Six Steps to MDM Success
Steps you should take before embarking on a master data management initiative.
My daughter learned a long time ago not to ask, "Are we there yet?" on long car rides, because when she did, no matter where we were, we'd respond, "Absolutely. Let's all get out of the car."
She's going to need a lot of therapy.
Anyway. I'm beginning to think that's what vendors are doing with multi-domain MDM - the capacity to support multiple types of master data, such as customer, product and financial master data. Vendors say they're already there, while analyst groups counter that they're very much not.
What's interesting about the article is it actually quotes two conference attendees, an IT manager and a data quality manager, rather than analysts. Based on their own experience with MDM, both immediately "got" how hard it would be to move from mastering one data domain to mastering multiple data domains.
Paru Mahesh, the data quality leader, specifically thought the integration between the transactional systems and the multi-domain MDM would be challenging.
"How do you build those insights? And how do you model the data so you can grow those insights," he asked.
Multi-domain MDM isn't the only issue tripping up the industry. Recently, experts like Gartner and Aaron Zornes, founder and chief research officer at the MDM Institute, have offered these criticisms of MDM deployments:
Organizations are tackling too much. In an effort to steer people away from the idea that MDM is a tool you buy, analysts and, yes, journalists often call MDM a discipline. Well, that seems to be backfiring, too, as companies are becoming overwhelmed by the idea of creating an enterprise-wide "discipline."
To combat this, Gartner suggests companies focus on building an MDM program. The theory seems to be that discipline is just too big and all encompassing, whereas programs are built one step at a time around business goals.
What this means on a very practical level is you shift from saying, "We're going to fix data quality" to "We're going to de-duplicate our customer data by 60 percent," the article noted.
Companies are building MDM in silos. On the flip side, some companies aren't going broad enough with MDM, and it's being deployed in separate hubs for managing different sets of master data. That in and of itself isn't a problem - but then no one is linking these hubs together. That's creating "master data silos," warns Zornes. And that further contributes to MDM problem number four
No governance. Zornes blames vendors for failing to add tools to MDM solutions that support governance. He says this is forcing companies to ad lib and rely on manual data governance. You can imagine how well that's going.
It's a problem that could be addressed by the technology, he contends, if vendors added workflow and business process management to link data governance and MDM.