The ‘Gotcha’ Problem with Master Data Management Tools

Loraine Lawson

Without formal processes, governance, and without formal technologies and software, it collapses of its own weight. You could get an MDM system up and running, but it’s not sustainable economically or politically unless you have some processes that are aided by the software to facilitate all this political administration.

Along come the vendors and everybody gets hot and excited about this data governance word. Every vendor started putting data governance as a descriptive term for its MDM product and it became about as usable as, “New Lemon-Scented.”  It didn’t have a lot of value. Nobody would agree on what it was.

This is circa two years ago. It turns out what 99 percent of the vendors were providing was simply a data steward console. A data steward console is some GUI that the users can sit at and when something goes wrong in the system, i.e., we get some data rejected because it didn’t hit all our rules right, then we manually look at it and go, “Ah-ha, here’s how to fix it.” We ask our subject matter experts, where do we put this guy in a hierarchy of companies. We find a new rule that we need to put into our system and we’ll fix it.

A data steward console is a good thing, but it wasn’t the data governance that organizations really needed. It’s all reactive. It’s passive. It’s not necessarily proactive or integrated data governance, okay? It’s what we call passive data governance.

My customers — who are global 5000, $1 billion or more in revenue companies — their IT organizations say, “We need active data governance. We need something up front that will let us design the system, manage it, do continuity and maintenance on it so that it can be sustainable. Otherwise it’s going to, again, collapse of its own weight … as  manual batch processes, uncoordinated, don’t work.”

So other vendors started saying, “Well, we have active data governance. We have a design tool. We can design your MDM hub.”

But what they didn’t tell you was that after they designed the hub, it didn’t do anything with the design. It didn’t integrate with the MDM hub. All it did was take your processes, your people, your data governance councils, committees, whatever, and all their deliberations and say, “Here’s where we get this field. Here’s where we get that field. If we get a collision on two updates, this is how we resolve it. How we back it out or go forward with the update.”

But then it stops. It didn’t generate the actual MDM rules that would do the data quality management within or relationship management, et cetera, within the actual MDM hub.

Every vendor had something that was a front end, but it wasn’t integrated, and that’s where we wanted integrated active data governance. We want something that actively designs a system and that’s proactively integrated, that actually takes those rules that we determined for our data quality, our data governance, and places them in the hub and manages their execution.

We like to create cute labels because that’s what analysts like to do. So we said, “Okay, what is it when somebody says one thing but does something else?” That’s “passive aggressive.” We’re going to say, “Okay, there’s another class of vendors — a lot of them, actually, most of them, who say they are active data governance, but really they do the other thing, passive data governance.  So we’re going to call them passive aggressive data governance.”

Now the good news is IBM has something called Data Governance Policy Manager, Oracle has Data Governance Framework, Informatica has the Informatica Data Director.  SAP has SAP Master Data Governance, MDG, as they call it. And they have it very specifically well tuned.  SAP has MDG for finance, MDG for vendors, MDG for materials, and MDG for customer.

Lawson: So most of them, have they corrected the data governance technology? Are they integrated now truly where they can do more than create a process of governance?

Zornes: Not to my customer satisfaction, no. It’s gotten even messier. That’s just the data governance picture, so let’s take it to another level.

MDM is going to be pervasive. You hear vendors talking about MDM everywhere. MDM is the foundation of every application package that’s coming out now. SAP, when it releases Business by Design for Aerospace or for Chemical, underneath it is SAP MDM and a bunch of other stuff. When Oracle releases Fusion CRM, underneath it is the Fusion MDM product. When Microsoft releases Dynamic CRM, underneath it is the Microsoft Master Data Services. When Teradata releases their Aprimo Database Marketing package, underneath it is the Teradata MDM solution.

That’s great, because it lets the vendors deliver value and evolve their packages a lot faster than in the past. That’s the good news.

The bad news is none of us buy our software from one vendor, right? Some companies are wall-to-wall SAP or Oracle, but they're rare. So most of us have some SAP, some Oracle, some IBM, some Teradata. So each of those packages that we bought or were upgrading to, came with its own MDM.

And guess what? Each of those MDM hubs brings with it its own governance. So it’s like what we now call “the Russian Babushka effect”: a doll within a doll within a doll.

And to make it worse, or better, however you want to look at it, you know, Oracle gives away DRM, their reference data solution. They bundle it into all of the deals that they do, so it’s basically given away. Microsoft bundles in their Microsoft Master Data Services, their MDM product that they bought and repackaged. Microsoft gives away theirs with SQL Server 2008 release too, it’s bundled in.

So you have a bunch of MDM coming in through the back door, through the finance department, the HR department, engineering — whatever, where they've got free MDM and they're building their own solutions because MDM is a powerful tool. But it’s yet a different brand probably than what they're using in central corporate or enterprise IT.

By the way, Oracle’s not going to use IBM’s MDM and SAP is not going to use IBM’s, you know, they're all going to use their own MDM, right? That’s the way of the world.

MDM was supposed to simplify and modernize all that for us, and it did. But the gotcha is that we don’t buy it all from one vendor and as a result, we’re now getting multiple MDMs and at the same time, each of these MDMs is bringing its own governance with it. So how do we get some governance to sit on top of all these MDM hubs and orchestrate it?

Add Comment      Leave a comment on this blog post
Feb 15, 2013 12:50 AM FX Nicolas FX Nicolas  says:
Excellent analysis! I think you point out the huge gap that exists between Governance and actual MDM implementation needs. Adds this to fast changing requirements and you get a deadly combination of factors. Reply
May 7, 2014 5:27 AM Willi Santiago Willi Santiago  says:
Hi- Excellent review. I would also add that Portfolio must be on board. No MDM program is successful unless the Application Portfolio is using it. Often Application owners have little incentive to spend $ integrating with an MD system when they have a working local database. Their immediate customers aren't concerned with Enterprise Data management and quality. The executives have to be behind the MDM program and realize that "if you build it, they will come" doesn't really work. (I could say the same for SOA systems) Cheers! Reply

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.