The Business of Data: How to Put 2 and 2 Together

Contributed by: Jasmien Lismont, Jan VanthienenWilfried LemahieuBart Baesens, Seppe vanden Broucke

This article first appeared in Data Science Briefings, the DataMiningApps newsletter. Subscribe now for free if you want to be the first to receive our feature articles, or follow us @DataMiningApps. Do you also wish to contribute to Data Science Briefings? Shoot us an e-mail over at and let’s get in touch!

In 2010 Davenport, Harris and Morison already wrote that “There may be no single right answer to how to organize your analysts, but there are many wrong ones”1. Modern organizations are investing more and more in data and analytics2 and as data comes in larger volumes, faster and in new formats — think of Tweets and clicks — every large company finds itself considering something as a “big data strategy.” But what does this entail? One might wonder if these investments are really worth something if the right data culture is not in place to actually do something with all these new data-driven insights.

Put some bright, analytically skilled people in a room with some software and some data, and chances are small their results will bring you actual value. The problem already starts with the data. Because getting that data together in that room with your data scientists is harder than it seems. Data is not just there, and you need to think about which data you want to collect, at which level of detail, in which format and how you are going to store, share and update it. A thoughtful data governance is thus an important starting point. Once you have the data, any statistician is able to find something. But is that something the right thing (“not everything that counts can be counted, and not everything that can be counted counts”)? And if it is, how do we share and sell it to our business people? In other words, there are a lot of elements that need to come together – data, tools, hardware, techniques. More importantly is the fact that also a lot of people need to be aligned – data scientists, but also business, IT and others – who all need to come together to turn those elements into valuable decisions.

These people hence need to be organized and coordinated. Which means that companies first of all need to determine how they’re organizing their analytics and consecutively continuously need to check if this organizational format is still the right one for their company and analytics. Because not only the world of technology is changing, also your business is.

By means of a worldwide, cross-industry survey of senior-level executives executed in our research team, we were able to identify modern trends in the organization of analytics. Furthermore, we hope to extract these findings and generalize them by means of a framework.

We observe that there are many formats in which you can organize your analytics. More often than not, companies opt for a mixed approach; popular formats are centralized teams or centrally coordinated centers of excellence which can easily be combined with analytics organized on a departmental level. Don’t be afraid to share your data scientists with other departments or business units. It might help your firm to grow cross-departmental with regards to the analytical maturity. Moreover, it stimulates your data scientists. There are already a lot of companies that apply rotational deployment of their business employees across business departments. However, the use of this “training method” is less often observed for data scientists.

We also observe that companies often turn to external consultants or third parties for (additional) analytics work.

Why is that? On the one hand, it seems that there aren’t enough academically trained data scientists on the labor market – especially if you also want them to possess business understanding, presentation skills, etc. A couple of years ago, being a data scientist was claimed, even, to be the sexiest job of the 21st century³. No wonder it’s not always easy to find a good analyst when they’re highly wanted and not wide-spread. On the other hand, the market is becoming commercialized and analytics is supposed to become easier and more automated. Furthermore, there‘s also an economical aspect that comes into play. If we outsource our analytics to consultants or a third party, we get access to expert knowledge which is possibly even cheaper. Moreover, it can relieve our own data scientists’ team of elaborate, time-consuming tasks; or in-house data scientists could be trained by external parties with specific knowledge and expertise. A lot has been written about outsourcing, but always keep in mind that your data is something unique to your company which can bring you valuable data-driven insights. Thus look at outsourcing as another part of your analytics organization which needs to be fit in according to your strategy. Like with many things, ideally you go for a mix. Can’t find your guy or girl with a PhD in data analytics? Go for a physicist and a marketer and combine best of both worlds. Eventually, you want employees you can trust, so focus on the skills that are hard to train and dare to think out of the box. As the analytical environment changes, so does the typical data scientists team.

To provide some concluding guidelines:

  • There’s no ideal organization format for your analytics, but if you want to derive results out of your data and analytics, you need to organize it.
  • Once you have decided on an organizational format, don’t stop there. Evaluate!
  • Think about the qualities you seek in your data scientists and try to combine these in a team instead of in one person.
  • Don’t forget about training your data scientists. Rotational deployment might be an interesting format.


  1. Davenport, T.H., Harris, J.G., Morison, R. 2010. Analytics at work: smarter decisions, better results. Boston (MA): Harvard Business Press. Chapter 6, Analysts; p. 91-119
  2. Columbus, L. 2015 Mar 3. 56% Of Enterprises Will Increase Their Investment In Big Data Over The Next Three Years. Forbes.
  3. Davenport, T.H., Patil, D.J. 2012. Data Scientist: The Sexiest Job of the 21st Century. HBR. 90: p. 70-76