data strategy for mdm · 2008-03-11 · mdm. phase 3: sustain and enforce data quality. benefits...
TRANSCRIPT
11
Data Strategy for MDM Implementation
임정혜
한국인포매티카
2008.3.5
2
Agenda
• MDM Overview
• Informatica’s Value Prop for MDM
• Case Studies
3
What is Master Data?
Vendor
• You are Master Data
• You buy from Master Data
• You buy Master Data
• What you do, or create is Transactional Data
ProductCustomer
Campaign Order Invoice ShipmentReceive Payment
Request a proposal
Master DataTransactional Data
4
Sales
Contact ManagementSales/Pipeline Management
CompensationWin/Loss AnalysisSales Information
Call Center AutomationTeleSales Scripting
Email Routing & Management
Online Communications
Finance
Sales/Revenue ForecastingPurchasing
AR/AP systemsFraud IdentificationAudit Fin Review
Marketing
Market Plan DvmntOutbound Message Mgt
Competitive TrackingCustomer TargetingCustomer ProfilingSegment Analysis
CustomerSupport
SystemsFinance HR
The Challenge of Managing Master DataINFA is no different than Customers
Organizations
Customer Service FinanceSales Marketing
• Multiple Lines of Business
• Multiple Applications
• Multiple Sources of Master Data
5
Master Data and Data Quality Rapidly Gaining Importance in Enterprises
“By 2005, Fortune 1000 enterprises will lose more money in operational inefficiency due to data quality issues than they will spend on data warehouse and CRM initiatives.”
Source: Gartner
Corporations are increasingly focusing attention to master data management and data quality due to increased pressure for corporate transparency, accountability, and performance on a enterprise-wide basis.
50% of enterprises surveyed maintain master data separately in 11 or more source systems50% of enterprises surveyed consider master data management to be a high or top priorityOver 80% of enterprises surveyed plan on centralizing master data management
Source: Tower Group
6
Source: Gartner Dataquest (December 2006)
0
100
200
300
400
500
600
700
800
2003 2004 2005 2006 2007 2008 2009 20100%
20%
40%
60%
80%
100%
120%
140%
160%Revenue
Growth
$102.9m
$43.2m
$274.1m
$189.5m
$386.3m
$485m
$585.3m
$706m
MDM Market
0
200
400
600
800
1,000
1,200
1,400
2005 2006 2007 2008 2009 20100%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Revenue ($M) Growth
Revenue (Millions of Dollars) Grow th
$322m
$214m
$638m
$462m
$845m
$1061m
Customer Hub, 2003-2010Significant growth ahead
Product Hub Market, 2005-2010PIM seems to be an even larger market
7
Generally Accepted MDM Principles
• MDM is an APPLICATION• Even if homegrown
• Composed of:• Data Models• Business Processes• User Interfaces• Role based users• Integration with the rest of the enterprise
8
MDM Application S/W Vendor Landscape
Tools To Build an MDM Application
KalidoOrchestra Networks Stratature Teradata
Product Information Mgmt
Teradata
Sterling Commerce
SAP
Customer Data Integration
Financial
Hyperion
Microstrategy
Microsoft
SiebelIBM/WPC
IBM/WCC
Initiate
Siperian
Tibco
Cognos
Oracle
IBM
SAP
Sunguard
GXS D&B
SAP
9
What Does an MDM Application Look Like?SAP
10
Agenda
• MDM Overview
• Informatica’s Value Prop for MDM
• Case Studies
11
Informatica Positioning and Value Prop for MDM
• We do not provide an MDM Application
• We complement MDM Applications
• We enable homegrown MDM initiatives
• DI and DQ are critical for MDM project success
12
Mainframe
Applications
Databases
XML
Messages
UnstructuredData
Data Model
Enterprise Connectivity
Address Validation
Data Profiling
Bulk Data Movement
InformaticaComplementing MDM
Hierarchy Management
Event Management
Cross-reference key management
Governance / Audit
DQ Scorecarding & Reporting
Info
rmat
ica
MD
M A
pp
Fuzzy Matching
Pattern Matching Changed Data Capture
MatchingDe-Duplication Publish
Validation
De-Duplication
Matching
Web Services/XML
Validation
FeatureOverlap
Data Standardization
13
Master Data ManagementNew and Grand Vision?
Mainframe
Finance
Sales
Manu
HR
Marketing
Supply
Client/Server / Custom Apps
ERP
SCM
HRM
CRM
Packaged Apps
MDM!
14
CDI PIM FI Hub EDWHomegrown
…or Same Approach with Different Name?
ERP
SCM
HRM
CRM
Mainframe
Finance
Sales
Manu
HR
Marketing
Supply
Client/Server / Custom Apps Packaged Apps
15
Informatica Value Proposition
• Interacting with external systems
• Monitoring and reporting data quality
• Ensuring compliance with rules for data governance
• Integrating with SOA strategies in terms of operating at the heart of "information as a service"
“All the MDM vendors are weak in…”
Gartner Report on MDM
Data Access & Delivery
Data Quality
Data Audit
Data Services
INFO
RM
ATICA
16
Informatica Value PropositionBloor Research Whitepaper
“Put Simply, if you ignore data integration or do not treat it as being of sufficient importance then your MDM project will fail.”
Philip Howard, Bloor Research - October 2006
17
Packaged Applications
Relational DB
Flat Files
Messaging and Web Services
Mainframe and Midrange MDM
…
…
• Interacting with enterprise-wide systems• Profile Data to identify problems• Analyze and share Results
• Standardize and Cleanse• Supplement and Enrich• Test and Load
• Monitor on-going quality• Enforce processes• Scale to other master data types
Assess
MDM Requires High Quality Data“As with any data-driven application, the quality of the data determines the level of success.”
…
18
InformaticaProven Data Foundation for Enterprise Projects…
Packaged Applications
Relational DB
Flat Files
Mainframe & Legacy
EDW
Data Migration
Data Warehousing
ODS
Application Data Synchronization
B2B IntegrationPartners/M&A
Analytics…
• Enterprise-wide data access• Secure and Scalable Platform• OEM technology for Siebel
Analytics, Hyperion, JDE, etc.
Data IntegrationData Quality
19
Packaged Applications
Relational DB
Flat Files
Mainframe & Legacy
EDW
Data Migration
Data Warehousing
ODS
Application Data Synchronization
B2B IntegrationPartners/M&A
Analytics…
…Seamlessly Extended for MDM
MDM
MDM
MDM
20
Data Strategies for MDM Implementation
• Data profiling
• Quantify data quality
• Data Quality Scorecards
Phase 1: Assess Data
• De-risk MDM project
Phase 2: Cleanse & Migrate
• Fix data issues
• Harmonize data for load into MDM application
• Successful MDM data migration and go-live
Phase 3: Sustain and Enforce
• Reuse of data quality rules
• Centralized rule maintenance
• Increase user confidence and adoption
Phase 4: Expand Scope
• Repurpose resources and processes
• Scale project success
Value
ProjectBenefits
MDM Project Lifecycle
21
Phase 1: Assess and Quantify Data Issues“You can’t manage what you can’t measure”
Relational DB
Flat Files
Enterprise Data Profiling
Analyze and Score Data’s:Completeness
Conformity
Consistency
Accuracy
Duplicates
Integrity
Benefits• Proactively assess
and discover issues
• Quantify and measure data quality
• Identify and plan for corrective action
• Manage data assets
Packaged Applications
…
22
Packaged Applications
Relational DB
Flat Files
Phase 2: Initial Load / Pilot Project
MDM
…
1. Access source systems
2. Manage Data Quality/Transform Data
4. Target MDM
3. Validate/ Test
Iterate
Directly leverage Assessment results to plan and implement data prior to MDM load
Benefits• Manage Data Quality
• Standardize, Cleanse, Enrich, De-dupe, Validate• Reduce burden on MDM app to clean legacy data
• Bulk data movement
• Complex transformation (no-coding)
Enrich
Hierarchy Management
De-dupe (cross-system)
Merge
Cross Reference Key Management
Audit
Stewardship Tool
Workflow management
23
MDM
Phase 3: Sustain and Enforce Data Quality
Benefits• Ongoing reuse and enforcement of Data Quality Rules
either:• External (to MDM App) – DQ for on going
synchronization between data sources and MDM app• Internal (to MDM App) – DQ rules repurposed at the
point of user entry within the MDM app
Assess
Relational DB
Flat Files
Packaged Applications
Bad: Fix Suggested or Record Rejected
Data Quality Service
Good: Committed
Master DataEntry
Master DataEntry
Data Quality
24
Packaged Applications
Relational DB
Flat Files
Phase 4: Enterprise-Wide DeploymentScale Across Systems and Master Data Types
1. Access source systems
2. Manage Data Quality/Transform Data
Synchronization/Bi-directional
Integration
4. Target MDM App
3. Validate/ Test
Iterate
MDM
Benefit• ROI – Reuse of assets and processes• Control Scope – Phased implementation by
sources/applications• Modernize application landscape –
framework to retire legacy applications• Scale MDM scope across data entities
Messaging and Web Services
X
X
Product
Vendor
…
25
KPNMDM Architecture and Project Phases
Customer Registration
System
Sales(SFA)
Customer Data Cleansing Platform
PowerCenter
IDQ
Enterprise Reference Data
ERD
CKRFlat FilesInitial and Scheduled
Loads
ERD Data Load
Data Quality Reports
MDM Staging
Error and Audit Logs
Reference Data for Data Quality Rules
EIMInitial / Bulk
Load
MDM
Cur
rent
Sco
peFu
ture
XML via EAICustomer
Insert / Update
XML via EAICustomer
Insert / Update
XML via EAICustomer
Insert / Update
XML via EAICustomer
Insert / Update
IDQ Match and Merge
XML via EAISFA Insert /
Update
XML via EAISFA Insert /
Update
Phase 1 – Profiling and Assessing DataPhase 2 – Cleanse and Migrate
Phase 3 – Sustain and Enforce
Phase 4 – Expand Scope
26
Informatica Data Quality AssessmentBenefit for MDM customers
• Identify project risks and challenges• Plan and Scope projects based on true state
of data• Promote collaboration between IT and
business owners in identifying and resolving data issues
27
Agenda
• MDM Overview
• Informatica’s Value Prop for MDM
• Case Studies
28
ref123
• Drivers for MDM project:• Lack of data consistency across enterprise landscape (including 8 SAP instances)• Need long term strategy to replace limited MDM solution
• Project Type: • Customer and Supplier (Phase 1); Product (Phase 2)• Currently running a ‘homegrown’ MDM solution - migrating to SAP MDM
• Key Message: • Profiling and Data Quality Management is essential to a successful MDM project.
DuPontDavid O'Donnell
29
DuPontMDM Architecture and Project Phases
Non-SAP
Legacy (+50)
DSAP
W
Q
GT
Profile Reports
IDE
C
KPI
BW
PowerCenter
Keystone(R/3 Instance
for MDM)
Bi-DirectionalLoad & Syndication
IDQ*DQ Rules Repository
Data Quality “Baseline” Scorecards
•Material (MRO)•Chart of Accounts•Vendors•Materials (all types•Customer•BOM•…
IDQ*WSXML XML
SAP PI
SAP MDM
(Pilot Phase)via SAP PI
30
ref123
“MDM is a journey”Geert van Den Berge
Data Management OfficerPhilips Consumer Lifestyle
ref123
PhilipsPeter De Gram
31
ref123
• Drivers for MDM project:• Harmonize product information via GDS (Global Data Synchronization) initiative –
Focus on top 20 business partners (Walmart and Carefour largest ones)• Customers refuse deliveries and penalize Philips (Philips CES losing 25M euros/year)
• Project Type: • Product first (Customer, Employee future phases)• Initial strategy to measure and quantify data quality within business apps and fix there
first• SAP MDM (>28 instances of SAP)
• Key Message: • Data Quality Management can help business create ROI
• e.g. 2.5M euros/year in FTEs just to maintain master data in spreadsheets today)
• IDQ success will help justify and lay foundation for MDM strategy
PhilipsPeter De Gram