Blog post

What’s Happening in Master Data Management (MDM) Land?

By Andrew White | October 05, 2017 | 3 Comments

One of my colleagues shared a short note published by a vendor on LinkedIn that commented on how Gartner’s views the current MDM market.  Of course I had to read it and in reading it I saw an opportunity.  Firstly all vendors have a unique view of the market.  That is useful and important – but it is only one view.  I myself worked for a vendor for 9 years before joining Gartner so I was very familiar with this idea.  But once I joined Gartner, and now after 16 years, I appreciate more fully that what we see is different.  Our lens tends to be much broader and much wider since we touch to many more organizations, more regions, more users, than many vendors added together.  So i comment on this vendor’s point of view not from any negative criticism but from an alternative perspective.

Ramon Chen (Chief Product Officer) of Reltio noted in his post ( that MDM is positioned in our recent Information Governance and MDM Hype Cycle in the “trough of disillusionment”.  If you read the note you will note however that technically MDM is out of the trough and just nudging up the first parts of the ramp up to the “plateau of productivity”.  But let’s not quibble about that.  What was interesting were Ramon’s observations on the MDM market.  Here are his comments and my initial thoughts on his comments

  1. The landscape is littered w/failed projects that use legacy on-premises MDM tools that are 12+ yrs old

Partially true.  There are many failed and struggling MDM implementations.  The part that is not totally true is the age of the technology.  This is because the main reasons for the failures have little to do with technology.  They range from:

  • Tackling too much data at the same time.  There is huge confusion in the market about what is master data and what is not (e.g. application data)
  • Assuming MDM was a ‘data’ program whereas those that know how to ‘do” MDM learn quick that it is about business process/outcome improvement, much less about data per se.
  • Not aligning to an (business) outcome – which is most easily noted when IT says it is “doing MDM” and business leaders are not even involved
  • Not establishing effective (as in right-sized) governance (too many folks, too big a program, not designed to operate at the speed of business)
  • Not even attempting operational governance (i.e. information stewardship)
  • Conflating MDM with a data quality tool or project or an entity resolution workload or a data matching effort or relationship discovery.  All are are useful.  None ‘are’ MDM; they may contribute to it.
  1. Gartner continues to ignore successful Cloud MDM deployments happening today

Not at all.  There are just few of them.  Also be clear about the last response to the first question.  Just doing entity resolution or graphing or discovering some relationship concerning customer data is not MDM but helpful to those implementing it.

  1. Coincidentally the companies they recognize as MDM leaders w/12 year old tech aren’t truly multi-tenant cloud enabled

What has that got to do with a business outcome?  Multi-tenant is an important architectural aspect for vendors only.  It actually removes one of the cost-penalties that vendors incur in the cloud.  Single tenant SaaS leads to higher costs to vendors and they would then pass this on to clients.  With multi-tenant, costs are lower to the vendor and so costs to clients can be lower.  This has no impact on the quality of your MDM program.  In fact, there is a small argument that multi-tenant apps would lead to a standardization of the business processes the apps would support.  Thus the value of MDM might be more easily achieved by your competitors using hte same MDM solution.  But that’s a nit.

So there you have it.  What do you see “out there”?

Comments are closed


  • Ramon Chen says:

    Hi Andrew,

    Thanks for taking the time to start this discussion based on my linkedin post. I agree with you that failures are not all to do with the age of technology. I would however respectfully suggest that multi-tenant cloud is not just a vendor cost benefit. It’s about agility, ala no impact upgrades 3 times a year to our customers, vs. the long upgrade cycles and large migration projects that take place every 18 months, that hinder businesses’ ability to respond to competitive pressures, and market conditions. The use of a logical data model (and the polyglot capabilities of using NoSQL, Graph and document stores), vs. fixed relational is similarly about being able to be agile, and to scale to internet-scale volumes elastically in the cloud. Contrasting with how challenging it is to create a relational model that will stand the test of time, and requires multiple changes often before the project is deployed.
    Thanks again for replying and posting here, I believe it’s an important discussion point, and as ever, there are multiple points of view. Companies should select what’s best for them. My LinkedIn post was to surface that MDM alone has not fulfilled its potential, and that there are opportunities today to look beyond.

  • Matt Gagan says:

    Thank you for the post Andrew. Focusing on your reasons for MDM failure: most of these I tend to agree with. A couple of comments, though:

    1) Re. your first-listed reason: Master data, or not Master data:

    We can constrain ourselves to defining MDM in the same way that folks have limited themselves over the last decade or so, basing our definition of master data on the limitations imposed by the serving technologies – meaning, “if there’s too much of it, or if it changes too quickly, I can’t handle it, so it’s not master data”.

    …Or we can ask why MDM solutions even exist – what use are they? The answer is often to be able to support a 360º View of the Customer/Product/etc. To that end wouldn’t it be better to consider a solution where the master data entities themselves, as well as a relevant set of application data, are brought together into a solution where state-of-the-art matching and merging take place, resulting in an operational data store containing the reliable set of necessary data to fulfill the 360º View, delivered in context, via a modern web UX?

    2) Re. your third reason: ‘Not aligning to an (business) outcome – which is most easily noted when IT says it is “doing MDM” and business leaders are not even involved’:

    This is exactly what a modern data management PaaS solution addresses, by being the seamless gateway to data-driven applications, including as described above.

  • Andrew,

    While Ramon makes some interesting observations, I couldn’t agree more with your points about MDM being about business outcome much more than data. I have seen many projects go by the wayside because the team lost perspective along the way; what clearly started as a strong business initiative to address a pressing business need, shortly became yet another data project. By the time the project was eventually delivered, there was a substantial mismatch.

    To be clear, every company today is in the data business. Whether they manufacture cars, bake cereals, sell shirts, they all depend on the availability and accuracy of data. It is a necessity, it is a matter of relevancy, it is a matter of survival… Yet, I would recommend that all of us practitioners bring a different viewpoint to the table stating it is not merely about data, whether consolidated or mastered; it is about managing outcomes, not IT plumbing, powering business users to shape their future.

    If MDM is to solidly establish itself in the “Plateau of Productivity,” we vendors must take the lead bringing in a business-first mindset to our customers and guiding them all the way.