Testing new direct marketing offerings: The interplay of management judgment and statistical models

The launch of a new product or service via direct marketing is nearly always preceded by a test of that offering. Such a ‘live’ test, conducted with a subset of the entire list of customer prospects, can sometimes be useful in a ‘go/no-go’ decision regarding a full-scale launch of the offering. More commonly, the test is used to direct the offering more effectively towards the market segments that appear most promising. Specifically, tests results are used and useful to determine whether a particular rental list of customer prospects should indeed be rented, and (for both rental and in-house lists) which specific customer segments should be contacted with the offering. This paper examines the effectiveness of managers' decisions related to designing a test and interpreting test results both conceptually – based on the literature of heuristics and biases in expert judgements – and empirically, for two new direct marketing offers. The paper describes how an interplay of management judgment and statistical models can lead to increased profits for new direct marketing offerings.