Promoting target models by potential measures

TL;DR

This paper introduces a potential measure for evaluating target models in direct marketing, emphasizing its advantages in fostering stretched targets, identifying improvement areas, and enabling automated model assessment.

Contribution

It proposes a new potential measure for target model evaluation that is cut-off invariant and potential oriented, improving upon traditional economic benefit measures.

Findings

The potential measure can be calculated effectively.

It supports automated evaluation of multiple models.

It helps identify models with high improvement potential.

Abstract

Direct marketers use target models in order to minimize the spreading loss of sales efforts. The application of target models has become more widespread with the increasing range of sales efforts. Target models are relevant for offline marketers sending printed mails as well as for online marketers who have to avoid intensity. However business has retained its evaluation since the late 1960s. Marketing decision-makers still prefer managerial performance measures of the economic benefit of a target model. Such benefit measures have merits but are unfavorable in other respects: They constrain leadership by stretched targets since they do not tell us how good a model could be. And they require a predisposed decision regarding cut-offs. Since this is based on earlier optimizations it virtually means sticking to traditions. Hence it is recommended also to use cut-off invariant and potential oriented performance measures for the model evaluation. This has three advantages: sustaining stretched targets, identifying improvement potential and sup-porting an automated evaluation of many different models. This article introduces a concrete po-tential measure and shows how to calculate it. It is especially recommended for direct marketing businesses churning out many specific target models at short intervals.