Predictive Analytics Q&A with Tom Davenport

November 17, 2011

Tom Davenport Predictive Analytics1 photo (predictive analytics)Tom Davenport is the President’s Distinguished Professor of Information Technology and Management at Babson College, the co-founder and research director of the International Institute for Analytics, and a Senior Advisor to Deloitte Analytics. He has published widely on the topics of analytics in business, process management, information and knowledge management, and enterprise systems. His most recent book is Analytics at Work: Smarter Decisions, Better Resultswith Jeanne Harris and Bob Morison. He wrote or edited twelve other books, and has written over 100 articles for such publications as Harvard Business Review, Sloan Management Review, the Financial Times, and many other publications.

– What will facilitate the greater adoption of analytics?  Improving analysis skills, easier to use analytics tools, awareness of the value of analytics, availability of data?

– All of the above, I think. Right now the two fastest-moving drivers are the greater availability of data and the “marketing” of analytics by researchers and writers like me, and vendors alike. And I expect that both of these drivers, and the other two as well, have a long way to run.

Q – “Gut feelings” are still widely trusted in organizations everywhere. How do you marry “gut feelings” with “analytics-driven” decision making?

A – The two modes of deciding are not quite as alien from each other as many people think. First, even the most scientifically and quantitatively-focused analyst knows that an intuition about what’s going on in your data—we call it a hypothesis—can be a very helpful guide to understanding and modeling it. Of course it’s important to still test your hypothesis to see if it’s valid. And intuitive decision makers who have a lot of experience are actually letting their brains analyze a lot of data. If you don’t have much experience in a particular decision domain, you should treat your gut as the last resort.

Q – It’s believed by some that predictive analytics is only available to the largest, most sophisticated organizations. Do you believe that’s true and if so what will make it more accessible to a wider range of organizations?

A – No, I think that predictive and prescriptive analytics (the latter refers to optimization and randomized testing) can be done by any organization. Almost every company has some data that can be analyzed. The software is getting cheaper all the time, and there are even open-source options that are free. So the only real limiting factor is awareness of what’s possible and the skill to pull it off. Even that is more accessible through on- and off-shore outsourcing.

Q – How can organizations improve decisions? Will the move toward self-service, more collaborative and social tendencies of today’s analytics teams and technologies improve the decision-making process? 

A – This is the key question and what analytics are intended to do. They are one tool for that purpose. But there are a lot more. Another, as you suggest, is getting more input and participation in decision processes through the use of technology—social media, prediction markets, and so forth. Another is to take the learning from recent advances in behavioral economics and neuroscience, and embed them in our decision processes. In effect, we have both the need and the opportunity to reengineer our decisions. I’m not sure we have the will yet, however.

Q – What are your predictions for the future of business decision making?  What do you predict will happen to the industry in the next five-10 years?

That’s an interesting way of putting the question. I don’t think it’s really an “industry” yet. There are a few assorted toolkits. I think in the future we will try to connect the tools much more closely to the decision being made. We’ll have thousands of “analytical apps” to choose from, and each one will guide a decision maker through the process of making an effective quantitative decision. There won’t be the need to know what all the tools can do or where all the data is; all you’ll have to know is what decision you need to make, and the rest will be structured for you. Maybe we’ll even have software that keeps track of all the decisions an organization needs to make, who makes them, and what process they used to make them. This would be possible today, but again, I’m not sure that most executives are ready for that level of accountability. But the organizations that adopt such tools will simply make better decisions than those that don’t.