Data Governance:

Building a Framework for MDM

Aaron Zornes
Chief Research Officer
The CDI-MDM Institute

Daniel Teachey
Corp. Communications Director
DataFlux
 Data Governance
Building a Framework for MDM

“84% of businesses surveyed believe that poor data governance

can cause: limited user acceptance, lower productivity, reduced
business decision accuracy, & higher TCO”

Aaron Zornes
Chief Research Officer
The CDI-MDM Institute
The-CDI-MDM-Institute.com a.k.a. tcdii.com
 Working Definitions
Data
DataGovernance
Governance(DG)
(DG)
Formal orchestration of people,
process, & technology to enable
an organization to leverage data
as an enterprise asset.

Master
MasterData
DataManagement
Management(MDM)
(MDM) Customer
CustomerData
DataIntegration
Integration(CDI)
(CDI)

The authoritative, reliable Processes & technologies for

foundation for data used across recognizing a customer & its
many applications & relationships at any touch-point
constituencies with the goal to while aggregating, managing &
provide a single view of the harmonizing accurate, up-to-date
truth no matter where it lies. knowledge about that customer to
deliver it ‘just in time’ in an
actionable form to touch-points.

Data governance is a prerequisite to efficient & effective

information integration & business intelligence –
two key differentiators in highly-competitive industries
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Data Governance
Strategic Planning Assumption

During 2007, enterprise-level data governance will be

mandated as a core deliverable of large-scale CDI-
MDM projects delivered via RFPs
Through 2007-08, major systems integrators & CDI-
MDM boutiques will focus on productizing their data
governance frameworks while most CDI-MDM
solution providers will struggle to link business
process design with process hub architecture
By 2008-09, both corporate & LOB data stewards will
be a common position as Global 5000 enterprises
formalize this function amidst increasing de facto &
de jeure recognition of information as a corporate
asset.

© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com CDI-MDM MILESTONE

 Why Data Governance? Why Now?

Businesses have been governing data for 20+ years,

however, only a rare few are doing it well today
Many companies historically assigned DG to a data
management group whose job is to integrate &
manage data
Contemporary DG challenges are far greater
Break down functional stovepipes
Integrate processes across the enterprise –
including corporate technology, all LOBs,
functional areas & geographic regions
Engage all levels of management
Based on recognition of issues at hand, an improving economy, &
increasing regulatory requirements, businesses are now recognizing
the opportunity to take a more strategic view of data governance
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Why Data Governance? Why Now? – cont’d

Once you know what data is worth, you need to calculate

probability for risk in a business processes
When you understand value of data & probability of risk, you
can evaluate how much to spend to protect it, manage it,
and invest in adequate controls
This is basis of modern underwriting – assets, risk, controls
Doing this systemically requires a combination of
organizational structures, business processes, & technology
– a “data governance blueprint” for:
Data quality
Information integration
Business intelligence
IT management must work with business leadership to design &
refine “future state” business processes associated with data
governance commitments
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Data Governance Maturity Level
50%

40%

30%
FSP
20% Non-FSP

10%

0%
Basic Foundational Advanced Distinctive
Source: 2006 -07 CDI-MDM Institute survey of 50 Global 5000 IT organizations

BASIC (“anarchy”) – App-centric approach; meets business needs only on project-specific

basis
FOUNDATIONAL (“IT monarchy”) – Policy-driven standardization on technology &
methods; common usage of tools & procedures across projects
ADVANCED (“business monarchy”) – Rationalized data with data & metadata actively
shared in production across sources
DISTINCTIVE (“Federalist”) – SOA (modular components), integrated view of compliance
requirements, formalized organization with defined roles & responsibilities, clearly defined
metrics, iterative learning cycle
Overall, Financial Services Providers are leading the way for non-FSPs
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Data Governance Challenges
Data Governance
Adjudicating between In Practice
centralized vs.
decentralized data CIO
Business
Exec
stewardship
Evolving key stakeholders
from “data ownership” to Chief
Data
VP of
Infra-
VP of
Appli-
Dept’l Dept’l Dept’l

“data stewardship”
Exec Exec Exec
Officer structure cations

Upgrading notion of “data” Data Governance

to “policy/process” Steering Committee

Corporate BU BU
Data Data Data
Steward Steward Steward

Data quality-enabled information integration is fundamental

to ability to "leverage data as an enterprise asset”
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Data Governance Enables
Enterprise Information Integration
Privacy
Preferences

Pricing
Bundles
Discount
Policies

Compatibility

Process /
Next Best
Offer
Policy Hub

Legacy

Upgrade/
Downgrade
Shipping
Approvals
Eligibility

Methodology is needed to bind process steps, skills &

software to produce data governance deliverables
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 “Top 5” IT Drivers

Developing architecture best practices & 31%

standards 47%

Monitoring to improve data quality 56%

46%

Building governance infrastructure, technology 56%

& supporting organization 60%

Developing standard metadata mgmt 63%

47%

Defining processes & business rules for 81%

ongoing governance 67%
Non-FSP
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
Source: 2006-07 CDI-MDM Institute survey of 50 Global 5000 IT organizations FSP

IT organizations cannot “go it alone” –

data governance experience & accelerators are needed
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Data Steward Residence & Specific Role

Developing architecture best practices & 31%

standards 47%

Monitoring to improve data quality 56%

46%

Building governance infrastructure, technology 56%

& supporting organization 60%

Developing standard metadata mgmt 63%

47%

Defining processes & business rules for 81%

ongoing governance 67%
Non-FSP
0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
Source: 2006-07 CDI-MDM Institute survey of 50 Global 5000 IT organizations FSP

DS is a business-oriented function – concept is embraced, active, generally

project-oriented & within the BU (not within IT ) – clearly, business is
stepping up more on a project basis to take responsibility for the data
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Survivorship of Trusted Sources

User-originated master data update 13%

Rules engine of legacy system 13%

Not applicable
16%

Rules engine of enterprise CRM suite 19%

Rules engine of data integration tool 23%

0% 5% 10% 15% 20% 25%

Source: 2006-07 CDI-MDM Institute survey of 50 Global 5000 IT organizations

Rules themselves are in many different places & need to be extracted

out into horizontal layer & then replicated back to individual units –
this is “true data governance-driven MDM” in the long run
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Basic Data Governance Key Steps

Establish formal data governance process that maps

to the business architecture
Differentiate between data stewardship & ownership
while defining, staffing & communicating
Define & continuously measure adequate, agreed-upon
data quality metrics
Design/select, & implement appropriate architecture
components to improve data quality
Empower & hold accountable “up stream” data
stewards for their domains

Formalize the treatment of “data” as a corporate asset

© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com

 Recommendations
Data governance is difficult – enterprises need to
embrace a blueprint for success
IT management must understand impact of data
governance on enterprise information integration plans
All businesses embrace political structures
Data governance is necessary to support politics of both
centralized & distributed information integration
Without firm commitment to corporate data governance,
BU-specific data governance projects run risk of becoming
“yet another master data mart”
Savvy IT organizations begin with data stewardship &
business metadata definition
Data governance is a key enabler for data quality-driven
information integration –
IT management should get “on message”
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
 Aaron Zornes
Founder & Chief Research Officer
The CDI-MDM Institute

The-CDI-MDM-Institute.com a.k.a.
www.tcdii.com

Authoritative

Independent

Relevant
© 2007 The CDI-MDM Institute The-CDI-MDM-Institute.com
Data Governance:
Building a Framework for MDM

Daniel Teachey
DataFlux
About DataFlux

Recognized as a leading provider of data quality

and data integration solutions
– Intelligent Enterprise Readers’ Choice Award, 2006
– CRM magazine “Market Winner” in Data Quality, 2006
– CTO and co-founder Scott Gidley named an “Innovator to
Watch in 2006” by Infoworld magazine
Acquired by SAS, the world’s largest privately-
owned software company, in 2000.
Data Quality is the Foundation

Master Data Management

– Unify your data to create a single view of enterprise information.

Data Governance
– Enforce business standards and rules across the enterprise.

Compliance
– Meet enterprise, industry and government regulations.
Data Management Landscape

Corporate Data
Governance
Enterprise Data Maturity Model
Our Approach

Solutions

Accelerators
Customer Data Product Data

Data Quality Integration Platform

Data Quality
Metadata Discovery Data Monitoring
Data Integration
Data Profiling Data Enrichment
Data Visualization
Data Monitoring: Key for Data Governance

Create business rules and enforce them in batch or real time

– Business rules can be simple…
• If address is invalid, email an admin)
– …or more complex
• If loan-to-value ratio in table A is .8 or above and table B does not show a PMI,
then email an admin and enact a corrective routine
Data monitoring options:
– Audit – Measure data quality over time
– Trend – Monitor new data as it enters a data store
– Alert – Notify user or application when user-defined rule is broken
Data monitoring is critical for data governance initiatives
– Support the codification and reuse of business policies
– Provide a way to ensure that data meets business standards
Data Governance Success
Mortgage Provider

Business Issue Existing data quality controls developed and maintained by

programmers. Rules duplicated across different programs and
different groups. No central repository of rules. Very difficult to
control internal processes.
Mission Meet standards for data quality across the enterprise.
Implement reusable data governance controls to monitor data
movement, business data quality, accounting, etc.
Results Established a “use once, reuse repeatedly” model to centralize
all business rules for data quality. Monitored internal sources
to validate the quality of the business processes used to run the
business. Generated better reports to federal and industry
agencies.
Data Governance Success
Chemicals Company

Business Issue Enforce business policies for product and inventory

classifications (reduce manual rework and execute automatic
consistency checks)
Consolidate data to a “master data” source. Needed to
analyze, improve and validate data quality along the way
Mission Meet or exceed data quality goals across the enterprise.
Implement reusable data governance controls to monitor data
movement, business data quality, accounting, etc.
Results Established a “use once, reuse repeatedly” model to centralize
all business rules for data quality. Monitored internal sources
to validate the quality of the business processes used to run the
business. Generated better reports to federal and industry
agencies
Monitor Results
Data Governance Dashboards
Data Governance:
Building a Framework for MDM
Aaron Zornes
Chief Research Officer
The CDI-MDM Institute

Daniel Teachey
Corp. Communications Director
DataFlux
Thanks for Watching