Do We Still Need a DM Strategy? Part 1: Clarifying the Landscape.

Some professional conversations stay with you longer than expected because they raise a question you continue thinking about afterward.

This happened to me recently when Sathya Ramamoorthy, Director of the DAMA Toronto Chapter, invited me for an interview. I want to thank her for this opportunity sincerely.

We discussed data management (DM) strategy. But as often happens with this topic, the conversation quickly broadened. We touched on the relationships between data, data management, data governance, and AI strategies, and on the role of business in making strategy work in practice.

In this article, I will summarize the key insights from our conversation. However, the article will mainly focus on a question that came to my mind while I was preparing for this meeting:
In an era of tremendous unpredictability and rapid technological development, does an enterprise need a DM strategy at all? And if yes, how should an enterprise deal with it?

I don’t believe an enterprise can develop a three- to five-year strategy and unthinkingly follow it without quickly adapting to constant changes in the business environment. I will share my professional viewpoint on the topics I mentioned. However, I would also love to hear your thoughts.
I strongly believe that in the era when AI-generated content dominates media, the exchange of ideas between people becomes even more valuable.
So, let me start by sharing my current viewpoint on a DM strategy. I already wrote several articles on this topic several years back, so if interested, you can read them here.

Data Management Strategy Is a Long-Term Plan

A DM strategy is a long-term, high-level plan that defines the role of data and the data management function in achieving key business goals. It also explains how the organization plans to do it. In my view, the strategy answers three key questions.

Why does an organization need a data management function?

You can answer this question by defining how data management can help the business achieve its goals and survive in the long term. In practice, this means explaining which business risks data management helps reduce, which opportunities it helps create, and which decisions, processes, and products depend on trusted data. Without this link to business value, data management easily becomes a technical or administrative topic instead of a business capability.

What does data management mean for the organization?

In my experience, a set of core data management capabilities is required for any data-related initiative.

I mean enterprise architecture, data governance, metadata management, data security, and analytics.

However, the scope of these capabilities depends on the organization’s needs and resources. A small organization may start with a limited set of practical activities, while a large enterprise may need formal capabilities, roles, policies, tools, and performance measures. The key point is not to implement everything at once, but to understand which capabilities are necessary to support business priorities and reduce data-related risks.

How can an organization implement the data management function?

The DM strategy must set clear goals and objectives for each sub-capability, taking into account their interdependencies.

A high-level roadmap then demonstrates the plan, priorities, and resources needed to achieve these goals. This is important because data management capabilities rarely work in isolation. For example, data governance depends on metadata, data quality depends on architecture and ownership, and analytics depends on trusted data chains.

So, the roadmap should not only show what needs to be implemented. It should also show the logical sequence in which capabilities must be developed to support each other.

A Data-Related Strategy May Have Different Names, but the Essence Must Remain Untouched

Once we define what a DM strategy should answer, another question immediately appears.

Data strategy, data management strategy, data governance strategy, and AI strategy. Are they different? And what is the relationship between them? This is a question many professionals have asked me for a long time.

I understand why this question appears so often. Organizations use different terminology, industry frameworks do not always use the same definitions, and professionals often bring their own interpretations based on their roles and experience.

Let me share my view on it. Let’s start by discussing data strategy and data management strategy.

Data or Data Management Strategy Must Define the Business Value of Data and How Data Must Be Managed to Deliver This Value.

First, the fact. DAMA-DMBOK2 by DAMA International uses these two concepts interchangeably.

Years ago, in my book The Orange Data Management Framework, I argued that data strategy discusses the business value of data for the organization. In contrast, a data management strategy defines how data must be managed to achieve that business value. These are definitely two different but complementary topics.

Since then, I have had many discussions with professionals worldwide. My view is that it doesn’t matter what you call the strategy. The most important point is that it must define the business value of data and explain how an organization must manage data to deliver this value.

This is why I would not spend too much energy debating the label. The more important discussion is about the content of the strategy. Does it clearly explain why data matters for the business? Does it define how data should be managed to deliver value? And does it create a practical bridge between business ambition and the data management capabilities required to support it?

Data Governance Strategy Does Not Make Sense Unless an Organization Substitutes the Term “Management” with “Governance.”

Since 2018, I have authored dozens of publications (a couple of examples are here) and, in many workshops at international conferences, discussed one of the biggest challenges in the worldwide DM community: the lack of aligned definitions of data management and data governance. If you are interested in diving deeper into this topic, I recommend reading my series of articles on this subject.

Now, in short.

DAMA-DMBOK2 and DCAM consider data governance as one of the data management building blocks. It defines the DM operating model, including the DM organizational structure, governing bodies, and roles. It also oversees whether DM capabilities operate in accordance with defined policies and processes. In this view, DM comprises multiple capabilities, such as enterprise architecture, metadata management, data quality, data security, and so on.

Gartner appears to have substituted “management” with “governance” and considers these capabilities part of data governance. Furthermore, dictionaries, such as Merriam-Webster, indicate that governance and management are synonyms. As a result, the community often uses these two terms to describe the same thing.

The O.R.A.N.G.E. DM Framework I have developed applies the common definitions of management and governance. Management plans, organizes, directs, and controls activities, while governance oversees and controls how management carries out these functions.

Therefore, data governance is a sub-capability of data management and operates at two levels.

First, at the enterprise-wide level, data governance sets up the DM operating model, defines the DM framework, and coordinates the activities of all other DM capabilities.

Second, at the capability-specific level, it ensures that each capability is operationalized by delivering the required outputs through defined processes enabled by policies, tools, roles, and other assets.

In short, a data governance strategy makes sense if you use “governance” as a synonym for “management.” If not, defining a separate data governance strategy is not necessarily required. A roadmap for setting up governance is sufficient.

Data Management and AI Strategies Must Be Aligned, or Even Integrated.

I’ve investigated different definitions of a data product or data asset.

Don’t be surprised, but there is no consensus on the definition of a data product in our professional community. The full list of possible components may include data, metadata, software, ML models, databases, pipelines, hardware, networks, and services.

So, let’s think carefully: what is an AI system?

It is a combination of input and output data, metadata, models, and software.

Finally, doesn’t the broad definition of a data product fully describe an AI system?

So, my conclusion, which is not always supported by the audiences I have talked to, is the following: if a data product and an AI system have the same components, why should they be treated differently?

Practice demonstrates that organizations choose different approaches: from fully integrated to fully separated functions. Those interested in this topic can read the series of my articles and my book, Aligning Data and AI Governance.

My view remains the same: data management and AI strategies must be integrated or at least aligned, because AI is an enabler of existing business and DM processes.

And now, the core question is this: do we still need a DM and/or AI strategy at all? And if yes, how can strategic planning adapt to the rapidly changing business and technological environment?

And if yes, how can strategic planning adapt to the rapidly changing business and technological environment?

This is where the discussion becomes even more interesting. Because the real challenge may not be the strategy itself, it may be how organizations work with strategy after it has been approved.

Should a DM and AI strategy remain a three- to five-year document? Or should it become a rolling management mechanism that is reviewed, tested, and adjusted more frequently?

This is the question I will explore in Part 2.