissions & area of responsibility for each party (field, instance, container level), how parties interact/ tasks, & auditing/ history…

Sequence of interactions

Permissions

Responsibilities

Page 15: Understanding Reference Data with Aaron Zornes

“Hierarchy Management Over Reference Data Sets”

RDM solution should harness relationships between reference data sets & ex isting party or thing data to create hierarchies

SIC Codes

Customer & SIC Code Mapping

ICD-10 Codes

Active Ingredients & ICD10 Mapping

Active Ingredients & Product Mapping

Viewing customers by industry classification

Viewing drugs by Active Ingredient interactions and ICD10 Codes

RDM Top 10 Eval Criteria #5

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 16: Understanding Reference Data with Aaron Zornes

EMEA OPS DEU Cost Ctr

FRA Cost Ctr

APLA OPS

TUR Cost Ctr

JPN Cost Ctr

MEX Cost Ctr

NA OPS CAN Cost Ctr

USA Cost Ctr

“Versioning Support” (a.k.a. time travel)

EMEA OPS

TUR Cost Ctr

DEU Cost Ctr

FRA Cost Ctr

APLA OPS JPN Cost Ctr

MEX Cost Ctr

NA OPS CAN Cost Ctr

USA Cost Ctr

Cost Centers (as-of 2012 Q2)

EMEA OPS DEU Cost Ctr

FRA Cost Ctr

AP OPS TUR Cost Ctr

JPN Cost Ctr

CALA OPS MEX Cost Ctr

NA OPS CAN Cost Ctr

USA Cost Ctr

RDM solution needs versioning & “as of” / effective dating to support recall of reference data values, relationships or hierarchies.

(versioning has *major* implications for analytics/ BI!)

Cost Centers (Current)

Cost Centers (Effective 2013 Q1)

RDM Top 10 Eval Criteria #8

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 17: Understanding Reference Data with Aaron Zornes

Reference Data Management Strategic Planning Assumption

During 2012-13, reference data will emerge as a key entry point for enterprises & in turn influence choice of MDM for Customer, Product & other domains

Concurrently, every MDM vendor will rush to market RDM solutions to apply MDM approach for centralized governance, stewardship & control

By 2013-14, large enterprises will also mandate that Reference Data be part of MDM platform native entities

By 2015, RDM will be commoditized via the efforts of MSFT & ORCL especially

MDM MILESTONE

Managing “simple” reference data will prove to be a key sales entry point for MDM vendors

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 18: Understanding Reference Data with Aaron Zornes

Competition for Multi-Domain RDM

Custom-built, manual solutions Hierarchy management system adaptations

Do not readily support publish-subscribe, classification mapping, etc.

Custom MDM domain type Lack of data modeling flexibility, rudimentary lifecycle

management capabilities & limited data governance features, esp. authoring & workflow

Multi-domain RDM RDBMS vs. semantic/OODBMS

Purpose-built or industry-specific RDM Premium priced real-time RDM solutions do not represent good

economic sense

© 2012 The MDM Institute www.The-MDM-Institute.com

Seek out multi-domain RDM solution providers that understand & have experience addressing complex ity of reference data

Page 19: Understanding Reference Data with Aaron Zornes

“Top 10” RDM Technical Evaluation Criteria

1. Administration of diverse reference data types 2. Ability to map reference data 3. Management of reference data sets 4. Architecture/performance 5. Hierarchy management over

reference data sets 6. Connectivity/integration 7. Import & export 8. Versioning support 9. Security & access control 10. E2E lifecycle management

© 2012 The MDM Institute www.The-MDM-Institute.com

Re-Cap

Page 20: Understanding Reference Data with Aaron Zornes

© 2012 The MDM Institute www.The-MDM-Institute.com

MDM Institute Field Reports – RDM

Aprimo LRDM (Teradata)

DataFlux qMDM

IBM MDM RDM Hub Informatica RDM

Kalido

ASG ROCHADE (Metadata-driven RDM)

Microsoft RDM (to be announced)

Orchestra EBX5

Profisee

SAP MDG-R

Oracle Hyperion DRM

Software AG WebMethods OneData

** General-purpose or multi-domain RDM, not industry-specific or real-time RDM solutions such as capital markets, pharma, e.g., AIM, Asset Control, Eagle, Golden Source, Kingland Systems 360 Data, &RSD

Page 21: Understanding Reference Data with Aaron Zornes

MDM Institute’s Field Reports on RDM

© 2012 The MDM Institute www.The-MDM-Institute.com

Page 22: Understanding Reference Data with Aaron Zornes

Field Report: Orchestra Networks EBX5 for RDM

Strengths Robust solution for centralized DG,

mgmt, stewardship, & distribution of enterprise reference data

Enterprise-scalable RDM1 Strong taxonomy support &

mappings Model-driven ease of deployment,

implementation, & use (built-in process flows + semantic database underpinning)

Support for temporal reference data

Cloud-based, SaaS option

Caveats Nascent North

American market presence

Shortage of EBX-knowledgable consultancies

Vulnerability in rapidly evolving market crowded with mega vendors & other nouveau MDM vendors

Under invested in marketing © 2012 The MDM Institute www.The-MDM-Institute.com

1 – BNP Paribas, Crédit Suisse, Michelin, …

Page 23: Understanding Reference Data with Aaron Zornes
Page 24: Understanding Reference Data with Aaron Zornes

Technip: MDM / RDM essential to delivering multi-billion € oil & gas projects

• Projects require coordination across multiple company and functional areas – Up to 16 Technip companies can be involved for one project

• Data coherence, sharing and timely availability are key success factors

• Private and Public reference data

Page 25: Understanding Reference Data with Aaron Zornes

Implementation: Hub and Registry

Page 26: Understanding Reference Data with Aaron Zornes

Adaptation / Customization essential to supporting downstream applications

Parent

Same structure, Different values

Child inherits structure, but not labels. Good where same hierarchy is used globally and only labels are changed

Different structure, Same

values

Child inherits values, but not structure. Good when hierarchy is customized to fit functional area.

Different structure, Different values

Child partially inherits structure and values. Good where hierarchy and labels change overseas, such as a foreign subsidiary with a different product hierarchy

Page 27: Understanding Reference Data with Aaron Zornes

Benefits realized in every functional area

Page 28: Understanding Reference Data with Aaron Zornes

Bottom Line

RDM is more than “reference tables”– i.e., also complex mappings (logical & physical) between different representations, data domains, versions & hierarchies

RDM impedance mismatch = inconsistent reporting, regulatory noncompliance, transaction failures & systemic failures

Central role of reference data means RDM can be expected to become “starting point” for many organizations planning MDM & MDG

Majority of RDM solutions do not address notion of "temporal" reference data or provide governance

Market misconception/dogma that “RDM *must* be in same stack as multi-domain MDM”

Buy, *don’t* build, RDM © 2012 The MDM Institute www.The-MDM-Institute.com

Page 29: Understanding Reference Data with Aaron Zornes

Aaron Zornes Chief Research Officer The MDM Institute [email protected] www.linkedin.com/in/aaronzornes @azornes

Conrad Chuang Sr. Product Marketing Manager Orchestra Networks [email protected] www.orchestranetworks.com/rdm @onmdm

© 2012 The MDM Institute www.The-MDM-Institute.com

Q&A

Page 30: Understanding Reference Data with Aaron Zornes

© 2010 The MDM Institute www.The-MDM-Institute.com

Page 31: Understanding Reference Data with Aaron Zornes

© 2010 The MDM Institute www.The-MDM-Institute.com

Page 32: Understanding Reference Data with Aaron Zornes

MDM & Data Governance Summit™ Conference Series

© 2012 The MDM Institute www.The-MDM-Institute.com

“More MDM programs get their successful start at MDM & Data Governance Summits than anywhere else”

MDM & Data Governance Summit Singapore Marina Bay Sands Resort ▪ December 4-5

MDM & Data Governance Summit Shanghai Shanghai International Convention Center ▪ March 2013

MDM & Data Governance Summit Europe Radisson BLU – London ▪ April 15-17, 2013

MDM & Data Governance Summit Asia-Pacific Four Points Darling Harbour– Sydney ▪ May 20-21, 2013 MDM & Data Governance Summit San Francisco

Hyatt Embarcadero – San Francisco ▪ May 2013 MDM & Data Governance Tokyo

Belle Salle Kanda– Tokyo ▪ June 14, 2013 MDM & Data Governance Summit Canada

The Carlu – Toronto ▪ June 2013 MDM & Data Governance Summit New York

Marriott Marquis NYC Times Square ▪ October 2-4, 2013

Page 33: Understanding Reference Data with Aaron Zornes

• Orchestra Networks is a leading Reference / Master Data Management vendor.

• Sole focus is MDM/RDM Platform: EBX5 • Company founded in 2000 • Stable, privately-held

About Orchestra Networks

www.orchestranetworks.com/rdm


Recommended