Who said you can’t mandate data governance?  Who says it has to be some smarmy, politically correct, incremental pitter-patter of a tiptoe into the bomb-laden lion’s den of IT super-forces?  How come so many other rules are blindly followed, up and down the corporate ladder?  The police have rules, don’t they?  The IRS has rules. The NSA (well sort of). Your local bank. The post office, all the software programs we use and every church and non-profit organization has rules.

Just because the rules exist, however, that doesn't mean that everything is working. So why can't we just cut to the chase, call a broken process broken and fix it by fiat?

As a matter of fact, a good way to get the organization’s attention is by singling out data quality that really is bad – that may even cause harm – to launch a data governance initiative (see the SAS e-book “When Bad Data Happens to Good People” for more on how this happens).  And it’s also a great idea to show a broken process that causes countless hours of wasted time when people should be focusing on more productive tasks.

Virtually every company where I have engaged as a consultant has told me about the amount of time their analysts, IT staff and other users of data spend each week (or month/quarter/cycle/year) reconciling data when report results don’t match.  Sound familiar?

The toughest thing about starting data governance isn't just writing the rules and making sure everyone follows them.  It starts with the way things have been cobbled together over time.

Nobody designed this convoluted process intentionally – it evolved from many varied and repeated attempts to make the data ”work” for a specific reason.  It may just be one person’s reason, but it’s a reason nonetheless.  We love to talk about data silos as the chief issue to tackle with data governance, but quite honestly, if there is a valid business reason for the silo, then it won’t be going away anytime soon. Those workarounds persist in order to get the job done. They often reflect the way the organization operates. If you made a rule that won't let me get my job done, then I’ll figure out a way to go right on doing it!

Data governance by fiat is a laudable goal, no doubt about it.  You can design best practices and impose them all over the place, and you can come down hard on the people that insist on hiding in wormholes just to preserve their little analysis.  I've seen more than one organization take this approach. Unfortunately, I wasn't there three years later to see the aftermath.  And I will put cold, hard greenbacks on the table and bet you it hasn't worked.

Why does it fail? They're probably still thrashing people out of the corners to justify their analysis (or a report or data handling process). They are still quibbling over which standard to use for which purpose or trying to make some simplistic piece of software catalogue all the different business rules for each little pocket of the organization.   They haven't been able to convince the IT chiefs that their way is the way of the future.  Or even better, they have been able to clamp down on items that are required for regulatory purposes but still have oodles of wasted time and effort trying to keep the lights on.

That’s my view from the trenches. If I'm wrong, I'd love to know about it.  I DO think data governance by fiat can work, just not without the hard work of peeling apart and re-engineering so many different data supply chains that – like cockroaches under the foundation – will always remind you that the world is indeed complex and three-dimensional.