When it comes to succeeding with data quality, you might gain an edge by avoiding a centralized approach, argues one data governance director.
Alan D. Duncan is the director of data governance at the University of New South Wales, Australia. In a recent MIKE 2.0 blog post, Duncan reacts to a survey finding that a “lack of centralized approach” is linked with inaccurate data. He questions whether it’s really lack of centralization or actually a complete lack of any structure.
Duncan’s premise, as he explains in some detail for InformationAction, is this: The social and cultural character of your organization should shape how you handle data governance. That means there will be a many different ways to structure governance, but broadly speaking, he identified three:
- Centralized governance for individual business units that share common processes across the company
- Federated governance for orgs like universities, where the business units might be geographically diverse, but have “the same overall functionality and responsibilities”
- Distributed governance for companies where business units operate independently
In a recent post, he extends his reasoning to data quality, except he argues that a distributed approach that relies on end users provides more advantage (right now).
He gives four good reasons, but this stood out to me: Pushing data quality to the end users is much easier to do when you don’t have the money or the support for a strong, decentralized approach.
It’s hard to argue with the idea that accomplishing something would be better than accomplishing nothing. Also, other experts agree that data quality should “belong” to the business, so from that standpoint, it makes sense.
Although… it does make me wonder, where are these places that end users care more about data quality than executives?
The reality is, as Duncan notes, information is getting darned hard to manage and the information has a lot of moving parts. It might help to remember that in the end, you’re probably going to need both a top-down, centralized approach, as well a bottom-up approach — basically, you need data quality and data governance to be “all around” and comprehensive. That’s the big Nirvana of data management.
But you also don’t want to spend so much time on the how that you never do it at all. OCDQ blogger and data quality expert Jim Harris calls this the “buttered cat paradox,” where organizations are so unsure where to begin that they perpetually spin, yet accomplish nothing.