Blumetra Solutions is a boutique consulting and professional services firm dedicated to customer success with Informatica MDM cloud software. We were selected by Informatica Corporation to help them implement their own customer/contact master data in their new cloud SaaS product. Here are some lessons we’ve learned from this experience and the best practices we recommend to our MDM customers:

1. Be Clear on your Destination, before you start climbing. (“Strategize and Plan”). This includes a rigorous discovery, roadmapping and prioritization exercise to determine:

1. What are the business objectives of each enterprise data initiative, including master data management? How will success be measured? What are the expected business benefits and how do these translate to annual financial value, payback period and ROI?
2. What is the current state of data governance (people, process, technology) and where are the gaps that we need to address as part of our data initiatives?
3. What are the adoption and value targets for each enterprise data initiative? How will adoption and value capture be measured and increased post-go-live?
4. Key data domains to start with – based on business value, data quality and availability, subject matter expertise availability and more
5. Which “upstream” data sources to prioritize, based on availability, data quality, business impact and other factors
6. Which “downstream” system integrations to prioritize – i.e. which systems should Informatica MDM “publish” to, and which data elements to include in this “publishing”
7. Which reporting and analytics capabilities (e.g. predictive modeling) are dependent on MDM and other data improvement initiatives?
8. Which IT and business processes will need to change for the enterprise to benefit from the improved master data – and when “change management” discussions should begin and with which IT and business stakeholders?

2. Choose the first Project, be clear on requirements, and choose an architecture and design that can scale with your future needs (“Specify, Design and Architect”), including:

1. Decide whether MDM should be centralized or federated and where the “single version of truth” should reside – may be the Informatica MDM system itself and/or the CRM (e.g. Salesforce) system and/or other financial system (e.g. Oracle Financials). More on the meaning of “centralized” and “federated” later in this essay
2. Determine the right set of entities and attributes to be mastered (e.g. customer entity, with attributes to include name, address, other contact info, key demographics, and purchase history)
3. Determine and prioritize requirements and use cases across departments – and ensure there is a clear business and technical owner for each
4. Find the right “vocabulary” for requirements – can be a combination of “functional requirements” (system must be able to do X), “non-functional requirements” (system must give response within Y seconds), “user stories” (data steward needs to be able to enter, search for, retrieve, and/or update the following information), “functional stories” (the following data from the CRM system needs to flow to MDM in real-time, or daily/monthly, etc.)
5. Determine the hierarchies based on the business needs, and make sure these are aligned with clear use cases – e.g. sales may have a different view on organizational relationships (e.g subsidiaries) and mappings of individuals to organizations (e.g. departmental contacts, purchasers, recipients of marketing collateral, etc.) than marketing or finance
6. Right-sizing the initial (“minimum viable product” AKA MVP) release – to give early business value without incurring undue technical debt that will impede or delay future enhancements

3. Build (and test) it right.We have found that the following are the most critical aspects of success during this phase, where success includes high quality, accuracy, stakeholder satisfaction and timeliness:

1. Clearly defined and documented requirements from the previous phase with sufficient deep-dive for business and IT stakeholders so they can make informed “sign off”
2. Clearly defined high-level architecture and design
3. A clear, simple, well-governed process for changing requirements, architecture and design if needed – including limiting changes as much as possible until after the initial “MVP” go-live.
4. Clear agreement to business/technical rules such as “match/merge”, “hierarchy management rules”, “data quality/cleansing rules”, and user workflows (who needs to approve which type of data change)
5. Clear and continuous communication between business and technical stakeholders throughout this process – including regular demos (every 2-4 weeks, typically aligned with the end of each Agile “Sprint”) of what has been developed so far, to reduce the chance of misunderstandings and missed requirements
6. Professional Agile processes and “ceremonies” – including backlog grooming, sprint planning, sprint reviews and retrospectives
7. Clear traceability between requirements, design elements, development tickets, test cases and UAT scripts – to reduce the chance of misunderstandings and increase the possibility of automation to speed up and reduce error

4. Manage deployments carefully– from both a technical and business perspective.

1. “Go-live” (i.e. deployment of a new MDM system in production) and “Cutover” (i.e. turning off of a legacy MDM system) are where the “rubber meets the road”. Careful planning and execution can make the difference between an MDM implementation which achieves real business and user adoption, and resulting business value, and one which “sits on the shelf”.
2. Beyond the technical aspects of deployment – e.g. “parallel runs”, validation to ensure accuracy – there are also key business aspects to “master” – namely, training and change management for a wide variety of stakeholders including Sales and Sales Operations Teams, Marketing and Marketing Operations Teams, Finance, IT. Of special note is to make sure that “data steward” roles, responsibilities and reporting relationships are clear – and to determine if any of these need to change post-go-live.
3. MDM initiatives are most effective when integrated with a broad set of enterprise-wide initiatives, including but not limited to data governance, data quality management and improvement and metadata (documentation) management.
4. Adoption and value tracking and “capture” is essential post-go-live – to make sure that investments are “paying off” in terms of the right set of changed stakeholder behaviors and business value (e.g. increased sales and marketing conversion rates, reduced operating costs, improved customer experience across channels, etc.)