To own data or not to own data, that is the question. This question comes up often when I am speaking with clients or groups of people during my Data Governance webinars and conference presentations.
To the disgust of some (people that choke when the term “data owner” is used), using the term “owner” instead of “steward” appears to be the rule rather than the exception. The semantics of “owner” versus “steward” appears to drive many conversations.
The definition of a Data Steward that I use (and a core concept of the Non-Invasive Data Governance approach) focuses on formalizing accountability for data resources.
A person is a Data Steward if they are held formally accountable for their relationship with the data. The relationships are 1) as a person who defines the data, 2) as a person that produces the data and 3) as a person that uses the data. I have been known to say that Everybody is a Data Steward and that people should get over that mere fact.
The truth is … pretty much everybody in the organization (or at least a large percentage of people) define, produce, and / or use data as part of their job. The reality is that these people, for the most part, are not being held formally accountable for how they define, produce and use data. Everybody is a data steward yet the stewards do not even recognize themselves as such. Getting the stewards to recognize themselves as stewards is part of the job of the Governance Administrator or the person(s) responsible for the implementation of the program.
Stewards do not, in fact, “own” the data, but rather they take care of it. Just like a babysitter takes care of the kids and (hopefully) returns them safely at the end of the evening, the Data Steward takes care of data for the period of time in which they are related to the data as a definer, producer or used of the data. When the babysitter leaves, so does their responsibility.
A person is a Data Steward if they are held formally accountable
for their relationship with the data.
Over the years, I could just imagine Chrissy (Friday night’s babysitter) arguing with Tracy (Saturday night’s babysitter) over who owns my kids. Although there may have been times when I wished this were true (not really), the truth is (besides for the fact that my wife and I seldom went out two nights in a row), the babysitters only have responsibility for the kids when they are in charge of taking care of them. Just like the data steward only has accountability when the data is under their “watchful eye”.
Ok, ok… that is a silly comparison. Or is it? When a babysitter arrives to watch the children, they are basically charged with SPECIFIC accountabilities, not ALL accountabilities. They are supposed to keep the kids safe, keep the kids happy, get them in their pajamas, and force them into bed (those with little kids can relate to the last action). Predefined actions are the basis of their accountability. The babysitter is not responsible for seeing that homework is completed (for example … that is a parental after school chore).
The babysitter is not responsible for teaching kids their ABCs or educating them about right from wrong. They are accountable for actions that are plainly defined ahead of time so there are no questions about what their responsibilities are. Just like data stewards should be.
The people who are responsible for defining specific data must have formal accountabilities that are related to defining that data. These individuals only have the accountability for the data that they define. Responsibilities for defining data include the creation and maintenance of data definitions for the company, integrity and quality of the definitions, following data definition standards, and communicating concerns, issues, and problems with data definition to the individuals that can influence change.
The people who are responsible for creating, modifying, or deleting specific data have accountabilities that are related to these actions. Responsibilities for creating, modifying, and deleting data include the integrity and quality of the data handled by that department, completeness and timeliness of data, management and control of data, and communicating concerns, issues, and problems with data management to the individuals that can influence change.
The people who are responsible for consuming specific data (using the data for decision making purposes or reporting information to internal and external entities) have accountabilities that are related to this action. Responsibilities for consuming data can include accountability for data usage, communicating new and changed business requirements to individuals that will be impacted, and communicating concerns, issues, and problems with data consumption to individuals that can influence change.
Stewardship often fails because of complexities that are not discussed when the stewardship part of governance is being defined. One complexity is to plan for how you are going to account for everybody (or almost everybody) being a data steward. Another complexity is to plan for different levels of stewards with different levels of responsibilities like Operational Stewards and Tactical Subject Matter Experts Stewards. Planning to handle these complexities are the true guts of data stewardship.
Planning to handle steward role complexities
are the true guts of data stewardship.
These items should be considered when defining Data Stewardship as part of your Data Governance program:
- Definition of roles and responsibilities (in other words, the accountabilities that come with the actions)
- Procedures for collecting information about who acts on data and how that information will be kept up to date (hint: it is like hitting a moving target).
- Selling the need for stewardship to the organization and making certain that people in the company recognize the importance of formalized accountability
- Procedures for stewards to go head-to-head to hash out data related issues (resolving data related issues is the ultimate goal, isn’t it?)
Companies squabble over the semantics of whether the steward actually owns the data. “Ownership” implies that the steward can do anything they want with the data (“heck, I own it, don’t I?”). People don’t own data; they take care of it. Like a babysitter.
If we aren’t going to use the term “steward” (instead of “owner”), I’d rather see us use the word “babysitter”. Consider the steward to be the Data Babysitter. They have responsibilities and accountabilities tied to the actions that they take when, in fact, they are on the clock.
By the way, my wife’s lease expires next month on the car she is “watching” for the dealership. And to think, she has been the car’s steward for the past two years. Argh… the babysitter analogy is better.