Header Background Shapes Blue

Back to Basics: Data Governance, Stewardship, and Management

Data Governance, Data Stewardship, and Data Management are terms so frequently used and exchanged in the field of information management that they have lost their meaning. Practitioners, on the other hand, assume that the rest of the crowd understands all their subtlety. Well, they don’t. These terms make sense if you use a simple metaphor.

Take a country. Most countries have many inhabitants living, acting, and having an effect on the happenings of the country. Usually, they are bound to rules and laws which are decided on by a government or by social agreement, and enforced by bodies like the police force, or a responsible parent. If you understand this, you understand data governance, stewardship, and management.

  • Governance is how the government is organized (aka the Constitution), and what it produces (decisions & laws)
  • Stewardship is about what the police do; that is, enforcing the aforementioned laws
  • Management is what you, as a citizen (or a responsible parent) do on a daily basis, that is living and acting as happily as possible in the limits given by the laws

When reading the Wikipedia article about political systems, or playing a game like Civilization, you realize that human organizations grow along with their demographics and size: from bands to tribes, then cities, countries, and even empires. As the system scales up, it requires a different organization. As I have quoted before, “More is Different”.

Unfortunately, in the world of information technology, data has grown and still grows faster than the organization required to handle this data. As we start to take care of the homeless data in Excel spreadsheets, bottomless data lakes are spawning. As we try to figure out data quality rules and processes, artificial intelligence is starting to use this data in ways we cannot even manage. Is there a solution to this problem?

Back to Basics: Data Governance, Stewardship and Management – Governance In Action

We have discussed how data has grown faster than the organizations bound to manage the data. The growth of the data demographics too frequently leads to a state of emergency in the land of data and to an urge to quickly set up “Data Governance”, which can be summarized as: “Let’s write a constitution and name a bunch of politicians in charge and all our problems will be solved!”.

Governance aims at producing its own organization (the equivalent of the Constitution), and the various laws to organize the rest of the people. Decisions taken at the governance level do not immediately change the way you breathe and eat, but are critical to how things are and work, and affect the way you live in the long run. For example, see how the concept of citizen is described in human history (from the Greeks to the modern age), and you’ll see that the simple definition of this single concept is critical to the way everyone lives in a country.

In the context of data, Governance issues commonly agreed elements, such as:

  • Commonly agreed terms in Business Glossaries – E.g.: What is a Citizen? a Crime? an Offence?
  • Data Quality Rules, meaning what is lawful or not – E.g.: “Stealing is not good”
  • Processes – E.g.: “Stop at the red light”
  • Metrics defining how well things work (or not) – E.g.: “GDP is the total value of things produced in the country”
  • Roles and Responsibilities – E.g.: “The emperor tells all the rules and is never wrong”

Thanks to governance, you know how things are defined and how they should work. Well…

If you lock a bunch of politicians in a room for a couple of weeks, they’ll come up with a stack of papers with words on it. They’ll call them laws. Terry Pratchett (one of my favorite authors), wrote the following quote:

“It is the nature of Great Big Things that if the money isn’t spent on them, it isn’t spent on smaller scientific projects either. Small projects don’t advance bureaucratic or political careers as effectively as big ones.” – Terry Pratchett – Johnny and the Dead

Governance alone is one of these Great Big Things. Very rewarding, but both time-consuming and expensive, and also rarely efficient by itself. We will continue by explaining this bold statement in the next episode.

Back to Basics: Data Governance, Stewardship, and Management Part 3 – Stewardship

In our previous episodes, we discovered how data has grown faster than the organizations bound to manage the data, and how governance could help define the rules to manage this data.

Governance alone, although very rewarding for certain careers, is inefficient by itself if not applied. Besides (and unfortunately), the real world is still out there, with real-world issues, and really complicated (or more numerous) folks. These folks, and the world in general, are not always kind enough to comply, right?

“That was their law. The strongest man led. That made sense. At least, it made sense to strong men.” Terry Pratchett – Nation

To make sure that everyone understands agrees with, and eventually follows the rules, there is an obvious need for intermediate parties willing to explain or enforce the rules. Enter the Data Steward: He or she is in contact with the crowd, understands it, understands the laws, and wants to apply them. In an old Midwest town, it is typically the responsibility of the mayor, the teacher, the judge, and the sheriff to explain the rules and processes and have them enforced (preferably in that order). Wait? Why are we now in a Midwest town? We were with the government! Sorry for the aside – the laws have now traveled from Washington. Now they have their reality check. Let’s proceed.

The data steward is the person in charge of the data. He or she is in charge of the rules applying to the data. The mindset of a steward character can be summarized with one sentence: “Even if it’s not your fault it’s your responsibility.” With this mindset, a steward will enforce any rule to make his or her small town a better place. They may be able to make up and enforce the rules all by themself (in an early stage of civilization), or enforce the rules decided by the governance body.

Deciding and enforcing rules locally is fine on a small scale, and when the local town does not communicate much with the nearest towns. This is how it worked in the early stages of human organizations. The shaman/chief would make the rule for the tribe. At a larger scale, a central body (king or parliament) decides for the country, but the sheriff still applies the law locally. This metaphor applies to data governance and stewardship. You can enforce data stewardship without governance, having one person “in charge” taking care of the decisions locally, but only at a small scale. On a larger scale, you need to make decisions more or less consistent everywhere.

So you need governance before your “small data project” becomes “a lot of small and disorganized projects”.

Having said that, we still are missing the last part of our equation: data management.

Back to Basics: Data Governance, Stewardship, and Management Part 4 – Management & Collaboration

In our previous episodes, we have defined data governance and stewardship. It is time now to explain data management. We all (well, most of us anyway) agree that data is an asset, rules the world, is the new oil, and so on. Now, ask yourself the two following questions:

  • First: “Who can use the company’s checkbook?”
  • And then: “Who can export company data to Excel?”

In both cases, someone is moving (and possibly stealing) corporate assets, right? Data, unlike money, is too frequently treated lightly. And it is way easier to alter, duplicate and move than actual money. Yet, you cannot lock data as you would lock a checkbook in a drawer, because everyone needs data to do their job. In a modern enterprise, data management is everyone’s work. When one creates, modifies, or moves a piece of data, he is actually “doing data management”. The big question is: “How does he do it?”. Is it in an ungoverned way (without definitions, rules, roles, and privileges)? With or without the required help to understand the rules and see them enforced?

As you cannot put a cop behind each citizen, you cannot put excessive over control all the data. Some critical data (credit card numbers, for example) require strict control. For most of your data, you may, by governance and stewardship, enable responsible management by all your data citizens.

“’Make them think. Tell them what’s got to be done, and let them work out how.” Terry Pratchett, Nation

To use with our real-life metaphor, we (well, most of us anyway) do not need a cop to tell us every 5 minutes not to steal. We just don’t, because we were educated, because we do not need to, and because we understand what is at stake. The right balance of governance and enforcement turns us into good citizens.

A last point before we close this blog series.

In real life, we can send feedback to our governance bodies to tell them what works or not, and they sometimes listen to us. Seeing that someone up there is listening is critical to understanding and accepting governance decisions. In the data world, comments, complaints, metrics, ideas, etc, should flow back from the crowd of data citizens and stewards to the data governance bodies, who should adjust the governance directions and decisions to the reality of the field. When this circle is in place, you truly enable shared and collaborative data governance.

Data governance, stewardship, and management follow similar principles in the real world. They should aim at working exactly the same way, or even better.