When Data Mining Makes You Smart, Should You Play Dumb?

If your data mining tools are successful at giving you business intelligence about your customers, should you act like you have that intelligence?  Or should you play dumb?

I started thinking about this after reading an article in this weekend's NY Times called "How Companies Learn Your Secrets" by Charles Duhigg.

The article describes how Target stores have had success at mining their customer data to figure out which customers are likely expecting a baby.  In fact, they even go as far as predicting the due date.  It turns out people change their buying habits during periods of major change in their lives, and having a baby is about the biggest life change that comes around.  However, if you wait until after the baby is born, you are already too late.  The changes in buying habits have already occurred.  Target tried to influence those buying habits during the second trimester of a pregnancy.

And they were good at it.  Too good in fact.  When women got coupon books with coupons for pregnancy and baby needs, they got spooked.  Their response was, "How did Target know I was expecting when I haven't told them?"  They wouldn't use the coupons.

So Target dumbed things down a bit.  Instead of sending customized flyers with just baby products, they added random coupons for things the expectant mother would not be interested in, such as a lawn mower.  When the woman thought she was getting the same flyer as every other house on the block, she would gladly use the baby coupons.

It seems like business intelligence is becoming more like its namesake, military intelligence.  When you know something about your adversary, it is only helpful to you if they don't know that you know it!  When the British broke the German Enigma code, they obviously wanted to use it in their war efforts, but they were also careful not to demonstrate their knowledge to the Germans.  That meant sometimes you didn't use that intelligence.  The most famous instance (still debated about its truthfulness) was when Churchill knew Coventry was to be bombed based on Enigma intercepts, but did not take defensive measures which would tip off the Germans.

Now that we are achieving some success at business intelligence, we have a new decision to make:  When do we choose to NOT use our knowledge about our customers in order to keep them buying from us?