Three Mini-Rules of Thumb

It’s an easy rule to apply in practice: If a list earns a 2-percent response on a test, you should project a lower response rate on rollout, perhaps 1.75 percent, perhaps less.

What’s harder is to understand why it works this way. Nobody knows the exact reason, but here are two possibilities:

1. “Evil list managers.” This explanation darkly suggests that list managers provide “better” names on list tests, to encourage rental income, then deliver the “real” names on rollout. This explanation has a certain paranoid appeal to it, but it lacks credibility in practice—even if a list manager wanted to do this, he/she would have a hard time getting the data processing house to figure out how to make it happen.