“Data strategy outlines when, where, and how much we should invest in people, processes, and technology to meet and exceed our strategic goals”
You have decided on a business strategy that requires you to make investments in data. You may have had some successes with data projects already or you may just be at the start of your data journey. Either way, you need to up your data game to successfully execute your business strategy. It sounds to us that you’re in need of a data strategy!
Where to start?
Data is the new oil. But as with oil, the value of data lies in what you do with it. Data becomes an asset when turned into actionable insights that support the decision-making in your organization. Your first data project may require only an analyst capable of extracting relevant insights from the data you already have available. But as a leader in your organization, you realize that sharpening your competitive edge will require your whole organization to become more data-driven.
To proceed, you will need a solid technical infrastructure complemented with standards and policies that ensure the availability, usability, integrity, and security of your data. In addition, you will need a work force with diverse and complementary skills, and the adoption of a data-driven mindset by all employees. To make all of this happen, you will need a roadmap that tells you where to invest at each step in your journey to become more data-driven. In other words, you will need a carefully considered data strategy. To explain how the intelligent data-driven organization relies on a data strategy, let’s draw the analogy with a soup kitchen.
The soup metaphor continues
In our interactions with clients, we use an analogy between building algorithms and making soup to explain the process of data science to decision-makers. Because this analogy has proved so successful, we’ve adapted it to the concept of data foundations, where it again helped us to easily explain the rather technical process of ingesting, processing, and serving data. Therefore, we’d like to use another soup analogy to explain the concept of data strategy.
The soup kitchen
Imagine your goal is simple: you want to make money selling soup in a restaurant. To start, you’ll need a chef that can turn ingredients into soup. In organizations that want to become data-driven, the business intelligence analyst or data scientist can be seen as the initial chef, the data as the ingredients, and the business insights as the soups that are served.
At some point, your restaurant may grow in popularity, and you may want to start serving more types of soup to your expanding clientele. Similarly, you may have successfully carried out an initial data science project and now want to do more with data. In the following, we’ll use the metaphor of the soup kitchen to describe what you’ll need to continue your data journey. Afterwards, we’ll explain how data strategy helps you successfully navigate your journey.
In a soup restaurant, expanding your menu means you’ll need more ingredients and more fridges to store them. Your chef could probably use some assistants as well: A purchasing manager who makes sure the fridges are stocked and kitchen staff to chop the ingredients and put them in place for the chef to do his or her magic.
Likewise, when you want to extract more insights from a wider variety of datasets, you’ll need technical infrastructure to ingest, process, and make data available to your analysts, while regularly storing the data in raw or processed form. We refer to these capabilities as data foundations, because they lay the groundwork for an efficient scaling of your data usage.
Every workplace has some rules and responsibilities, and a restaurant is no different. Some of these make sure that the food that leaves the kitchen tastes good, and above all, is safe to eat. For example, the purchasing manager checks the food delivery for ingredients that have gone bad. Anyone who stores food in the fridges, labels and dates it. And kitchen staff taste everything they prepare before it leaves their station. Other rules and responsibilities make sure the kitchen operates like a well-oiled machine, even during busy periods. For example, workstations are organized to avoid wasted movement. While walking in the kitchen, staff announce their presence to avoid collisions. No one touches another chef’s knives, and “Yes chef” is the only right answer.
For organizations that work with data, rules around who can access which data – enforced by access management tools – will keep data secure and private. They will, however, also reduce operational friction, as employees no longer have to put in requests to the IT department to retrieve certain data. Data should be accurate and available, and when they’re not, it should be clear who is responsible for fixing the problem. Finally, good documentation practices ensure that data is findable and understandable. We refer to these rules and responsibilities as information management. Together with data foundations, they are essential for growing your data operations in a fast and consistent manner.
Above, we have described what happens behind the scenes of organizations that successfully use data to execute their business strategy. However, we have yet to define data strategy. Whereas a business strategy outlines targeted outcomes, a data strategy specifies how the use of data contributes to achieving these outcomes, and where investments should be made to realize the desired level of sophistication of data usage.
Let’s return to the soup kitchen. Imagine we want to increase our revenue by attracting new customers. Our business strategy could then consist of offering more types of soup. For this, we’ll need to buy new ingredients, and depending on our current setup, we may need more fridges, more kitchen staff, chefs with other skills than our current chefs, etc. However, if our business strategy consists of selling greater volumes of the soups we currently have on offer, we should instead first invest in fridges and kitchen staff.
In a similar fashion, our data strategy outlines when, where, and how much we should invest in people, processes, and technology to meet and exceed our strategic goals. For companies at the beginning of their data journey, this will start by selecting one or more high impact use cases. These use cases will help prioritize certain investments over others. Often, one of the first items of business is the recruitment of an analyst or a data scientist. When carried out successfully, these initial use cases will whet the organization’s appetite for further data projects. Companies with greater levels of data maturity will need different types of investments, depending on their current data capabilities and strategic goals. For example, they may want to segment their customer base to better allocate their marketing budget. For this, they should start collecting data on their customers’ buying behavior. Or, they may want to react more quickly to price changes by competitors. For this, they may need to invest in faster infrastructure. In a final example, they may have been haphazardly adding data to their customer relationship management system, and now find themselves in a situation where important business concepts have multiple definitions and many of their records contain duplicate or incomplete information. Before doing anything else, they should invest in proper information management.
What distinguishes a good from a bad data strategy? A good data strategy will function as an easy-to-understand roadmap, by providing a sequence and prioritization of investments in your data capabilities that is supported by all relevant business stakeholders. Ultimately, however, the proof of the pudding is in the eating. A good data strategy will enable better and faster decisions and will put an organization in a situation where the value of their data only increases with time.
Let’s cook together
Did our soup story inspire you to discuss how we can take the next step in your organization? Contact us and we will open our kitchen to share our favorite recipes. We would love to help you take the next step in becoming The Intelligent Data-driven Organization!
I want to thank Hendrik De Winter for writing an initial version of this article and Nick Janssens and Geert Verstraeten for the discussions and feedback that led to this article. Any remaining errors are mine.