Business Intelligence Network business intelligence resources

Blog: William McKnight

Main | Excel makes a comeback... or did it ever go away? »

BI Trends: Master Data Management

At Scott Humphrey’s Fourth Annual Pacific Northwest Business Intelligence Summit the last weekend of July, I shared a panel with Claudia Imhoff, Jill Dyche and Colin White (covered here by Claudia as well.) Between the panel and the esteemed attendees from the who’s who and the up-and-coming in business intelligence, we covered the bases of business intelligence today and looked ahead to the future. I view it as the best barometer for the BI market.

This was my third year on the panel and it was interesting to see the transformation take place in the topics over the course of those three years. Among the hottest topics we covered, which received no mention 2 years ago, were master data management (MDM) and its specialized cousin, customer data integration. I actually introduced the topic of MDM and here’s where we got to…

MDM is an industry on the upswing due to the need for that “single version of the truth” we’ve been talking about in BI circles for a while combined with the newfound importance of information in business execution. But the need for the concept is not constrained to BI – the entire enterprise needs it. MDM can either refer to a concept of management at the enterprise level (i.e., SAP) or at the distributed data warehouse level (i.e., Kalido). It depends which level you are driving for consistency at.

Many shops find they now have a distributed data warehouse by virtue of M&A or a distributed, global business environment. Rather than combine warehouses, they look to distribute (“federate”) shared dimensions. The solution here could be MDM at the warehouse level.

More common are the approaches that treat the warehouse as just another target of the master data. Indeed, data warehousing has struggled to “close the loop” and clear its cleansed data back to the operational environment. So placing the MDM earlier in the data life cycle of a corporation makes sense.

Regardless, it is the business intelligence staff who usually take the MDM charge, just like we have done for data quality, data stewardship, and other subindustries arguably operational in nature. We should be information experts first, application second.

The trend in the BI industry now is very clearly the removing of the walls constraining business intelligence to the post-operational data warehouse world.

  Posted by William McKnight on August 27, 2005 8:55 AM |

Post a comment