Think of approaching data governance effort as you might work on a craft project with your kids. If you clean up the mess as you go along, everyone gets a job to do and you end up in a much better state of mind. If you leave all the clean-up to the end, it may all fall to you and completely overwhelm you.
In data governance, the more you set and define your targets for documentation upfront, the easier it will be. In addition, you’ll make better decisions as you go, versus ending the process and realizing, “Yeah, I guess we should have locked that down.”
Privacy and user-driven security, for example, are much trickier to pull off if you never captured anyone’s user credentials at the beginning or considered how it all might work. The same is true of team structures. If you carefully define the things you’re going to measure up front, then the documentation that goes into your requirements should evolve into your data governance. You’ll naturally want to keep track of your lineage and properly define how everything was created.
Due diligence, agile process, better results
Of course, in the heat of the moment, we’re often forced to produce and we can't always think about all the implications of making a bad choice.
This is where striking the right balance in data compliance is key. I’ve seen lots of projects that never really got started because they spent so much time defining all the requirements. I think this is why people are moving to agile development in business intelligence. You come up with a minimum viable product, iterate fast and rapid and often, then rinse and repeat. If you try to build it all at once, especially if you hadn’t considered data governance at the beginning of building, then you might have to crack the whole thing open and go back.
Building the right data governance team
Who should be on the data governance team? There definitely should be an executive sponsor, without a doubt. Because data governance documentation is too easy to skip if the pressure is on, it really needs to be a top-down driven effort. Otherwise people may make the determination that it’s optional.
Beyond executive sponsorship, you probably should include at least one person from each key functional area. This includes the people that are building the data, but also from among the end users who are going to consume it. This gets back to a point I make with my team, that the people who design and build databases do not speak in the end user’s terms. So if you don’t have every function represented in a data governance effort, you’re going to have a language divide.
And don’t fool yourself: you’ve got to sell data governance to people to get them to join a team. For example: “Hey, this will be great! You'll spend less time doing repetitive things. You know how you sometimes keep answering the same question over and over again on how to calculate this measure? This will help. And think of how we can all sleep better at night knowing that we're not at risk for exposing information that shouldn't have gone external?” Whatever it is, spin it in a way that's in that person’s best interest.
So choose wisely as you define your targets, document your work and define your team. Look for individuals who make the big decisions as well as a sounding board of users to make sure you've heard everybody's opinions. It goes back to the kid’s craft example: if you assign too many people to clean up without enough oversight, you may end up with a big mess and an unhappy group on your hands. Choose the right people, process, and technology and you’ll be pleasantly surprised at what you can do together.
Why governance must go on
Yes, governance is a monumental effort. It's a struggle because it takes place against a backdrop of constantly rushing to deliver new things. It’s the documentation step of every phase that feels like it frustrates your impulse to open data up to everybody. As soon as you try, governance closes more of it off.
Yet good data governance is always going to be worth it. Why? Because in many organizations, data is all you have, and therefore it has to be as protected as it is accessible. If you make decisions with data, strong governance builds to better ones. Governance also ensures that your data is compliant as it is appropriately distributed.
So it’s a false choice to say that you’re either going to be able to keep up with the ever-changing needs of the business at an ever-increasing velocity, or you're going to take the time to do it all right and govern use of your data. Success today dictates that you do both.