What is data governance? | Amazon Web Services
1 Video View·Dec 9, 2024
Learn why business and IT need to work together, and how Amazon's working backwards method can help you.
Hi, and welcome to our Masterclass for Data Governance on AWS.
I'm Kevin Lewis.
I am in the AWS Professional Services Organization,
working on data strategy and governance.
And I have been in consulting for about the last 15 years or so.
Before that, I worked for a large retailer
where I was responsible for my own data management program.
And, you know, in my time in consulting,
I've visited, geez, probably well over 100 customers,
either for short conversations or long-term engagements.
And there are some things that we've observed
that are very consistent patterns of things
that cause organizations to run into trouble
as they're implementing data governance.
And really significant elements of success,
you know, things that work very consistently.
And there's a lot of things that I wish I knew
when I was starting my own data governance program
that I had to figure out over time.
And I'd love to share those with you over the course of these videos.
So let's start with, you know,
what is this thing called data governance?
What is it about?
Now, if you look for definitions,
you're going to find a lot of things like,
you know, decision rights and policy and, you know,
access and all of those things.
And the definitions you find are, of course, accurate,
but they may not necessarily be practical.
So for our purposes, let's talk about what data governance
is in a way that can help us succeed with implementing it.
And so let's think about it this way.
Let's think about data governance
as making sure your organization's data
is in the right condition needed to succeed
with business initiatives and business operations.
And we'll find that in order to get data
in the right condition for business initiatives,
it takes a partnership between the business and IT.
So for example, let's say I'm with an insurance company, right?
And I want to improve claims adjudication.
All right, well, we know that in order to process claims,
information has to be captured adequately
within the claim information.
Now, we know that it requires a lot of knowledge
about the business of what a claim is and how it's processed
and the kind of information
that's necessary to effectively adjudicate a claim
and the kinds of data issues
that would slow down the process of adjudicating a claim.
So because getting that data in the right condition
is not purely a technical issue, it's also a business issue.
That's why we need a partnership between IT and the business,
and we call this data governance.
So that's a more practical way to think about it.
And so let's talk a little bit about the people
that are involved in this process.
So we'll hear the term data steward.
So what do we mean by a data steward?
So a data steward is a person from the business
who has detailed knowledge of the data
that's needed to support targeted business initiatives.
So if I am in this insurance company
and we are trying to improve the quality of claims data
in order to improve adjudication of claims,
well, then I need someone in that project day
to day to do some specific things in partnership with IT.
We need to evaluate the quality of the data.
We need to develop data quality rules for monitoring the data
and taking action when there are issues.
And we need someone to work with IT to leverage technologies
in order to accelerate some of these processes.
So this is the role of the data steward
within the data governance process
and especially within the projects
that leverage data for business needs.
So that's the data steward role.
And so you can see why it's so important
that the data steward be someone from the business.
In fact, DEMA, it's a very popular,
very well-known data management organization.
They have a phrase that I like.
They say data stewards are found, not made.
And I like that because really,
there are probably already people in your organization
who are doing this kind of work.
They're just not officially sanctioned.
They may not have the right formal roles and responsibilities.
But there are already people in the organization
doing this kind of work.
So one approach to finding data stewards
is to look for the people who are doing this kind of work already today
and then wrap some formalization around that role.
So let's talk a little bit about another role called the data owner.
So a data owner would be somebody who, more at an executive level,
would do things like make policy around the data.
So they would decide, for example, who has access to claims data,
who has access to customer data.
They can resolve any issues about definitions of terms
like what is a customer, what is revenue.
These are things that have to be figured out
so they can help to resolve some of those things.
And so again, this is a role that comes from the business.
And it's more of an executive level role.
So that's the data owner.
So there's a clear relationship between the data owner
and the data steward where the data steward
is more day-to-day involvement in projects.
And the data owner is more at the policy level.
Now let's talk about the role of IT.
So the role of IT would be to help navigate all of the systems
that both produce and consume data.
But they also want to enable data stewards
with the right tools and technologies and capabilities.
So we'll find that things like data quality management
and master data management and data access
and all of those things can be enabled
and accelerated by modern technology.
So this is why when we're building a data governance program,
we want to make sure that there is roles for the business,
roles for IT, and that we're working together.
And again, going back to the definition of ensuring
that we are preparing the data
to support the success of business initiatives,
we call that at Amazon, we call that working backwards,
working backwards from the business initiative
to decide how to prioritize what to do with data governance
and the underlying technology and so on.
So this group that works together from the business
and IT needs to ensure that they are,
in fact, aligned to those business initiatives.
Now let's talk about another level of this.
Another level of a data governance program,
in addition to helping succeed with individual business initiatives
and the use cases within them,
is to look across business initiatives,
to look across the organization.
So for example, reusing data.
So going back to our claims example,
you're going to find that within an insurance company,
there are many, many uses for claims data,
not just for claims adjudication,
but things like maybe loss prevention, fraud detection,
things like that, understanding the cost associated with claims
and many other things.
So the data governance program can help look across business initiatives
and ensure that when we are preparing data
to be useful for a business initiative,
that it's actually being deployed in a way
that it can scale and be shared across a wide variety of business initiatives.
And another way that it can be beneficial to look across initiatives
is to think about the level of integration required.
So bringing data in from a variety of sources
to be used in multiple business initiatives.
So let's take, for example, a customer experience for a retailer.
So you need to integrate data about the customer
from across various channels.
You need to integrate data of transactions and touch points
for web interactions and physical stores
and all of those things need to be pulled together.
Now, that doesn't happen automatically
because all of this data comes from a variety of different systems.
You have to do certain things to make sure the data fits together
in order to serve the needs of something
like a customer experience initiative.
And that requires, again, business and IT to work together
to make sure that the data can be linked together.
It's of high quality, it's coherent.
And if you're doing all the hard work of pulling this data together
from a variety of sources,
why not leverage that for initiative after initiative after initiative?
So your data resources get better and better and better over time.
What we have seen in real life is that very often
a business initiative will start up
and it will try to thread together data
from a variety of different sources for their purposes.
And then the next initiative will come along and do it again
and then do it again and do it again.
So having an effective data governance program
means that you are putting in place all the processes
and the procedures and strategies so that you can do this
just as a natural course of doing business.
So in our time together, we're going to talk about
some more of the details of sort of how to make this happen.
We'll talk about some very, very common strategic issues
and some very subtle ones that can really get you into trouble
and we'll do our best to shine a spotlight
on the difference between the right way and the wrong way to do things.
We'll talk about the distribution of responsibilities
within the roles that are involved in data governance
and how to do that effectively and lots more.
So we'll talk about underlying technology.
We'll talk about some of the detailed processes that are involved.
And importantly, we'll talk about how to sustain the program going forward.
So once you do all this work to get the program started,
how do I keep the program healthy? How do I keep it going?
How do I make it a part of just the way the organization does business?
So we'll be talking about that and thank you for spending time with me.

