Transcript Suggested Outline for TDWI Presentations

MDM for a Consistent View of
Customers
JB Sastry
Chief Architect, DW- GE Money
May 15, 2007
CDI/MDM initiatives are NOT IT initiatives
They need to solve prevailing or future business problems
Sponsorship is critical
It is a progressive roadmap
Each step has validation-quality cycles associated
Our Customer Centric Enterprise
Compliance Management
Framework
Legal
Barriers
Contractual
Obligations
Customer
Centric
Processes
Strategic
Imperatives
Examples of contractual barriers – retailer contracts; no mix and merge of specific retailer data
Legal examples- can not use house hold info for adverse action
Can not use customer data from Acxiom for adverse actions
Define to ensure what exact customer oriented processes are being catered to: How do thes efit into
over all organizational strategies
Make the investment work toward integrated joint strategies- avoiding business functional siloes is
more of a dream than a reality
2
Customer- 360 [ x ‘n’]
Sales/Mktg
Profitability
Finance
and Capital
Mkts
Relationshi Addresses
p-1
Legal &
Compliance
Contacts
Relationshi
p-3
Life Evt
Scores
Self Svc
Apps
Risk Mgt
Relationshi
p-2
Channel &
Contacts
Mgt
Fraud &
Money
Laundering
Triggers
Delinquency
Mgt
Many customer touch points
Many customer management aspects
Many customer oriented biz functions
3
Examples of Customer Centric Processes
Risk Management:
Vintage Analysis
Channel/Contacts Mgt:
Preferences and Contact
Optimization
Delinquency Mgt:
Treatment Optimization
Preference & Privacy Mgt:
Extended
Operationalization
Fraud and Money Laundering:
Suspicion Triggers
Increasingly and stunningly the customer centric information is finding place in
predictive models
Analytical models are becoming sophisticated and processes richer due to
customer level data injection
Creative applications are emerging suggestive of DW- Info Mgt Maturity rise
4
Key Goals of Master Data
A solution that collates, maintains and provides
organizational Master Data to operational systems
as needed in a standardized manner.
It has built in data governance and synchronization
mechanisms to ensure appropriateness is
maintained consistently.
Some industry definitions
Key words- standardization, governance, synchronization
5
MDM Characteristics
Without enterprise-wide
integration, customer centricity is
a lost cause
Enterprise-wide
Data
Integration
Synchronization
(With OLTP)
Guarantees
currency
By far the most complex piece of the puzzle:
App and Data Integration come together.
Watch out for complex data flows across
multiple tiers
Access
enables
Agility
Multi-tiered systems facilitate
access points in line with
consumption
Often, access is the main leverage
used by biz siloes to proliferate
“irrational data exuberance”- to
steal a phrase from the
redoubtable mr. greenspan
Quality and
Data Governance
ensures
Reference
Value of information is
compromised (enhanced) by data
quality
Data ownership and stewardship policies a
must to avoid unbridled chaos
6
Extracting the Master Data
Providers: Source systems,
Comm Channels, Web Logs,
Bureaus…
Data
Processors: Rules
Engines, CDI, App work
flow, SOA,…
Process
Enablers: Meta Data,
Data Quality/Profiling,
Master Data …
Information
Integrators: Strategic
Tools: DWs, ETL,
information integration, Tactical
Referencing, Behavior Modeling,
Strategy Bldg, Optimization,.…
OLAP, Data Mining,…
7
A Multi-tiered Data Mgt Solution
Analytics
8
The DW-MDM Bridge
Operational Data systems that
are refreshed from DW
servicing the Master Reference
Data Tier
DWH
ETL Mappings
Three forms of Master DataAnalytic
Operational
Actionable- e.x. Contacts; Offers
9
The Information ‘Package’
Meta Data:
Lineage, Linkage,
Definitions, Logistics and
Morphosis
Raw Data
x
+
=
Information
Data Quality:
Accuracy, Timeliness,
Relevance,
Completeness,
Trustworthiness and
Meaning
10
Corner-Stones of Customer MDM
Extremely diverse ‘aspects’ of
behavior dependent upon the
particular biz function’s outlook
Standardized tools and processes
for cleanse-match-merge routines
Contact mgt, cost mgt (delq),
cross sell, ….
Identity
Behavior
Meta-Information
Mappings
Organizational motto of “Know thine
data”
Structured approaches
Meta information could be exhaustive- a
thankless information assembly that is
only sporadically used
Data Stewardship crucial for success
Naming conventions
Often the most neglected area of DW
builds due to typical cost/time over runs
11
Customer- 360
View
Account
Level
Consolidation
Divergent, Atomic
Data Streams
Cleanse- Dedup- Match- Unify
Integrating the Customer Information
Customer- Demographics,
Contacts, Preferences, …
Master Ref
Acct Financials, Trending and
Tracking
Actionable
BI
Behavior Modeling & Strategy
Rules
Mostly this means customer level analysis but account level
execution and tracking
Integration funnels on both sides of the account-level data
usage
Expensive and painful system builds call for iterative
production implementations for self-sustained projects
Aim for ‘Bursts of ROI’ realization
12
Getting Organized for MDM
Evolve a Solution Framework
Find the Champion
Encourage Sponsorship
Create Roadmaps
Data Stewardship
Governance Board for change control
Build Domain Expertise
13
Contact Information
• If you have further questions or comments:
Brahmaiah S Jarugumilli [JB]
[email protected]
14