Data Governance Next Practices: The 5 + 2 Model by Jill Dyché.
From the post:
If you’re a regular reader of this newsletter or one of my blogs, odds are I’ve already ripped the doors off of some of your closely held paradigms about data governance. I’d like to flatter myself and say that this is because I enjoy being provocative and I’m a little sassy. Though both of these statements are factual, indeed empirically tested, the real reason is because I’m with clients all the time and I see what does and doesn’t work. And one thing I’ve learned from overseeing dozens of client engagement is this: There’s no single right way to deliver data governance.
Companies that have succeeded with data governance have deliberately designed their data governance efforts. They’ve assembled a core working group, normally comprised of half a dozen or so people from both business and IT functions, who have taken the time to envision what data governance will look like before deploying it. These core teams then identify the components, putting them into place like a Georges Seurat painting, the small pieces comprising the larger landscape.
Well, I don’t have any real loyalty to one particular approach to data governance over another. And I am very much interested in data governance approaches that succeed as opposed to those that don’t.
What Jill says makes sense, at least to me but I do have one question that perhaps one of you can answer. (I am asking at her blog as well and will report back with her response.)
In the illustrations there is a circle that surrounds the entire process, labeled “Continuous Measurement.” Happens twice. I searched the text for some explanation of “continuous measurement” or even “measurement” and came up empty.
So, “continuous measurement” of what?
I ask because if I were using this process with a topic map, I would be interested in measuring how closely the mapping of subject identity was meeting the needs of the various groups.
Particularly since semantics change over time, some more quickly than others. That is the data governance project would never be completed, although it might be more or less active depending upon the rate of semantic change.
I am sure there are other aspects to data governance other than semantic identity but it happens to be the one of the greatest interest to me.