Data Governance and the Enemy of the Good

by Jason Williscroft

This Monday, 18 April, marks the beginning of Enterprise Data World in San Diego, arguably the most important North American Data Governance conference of the year.

I’ll be there, as will be most practitioners of note in the data world. We will be in the minority, though. Well over half of the attendees at EDW will come from what I would call the client side:representatives of institutions large and medium-sized; from finance, health care, and diverse other industries; tasked with making sense and, perhaps, profit from the epic cataracts of data that flood their respective organizations every working day.

For some of these attendees, Data Governance is an opportunity to shine, a framework around which to solidify practices and systems that are already fundamentally sound. For most, though, it is a life ring: a method for navigating a sea of endless complexity, and transforming impossible tasks into merely extremely difficult ones. These folks are easy to spot… they’re the ones asking a million questions and taking bushels of notes.

The lessons taught at EDW and other Data Governance conferences are anything if not consistent: Data Governance is big. It transforms an enterprise. It is ontologies, metadata, data lineage, ETL, EDM, MDM, and a host of other TLAs, each requiring a total commitment from the boardroom and a whole new category of expenditures from the budget. The hardest part of doing Data Governance, one learns, is selling Data Governance to the people whose signatures will authorize payment for all this stuff and stand in mute accusation of them if the anticipated ROI fails to materialize.

Do these things, the attendees are told. Craft a plan that includes all these puzzle pieces, calculate ROI according to this formula, and your proposal will be accepted and your career safeguarded for another season.

All of which sets the stage for the phone call I had Friday afternoon.

My friend on the other end of the line is in charge of data management at a large retail organization. I met her last year at a Data Governance conference in Chicago. She has been tasked with getting Data Governance up and running at her firm… as were her five predecessors over the past decade. My friend is a year into her position and recently saw her comprehensive Data Governance proposal rejected by the very executive team that tasked her to set it up in the first place. Too big, they said. Too much money to spend on things that are not a priority right now.And then they told her to figure out how to achieve all the promised benefits of Data Governance anyway, but with virtually nothing in the way of additional budget, resources, or authority.

So at this point my friend is worried, and open to ideas. I offered to help her brainstorm.

First I asked her who had led the argument against her proposal. The guy in charge of the retail supply chain.

Then I asked what he stood to gain, personally, from the new capabilities outlined in her proposal. It turned out that a key element would enable more efficient sourcing of goods, so that for example the retailer would wind up spending $1.08 instead of $1.12 on a tube of toothpaste, with the increased margin going directly to the bottom line… after paying for Data Governance, anyway.

I asked her whether this executive’s anticipated marginal savings would add up to significantly more than the cost of instituting Data Governance around his area of responsibility. Well, no. But the long-term benefits are cumulative…

And then we both chuckled, because we’ve both heard that pitch before at any number of Data Governance conferences. Meanwhile, I was mentally climbing into that executive’s shoes: why in the world would I buy a program that might marginally increase my profit on the toothpaste I already sell when what I really need is to sell three times as much toothpaste?

But that isn’t the way Data Governance is sold. The consultants who teach organizations how to master their data are in it for the long haul: comprehensive organizational change requires massive capital projects that deliver their promised benefits long after the consulting fees are paid. Highly targeted projects are unlikely to make any consultant rich, particularly when run on a shoestring budget, with tight constraints on scope and duration and high expectations for immediate ROI. That’s just risky… at least, it’s risky for everybody but the client.

Long before I did my engineering with a mouse and a keyboard, I did it with an oscilloscope and a soldering iron. In those days, particularly during project season, engineering professors could be found wandering the labs at all hours of the night, delivering the same message to students endowed with far more talent and enthusiasm than good sense:

The Best is the enemy of the Good.

In other words: yes, we know your senior project is the construction of an autonomous robot capable of retrieving a beer from the refrigerator. And we do appreciate an element of style. But the robotic happy dance at the end of the process has taken you three weeks to program, it isn’t on your list of requirements, and meanwhile you have other projects to complete…

Big Data Governance is sold with the best of intentions: to remake an institution into its bestpossible self, data-wise. But an exclusive focus on long-term, big-ticket outcomes can obscure very real opportunities to remake just one piece of an organization into a merely better, more profitable version of its former self. Yet it is precisely those highly-targeted “small” wins whose cumulative effect is often to convince an organization that maybe Data Governance should be a high priority after all. Because in the eyes of an executive whose annual bonus depends on how much toothpaste he sells, selling three times as much toothpaste is not a very small win at all.

My friend and I will talk again, but I believe I already see the outlines of a path forward for her. Her first proposal, which distilled the best recommendations of dozens of top Data Governance consultants presenting at every conference in the country, was very much a solution in search of a problem. Her supply-chain guy rejected her proposal because it didn’t solve his problem, which when you think about it is a fairly reasonable response on his part.

My counsel to her was to focus on this fellow in order to find out what his problems actually are… and then to apply all of her considerable and hard-won Data Governance expertise to figure out how to leverage the data assets already in her firm’s possession to help that guy sell toothpaste like it’s made of hotcakes. In other words, table Data Governance as a capital project aimed at solving abstract problems, and instead use it in a targeted manner as a powerful tool for solving real problems.

This changes the conversation from “let me sell you this thing” to “let me help you be more successful”. It transforms the value proposition from big risk/unknown reward to limited risk/tangible reward.

That’s a recipe for success.

Previous Post Solid EDM: Why the Raw Table Matters
Next Post Guide to Continuous Integration in Data Management