Companies want to believe that all advertising ideas—like the children of Lake Wobegone—are above average. But the bell curve dictates that only 30 out of every 100 television ads we test at consumer-products giant Unilever will be superior. Forty out of 100 will be just average, and, understandably, we don’t want to put average ads on the air.

A version of this article appeared in the April 2005 issue of Harvard Business Review.