In-Depth
Q&A: Try a "Non-invasive" Approach to Data Governance
The path to data governance starts with small steps.
- By Linda L. Briggs
- 04/13/2011
Bob Seiner has a message for organizations contemplating data governance and stewardship -- you're already doing it. Your efforts may be somewhat inefficient and ineffective, says the long-time data management consultant, "but there are people in your organization who informally have responsibility around the data. Rather than giving people new titles and roles and adding to their existing work, my suggestion is, why don't we identify those people and put a program together around that?" In a nutshell, that defines Seiner's "non-invasive" approach to data governance.
Seiner has worked on data management issues for many years, beginning in the mid-1980s. A consultant and speaker on the topic, he maintains an online publication for data management professionals (The Data Administration Newsletter, TDAN.com), and KIK Consulting & Educational Services, his consulting and education Web site. Reflecting a growing interest in data management issues in the industry, TDAN.com continues to grow every year, Seiner says. In this interview, he discusses how an approach he has trademarked as Non-invasive Data Governance can help organizations move more smoothly to a higher level of data management.
BI This Week: You coined and trademarked the term "Non-invasive Data Governance." What do you mean by it?
Robert S. Seiner: Very early on, I came up with an idea that instead of trying to assign people into roles as data stewards, we should acknowledge that they already exist within the organization. There are people with responsibility around the data in every organization, yet it's very informal, ineffective, and inefficient in a lot of ways. As I started to focus on it, I realized that if we could formalize people's behavior around the management of data, which is truly what governance is. We don't need to go around with a stick hitting people over the head, telling them what to do. We need to help them to become more efficient and more effective. I came up with the term "Non-invasive Data Governance" as the description for the approach that I use.
To a client, they tell me that the noninvasive approach is the only way that will work in their organization. I've banked on that and used the approach to help my clients become successful.
With your approach, are data stewards people who organically grow into that role but don't necessarily have that title within the organization?
I don't give people the title of data steward. My approach is very different from many people in the industry, and I'd prefer to keep it that way.
Almost anybody in the organization could be considered to be a steward: somebody who is creating reports, somebody who is defining data for a new database, somebody that is on the front line producing data at a retail store, or somebody who is running the checkout line -- when they enter data into the system, they have some level of accountability as to the accuracy of the data they enter. I've set up an operating model of roles and responsibilities that allows for the fact that almost anybody could be a data steward.
We don't need to tag every one of them, tell them they're a steward, and give them more work. There is a role that I think is very important for organizations -- what I call domain stewards. Those are people who have responsibility for a subject matter of data across the enterprise rather than just in their business unit.
To be honest, that's the largest hurdle for organizations: The ability to identify people with some level of accountability and responsibility for a subject matter or a domain of data across the enterprise. That doesn't usually exist.
Here's an example: I'm working with a client -- a large university. The domain steward for student data is the registrar, and the domain steward for financial data is the controller. These are people at a relatively high level of the organization that have a vested interest in, and the most knowledge about, how student, finance, and HR data are used.
I don't suggest to my clients that they change their titles. I just tell them that they should acknowledge themselves as subject-matter stewards, enterprise stewards, or what I call domain stewards.
Where should an organization begin in trying to get a handle on data governance? It sounds like part of the process begins with identifying what you refer to as "domain stewards."
What I suggest is rather than picking your data or domain stewards, or assigning them, you first identify who they are.
So you find people who are already doing that job?
Yes, who are already doing it to a certain degree. That doesn't mean it doesn't come without any pain, or that there's not some work that needs to be done. Obviously, this is hard work, but people are much more willing to listen if you tell them, "Hey, we're not going to club you over the head and give you more work to do. We know you're already busy 150 percent of the time."
Instead, we come to them and say, "Now you have been assigned to be a data steward." We need to set some priorities as to what's important and what's not important. My suggestion is that we identify them and we record them -- and that's a big piece of it, to record information about who does what with data across the organization. If you don't have an inventory of that, you're going to have a very difficult time putting a governance program in place.
You had talked about making it enterprise-wide. How challenging is that? Are there political considerations?
If you have a very large organization, without a doubt there are political situations. In the operating model of roles and responsibilities that I share with my clients, it's at the operating level where almost anybody could be considered a steward. In the middle tier of a pyramid diagram, there's the tactical layer. That's where the domain stewards reside -- the people who have enterprise responsibility. In order to cut through the politics, you need to have a strategic level as well. You may have heard the term "data governance council" used before. The strategic level consists of individuals that represent the different parts of a business.
Certainly, you need to convince upper management that they need to do this. [I tell clients,] "If you go to senior management and say, 'Data governance is this huge, challenging effort. It's going to take a lot of time and it's going to take a lot of money,' they're going to prioritize it among many other things that are important to them. Instead, go to them and say, 'We're already doing governance in pockets. We're doing it informally. We need to put some formal structure around it and somebody needs to have the responsibility for moving governance forward in our organization.' In that way, you can cut through the politics because again, they're willing to talk to you about how you can help them with their daily job. Yes, there are politics, but you have to find a way to be able to cut through the politics. That's what I've found to be most successful.
You've worked in this field for a long time. What changes are you seeing in how data governance is approached, and how important it is to companies? I seem to hear more and more talk about data governance these days.
Almost every organization is at least talking about governance, but there's a lot of confusion out there as to what it really means, and what the options are in terms of approaches to putting governance in place. I'm certainly very busy. It's come to a point now where there are several data governance conferences a year, conferences that are completely focused on data governance. They're being attended by people who are new to it, who have started to put programs in place and want to learn from others.
Is the onslaught of technologies such as analytics and real-time data access changing data governance?
It's bringing more awareness to the need for data governance. The TDWI conferences, for example, tend to focus heavily on analytics. You can perform analytics until you're blue in the face, but if the data you're working with isn't any good, the results will be decisions made with poor-quality data. You're only going to get answers to your questions that are as good as the data you have.
In terms of return on investment, where do clients typically see the ROI in implementing good data governance?
In my opinion, you can have immediate turnaround. You need to take inventory, and you need to know who the stakeholders are, and who governs certain types of data across the enterprise. Then, if a business rule changes that's associated with that data, you can immediately identify to whom you need to communicate. You don't need to wait for the program to churn for 18 months to start showing some direct ROI.
The bigger question is, how do we measure that ROI? Typically, there are different ways of measuring the success of data governance. One is the business value that it brings. There are many different areas of business value you could focus on; there's also how well governance has been accepted into the culture of the organization.
What advice do you have for getting top management to support data governance?
It's really a matter of awareness. If they view data governance as a huge challenge, they're going to back off. If it's sold to them as something we're already doing to a certain degree, but we need to put some formality around it, they're going to be much more willing to listen.
So buying a tool might be part of the solution, but isn't the whole solution.
A tool can certainly help to enable your data governance program. … Buying those products is great, but it cannot be your data governance solution, just like buying a data warehousing tool suite does not build you a data warehouse.
We've covered a lot of ground. Any parting thoughts on data governance issues?
Here's one thought: If you're not thinking about implementing formal governance yet, you will be eventually. There's more than just one approach out there. There's more than the command-and-control approach, or what I call the two-by-four approach, where you assign people roles and give them more work to do. There are other options, and one of them is doing [governance] in a non-invasive way.
The TDAN.com Web site contains articles I've written that describe what Non-invasive Data Governance is and why you might want to consider that as an approach. Your readers should check the articles Non-invasive Data Governance Explained and Messages for Management -- they discuss what we need to share with management so that they're not threatened. The message is: we're doing it already. Let's just do it better.