In which Jill repeats the question.
I do a lot of blogging and webcasts and on-line stuff. I'm on Facebook, Twitter, and I've even been known to Blip on occasion, but that's another story for another day. All that said, I really do prefer in-person events. There's nothing like speaking to an audience of engaged individuals who actually got in their cars or hopped aboard airplanes to learn and share ideas. Yes, sometimes they're e-mailing or tweeting as I talk, but I like to think that they're simply relaying the information they're learning to colleagues who couldn't join them. In any case we get to see the whites of each others' eyes, engaging in data debates over our lunchtime chicken parm and hanging out in the hallways doing what professionals do when they share the common challenge of hard work.
So I'm having a ball with the folks at Informatica and Initiate Systems with the MDM Forum, an 8-city road show we're doing together this spring. You should plan on coming. But if you can't swing it, here are some of the more interesting questions audience members asked in our Chicago and Columbus events. And, of course, my answers.
Question: Why can't my data warehouse do MDM? We have metadata. What's left?
Answer: Lots. There are different functions within the MDM stack--some of them might be supported by your data warehouse. But if you designed your data warehouse to support analytical applications with structured, cross-functional data, odds are some of them aren't. Business rule support, real-time matching, merge processing, data validation, data transformation, source system interfaces, workflow, and other functions aren't normally packaged within data warehouse solutions. But many are nevertheless native to MDM tools.
Initiate has a new white paper available that delves more deeply into why MDM is different than data warehousing (and CRM, and EII, and other solutions). You can download it here.
Question: Should the BI team own MDM?
Answer: It all comes down to company size. The larger the company you work for, the more likely it is that MDM will be run by a separate team of individuals who are charged with availing data as a service to various operational systems. The greater the number of business applications and systems a company has, the more likely a discrete and highly-skilled MDM development team will be necessary.
That doesn't mean that in large corporations the MDM won't interact or depend on the BI team (and the Data Governance Council and the IT Architecture Team and other functions within the company). Like most strategic initiatives that involve technology, MDM should be collaborative.
Indeed we're finding that in many of our clients MDM efforts will start in the data warehouse or BI group. (Shout out here the current clients whose BI teams are creating MDM Masterplans.) After all, they know the data and are often the only group that truly understand the drama that is integrated data.
Question: Is MDM destined to become one more of the many big-deal projects that starts with a bang and ends with a whimper?
Answer: You tell me. If you always did what you always did, you'll always get what you always got. (Got that one in Weight Watchers.)
Does your company have Bright Shiny Object Syndrome? Then hurry up and find your next semi-informed executive sponsor, convene a kickoff meeting before you're ready, define your requirements through your data model, tire-kick randomly-selected vendor products, guess at data governance, and generate lukewarm support from stakeholders who really aren't. Or decide on a staged and deliberate approach to MDM and watch your project--and by extension, your job--become foundational.
We're in the Twin Cities and Atlanta this coming week. Hope to see you there!
Posted April 13, 2009 6:00 AM
Permalink | No Comments |