Data science, AI and machine learning. Just a few of the topics that are currently booming in many organizations. Data is said to be the new oil, or gold or whatever other high value resource you can think of. And AI and machine learning are the mining tools that should turn data into competitive advantages, revenue increases and cost savings. While this is without doubt true to some extent, simply adding a data science team or department to your organization will get you only a fraction of those benefits. The reason is that many organizations are simply not ready to apply data-driven decision making throughout the organization.

Being ready for organization wide data-driven decision making does not come from having a group of data scientists. It comes from having a clear connection between business problems and analytics solutions, which in turn originates from the capability to translate business questions into analytics questions. There is no quick fix for this and dealing with it entails all aspects of the traditional “clover”: organization, process, technology, and people. While all are important, I would like to focus on two aspects most important to addressing analytics awareness: organization & people.

There are several ways to organize the data science capabilities in an organization, but broadly speaking it boils down to the choice between centralized and decentralized. Organizing data science more centralized has the benefit of increased strategic coordination and development, while more decentralized has the benefit of improving domain expertise and business intimacy. Regardless of the organizational model, analytics awareness will not be there automatic and requires considerable effort.

To create analytics awareness, you need to build the capability of translating business questions into analytics questions such that these analytics questions can be solved by your data scientists. This can be done in two ways that match with the two organizational design options.

  • In an organization with decentralized data scientists, you can teach the data scientists to make the translation. Being part of business units allows them to spot business questions and proactively address the analytics questions that might be in there.
  • In an organization with centralized data scientists, it makes more sense to teach the business to do the translation, because the data scientists are typically not involved in the day-to-day dealings of the business units.

At BearingPoint we help organizations build analytics awareness in various ways. First, we help (re)design the organizational structure to maximize interaction between data scientists and business users. Second, we deliver “inspirational proof-of-concepts”. This means that we provide the analytics translation as a service to build one or more analytics models together with the business users to inspire them and in the process train them to ask the right questions to their in-house data scientists. Third and finally, we design and deliver training curricula tailored to the specific situation of our clients that will develop analytics awareness, or even development skills.


Michiel Musterd