Bewitched, Baffled and Bewildered

http://insideanalysis.com/2014/04/trends-in-predictive-analytics/

quote

“Stagnant deployment. Ironically, despite the hype around predictive analytics, the percentage of companies that have implemented the technology has remained flat.This is hard to explain.”

Here is my explanation –

A. It takes time for people to trust 18 months of machine learning over 15 years of experience.

B. Coin toss will be 50/50 right – if machine learning is 53/47 right, What kind of leverage does it take to justify 3% improvement? Chances that 10,000 clients will default on 500,000 $ mortgages is frightening. Save 2 percent on default it pays the bills for months. 50$ shirts sold to 2,000 more customers is insignificant. The odds are the same, whether it is machine learning or not.

C. Someone may claim predictive analytics is 80/20 right. In that case, they have a rule to hold on to AND turn off the maintenance on predictive analytics. That is probably why prediction is not being adopted all that fast.

Prediction is really no more than bucketing items on a single score. It certainly will not tell us what-to-do to avoid falling in an undesirable bucket.

Is it time to look at Prescribing Action.