ABSTRACT
Marketers use decile analysis to assess predictive incremental gains of their response models over responses obtained by chance. There are many perspectives, which account for the confusion matrix branding. In marketing, the perspective precision, unbeknownst to marketers and apparently not in the literature, is the underpinning of the decile analysis. This chapter presents four widely used perspectives of model performance: accuracy, specificity, sensitivity, and precision. Accuracy and precision values without corresponding baseline values are subjectively qualified as excellent, good, poor, or whatever hyperbolic adjective the statistician chooses. Marketers use the decile analysis. The chapter focuses on the proposed performance metrics, model decile-analysis precision and chance decile-analysis precision. These metrics allow for singular decile analysis assessment of a given model. Decile-analysis precision is the underpinning of the Cum Lift, which is the chief factor in the success of the decile analysis.
