APPROACH FOR MDM

& DATA GOVERNANCE

APPROACH FOR MDM &
DATA GOVERNANCE
APPROACH TO MDM (MORE THAN TECHNOLOGY..)
Data governance sets the policies and practices that address how data Data stewardship is the set of operational activities that
management decisions are to be made (work allocation & issue resolution are performed according to policies and practices
mechanisms) and monitors compliance. established as part of governance

Organizational change management

Governance
facilitates successful implementation of the data
governance and data stewardship Data determines what information is considered to be
the enterprise view of a person, product or vendor to
Change Data be managed, monitored and exchanged
management stewardship

Data
Platform

Platform is architecture (data model, integration model ..) and technology that supports storage, data integration, data quality, metadata
management, reporting, data profiling, workflows and data management etc.
MDM Complexities – to be dealt with and which we helped address..

• Multiple ownerships
Business / Functional • Clear understanding and demarcation of Change management and Data Governance initiatives
Complexity • MDM get diluted to address tactical priorities and loses larger view from Sr. Management

• Master Data Governance - managed in silos – separate LOBs having people doing redundant jobs
Technical & Architectural • Reference Data Governance – not an area of focus
• Managing multiple processes around master data & reference data across multiple regions for
Complexities different entities

• Multiple LOBs, tech groups having different MDM expectations

Implementation / Rollout • Prolonged time commitments & re-alignments leading to stretched schedules and spill overs
• Hybrid/iterative approach considered cumbersome
Complexity • Business requirements handled outside of business groups

• Data model design being independent of a MDM related program – entities & relationships identified
Governance / Data from a DW/legacy perspective
• Opting for a customized data model
Management Complexity • Not having proactive data governance
Wipro Mpower – Full lifecycle MDM methodology
MDM Dimensions
Business Data Technology Governance
• Business drivers &
• As-is data management
imperatives • Key master Data entities Master data landscape
processes
• Key business use cases • Data quality assessment assessment
• Data ownership
Assess • Enterprise initiatives

MDM • MDM Business capabilities • Master data definitions • MDM architecture Evaluate data governance
Strategy • Business sponsorship • Key attributes • Solution options organization models
Envision

• MDM business value & ROI

MDM tool evaluation , POC Draft data governance
Blueprint • Linkages with enterprise Conceptual data model
and recommendations organization , RACI and metrics
programs Org.
Change
Mgmt.
Define master data conversion & cleansing strategy , Infrastructure readiness , Communication and training plan ,and data stewardship
Prepare
processes
MDM
Realization
Master data business processes , Logical data model , data Incremental roll-out (Global &
Rules and KPIs integration and DQ rules Data governance roll-out
Implement Local)

Measure Data quality improvements , process improvement and continuous monitoring

PROGRESSIVE APPROACH FOR MASTER DATA MANAGEMENT:
MDM-
SST
MDM-
SSR

DQR Centralized Master Data

Management
(Single Source of Truth)
Consolidate View of Master • One-Stop Master Data
DQM Data management solution for
across organization
(Single Source of • On-Line Master record creation
Data Quality Remediation Reference) • Unified set of policies and
• Standardize tools and business • Off-line Golden record creation with procedures
processes for DQ improvement linkages to source system records • Unstructured Data
• Group-wide Data Model Data • Golden Record available for Management
Data Quality Monitoring creation and mapping to source management reporting • Full Stack Data Governance
• DQ Monitoring tools systems • Syndication of golden records for capability
• • Product Classification & bulk correction at source • MDM tool implementation
DQ reports for data stewards
Hierarchy/Taxonomy Management • MDM tool implementation
• DQ Dashboard for
• Data cleansing through DQ tools
Management
and reference data sets
• Data Validation and correction by
data stewards

Data Quality Monitoring & Improvement

Data Governance
KEY CONSIDERATIONS FOR MDM PILOT IMPLEMENTATION & OTHER
REGIONAL ROLLOUTS

MDM Portal Portal Localization, Regional / Country user

onboarding
Inbound / Outbound
Integration Country Specific
Source / Target
Integration
Data Model / Must
Data Model Localization
have attributes
MDM Rollout

Data Stewardship Local Data Stewardship

Rules

Data Quality Country Data Preparation and

Rules Cleansing

Country specific
CRUD Rules
CRUD Rules

Training and Training and Change

Change Management with localized
Management content
 MDM Consulting and Strategy
Our strong consulting approach will help customers in –
1. Building MDM & Data Governance roadmap for reduced financial risk 1

Governance, Glossary, Catalog &

The MDM Strategy delivers:
detailed roadmap
exposure
2

2. Define & assess Product Vendor strategy and POC for a complete Business Services for the CDO org
Enterprise architecture
implementation roadmap 3 roadmap for business
Metadata, Lineage, masking & leveraging IT
3. Helps CDO in Identifying & quantifying MDM business value to get security
Deliver ROI and TCO via
business buy-in for the program 4
value based themes
Source system Reporting
4. Defining technical solution implementation roadmap and calculating the strategy priorities Client-optimized MDM
5 technical architecture
TCO/ROI for the engagement Infrastructure Service advisory for
data strategy
5. Building a project cost element structures & cost model, technology 6

Migration, DQ & Integration

migration and upgrade framework strategy

6. Help build POVs on opensource, data lake integration and Big Data 7

Assessment and Data use cases

quality, matching & integration use cases
7. Current state assessment on technology, business processes and
governance practices across geos and departments
Wipro’s MDM Reference Architecture has been the foundation of successful master data
8. MDM Solution blueprint and implementation rollout strategy consulting gigs; successfully utilizing the mPower MDM strategy ( Assess, Envision & Blueprint)

9. All domains master data architecture definition and implementation

roadmap
A Sample MDM Roadmap..``
Rollout Additional MDM Domains Consolidation/ Analytical MDM HUB
1) Reference Data Management (RDM)
1) Consolidation (analytical) MDM HUB
2) Product enrichment, B2B enrichment
2) Address urgent MDM needs – Prioritize sources
3) MDM interfaces to external 3rd parties
3) Deploy new MDM Hub with one-way feeds
4) Private and (later) public cloud integration
4) Manual update process using MDM Hub UI
5) Cross-sell, up-sell, campaign management analytics
5) Data steward training on MDM metrics

3
5

Fully Synchronized MDM to Source System

1) Rolling deployment of real-time, synchronized interfaces
2) Bi-directional Source-MDM-Source interfaces
3) Automated matching with data steward workflow 2
4) Advanced data steward MDM metrics
in MDM Hub

4
MDM and IT Governance
MDM Strategy and Business Case 1) Create governance organization and PMO CoE
1) High level requirements, source system analysis 2) Select IT admin, stewards/ global representatives
2) High level business case/ ROI (detailed extra) 3) Create new MDM, governance workflow
3) MDM governance recommendations applications
4) MDM road map and approach recommendations 4. Data profiling exploration & proactive DQ
1 scanning tools
MDM BEST PRACTICES
MDM Best Practices based on our MDM engagement experience
1. Build Business Case 2. Planning 3. Program and Data Governance
 Analyze business problems that necessitate need  Seek commitment time from business teams for  Keep Data Governance as an integral component
for MDM implementation Requirements, Design Review, demos along with of the implementation
development, User Acceptance  Establish roles and identify individuals
 Define quantifiable returns
 Focus on Continued Value Delivery – Source  Enable change management along the
 Poor data quality is often not noticed & the
Level Improvements, Smaller Iterations implementation path
impact of this is not known unless exposed
 Build ‘POC’ and demo the ability of the system to  Keep Business teams ‘engaged’ during the
 Build consensus amongst business community ‘eclipse’ period
address One or Two use cases
about the data problems and the need to address
them  Plan to familiarize the Core Business team on  Engage with Business users and address criticism
MDM Terms and functionalities and concerns
 Gain senior management commitment for
continued support for the implementation  Avoid ‘Big Bang’ Approach  Clearly define a business glossary for all MDM
entities and attributes

4. Technology 5. Business Value Delivery

 Realize the limitations of IDD Interface and native workflow capabilities  Identify key data metrics that reflect the performance
of the tool improvements – Reduction of DQ error, reduced cycle time
 Create an instance of the MDM Reference architecture by identifying for creation
specific tools and technologies  Enable data governance implementation alongside of the tool
 Gain understanding of Product Roadmap to design functionalities implementation
 Adopt data quality principles and business rules in existing
 Validate the Product Availability Matrix with the organization standard
application
to ensure compatibility
 Plan for incremental delivery – preferably monthly releases
 Sizing of the environments and detailed ORS assessment will help save
post the implementation of the main phase
potential delays and challenges at later point
 MDM solution implementation
By leveraging our Best Practices for the implementation
1. Conceptual Data Model 2. Data Enrichment,
Standardization & Validation 3. Data Matching 4. Data Integration

• Make sure the match requirements are well • Use 3rd party party tool for data enrichment • Plan adequate iterations for doing • Consider real-time data ingestion if
thrashed out as they will drive the model like Address Doctor for address enrichment, match rule tuning source provides data in real-time
and also the major effort required towards a and standardization
successful implementation • Leverage exact match column in fuzzy • Create business services for course
• Standardize all reference data elements like match rule for better performance. Limit grain MDM APIs by orchestrating
• Keep the database at a 3NF form and don’t country code, state code, phone numbers, trust configuration for important fine=grain services
normalize it completely as it limits flexibility, salutation columns only
performance and functionality • Profile source data and define
• Use Appropriate DQ tool’s cleansing • “BOT” can learn and take actions on appropriate data quality rules
• During design, it is useful to create a mix of capabilities & reference tables in MDM to such records by learning from actions
logical & physical model that contains key • Evolution of MDM architecture from
store “Standard” reference values taken by human data steward in past
attributes & related notes one style to another will require
• Implement robust validation rules in the • Consider data model design between changes to the data integration
• Articulate hierarchy and relationship application where data is generated like 2NF & 3NF; database hardware patterns
definition as they have direct impact on POS, Portal, etc. capacity, data volume must be
MDM data model and configuration • Define integration layer to handle
• Load only valid data in MDM. Data should be considered during design stage
unplanned exceptional scenarios /
corrected in source & re-loaded in MDM Data rejections

Data Management – Best Practices

B. Data Modelling C. Data Publish D. Data Stewardship

• Match requirements to drive Data model – A key • Establish the integration pattern such as service based • Setup VIP and Load Balancer in a clustered
value based model supports ‘OR’ conditions vs. Message Queue VS Batch or a combination of them environment for distribution of loads across
whereas column based model supports ‘AND’
conditions in matching
• Define reusable services that can be leveraged across nodes
the enterprise
• Avoid consolidating child tables unless there is a • Enable SSL to enable data security while data
specific business case
• Define a wrapper over MDM provided services and
in motion
expose wrapper to consuming applications
• MQ based integration can trigger a lot of messages • Consider using OOB; security model to
depending on type of events. This must be a enforce data authorization instead of
consideration when taking this approach customization