Uplift Model: Building a Database Model to Assess the True Impact of a Test Campaign

At the beginning of everyday for the regression modeler, whose tasks are to predict a continuous outcome (e.g., profit) and a binary outcome (e.g., yes-no response), the ordinary least squares (OLS) regression model and the logistic regression model, respectively, are likely to be put to use, giving promise of another workday of successful models. The modeler is unwittingly assessing the model against the chance model, namely, a randomly selection of individuals. A stronger model would be one that assesses the true impact of a test campaign. The purpose of this article is to introduce such a model called variously uplift model, incremental model, true lift model, or net model. I call my uplift model the machine-learning GenIQ-TAC Model (test and control), which builds a model, for either outcome type, that guarantees to assess the true impact of a test campaign a given a control campaign. Two cases studies are discussed.