Some time ago, a colleague sent me a link to an article (https://www.techrepublic.com/article/forrester-functions-enterprise-data-organization/), with a note: “saw this article and thought of you.” I have long advocated ideas along the lines of what the article explains and agree with the general sentiment. However, the article addresses only half of the story.
It is missing the raw-materials perspective of data. This is not trivial, and it is routinely what gets neglected the most. The absence of the raw-materials perspective is a major gap I see repeatedly in organizations of all sizes and sectors. Not surprisingly, many people and organizations still struggle with data quality.
One would think that with such a common struggle, it would be obvious to every level-headed leader to adopt good, sensible practices. However, people are still singularly focused on making something out of data without paying enough attention to what goes in: data itself.
Imagine if manufacturers of anything took whatever they found for raw materials or parts and left everything to those running the manufacturing process. There would be material shortages, missing parts, parts not to specs, impurities in raw materials, costs out of control, bad quality, regulatory and contractual issues with materials, and supply chain issues, just to mention a few problems. It would be a giant mess.
It is hard to do two things equally well
I have discussed how data, analytics, and technology should be three separate roles at various times (see this article on data roles and responsibilities for example). The question is less commonly whether analytics and technology should be separate (although it is a question). Rather, the debate is often about which of these two should have data responsibilities and accountability. It reflects a lack of recognition that data must be viewed in its own right.
The role of Chief Data and Analytics Officer (CDAO), accountable for both data and analytics, has recently gained popularity. The idea is not new; I have seen it going back almost two decades. Something similar also happens with technology and data, in which case IT is assumed to have data accountability. That said, their data responsibilities tend to be more limited and less formally defined, or worse, assumed. The data accountability for the CDAO is much more formally and explicitly defined.
I am not a fan of the idea. Never been—because it is not realistic to pull it off well.
There are indeed cases in which combining the D and the A are necessary for good reasons. Most of these are organizations that are too small or that are too early in the stages of data and analytics maturity to have two separate roles. It is not unlike how the baker in a small mom-and-pop cookie shop handles both the ingredients and the baking. It makes no sense to have a separate person dedicated to sourcing flour.
However, at some point, organizations grow larger or more complex. The responsibilities become too much for a single role to handle both equally well. Have you ever tried to focus on two things equally? It is not easy! We naturally gravitate to one of the two things because we are humans. And we gravitate to the one we like better or feel more comfortable with.
All CDAOs I know of have grown into the role through analytics, not through data strategy and management. What invariably happens with data and analytics is that the A in the CDAO gets more attention than the D. Analytics is sexier and easier to show value. The raw materials view is the least sexy aspect of the data realm.
Role definitions for data must be value-based
In baking cookies, the tasks and skill sets involved are different between sourcing and managing ingredients and baking. You can define roles throughout the cookie supply chain based on tasks and objects without much confusion.
In contrast, many skills and task types are common throughout the information supply chain. Skill-based role definitions quickly lead to confusion and ambiguities. Instead, we have to think of the information value chain, and role definitions must be value-centric. Different roles bring different values to the table from the perspective of the consumer—the business.
Analytics and technology practitioners are trained for, comfortable with, and prefer their respective areas. As data users, analytics practitioners have the skills and the perspective for analyzing data quality. However, they are trained to leverage algorithms and other analytical methods, not to detect and maintain data quality. They analyze data quality only when they have to. Likewise, technology professionals are ultimately trained in something else—programming, systems, architecture, security, etc.
The lack of a value-based data strategy and management role is a tell-tale sign that there is no value attributed to the raw-ingredient view of data. A dual role just means they are OK with deprioritizing one of the roles because it rarely ever works to the satisfaction of anyone.
