Happy New Year, and welcome back to the Reservoir of Ideas. OK, enough pleasantries.
I’m going to spend the first part of this year exploring the following observation:
Many, if not most organizations have been struggling with the same Information Management and Data Governance challenges for decades.
Certainly, we’ve seen traction in some organizations and they’re reaping the benefits, but show of hands: How many of us would say that Data Governance has been broadly successful? Probably not many.
And it’s such a contrast with other data and analytics innovations through the years. I can remember when just having a Data Warehouse was a competitive differentiator, but now, and for many years, it’s table stakes. It’s the classic Innovation-Adoption Curve: innovators, early adopters, early majority, late majority, and laggards. Big Data, Cloud, and now AI have followed or are following the same pattern.
Data Governance has been a constant “presence” throughout. The words “Data Governance” are on all the architecture diagrams. They usually appear in a long rectangle underneath the producer to repository to consumer process flow, in a ring of one of the concentric analytics process circles, or at its hub. The implication is that governance is pervasive throughout and central to data and analytics. It is so important that it deserves special attention and its own space in the diagram. While well intentioned, instead of getting special attention, it seems to get minimal or none at all.
Illustrating Data Governance as a separate activity makes it easier to marginalize and ignore it.
We’ve completed all of the other activities and the data is moving and our customers are happy. Celebratory chicken lunches all around!! We don’t have time or resources as it is. Maybe we don’t need to do all this Data Governance stuff after all. What? You say that we do? Tell you what, we’ll incorporate it into the next phase. Translation: Never.
This scenario has played itself out countless times and little is changing. We’re still trying the same approaches and methodologies with the same non-success.
The first step of any recovery program is acceptance.
We must accept that Data Governance as a discipline has failed.
It’s stupefying. And frustrating. Everybody knows what needs to be done. We have fantastic resources and frameworks and books and practitioners and conferences and methodologies. The depth of knowledge is greater in this area than most anywhere else in data and analytics. We’ve seen the benefits to the organizations that have done it. But even then, it’s usually a struggle and the successes are often limited to certain specific areas. Maybe yours is one of the organizations that has been successful across the enterprise. If so, then congratulations!! Get in touch with me and let me know how you did it and I’ll add it to the list.
So, somebody asked us to put together a Data Governance program. Too often, we reflexively start by defining councils, meetings, and forums, assigning stakeholders with roles and responsibilities, and creating forms and process flows. It’s easy to go wild, building an enormous governance edifice. We value completeness. We want to deliver a comprehensive package. It’s our nature.
Newly created Data Governance processes are like a large backpack.
Nobody wants to be handed a hundred-pound sack to carry around with them on top of everything else they’re already carrying. Maybe they didn’t want the complete package after all.
In response, we tried shrinking Data Governance. Make that backpack as small and as light as possible. A friend coined the term “Minimum Viable Governance.” I have long advocated a similar minimum set of data curation requirements: definition, expected content, authoritative source, and security and privacy. Last year’s Data Chasm series discusses this, but these activities are still bolted-on.
What do we need to do differently?
Everything.
Maybe we’re thinking in the wrong direction. Maybe we’re thinking too small, too concentrated, and too separate. Maybe we should be working toward a more audacious, and perhaps a more effective goal of upending the entire application development culture. It’s not out of the question. Consider how the introduction of Agile methodologies have impacted development.
Yes, anything that we want to do will be additive to somebody. Information Management will always fail the primary corporate requirement where you can do anything you want as long as nobody has to do anything that they’re not already doing.
We need a systemic, transformative solution.
A solution that becomes part of the culture of the organization and outlives any individual executive advocate. A solution that everyone recognizes as necessary and just follows as standard operating procedure. A solution where the benefits are taken for granted and that would be sorely missed if it were to be taken away.
The need for Data Governance disruption is not a unique observation. It’s becoming increasingly evident to those in the field. Frustration at the lack of progress has been building for some time, and the conversation is now emerging from the shadows. My goal in these articles is to explore root causes and challenges, make recommendations for addressing them, and to lay out a framework for this necessary transformation.
Here’s the first recommendation:
Hang on. It’s going to be a bumpy ride. But I’m increasingly convinced that’s the approach we need to take. The alternative clearly isn’t working.
Stop treating Data Governance as a separate activity. Don’t illustrate it with a separate box. Instead, distribute the set of capabilities that we describe collectively as Data Governance into existing corporate processes.
I touched on this last summer when I discussed distributing responsibility for analytical data to the business process teams. There’s more coming.
What we’re doing now, and what we’ve been doing for decades, clearly isn’t working. It’s time to start thinking differently and doing differently. Hang on. It’s going to be a bumpy ride. But I’m increasingly convinced that’s the approach we need to take. The alternative clearly isn’t working.
Photo Credit: Brandon Doran, “Once the dust settles.” Flickr.com. Some rights reserved.
