Raden's Corollary to Simpson's Paradox

My friend Gary Cokins just penned a blog "Is Business Analytics Just a Passing Fad?" As an old friend of the family, Leon Trotsky, once said (before Stalin had him killed with an icepick in Mexico City), "Some questions are answered merely by being asked."

I think everyone already knows my feelings about this, that analytics are crucial (and even fun if you're a math geek like me). What makes them crucial is that everyone can do it now. Otherwise, no one would need it. When everyone was shooting from the hip, efficiency was a matter of degree. If everyone used crude models and unreliable data, then everyone should, more or less, work within the same margin of error. What separated competitors was good strategy, good execution, etc. So I suggest than once everyone starts using naive Bayes, EM and support vector machines, it will still be the strategy and execution. Analytics are just table stakes. It pains me to say this, but I believe it.There are exceptions. In my old patch, actuarial work, the (continuing) use of analytics is not a competitive issue or a game changer, it is an essential part of the "manufacturing" for risk analysis, reserving, etc. Probabilistic modeling for worse case scenarios like oil gushers or safety of nuclear waste over thousands of years have nothing to do with competition.

Based on my own experience as both a predicitve modeler and analyst/consultant, I really don't see business analytics thriving for more than year or two before it is displaced by something else that (we) analysts think is cool because -- and I modestly call this Raden's Corollary to Simpson's Paradox -- no one wants to learn advanced analytics but every organization wants to do advanced analytics. It's like automatic seatbelts; remember those? Designers thought they were the next great thing, but nobody really wanted them.

Gary quotes three former colleagues from Deloitte who are all quite sanguine about the future of business analytics, as well they should be. They'll make a fortune on it. So I would have to say that this Gang of Three (all fine people, I mean no disrespect) are what you would call, in business analytics, a very biased sample.My friend Gary Cokins just penned a blog "Is Business Analytics Just a Passing Fad?" As an old friend of the family, Leon Trotsky, once said, "Some questions are answered merely by being asked."