Master data management (MDM) spent the year in the Trough of Disillusionment, which is sort of like that fourth day of school being cancelled due to snow: The harsh realities set in and you realize how much boredom and work it’s creating.
But, the good news is that this is when we start to see what’s working and what’s not.
In both the Magic Quadrants for product and customer data management this year, Gartner included some interesting findings about how organizations implement MDM.
Gartner generously calls these “key themes and ideas,” probably because they’re talking about Gartner clients and don’t want to annoy them.
I have no such concern, so I’m going to cut to the chase and call them what they are: the most common mistakes made with MDM in 2013. I’ve added in a few more MDM mistakes mentioned elsewhere to round out my top 10 MDM mistakes of this year.
- No proper ROI analysis before investing in MDM software. Instead, it seems many organizations just reported that the software was part of a “strategic decision.” That, and $4.50, will buy you a medium coffee while you search for a new job.
- Poorly scoped “add-ons.” Many IT organizations have added data and functionalities to the MDM hub that should’ve been added to another business application instead. Gartner is careful to source this mistake to an “increasing number of end users,” which leads me to believe business users think IT messed up.
- Too little governance or stewardship by the business. It’s a classic data quality mistake: Allowing IT to control and manage the data rather than involving the business in governance.
- Failing to consider the business’ and end users’ needs before buying a solution. “A notable number of SAP and Oracle clients did not conduct a competitive analysis before adopting that vendor’s MDM solution,” Gartner states. “This approach could lead to end users discovering later that the adopted solution does not meet their requirements.”
- Neglecting security. End users also confessed to thinking about security “far too late.” As a result, they had to back track and assess security as an information policy.
- Relying on MDM superheroes. If organizations failed to do the head work on MDM, why didn’t it fail? Often, one or two “citizen stewards” or “citizen governors” stepped up to save the day. These key people saved the implementation, according to successful end users. The problem, Gartner points out, is that house of cards is likely to fall when the MDM superhero leaves the organization.
- Starting MDM within IT. This wasn’t mentioned in the Magic Quadrant reports, but Gartner analyst Andrew White did bring it up in a Supply Chain Quarterly article. Frankly, I think this alone would solve many of the mistakes already mentioned.
- Making governance a full-time job. “If it is, then the organization is focusing on the wrong things,” White warns.
- A related problem: Making governance a full-time job for someone who already has a full-time job. “The role of data steward, by the way, should not be an onerous one,” White wrote. “If it is, then the organization is focusing on the wrong things.” How much time governance takes will vary by organization, but think more in the 10 to 20 minutes a week range, rather than hours.
- Building in MDM rules that are so restrictive, end users can’t do their jobs, so they find or create “work arounds.” If you have end users using tricks to get out of the restrictions you’ve put in place, then it’s time to change the process to support compliance, not hinder work, White says.
“Six Dos and Don’ts for Successful MDM and Data Governance” offers more tips for finding success with MDM.