BeyeNETWORK - The Global Vision for BI and Beyond

Data Governance in 3-D

Published: August 28, 2008

Jill Dyché discusses the core define-design-deploy continuum for successful data governance initiatives.

 

". . . nothing contributes so much to tranquillize the mind as a steady purpose . . . "

Mary Shelley, Frankenstein

Remember those Saturday matinee creature features in 3-D? Me neither. We’re way too young. But from what I’ve read, it was those blue-and-red cardboard glasses that helped bring eye-popping clarity to the slimy green monster emerging from the deep. That lone monster, lumbering and forlorn, is generally misunderstood, much like a term we’ve all come to know and love. That term? (Cue stark organ music.) It’s data governance!

Like many movie monsters, data governance isn’t something to dread, but rather to understand. So let’s kick back with a bucket of popcorn (none of that fake butter flavoring either!), slide on our spectacles, and a take a look at our monster in glorious 3-D. Of course, in our case, the three D’s are very specific: define, design, and deploy.

As we’ve learned the hard way, the biggest mistake a company can make when launching a new data governance program is failing to define what, exactly, data governance is. There’s no better way to gauge the vagaries of data governance than to ask stakeholders from various lines of business to explain it. You’ll hear everything from “improving customer data” to “semantic reconciliation.” Remember the scene in the movie Frankenstein where, when offered a flower by a young girl, the monster got a little too rough and the townspeople, torches aloft, attacked? It’s the same with data governance: Diverse interpretations of data governance – the lack of consensus on the definition and value – are directly proportional to its likely demise. (Cue the anguished monster wail.)

Data governance is the decision-making process that prioritizes investments, allocates resources, and measures results to ensure that data is managed and deployed to support business needs. There’s your textbook definition. But what does it mean to you? It’s not enough to just understand that data is important. I’ve written before about the “kickoff and cold cuts” syndrome, in which deli food is consumed, heads nod, and theories abound. But there’s still no clarity beyond a general consensus that data is an asset, and...well, something should be done about it. After a while, the term “governance” can become a dirty word.

Regardless of the textbook definition, your stakeholders should agree on the need, pain or problem you’re trying to solve with data governance. And this leads us to our second D: design. (Don’t worry; the mad scientist laboratory is optional.) Designing data governance means taking a detailed look at how to frame your program prior to launching it. This means accounting for your company’s specific culture and organizational structure, as well as the decision-making processes and immediate business requirements that your data governance program can address. If you’re not looking at your governance program from all the angles, it risks becoming nothing more than an intellectual exercise. Designing data governance means establishing guiding principles, building a framework, defining decision-making and execution processes, and identifying decision-making bodies, among other things.

A well-designed program involves what I call a “core team” of visionary staff members – ideally blended from both business and IT – collaborating to envision and structure what data governance looks like before rolling it out. Do this before assigning a formal council and you’re a bona fide best practice. In the design process, the core team establishes data governance success metrics. It makes the top-down versus bottom-up decisions. It circumscribes data stewardship. It chooses the initial business project on which to dovetail the first data governance effort.

This brings us to that eerie third dimension known as deployment. I don’t usually like to ruin the ending, but here’s a little secret: this is actually the easiest step. If you’ve properly defined and designed data governance, you’ll create a “closed loop” between data governance (policy-making) and data management (execution). When you ensure that the policies are clearly established and communicated, and execution capabilities are mature, the results can range from more robust data quality, sanctioned data privacy and security rules embedded in the data definitions, or double-digit uplift in marketing responses as a result of reconciled customer data. Overall, the data governance deployment process is thus is incremental, deliberate, and planned, as shown in Figure 1.

Figure 1: The Three D's of Data Governance, Plus One

Figure 1 includes an unforeseen-yet-critical fourth D, which is deliver the benefits. As many of the early adopters have learned the hard way, enacting a data governance program isn’t enough. Depending on your organization, there will be varying expectations for data governance results. Encapsulating and communicating the discrete business benefits, cost efficiencies, and innovation potential of data governance will ensure that it lives on.

Understanding the core define-design-deploy continuum gives you your own set of 3-D glasses, with which you can take a fresh look at your company’s current data governance initiative. So, are you seeing things clearly, or are you still hiding under the covers, just a little afraid to turn off the light?

If you found this article helpful and would like to receive the latest data governance insights each month from Jill Dychè, please subscribe to the BeyeNETWORK's Data Governance Newsletter.