I’ve been working my way through Nate Silver’s The Signal and the Noise and early in the book he shares an interesting finding about the predictions of political pundits. Basically, there is a negative correlation between how often a pundit appears on TV and the accuracy of their predictions. The louder and more visible the talking head, the worse their predictions tend to be. I share this because in business media, prognostications continue to get bolder, louder, and more frequent. Erudite experts present vast databases and newfangled models that fuel their analytical crystal balls. However, the media’s love affair with data-driven predictions has not proven infectious in the Business Intelligence (BI) community. Past Aberdeen Group research has extolled the value and benefits of predictive analytics in a number of functions and industries. However, new data from Aberdeen’s 2014 Business Analytics survey shows that adoption of predictive analytics has not taken off as one may have, well, predicted.
For the past three years, Aberdeen has asked its BI community about current and future adoption of predictive analytics (Figure). From 2012 to 2013, predictive analytics appeared to be gaining traction and preparing to take off. Adoption within the community grew by 39%, indicating that many of the 2012 respondents that claimed they were planning adoption within 12 months actually followed through. In addition, the percentage planning to adopt within 12 months held steady while the percentage planning to adopt beyond 12 months shrank. This seemed to indicate that a segment of organizations were moving forward with implementations and would be adopters in 2014.
However, the 2014 data set indicates that the apparent gains in 2013 were either misleading or temporary. The adoption rate in 2014 is at 19% across all industries. Not only is overall adoption much flatter than it previously seemed, but the “2012 planned adoption beyond 12 months” numbers and the “2013 planned adoption within 12 months” numbers now look soft and overly optimistic. What does this say about the 57% of respondents in 2014 who claim to be planning to adopt predictive analytics either within or beyond 12 months? Also, those claiming plans to adopt predictive analytics beyond 12 months now represent the largest group in the community, when in 2013 they were the smallest. This apparent lack of enthusiasm in the community for predictive analytics is difficult to reconcile with the manifest successes of predictive analytics adopters in 2014.
Predictive analytics lack of accumulated traction in the BI community is surprising given the results and improvements organizations with predictive analytics have achieved. Stay tuned for the full report, which will examine the manifest benefits of being a predictive analytics adopter.