The models of data stewardship: how to run a data governance program

Even though data governance was born out of business intelligence (BI), many IT and business professionals still don't fully understand the data governance process, the roles and responsibilities of BI teams on governance programs, or how to set up a data stewardship function, according to industry experts Jill Dyche and Kimberly Nevala.

In this video interview from the fall 2009 TDWI World Conference, Dyche, partner and co-founder of Baseline Consulting, and Nevala, a senior consultant at Baseline, give tips on how to manage data governance programs and detail what they've defined as the five models of data stewardship. In addition, they discuss the benefits that BI teams can get from data governance programs, the need to measure the success of governance initiatives and the importance of giving data stewards the authority and accountability that's required to succeed in that role.

In this 11-minute videocast, viewers will learn:


  • What BI and IT professionals need to know about data governance
  • What the five models of data stewardship are, and what organizations need to consider before choosing one
  • The extent of the authority that data stewards typically have and whether it's enough for them to do their jobs effectively
  • The governance and stewardship issues that generate the most interest among BI and IT pros, and why
  • The major roadblocks companies run into when implementing a data governance or data stewardship initiative


Related resources on data governance and data stewardship:



Read the full transcript from this video below:

The models of data stewardship: how to run a data governance program

Craig Stedman: Hello and welcome to a Data Management video about data governance. This
is Craig Stedman, site editor for I'm in
Orlando at the fall 2009 TDWI World Conference covering the event.
Joining me today to discuss what they've defined as The Five models for
data stewardship are Jill Dyche, partner and co-founder of Baseline
Consulting, and Kimberly Nevala, a senior consultant at Baseline. Thanks
for speaking with us today, Jill and Kimberly.

Jill Dyche: Thanks Craig, nice to be here.

Craig Stedman: So I see that the two of you are teaching a full day class on data
governance here at TDWI for BI pros. A lot of the attendees seem to be pretty BI
savvy. What do you think they need to hear about data governance that they
don't already know?

Jill Dyche: Well what's interesting is BI has actually been the birthplace of data
governance for a long time but people still don't necessarily understand
what that means in terms of roles and responsibilities for data governance and they
don't necessarily understand that data governance is a process.
And so what we're teaching today in our full day workshop is how to not
only launch data governance, but what the role of a BI team would be.

Craig Stedman: What's the key message that you'd like attendees to take away from
your class at the end of the day, and that people watching this video
should come away with as well?

Kimberly Nevala: I think if we think about data governance sort of at its essence it's
a very all encompassing sort of problem. In that it sort of impacts every
business process, every system in your organization when done right, and
done in a consistent sort of methodical fashion. So we really want people
to understand and think about data governance as a program. Think about
something with a focus on process and culture more so than as a technology
problem. So that concept of change management and on iterative improvement is
really, really important because without that sort of considered,
deliberate, iterative approach people sort of get lost in the maze and
ultimately that means we don't deliver value with data governance and it
sort of becomes the proverbial dirty word.

Jill Dyche: Well and it's an opportunity for the BI team to then proselytize that
message. What we're finding is that often the BI teams are the best people
in the organization, because they understand integrated data better than
anybody else, to actually kick off the governance conversation. So they're
sort of the soft benefits around data governance for BI teams, it gives
them extra exposure. And it let's them proselytize their expertise.

Craig Stedman: I know you're going to talk about the five models for data
stewardship during the class. In a nutshell what are they and what do
organizations need to consider in deciding which one to adopt?

Jill Dyche: So five models of data stewardship, the first one is subject area data
stewardship which is essentially the assumption that a lot of companies
make. That we're going to have a customer data steward and a product data
steward. It's one of the five models. The second of the five models of
data stewardship is organizational data stewardship, where we have a
particular data steward dedicated to a particular organization. The next
kind is business process data stewardship. Which not a lot of people know
about but a very, very viable model of data stewardship for particular
companies with particular cultures. We've got application level data
stewardship which is essentially, stewarding the data for a particular
application or system. Finally we've got project level data stewardship so
very project specific. What's interesting about the five models is you may
start with one model and end up with another model.

Kimberly Nevala: The other thing I'd say about that is people always think about this in
the sort of academic fashion. That the stewardship model is subject area
data. I'm going to have a data steward for a customer. That's actually a
very daunting task when you start. Because when you think about the number
of places customer data is managed, stored within an organization, it's
pretty hard to get someone with sort of expertise in all of those areas out
of the box and so one of the things that we do encourage folks to think
about is sort of taking a hybrid approach. So look at your organization
and if we want to move towards master data management and subject area,
around customer for instance. Maybe that's our primary model, but let's
augment that by saying we're going to have customer data stewards assigned
from each of our functional areas. For instance if today we're very
functionally focused. So I'll have a customer data steward from sales,
marketing, and finance. Who have the most sort of inside invested interest
in customer data. So it allows me to sort of pay homage as it were to my
current organizational biases, and still start changing people’s perception
and thinking about data in a different way. So it's a great way to sort of
bridge the gap between the short term current state and the long term

Craig Stedman: How broadly has the role of data steward been institutionalized
within companies, and do most data stewards actually have the authority
that they need to do the job effectively? If not what can they do to get
that authority?

Kimberly Nevala: It was interesting there was a 2008 TDWI study, it was on roles and
salary. And a bunch of people responded, indicated that they had the de
facto role of data steward but the actual percentage of people who
had a title of data steward was like 1.3%, just a minuscule
amount. That's actually pretty representative of what I think we're still
seeing. In that, data steward as a formal function, is relatively new and
organizations that are putting that in place tend to do that when they
are sort of trying to, sometimes it's almost a kind of kick start mechanism. I'm
going to have a data steward and sort of try to promote this concept of
data management and data governance that way but I think what's much more
prevalent is that-- what we find is that data stewardship exists in an
organization, but it's informal. Because you already have business subject
experts. Or data subject experts for systems. So one of the most common
comments that we sometimes get is, people will come up and say, wow I just
figured out I'm a data steward. So I think it's more informal than formal.

Jill Dyche: Which is good news and bad news, right because, what happens with data stewardship is
sometimes it's just a check box. OK. Now we have a data steward with no
actual roles, responsibilities, formal accountability for the role and
then you're sort of setting those people up for failure. So one of the
things we're talking about in class today is what does it look like, what
are the five models. And how do you actually formulate some decision
rights and accountabilities around that role, so that everybody
understands what that person is actually delivering.

Craig Stedman: Do you expect that to change, and for things to become more
formalized going forward? Should things become more formalized?

Jill Dyche: Well part of it depends on the scope of the company. I mean the larger
the company the more likely the data stewardship will be formalized over
time and it should, because as data really becomes an asset that's managed in
it's own right, it really does have to have discrete roles and
responsibilities around it.

Kimberly Nevala: I think to the extent that an organization has a lot of risk or exposure
from an external regulatory or compliance management perspective to the
extent that you already recognize enterprise risk management or sort of GRC
units. Those are the companies that I think are going to move towards
very formal data stewardship in a short term.

Craig Stedman: What other new and interesting topics are you discussing during the
data governance class?

Jill Dyche: I think one thing is audit and measurement. So it's not just enough to
launch data governance. You also have to measure the success of it. And
that's something that's sort of new to some of these people who've been
working in BI for a long time, and they're deploying data, and they're
deploying applications and they're moving on to the next project and what
data governance needs to mandate is closing the loop between some of that
audit and some of those requirements. So this concept of measurement is a
big deal. One of the other things that we're talking about that's new is
workflow. Workflow is really endemic to data governance and here at TDWI
in Orlando, there are a lot of vendors downstairs. And they'll be talking a
lot about data stewardship dashboards, and a lot of that is about the workflow of particular decisions that have to take place in managing that data.
So that's another one that we're focusing on.

Craig Stedman: What issue typically generates the most interest among attendees
when you talk about data governance?

Kimberly Nevala: First and foremost is probably this concept of data stewardship. Again
because we hear it from vendors. It's sort of a hot topic out there, you
can't do governance without stewardship. But again sort of the issues
with how do we actually instantiate that within the organization, are
pretty overwhelming. A lot of folks are coming in and again they sort
of,try this approach of data stewardship means we have to have customer
data oriented stewardship or subject area, and have found that that
doesn't work, so that whole concept of what is data stewardship and how do
I actually make it work organizationally, sort of manage up and manage down
in the organization? What do I actually need to do to make it work? The
flip side of that is what do I need to do to get some respect for those
that are saying I'm a data steward today, and I'm doing a lot of work sort
of under the covers and people don't recognize it?

Jill Dyche: I think that's the theme actually for today. A lot of the people who
are here in this class and it's actually a pretty full class. There are
people that are thinking we need this. I know we need this, it's what
Geoffery Moore calls the visionary in an organization, where they may not
necessarily have the organizational authority to push it by themselves
but they're actually doing the missionary work and so I think what a lot of the
people here are asking is given my role, I'm on a BI team, or some sort
of data administration function, how do I actually push this messaging
through both up into my organization as well as downwards? And so I think
that's sort of the common theme that we're addressing today in a lot of our
discussions with people.

Craig Stedman: I know you also do a lot of work with companies on data governance
and stewardship initiatives. When your clients call Baseline for help what
are they looking for, what are the big roadblocks that they're running

Jill Dyche: Increasingly, and I think this is good news for everybody. People are
engaging us to actually do deliberate road maps for data governance. We
talk in the class about why people fail with data governance and data
stewardship and it's essentially death by meeting and so what we're saying
is you need a deliberate design for this stuff. And if you build a road
map and establish some of those processes, some of those decision rights,
some of those accountability metrics up front. It can save you a ton of
time and a ton of effort. So that's one of the things people are finally
realizing and that's normally how we engage in data governance.

Craig Stedman: OK. That's all we have time for today. Thanks again to Jill and
Kimberly for speaking with us.

Kimberly Nevala: Thank you.

Jill Dyche: Thanks Craig.

Craig Stedman: That wraps up today's video on data governance and data stewardship.
You can learn more about Baseline Consulting at their website, To learn more about data governance and to
watch additional videos and read news stories, feature articles and
expert advice on a variety of data management topics, please visit Thanks for joining us and have a great day.


View All Videos