Data Governance Matters

My friend Tom Jesionowski, who sometimes writes in this space, blew my mind last fall when he drew a little picture on a whiteboard. It was a square split into four quadrants - a very simple picture. It was what he said about it that was so exciting.
I’ll try to summarize:

In the upper right quadrant he put a category of business people: Those who make decisions. You know the type: analysts, managers, executives. They need good data to make good decisions. He called this the Analyze quadrant.

In the upper left quadrant he put another category of business people: Those who are paid to act. They’re the ones out there making sales, making products, offering services, and performing the business of business. These people are involved in interactions and transactions that generate data. He called this the Act quadrant.

Below them, in the lower left quadrant, he put a category of IT people and systems: Those who collect information from the Act quadrant. He called this the Acquire quadrant.

Finally, he labeled the bottom right quadrant the Integrate quadrant. To the right of the Acquire quadrant and below the Analyze quadrant, the IT people and systems who reside here are charged with integrating collected information and making it available for analysis.

Information constantly flows counter-clockwise around these four quadrants.
***How WELL it flows, and in WHAT SHAPE, often determines the quality of decisions that business leaders can make!***

You ever see a picture that makes an idea come alive for you? That’s the way it was when I first saw Tom’s diagram.
OOOOOhhhhhh…
Right! We’re asking upper right people to make decisions, but the lower right can’t give them information in the format they need, because they didn’t get it from the lower left, who was never told to collect it in a specific format by the upper left, because the upper left either didn’t care or didn’t know about the need, because the upper left and the upper right SIMPLY WEREN’T TALKING.

OOOOOhhhhh.

So maybe the answer isn’t to beat up on our Data Warehouse team, but to get the upper right folks to put pressure on the upper left folks to build in requirements that will meet the needs of ALL 4 QUADRANTS. Yeah - that’s it.

So then Tom and I had a great, great series of conversations about why this is hard - the usual topics about power, politics, communication, consensus-building, etc.

Ever since then, I’ve been hounding Tom to publish this so I could start using his framework in discussions. I mean, how great will it be to use verbal shortcuts such as “Since this data will eventually feed analysis, let’s get some UPPER RIGHT folks involved now, in the design phase.”

So I am PLEASED and PROUD and RELIEVED to say that Tom’s framework has been published in TDAN, the great publication that’s run by old friend Bob Seiner.

You should all check it out.
Tom Jesionowski – The Prime Business Decision Loop
http://www.tdan.com/view-articles/7349

In a week I’ll be in San Francisco at the Annual Data Governance Conference put on by Wilshire Conferences and DebTech International. I can’t wait! These folks put on an amazing time, with lots of rich content, case studies, and expertise. The best thing is that the event is designed to get people talking to each other – practitioners, academics, and thought leaders start sharing tips with each other on Monday morning and don’t let up until the event is over on Thursday.

Of course, I really can’t wait to see some of my buddies. The type of work we do can be pretty stressful at times, so it’s great to get together with others who “get it” and can relate. I’m looking forward to picking up some new ideas about power, persuasion, control, and consensus-building. I’ll be sure to share…

I met Mike Meiers at a conference. He attended a session I gave on communications for Data Governance teams, and he told me I was using a word imprecisely. I fell “in admiration” of him on the spot. I love word people! So I was really honored when he suggested we develop a talk that matched his actual experiences doing Data Governance, Data Quality, and Data Warehousing at Olmsted Medical Center against best practice models and buzzwords. The ideas is to describe both his experiences and a sort of “meta-process” around the choices we all have to make as we configure our programs.
Mike and I will be giving that talk in San Francisco at the Annual Data Governance Conference there.

While developing our presentation, we spent a lot of time talking about how you accomplish data-related goals when they aren’t sponsored by a formal program. How do you persuade people to engage when you don’t have power? How do you motivate participants? How do you reward them? If there’s no budget for new tools or new people – if you simply have to accomplish your new goals from within a business-as-usual terrain and climate, what do you do? Moreover, what model do you follow?

During our discussions, the term “guerrilla governance” was coined. Isn’t that a great term? I love the visual of passionate, unmarked “Data Stewards” streaming out of the hills, swarming through office buildings, leaving a path of accountability and controls in their wake.

“Who were those masked men?” office workers cry. “Who cares,” cry others. “They brought us standards!”

So now Mike tells me that he’ll be presenting a follow-on session this fall at a Data Quality conference. This time his title is “Guerrilla Governance” and he’ll be showing Data Quality professionals how to mobilize their forces, inspire through action, and be prepared for the time when they get money for tools, get acknowledged as governance teams, and get empowered to move from guerrilla status to a formal program.

Go Mike! Viva la revolution!

I had a very interesting experience last month with an organization that makes/distributes food and drink. You’d all recognize the name. Bright, bright people doing some ambitious things. It was a privilege to be asked to poke holes in their data strategy.

While I was with them, I saw a perfect example of the old adage that “Architecture is Destiny.” This organization lives and dies by its knowledge of consumer patterns and behavior. They need data. And they have it - lots of it - and they’re using it.

But they’re held back by an architectural decision made long ago. Maybe it made sense at one time to model “Reason to buy” and “Product” as a one-to-many relationship. But not now. I don’t know about you, but if I decide to buy a cup of coffee, I might have many reasons!

So here’s a major company struggling to analyze buying habits with data that is forced to conform with a one-to-many relationship instead of a more accurate many-to-many. They’re changing this, of course, but still, I can’t stop thinking about how they got here.

I keep imagining a scene from years back: A guy, hunched over in a gray cube, reading from a requirements document. Hands on keyboard, he’s ready to start building a data model. “Really?” he mutters to himself, as he reads the lines that describe the relationship between Reasons to Buy and Product.

“Doesn’t seem right…” he says to no one in particular. Then he shrugs, says “They’re the bosses,” and gets to work. In just a few minutes, he’s defined the relationship that today is limiting how the company is able to understand customer buying habits.