Data science should be making the operational decisions in your business more accurate and effective – the aim of which, is of course to enhance the top and bottom lines in your business. Is this happening? Well, sometimes. A survey commissioned by datascience.com and executed by Forrester confirms many suspicions. First of all it seems we are hooked on data. We can never get enough, and will invest heavily to archive every little scrap. It would seem however, that this data often does little other than consume disk space. The skills and will to turn this into actionable insights are often lacking. This is not always the case, since some businesses take the idea of actionable insights very seriously. They are usually smaller firms, and ones without the political and organizational legacies that might stifle such an initiative.
Also reported in this survey is the fact that many businesses use several platforms and tools, most of which do not integrate well, and as a result a certain amount of inertia is created as people try to move from data to insight to action.
What we are seeing here is just a rerun of what has happened with IT. Business managers and executives tend to be suspicious of people who talk a different language. In fact this problem became so obvious and visible that CSC published a document called ‘CEOs Are From Mars, CIOs Are From Pluto’, just to highlight the issue. Well, if CIOs are from Pluto then data scientists are probably from Betelgeuse, and the lack of trust, a common language, and dress code does nothing to help the situation.
Forrester found that those businesses that had managed to fully embrace data science were much more likely to be leading lights in their industries, and much more profitable. But for many businesses it will not matter how much they spend on big data, data scientists (grudgingly typically), and data science platforms, the result will be the same – lots of data, lots of interesting ideas, and very little action.
There is a solution to this, and it is one I have personally witnessed. Set up a skunk lab outside the main business. In the instance I saw, the senior management within the business knew that introducing something radically new would have almost no chance of survival in the midst of established ways of thinking, and so they deliberately set up an environment away from the main business – and it worked brilliantly. Feeding solutions from the skunk lab to the main business was tricky, but that is why senior management need to be strong sponsors.
Without an initiative of this nature data science could well become a curse – plenty of expenditure and very little return. With such an initiative the payback can be rapid and profound. It goes without saying that the data science team needs to be business focused, and as such it helps to have sympathetic domain specialists involved.
As the technology matures so data science will become easier. Several AI powered data science platforms are already available, although these serve only to make the job of creating models more efficient. Understanding the problem and translating it into something data science can solve is perhaps the most important task.
Eventually businesses will be able to purchase solutions for many data science tasks – recommendations, churn reduction, sales forecasting, marketing spend optimization, and so on. But even here, businesses will need strong leadership to deal with distrust and fear of the new.