In my final installment on the 5 Critical Success Factors for Initiating Master Data Management, I want to discuss why tackling Operational MDM is valuable, and when to do it.
A major contributor to disruptions in MDM projects is lack of stakeholder agreement with respect to what the team is trying to accomplish with MDM. It’s important to get everyone on the team clear on the two purposes of MDM.
The first purpose is to facilitate better reporting (Reporting MDM). The goal with Reporting MDM is to gather and aggregate data (sales, invoices, purchase orders, etc.) such that in an Enterprise Reporting Context, all of the data is included from a set of source systems. A Reporting MDM system does this by Matching Master Data records from each of these systems. Then, it provides Views of these matches (groups of master records) to subscribing systems and users to consume and use for their own purposes. It sounds simple, and in fact, it is pretty simple.
The second purpose is to improve the overall data quality in each operational system (Operational MDM). The goal with Operational MDM is to ensure that each representation of the same “thing” (e.g. Vendor) is the same in all systems which house master data for that “thing”. An Operational MDM system does this by Matching Master Data records from each source and then Harmonizing the records (i.e. makes all of the master records in a group “line up”). Finally, it Distributes the harmonized data back to the source systems. Imagine knowing that all representations of your most valued customers, are verifiably represented in a logically consistent way in all of your AR systems.
Visually, an Operational MDM synchronization process might look like this.
Once, we have those two concepts solidly understood, the question becomes: can we have both? Yes, you can have both. However, if delivering value to the business quickly is a consideration (it should be), I recommend that you tackle Reporting MDM first. Reporting MDM has fewer technology hurdles, initiates a Data Governance program, and delivers real value quickly.
Here is what Operational MDM will take:
- A Data Bus – you’ll need a integration solution which can handle connections to all of the LOB systems which you want to synchronize. My team uses Microsoft BizTalk Server for this.
- Subject Matter Expertise – you’ll need access to the people who understand the target systems extremely well. Often they will need to expose API’s to the MDM team, so that synchronization can be “real-time” (a change is made to MDM and the change event propagates to all of the affected systems)
- Business Process Review – your Data Governance team will likely need to consider the full lifecycle of the master data- creation, maintenance and archive.
In summary, Operational MDM is achievable and yields tremendous value. But first, build the foundation and “put some points on the board”. If you build a Federated Data Model, Keep MDM Separate, Flip the Script and Formulate your Governance Plan, Phase 1 will be successful, and you’ll get funding for Operational MDM in Phase 2.