Tuesday, May 29, 2012

Wonderful read

Bob Kalaba, one of the brightest and finest people I have ever known, once mentioned in passing that, "regression is an admission of defeat."  Jim Manzi, author of the wonderful Uncontrolled: The Surprising Payoff of Trial and Error for Business, Politics and Society, would agree.

Manzi argues that the better way to go in order to figure out what works (and why) involves randomized field trials.  Many businesses perform RFTs all the time.  Capital One does 60,000 of them per year.  With their web-based customer data and access, they can easily do this -- as do many others.

Private sector entrepreneurs are incited by competition to get discovery right.  Discover prices, product mix, opportunities, etc.  This is their great contribution. 

Can their public sector counterparts discover "sweet spots"?  Manzi says "yes" if they embrace RFT.  There is policy variability in the world, but it is not yet suitably evaluated.  Students going to school within and across districts encounter different classroom modes.  There are data here to be mined.

The fifty States should seek and get federal policy waivers in order to experiment.  But are we still in RFT-land here?  In passing, Manzi notes that different communities have sharply different policies re Wal-Mart and big-box stores.  Evaluate variations in outcomes.

But there is a problem here that the author overlooks.  Most communities are not consistent when it comes to a pro- or anti- big-box policy.  There might be a deal one day and no deal another, depending on who has lined up who in city hall.

Minor quibbles aside, this is an amazingly useful (and enjoyable) book.  Students (and others) get good background in scientific method and evaluation.  Business school or policy school students (as well as many others) should read Uncontrolled.