Oracle DRM Blog

In the old days, Master Data Management are much smaller in scope. It’s often just a bunch of scripts doing data quality checks or hierarchy management. Unfortunately the problem got bigger because of the complexity of businesses and the massive volume of data that are hard to mine. Imagine doing a dashboard of the top 10 customers. How do you categories one of the top 10? You see, the customer record probably originated from the Order Entry system with Ship To and Bill To details. So which one do we use for measure? Also if we consider “Dell” as a customer, how many countries do they operate in? And what if 20 of these exist in our reporting system with different spelling etc. And finally, how can you distinguish one company as a subsidiary of another, to ensure you consolidate and report the revenue properly? With these issues in mind, an array of solutions categorized by TDWI have been built by different vendor to tackle these problems.

Operational MDM
Upstream in the general flow of data, one or more MDM solutions are built into and/or used to integrate operational applications for enterprise resource planning (ERP), customer relationship management (CRM), supply chain management (SCM), financials, and so on. Since the applications don’t change frequently, master data and entity definitions don’t either. Furthermore, these applications and their business processes support transactions and sometimes operate in real time, so operational MDM must, too.

Analytic MDM
Downstream in the data flow, data warehousing has long involved some form of MDM, because of the balance between tracking data lineage (to ensure you have the right data) and repurposing data to create new structures (like aggregates and time series). Entity definitions change often due to data discovery, analytic business modeling, and other iterative practices. Analytic MDM is also seen in practices that resemble data warehousing, like customer data integration (CDI) and financial performance management (FPM).

Enterprise MDM
Today, MDM is practiced mostly in isolated silos or with a short list of applications that don’t step beyond the bounds of either operational MDM or analytic MDM. However, some organizations have moved to the next level with enterprise MDM, which is an autonomous infrastructure that can integrate master data across multiple IT systems and businesses. Spanning the whole data flow is daunting, because enterprise MDM must satisfy the diverse requirements of both operational MDM and analytic MDM. Yet, enterprise MDM is a worthy goal, because it extends beyond IT silos and organizational boundaries the general benefits of MDM, namely: well-designed entity definitions applied consistently.

The next time you contemplate about doing a Master Data Management project, make sure you don’t just think about writing a bunch of SQL scripts because the solutions could be much more comprehensive and with alot more impact to keep in mind.