Why You Shouldn't Rush to Define 'Master Data'

Loraine Lawson

How do you define master data?


In some ways, it's easy to dismiss this as a philosophical debate for analysts, with little bearing on reality. And certainly, as past posts and the readers' comments have shown, even analysts have a hard time giving you precise guidelines for nailing down master data or even agreeing to a definition.


"Who cares?" You may think. "I'll define master data however I want to."


You certainly wouldn't be alone in thinking that.


But David Loshin believes that's a mistake. In fact, he thinks it could be creating problems for organizations and their MDM initiatives, leading to practices that he calls weird, and I call a recipe for no ROI and MDM failure. Here's his list of signs you might be on the wrong path:

  • An MDM business case that's justified on supplementing another technology, often SOA.
  • Small starts and mulitipe MDM hubs across different divisions, all pursuing the same concept (talk about ironic wheel spinning!).
  • Focusing on master data quality-after you've already deployed your master data repository.
  • Trying to solve master data hub problems after-the-fact with data governance.


I'm paraphrasing here, but that kind of wrong-headedness you can avoid if you take the time to think through a definition for master data, according to Loshin.


I'll be honest, this isn't a quick and easy fix he's offering, but you can tell he's thinking deeply about the topic. He promises this won't be the last piece on it, either. So bear with him, because I believe while he's working through this, he's hitting upon some very useful signposts that will help you find your way and perhaps even wind up with a more efficient route to successful master data management.


For instance, he offers some traits that you can look for when you start on this process of determining what you want to cover with master data management. He then helps you narrow the field with a checklist of four items, any one of which means you're dealing with something that qualifies as master data.


But it's the next list in the post that may prove most helpful, because it's a rundown of the types of hard questions you should address about the master data so that you're adequately defining it, agreeing about it across the organization and-most importantly, perhaps-so you know how to govern the data. That means, among other things, knowing who's responsible for defining it and governing whether it meets the requirements for master data.


Look, no one ever said master data was easy, and despite what advertisements might say, an MDM solution is only a small piece of the puzzle. Master data takes work, discipline, and like all things worth doing, it takes forethought. Do yourself a favor and don't underestimate this step. MDM's too expensive and time-consuming to rush.

Add Comment      Leave a comment on this blog post
Oct 4, 2010 5:09 AM David Loshin David Loshin  says:

I don't recall actually using the word "weird."

Oct 4, 2010 5:33 AM Loraine Lawson Loraine Lawson  says: in response to David Loshin

My mistake. The word was "strange."

"I suspect that is the reason for strange behaviors such as these..."

Oct 6, 2010 1:58 AM Lindsey Niedzielski Lindsey Niedzielski  says:

Great post again Loraine. I really like your emphasis on "taking time" to define MDM, so many people forget to do this.  Once again, we have a posted your work on our community for IM professionals (www.openmethodology.org).  Look forward to reading your work in the future.


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.