We discuss the use of segmentation as a predictive model for supporting targeting decisions in database marketing. We compare the performance of judgmentally based RFM and FRAC methods to automatic tree classifiers involving the well-known CHAD algorithm, a variation of the AID algorithm, and a newly developed method based on genetic algorithm (GA). We use the logistic regression model as a benchmark for the comparative analysis. The results indicate that automatic segmentation methods may very well substitute the judgmentally based segmentation methods for response analysis, and come only short of the logistic regression results. The implications of the results for decision making are also discussed.