Modeling the Response Pattern to Direct Marketing Campaigns

Abstract

The authors propose a new model of response pattern to a direct marketing campaign. They estimate their model and four competing model forms using a new maximum likelihood procedure for response data from four different campaigns. Their proposed model fits the data very well. The authors then offer a method to use the model to estimate the eventual number of responses to a campaign using early return data and knowledge of past campaigns. Finally, the authors suggest an application to predicting returns to mail surveys.