Definition
MDM, is a centralized operational application that provides pristine master data in real-time and batch to the subscribing applications in support of business processes, backed by governance, and rich services such as data cleansing, data Enhancements, cross referencing, matching, de-duplication, hierarchy management, security and synchronization across all enterprise applications. Examples iClassic MDM, Informatica MDM,
DNB, Oracle MDM, SAP MDG, and IBM MDM.
DNB, Oracle MDM, SAP MDG, and IBM MDM.
Guidelines
- Establish an Enterprise Governance
- If Governance program is already established, Ensure the MDM initiative is sponsored by them to ensure
there is no ambiguity on the reason for the Master Data Management initiative - Perform Requirement Gathering
- Identify Integration Touchpoints
- Capture End to End Flows requiring Master Data
- Identifying Sources and Targets capabilities
- Define the Canonical Data Model
- Perform Source Data Analysis
- Establish KPIs on Sources to ensure they meet target Data Quality scorecards
- Define Data Quality - Match & Cleanse Rules
Auto Rules and Manual Review rules - Define the enterprise Master Data Key & identify the lifecycle events that can modify this key
Ensure modification of this key is well handled in the downstream applications
for example - if ERP cannot handle merges, then put in place holds to prevent such
modification if live transactions are booked on those modification - Define Test Cases for each of the Subject Areas
- The Test Cases should include CRUD, Merge and unMerge Operations
- The Test Cycle should include core MDM and Cross Application End to End test flows as well
- Identify the Consumers and Producers capabilities
- Ensure the Integration Framework is well understood if one is available and
the MDM's integration capabilities are mapped to ensure there is no gap
Understand the usage pattern for each of the consumers and producers,
so they do not misuse pattern for instance using Real-time APIs for batch loaders - Establish Security such as oAuth as per the Enterprise guidelines
- Ensure the Design Document is well defined, the following is must
- Solution Architecture
- Component Object Model
- Software Perspective
- Infrastructure Perspective
- Deployment Model
- Unit Test Strategy
- Integration Test Strategy
- Cross Testing Strategy
- Monitoring & Debugging Strategy
- Release Management
- Innovation Strategy
- Establish the Functional Design Document & Technical Design Document for each of the requirements
The functional & technical design should map into the Design document and the framework defined in the earlier stages - Ongoing Operations and Maintenance strategy should be well defined and need to constantly evolve
- Define Communication plan
- Define Business Glossary and Data Catalog
- Ensure the Glossary and Catalog is mapped to the Business Process and techincal processes
This will provide insights into Lineage for any business and technical impact analysis - Ensure process team / Data stewardship flows are well defined and documented