Entangled in strategies

Have you ever noticed that when you ask about data strategy, you sometimes receive a response about digitalization strategy instead? This can be confusing and frustrating, as it may not address your original question. Misunderstandings can easily happen when discussing complex concepts, even among professionals in the same field. Communication is challenging, and it's crucial to establish a shared language to ensure clarity. As data professionals, we often need to clarify terminology with business partners, even if we have experience in their industry. Organizations may use different terms to refer to common concepts, adding to the confusion. This is particularly true when it comes to data and digitalization strategies. To avoid further misunderstandings, it would be helpful to have concise descriptions of both mentioned strategies to guide the conversation in the desired direction.

So, what are digitalization strategy and data strategy? How are they different, and most importantly, where do they get mixed up? Here's my view.

Digitalization strategy is about the digital experience

Commonly digitalization strategy is defined as a plan to leverage digital technologies to drive digitalization and steer digital initiatives. It does this to transform business processes, products, and services to become more digitalized. However, there are far more detailed and wider definitions as well. Terms get cluttered when we find parts that are related but haven't been addressed. The natural instinct is to add them under the umbrella of existing terms rather than defining new ones. This commonly leads terms of the same general area to overlap. I'll do my best now to find the common ground that most of the definitions of digitalization strategy seem to agree upon.

Digital technologies and infrastructures are at the core of the digitalization strategy. Organizations need to build up a plan on what technologies they will use. Using every available technology, while it might be extremely fun, leads usually to some level of chaos. Selecting a technology doesn't mean that we must decide to use one for everything, but it usually is more efficient to use a very limited set of technologies for a specific type of need. For example, imagine that each business line would use its self-selected CRM. Doesn't sound good, doesn't it? The number of required integrations would be astronomical without any governance and centralized steering for selected technologies. 

A central architectural guidance is also needed to select data integration and processing patterns and tools. But these, in my opinion, are not strategy-level decisions. Where digitalization strategy might go quite a bit deeper are the digital initiatives aiming to produce new services or products. In such cases, it might be a strategic decision to create a new custom-built solution or to use and leverage something existing in a new and innovative way.

One core consideration with digital infrastructure and architecture planning is cloud transformation. This can also be seen as a general technical transformation because organizations are migrating old cloud solutions to newer ones. Concrete examples of cloud transformation digital infrastructure planning can include migrating on-premises data warehouse solutions to the cloud and upgrading ERP systems from on-prem to cloud versions or from cloud-hosted systems to SaaS-version. And the list goes on. Quite often, especially after organizational merges like acquisitions of other companies, digital transformation also includes transitions into central systems. And all of this requires networking and security. So, in a nutshell, the digital infrastructure roadmap is not only, or even most, just about the new solutions and digital technologies, but about how we keep the business running in the ever-changing digital environment.

Digital transformation concentrates on changing how the business operates and its customer experience works. It is about how users experience digital tools rather than what technologies are selected to produce them. Meaning that we go one level closer to the actual users. How people, internal users or customers, use digital services they are provided with. How we can improve the process of how people operate through digital technologies. When approaching a digital initiative, I would feel more comfortable considering first about digital transformation than infrastructure and technology. Acknowledging that digital transformation includes optimizing costs and processes, enabling new and streamlining existing business operations, and enhancing customer experience is also important.

As a one-line definition for digitalization strategy, I would propose the following: A plan on how the organization will leverage digital technologies to transform their business processes and operations to support their business strategy.

Data strategy is about utilizing data to support the business

The common ground in definitions of data strategy seems to be that it supports business in its decision-making. Here are some core principles to define it on a more pragmatic level.

Data strategy supports business strategy by transforming data into business insight. What kind of data do we need to support operational decision-making, and what data is required for new data initiatives? For me, this is the core principle of a data strategy. The responsibility of data strategy is to ensure that data is produced in the required format and quality. While it is not necessary to determine the technical details of how this is accomplished at a strategic level, decision-making should confirm that the requirements are achievable. Technical limitations are usually not the primary source of constraints on what is possible, as the availability of data and the business's ability to leverage results operationally are often the main factors. If mistakes are made in evaluating these real-life restrictions on a strategic decision level, it most likely will render the data project results unattainable or unusable.

Ensuring data capabilities and knowledge is also at the core of data strategy. The knowledge part has a few things included. Training people to use the available data and tools efficiently. Ensuring that data can be discovered and solutions are well documented is part of the requirements. Additionally, people participating in creating new data solutions need specialized skills and information to do the data development work. It is not necessary to acquire all knowledge internally, as external consultants can be engaged and compensated to contribute and address any knowledge gaps. Nonetheless, it is crucial to be aware of the knowledge that the organization should possess, the extent of such knowledge, and which parts are intended to be procured from external entities.

Data capabilities are a bit more architectural part of data strategy where, for example, an organization defines the data they want to be available for all employees, management, HR, and so on. On a more detailed level, strategy can define how the data needs to be processed and served to be viable for decision-making. This can include but is not restricted only to, real-time needs, the use of predictive algorithms, and accessibility. Defining data capabilities can also mean what technical capabilities and data sets are required for new digital products. Data capabilities and knowledge decisions concentrate on making sure that there is required knowledge to support data development and data usage, as well as defining the data sets technical capabilities needed to support business.

Data governance and data management are closely intertwined with data strategy. Although I don't consider these components to be a part of data strategy per se, the creation and updating of data strategy are part of data governance. On the other hand, strategic objectives are necessary to guide data governance and management in the correct direction. This creates a "chicken and egg" dilemma, which will be revisited later.

I would like to highlight that data strategy is often defined as a series of highly detailed decisions, such as determining which datasets are loaded into the data platform, specifying the types of analytical dashboards to be developed, or identifying particular algorithms to be utilized. It can extend to methods of operation and even individual KPIs to be established. However, I do not view this level of decision-making as strategic and therefore have excluded this perspective from the sources used for this blog.

To summarize, if I were to provide a concise definition of what data strategy entails, I might phrase it as follows: A plan aimed at supporting an organization's decision-making and digital products in alignment with its business strategy by identifying the required knowledge, capabilities, and data assets can be considered as data strategy.

The responsibility challenge aka. where things get confusing

Now there are a lot of definitions of the previously mentioned strategies that do not align with what I just have presented. Where I think the responsibilities of these strategies get most mixed up are the realm of data analytics and data culture, but before I go into these, I want to point out a few that are often seen under both, data and digitalization strategies. 

• In my view, technology decisions are clearly in the domain of digitalization strategy, and can even be considered tactical decisions. However, decisions about analytical tools, such as choosing to use Power BI as the main tool, are often included as part of the data strategy. If it were up to me, I would categorize the decision to use Power BI as part of the digitalization strategy. However, determining how the tool will be used, such as how workspaces are used to align with the business, is a decision that falls under data management rather than any strategy.
  There are times when a data strategy might decide on tooling. One example that comes to mind is the use of a data catalog. While the decision of needing, acquiring, and setting up a data catalog is a strategic decision, is selecting a specific technology also one? The point of this question is that there is a grey area on what decisions can be made on operational and tactical levels instead of the strategic level. In any case, the decisions of selected technologies should be transparent so that all parties are at least aware of made decision and if needed can validate them to be aligned with the overall strategies.
• Data governance, as previously mentioned, is the second one to be infused with strategies. For me, data strategy is part of data governance, not the other way around. Important yes, but for me, it doesn't have anything to do with strategic thinking. It’s a process. I can understand where the idea comes from, but having data governance should be given to any organization that is on the road to data-driven decision-making. If the data governance model were not in place and I had to assign it to a strategy, I would choose data strategy rather than digitalization strategy because it primarily concerns how data should be managed and utilized.
• "Everything" is the third and most common one. This is where under one term, everything related to data or digitalization strategies is mashed up, also regardless of whether it's a strategic or tactical decision. The lack of clear boundaries between what falls under data strategy and digitalization strategy can be confusing. As for myself, I prefer a more distinct separation between the two.

Now to the ones that need some extra consideration. It's difficult for me to determine whether data culture should be considered part of data strategy or digitalization strategy. I am not even convinced that it should be part of either, but data culture should be a separate thing on its own. Naturally, it can be supported by both strategies. For me data culture is about the working culture in the organization: How common it is to back up a decision with hard analytical data or if new problems are approached by trying to explain them first with the existing data and then, if needed, trying to find new data connections and/or sources that can explain them. So, data culture is about how a data-driven organization is through the different levels of decision-making, and how people use data to communicate in the organization. If I had to choose, I might be inclined to consider data culture as part of the digitalization strategy, because it is in the end more about people and how they operate, than how it is about how to specifically leverage the physical data.

Data analytics is another term that gets brought up in both strategies. In the case of data strategy, it means using data to make decisions by using data modeling, processing, algorithms, and analytics. For data digitalization, data analytics is defined less as how to use data and more as what it should be used for, like following trends, performance, and predictions to make decisions. I think if defined like this it is a bit clearer what part of data analytics is part of each strategy, but it might be worthwhile to think of better terms than data analytics to present these concepts.

In the end, I don’t mind at all if terms get mixed up or even, if absolutely necessary, they are combined as long as all of the core principles are addressed. I feel that in the end, we speak about the same things by sometimes using different terms or having slightly different meanings for the terms used. Only when the terms are combined, and an important part of the picture gets left out, we are going to run into trouble.

Data product strategy - The new kid on the block

Talk about a term without a standardized definition, here comes the winner of the punch. Due to the extensive range of definitions and interpretations of data product strategy, I won't even attempt to provide a summary. Instead, I will give the perspective that I feel is important in the context of this article so bear with me. Data product strategy is first of all a subdomain of data strategy. If the organization does its data development based on data products, then it needs a strategy for what it tries to accomplish with this way of doing. Data strategy should define if data products are used or not. The goals and intended purposes of data products, as well as the problems they aim to solve, should be defined by the data product strategy. 

As a core difference from a more standard data strategy, the data product strategy is based on customer needs. Instead of thinking about what is the most efficient way to serve data to users, it starts from a perspective of what data users most require and what would be the most desired way for users to access and utilize data. This requires defining who, and often also what (systems), are the users and what are the most valuable and important use cases for them. Also, the constant evaluation and measurement of data product utilization (analogically ROI) should be at the core of a data product strategy.

I wanted to bring up data product strategy in this blog because I feel it emphasizes what data strategy should be most about - how to produce the most value out of data. Using data products is a way to approach this pragmatically. It steers the way of thinking away from technical challenges and the most optimized and beautiful technical solutions into something that matters even more, producing value. Technological excellence is important to a certain level. It is required to produce well-functioning, reliable, and maintainable long-term data solutions. However technical excellence should not be the primary driver or consideration for data development. Ultimately systems are built to accommodate human requirements. If we lose sight of this idea, we are surely going to be lost.

A common pitfall

With so many strategies, also counting the one described in this blog, I think the most common pitfall is in the level of abstraction. Now, this can go in both ways, high or low, but the one that I have witnessed is that the strategy is left too abstract. What is described as strategy, for me seems more like a vision. A very high-level definition of what we want to do or achieve. For example, for a data strategy, we say that: we want to optimize our business processes by utilizing data and we want to create an advanced analytical product that helps us sell more. This is at best a vision. A bad one if you ask me. This is not by any means a strategy.

If a strategy doesn't define a single thing that gives any concrete measurement or goal to achieve, it's not a strategy. In the opposite scenario strategy can become a set of tactical decisions without any shared vision or strategy. It is challenging to assess the feasibility of any technical or architectural decision made in accordance with long-term goals in either of these scenarios. Without measurable or quantifiable long- to mid-term goals, evaluating decisions becomes extremely difficult for all levels of data governance, data management, and architectural steering.

I end my article with a quote generated by ChatGPT at the end of a long list of queries. I think this is the core message not to forget, no matter how we name or define our strategies:
"Overall, digitalization and data strategies are complementary approaches that can work together to drive business transformation and innovation. By developing both strategies and ensuring alignment between them, organizations can improve their ability to leverage digital technologies and data to drive business outcomes."