Much discussion exists about the role of a data steward and what it entails, and in my experience it is quite unclear in many enterprises. A large part of this lack of clarity seems to arise from confusion with two other roles - subject matter experts and data managers. We need to understand how these three roles fit together. Either we should justify them as separate roles or recognize that they boil down to the same thing. The fact that many enterprises have only a fuzzy grasp of what these three roles involve is a major risk for the successful operationalization of data governance. This is because individuals will be assumed to have accountabilities that they have not been formally assigned, and which they are not equipped to carry out. To the extent that these accountabilities are assigned by a data governance program, it is data governance that will be blamed for lack of results.
What are SMEs and Data Managers?
Let's begin by looking at what an SME is. This is an individual who has certain expertise in a particular domain. "Expertise" is usually broken down into knowledge and skills; either the SME knows about a particular topic or knows how to get something done. Since we are talking about data, the SME is typically someone who knows about a particular data topic in the enterprise or how to do a particular thing with data. It is important to recognize that a SME is an individual person, rather than a role. You do not become a SME by being assigned as one.
A data manager is a relatively new role that involves overseeing the processing of data in an operational environment. My experience with this role is mostly in financial services, where it is explicitly recognized and the title is used. Typically, data managers look after master data, such as Customer or Financial Instrument. For instance, in an investment bank there are likely to be a number of data managers involved in onboarding institutional customers. Their tasks will be to check that the data received about a new customer is accurate, perform anti-money laundering checks, "know your customer" checks, credit checks, politically connected person checks, and ensure copies of all relevant documentation such as contracts, certificates of incorporation, etc. are received and stored. There may be quite a range of additional tasks, too. Basically, a data manager is oriented to working with data rather than what the data represents, and ensuring this data is available to the rest of the enterprise for consumption. The fact that data managers exist in back office operational environments also needs to be fully appreciated, but that is a separate discussion.
What is a Data Steward?
This leaves us with data stewards. The term "steward" commonly means someone who looks after something belonging to someone else. Since the data in an enterprise does not normally belong to anyone who works in it, then anybody working with the data can be considered a "steward," which is not very helpful.
Nor is it helpful to rebrand a SME or a data manager as a data steward. All we are doing is introducing an additional title that is not based on a real distinction, which is bound to cause confusion. What about formalizing the data management aspects of a SME or data manager, particularly when the individuals concerned do not have either of these titles? Simply clarifying what SMEs and data managers do is helpful, but if there is no change in what they do, it is difficult to see how any improvements can be brought about. And if this is what a data stewardship program does, and if it really does not result in anything improving, then the program is not going to be successful.
I would argue that data stewardship is a role, but unfortunately a role that does not have an easy analogy to anything else that we are generally familiar with. To me, a data steward has the following characteristics:
A data steward has responsibility for a particular domain of the enterprise, and knows who does what with data within that domain, i.e., he or she has local knowledge.
A data steward coordinates the promotion good data governance practices within their domain.
A data steward assists in getting resolution to cross-organizational data-related issues.
Data Steward Differentiators
We can compare these three characteristics to the SME and data manager. A SME will not be oriented to knowing who is doing what with data outside of their particular expertise, which is likely to be narrower than the domain of a data steward. Nor is the SME likely to be too interested in this - after all, expertise is a personal matter. A data manager will be interested in who does what with data, but within the scope of their operational accountabilities and only within the silo he or she is responsible for. By contrast, a data steward will be concerned about getting to know who is doing what with data in general within a broader domain.
Neither an SME nor a data manager will be concerned with promoting good data governance practices within their domain, unless these are needed to solve a particular problem. Operational environments are rife with tactical decisions about data that solve a short term problem for those immediately affected, but which cause long term headaches for the enterprise as a whole. Yet we cannot expect SMEs or data managers to behave differently because they are not working to a "bigger picture" that a centralized data governance unit would provide. By contrast, a data steward will understand that there is a bigger picture and act to rationalize what is otherwise an unending series of tactical fixes.
Industrial Age management, with its division of responsibilities does not match Information Age needs where data is a common resource driving the business. Very often, the wrong people make decisions about data, or no decisions are made, leaving problems to fester. Neither a SME nor a data manager is really oriented to these concerns. Even worse, since data crosses organizational boundaries, there are cross-organization issues with data that are quite common. Since SMEs and data managers have no mandate to direct activities outside of their organizational silos, these issues tend not to get solved.
Data stewards, or more properly a data steward network, can deal with cross-organizational boundary data issues. Such a network will itself have to be coordinated by a central data governance unit, but it is probably the only way of dealing with these issues that would otherwise fall into the cracks, or discontinuities, of the overall organization.
There is good reason to separate out SMEs, data managers and data stewards. There is a lot more to these roles than can be discussed here, but the important thing is to understand the differences between them.
Malcolm Chisholm, Ph.D. has over 25 years of experience in enterprise information management and data management and has worked in a wide range of sectors. He specializes in setting up and developing enterprise information management units, master data management, and business rules.
This story originally appeared on
