On Net Lift in a (Direct) Marketing Environment

• Activity Number: 590
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics and Marketing
• Author(s): Leonardo Auslender*+
• Companies: SAS Institute
• Keywords: net lift ; matching ; net lift curves ; random split

Abstract

Net Lift analysis belongs in the area of Propensity Score Matching with applications in clinical data and causality analyses. The starting point is the recognition that most available data set are observational and are not amenable to traditional experimental design techniques. In the specific case of direct marketing, a typical campaign divides the likely targets into treated (treated, mailed, phone called, etc) and controls (or untreated). While in clinical data analysis (e.g., ER care), the decision to treat or not to treat a patient is determined or conditioned by clinical information, the marketing decision may or may not be similarly decided upon. The aim of the analysis is to ascertain the outcome behavior (Y) of a customer/patiennt unit if he/she had been treated (mailed, given medical care, etc) when the unit was not treated, and in turn the outcome behavior of an untreated one.