Thoughts From the Floor of the Business Intelligence Summit
By Robert Barker, Infoglide Senior Vice President & Chief Marketing Officer/Acting Chief Technology Officer
Infoglide Software’s Bob Barker is at Gartner’s BI Summit in Chicago this week. Here’s his report:
As Gartner’s BI Summit moves ahead this week, it’s interesting to note how thoughts about master data management (MDM) and business intelligence (BI) interact. It’s apparent that MDM is high on the “hype cycle” right now and, given the power of its main premise, rightly so. Yet it’s also true that current levels of Chief Information Officer (CIO) interest in applying powerful BI concepts to improve business processes have never been higher.
Master data represents a critical subset of your data, according to Gartner’s Andrew White. It’s the part that’s critical to daily operations - customers, products, etc. - and that demands a certain level of management for accuracy and uniformity if good operational decisions are to be made. The “right data” must be available if operational systems are expected to produce consistently good results.
That said, I worry that both proponents of MDM and BI are overlooking an important subset of problems - you know, the ones we write about here all the time. They’re the ones that, while sometimes benefiting from MDM and BI improvements to data quality, often draw extremely valuable information from the “dirt” that MDM efforts remove.
For example, say we have two records:
Robert E. Lee, 345 Main St, Anytown, US 12345
Robbie Leigh, 345 Maine Street, Any Town, US 12345
If we were just cleansing and integrating the data, those records would probably be seen as being the same entity so they’d be combined. But what if Robert E. Lee was a witness in an accident that resulted in an insurance claim, and Robbie Leigh was a chiropractor that provided services for an injured party in another insurance claim? Then, the fact that there are two distinct but seemingly related records starts to get interesting.
What if we cleansed the data and combined these records and then there was a subsequent claim with Bobbie Lee, 345 Mane, Anitown, US 12354 as the driver? It probably would have been more valuable to have kept the original records and linked them together. Having all three separate, “dirty” records would be important and valuable.
This process of identity resolution (or entity resolution and analysis) is starting to be recognized as an important subset of business intelligence. Gartner just recently added it to their chart of BI technologies. It’s just in the beginning stages of adoption though and has not hit the hype part of the curve like master data management. As I review the agenda for the BI Summit, there are three sessions on MDM, two on data integration, one on data quality, but identity/entity resolution is nowhere to be found.
MDM and data quality initiatives deliver significant value, but let’s remember to take advantage of the insights that inconsistent and ambiguous data can hold. CIOs and analysts are starting to realize the value of entity resolution technologies, and I expect the next BI Summit will have an item or two on the agenda that discusses this next step in the evolution of business intelligence.
Subscribe to this blog