It is clear that customers recognize a good thing when they see it. Some forward-thinking companies have realized that they can apply MDM to more than just customer master data. All shared data such as financial hierarchies, product information, HR, assets or cross-function reference data can benefit from master data governance. These initiatives are often much broader in scope than customer data integration because they impact a variety of business teams across the company and require active collaboration.
My favorite research note from Gartner this year is the one called “Different Approaches or Usage Patterns for MDM and Their Implications,” published in April by John Radcliffe. He describes two styles of MDM and explains why they are more complementary than competitive. Basically, the first style is a transactional hub, designed for B2C customer data. It is focused on real-time integration and data “match/dedupe,” but lacks governance capabilities, rich data modeling and business rules.
The second style described in Radcliffe’s note is called “Centralized authoring with workflow.” This name is a bit reductive, but it clearly illustrates that companies are looking for active master data governance. They want to use MDM as the main point of entry and/or governance for master data, and they want business users to be able to collaborate through a central user interface. That means they need highly sophisticated features such as workflow, authoring, hierarchy management, CRUD, version control, time management, security, semantic data modeling and business rules definition. However, when they open MDM to a wide population across business functions, it is imperative that they have an easy-to-use user interface that requires near-zero training. When that is available, MDM no longer needs to be the sole purview of a select few data stewards.
I am not alone with this line of reasoning. For some time now, Rob Karel of Forrester has also been promoting this style of MDM, focused on collaborative authoring and a model-driven approach. He calls it a process-driven approach to MDM.
The most recent Gartner Magic Quadrant on MDM for Customer Data highlights large players like IBM, Informatica and Oracle. While some of them claim to be “multi-domain” or “universal,” they still rely on relational-based technologies without any sophisticated data governance capabilities. That means that when it comes to data governance, you have to do it yourself. As the larger vendors try to increase their MDM message, solutions become more confusing, as explained in this article by Mark Brunelli.
In a previous blog post, I wrote about how solutions formerly know as CDI (Customer Data Integration) have structural weaknesses when applied to master data governance. They don’t provide any model-driven, semantic capabilities so it is impossible to generate sophisticated user interfaces or handle complex data life cycle patterns (past and future versions, inheritance, etc.). Adding third-party workflow tools is not enough.
The complexity of vendor offerings makes selecting an MDM platform more difficult than it should be. But starting with a transactional hub (a.k.a CDI) could mean embarking on a risky journey since it will be difficult to apply it to data domains outside of customer and other governance patterns.
Perhaps then, it would be best to build a foundation for master data governance based on the data that is common to, and most often shared by, the enterprise and then connect and deploy additional solutions / architecture patterns for specific data sets as the need arises.
Our customers are among those forward-thinking companies I mentioned earlier. They deploy our EBX5 MDM software as a corporate standard for reference and master data that requires strong governance. They manage corporate functions’ data (finance, HR, risks, procurement), products and services (from R&D to sales and supply chain), assets, reference data and some parties such as suppliers, B2B customers or partners.
In some cases, they complement their MDM platform with a specific solution (such as a CDI hub for the B2C customer master) or use our new relational mode for a hub-type implementation. But these are still exceptions in terms of MDM use case and architecture. In fact, those siloed MDM solutions are connected to EBX5 as consumers of some of the core corporate hierarchies and reference data.
It will be interesting to see what 2012 brings regarding the MDM space. One thing is clear, even without a crystal ball, we need to listen to our customers. They know what they need. Our job is to provide just that. I look forward to discussing it with you and our customers during the upcoming Gartner Master Data Management Summits in London or Los Angeles where Orchestra Networks will be a platinum sponsor.