Data's Conundrum: Everybody Wants Control, Nobody Wants Responsibility

Share it on Twitter  
Share it on Facebook  
Share it on Linked in  

Who owns the data?

The more I think about this question, the more complicated it seems to me.


One of the advantages of being an outside observer in this game is that, on occasion, I've got a clearer view of the conflicts than the people involved. And, from where I sit, there is unresolved conflict on both the business and the IT side over data ownership and its related issues, from stewardship to governance. And I think these unresolved issues might be central to resolving at least some of the never-ending problems with data integration.


I started to see the situation differently this week after writing about the UK CFO who resigned after a data integration glitch. You just don't see that headline-maybe you hear about CIOs being fired over data glitches, but you certainly don't see a CFO resign when technology was at fault. In fact, in the United States, you never see any executives resign without doing so begrudgingly and with an outrageous package that ensures they're financially better off post-employment.


IT Business Edge's Ann All had a question, though: What about the CIO?


"What about him?" I thought. "I guess he didn't want to resign." It seemed pretty clear to me from the CEO's remarks that the CFO chose to resign. But I tweeted the question, because what do I know?


This prompted Gartner Research VP Mike Rollings to write about the topic, and his post really got me to thinking about the duality of data ownership:


...I believe the reason the CFO resigned is because 'the CFO' is the answer to the question "who is responsible?" Yes, the data was flawed. Yes, the systems were implemented by IT and the CIO was the data custodian. But, ensuring that financial data is accurate is the responsibility of the CFO. In this case the data may not have been stewarded, but the CFO is taking responsibility like a data steward.


When experts talk about data governance, data stewardship inevitably comes up. "The business has to own the data, not IT," I've heard time and time again. Rollings' post exemplifies the best of this line of argument.


And it sounds good. But when you consider our UK CFO, it looks a bit different. Who in his right mind would want that kind of responsibility? If resigning-or, worse, being fired-is a very real possibility, even if it's a technology glitch that underlies the problem, no wonder business leaders aren't lining up for the data stewardship role. And no wonder IT wants to offload it onto the business.


Here's the rub about data, though: While no one wants to be responsible for it, everyone-including IT-wants to control it. Ask anybody who's tried to do an enterprise data integration project-there's a lot of data rice bowling (that's military-speak for a "jealously protected program, project, department, or budget; a fiefdom," as in "this is my rice bowl and I'm the only one allowed to eat out of it") that goes on.


Business units want to see everyone's data, but they don't want to share. As Tony Fisher, CEO of DataFlux, recently told DestinationCRM.com, data outlasts both applications and employees, so data management needs to span the enterprise. Ownership needs to belong to the organization, he argued.


Traditionally, this problem has been solved by IT taking back the data, at least on a technology level. Let's face it: Business units wouldn't share, and somebody had to do it. Sure, you may have gotten an executive sponsor, but when the rubber hit the road, IT was the driver for integrating the data and leading enterprise-wide data initiatives.


And it worked, to some extent ... except now, we've got data quality issues and IT begging the business to take responsibility for the data. In a recent post, Ann All explores this issue as an IT/business alignment problem. She quotes an Information Management article by Platon A/S Executive Vice President Martin ABC Hansen, in which he writes:

Data is not owned, but partly managed by IT, and business doesn't want to own anything that "smells" like IT. This is typically the reason ownership is not anchored, and hence, falls between two chairs.

Once again, it's a great point. Let's face it: IT is good at aggregating and integrating data, but it still doesn't actually create or control that data in any meaningful way.


And given the role of data problems in things like the near-collapse of the U.S. banking industry, the subprime mortgage problems, and perhaps even the recent foreclosure fiasco, who can blame IT for wanting to clarify its lack-of-ownership when it comes to the data?


But I can also see why business units wouldn't want it back, either.


This is what I mean by the unresolved conflicts with data governance. Everybody wants to control the data, but nobody wants to own it. Ultimately, I think this is a huge issue for data integration and enterprise-wide information management initiatives.