How to forecast the rollout response of a mailing list from a sample test in direct mail

Abstract

Testing new lists is the lifeline of the direct mail business because it provides access to new customers, which is necessary for future growth and profit. Mailers often lose money on list testing. This article proposes a procedure (method) that is cost-effective, and at the same time, provides adequate information about the universe so that the mailer can make the right decision whether to go slowly on a list or to jump from a sample test to a full run. Many articles have been written suggesting, on a theoretical level, how the rollout response can be predicted from a sample test, but no one has shown whether or not that method works when applied to real data. This article not only proposes a new method for predicting rollout response from a sample test, but also tests it with real data. Furthermore, it compares the results of the two methods. The results show that the method described here gives far better rollout predictors than the other method. This article also suggests a mailing strategy that can be used as a guide.