The Real Value of MDM Lies Beyond a Single Version of Anything

Loraine Lawson

Analysts and vendors talk a ton about this Holy Grail-a single version of the truth-when it comes to master data management, that it's just a given that this is a worthy and even achievable goal.


Technologist Robin Bloor raises worthwhile questions about this and other data-integration assumptions in a recent blog post appearing on a data-integration vendor Pervasive's blog.


It's a fun, worthwhile read. He starts out at a highly philosophical level, going back to Plato's classic argument that there was somewhere an ideal tree that defined tree, and works his way back down to the world of business intelligence and the equally impossible task of achieving one definition of "customer."


Those of you with liberal arts backgrounds will love this, of course. But he's making a larger point that's worth consideration by the very business-minded and tech-minded among you, so I'd urge you to stick with him through the philosophical discourse.


The questions he pulls out for discussion are:


  • Is complete data integration desirable?
  • Is complete data integration possible?


For me, the big question he's tackling here is whether the real goal is a single version of a customer and master data management's ability to actually deliver it. Ultimately, he argues, the value isn't in achieving a single version of customer, but in achieving "a data rich version of the customer":

And if we successfully unite a wide variety of data sources without encountering contradictions, we will get a very rich set of data around any given entity, whether it is a product, customer, business partner or whatever. But, if we gathered all the data surrounding any given entity, we would never have reason to use it all at once in any given context. The benefit would not be in having a single version of customer but being able to generate many different views of customer.

It's an excellent point, and arguably much more valuable to business users than any mythical data Holy Grail.

Add Comment      Leave a comment on this blog post
Sep 7, 2010 6:48 AM Paige Roberts Paige Roberts  says:

Robin really knocked it out of the park on that post. "Unambiguous and data rich" version of any entity is far more valuable than "single version." Puts master data management initiatives in a whole new light.

Thanks for the insight on it.


Sep 8, 2010 11:12 AM Jake Freivald Jake Freivald  says:

Great post. I hadn't seen Robin's blog entry, so thanks for pointing it out.

In reading the part that you quoted above, I was struck by its similarity to the way some people used to talk in similar ways about data warehouses: Warehouses were the "single view of the truth", data marts were created from them so you could have single-subject views of the same data. It sounds like, in Robin's opinion at least, we're trying to get the same sort of benefit from MDM. That raises the question, though, of how to be more successful with MDM than we were with large data warehouse / data mart infrastructures.

I think Robin misses one of the complexities of any data integration process when he says that metadata describes data rather than holding it to rigid definitions. Metadata assumes or implies an ontology, a set of objects that describe how the world works, and if two different systems have different ontologies then the metadata itself will be contradictory, too. That's part of what made data warehouses hard to reconcile, and it makes MDM complex (though not as bad as data warehouses) as well.

Sep 9, 2010 7:00 AM Tas Ranson Tas Ranson  says:

I think this just shows a misinterpretation of what data management is about.  It is not about getting all the data in one place, but about having no contradictions in the data.  One version of the truth does not mean a sinlge place to store data, it just means that all data stores are consistent.  In simple contexts this means one master source and any number of slaves, but in reality bidirectional syncronisation of data is a very achievable, so that doesn't limit us to one master.  We just need to know which sourcess are authoritative, and configure accordingly.

Sep 15, 2010 8:46 AM m ellard m ellard  says:

Fun for the liberal-arts minded among us! A thought or two -

From a customer-service perspective, let's say the Jane Doe has a problem with her purchase online. She gets a hold of customer service and - if she's not been combined into one record. She may be a VIP customer in the store and get average-joe treatment (or worse) when she complains through online channels. The whole experience leaves a bad taste, and the in-store relationship becomes damaged.  In this case, it boils down not to wanting to understand the customer to be more effective pro-actively, but instead, re-actively - to know what the customer believes you should know about him/her.

This also calls to mind the story of the blind man and the elephant - he separately feels an elephant leg, an ear, the tail maybe - and, if his sense of smell is not so keen (or he's not smelled an elephant before), coming at the elephant from these disparate angles may not enable him know he has an elephant.

Take Jane Doe again. She decides that she loves shopping online and suddenly goes to the store less. But she buys the same amount of stuff. She should still be a VIP (an elephant), but it takes that cumulative data to see that.

I'm not sure a single view is just a slogan - though it's certainly a lofty ambition - you may never use all the data at once - but having the ability to do so has tremendous value.

Sep 17, 2010 1:17 AM Lindsey Niedzielski Lindsey Niedzielski  says:

Great post Loraine. Thank you for sharing Robin's distinction between an ideal customer and a data rich one. We have a community for IM professionals ( and have bookmarked this post for our users.  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.