A Screening Tool Based On Neural Networks

Abstract

Countless words have been written about Customer Relationship Management CRM. Most of those papers focus on CRM's main strategies -identification, differentiation, interaction and personalization. These strategies in themselves are perfect, but we should not fail to discuss another crucial topic: what must be done after identifying the customer through a segmentation analysis? In other words, how can we know for sure that our marketing campaigns have been successful in reaching their target -that is, the customer? Our goal here is to introduce a tool that will allow us to identify a customer's interests through a segmentation algorithm; help create ideaslprototypes for products, services, advertising, or promotions; and test marketing actions aimed at specific clusters in order to find out if those clusters have been reached successfully. This paper presents a model based on a neural network which was created to control the response obtained from customers who have interacted with the company after the campaigns. The main idea is to reduce a lengthy questionnaire by using the principal component analysis -PCA. The next step will be to use the main variables to classify -what we call clustering -and to understand the new customer. The present work was based on a real marketing research (both qualitative and quantitative) conducted by a Telco and Health Company. The idea was to test customers' response following a few marketing actions aimed at specific clusters. We come to a close by examining the pros and cons of other tools, providing references to other areas of knowledge (such as education, pre-employment tests, and investor's profiling) and presenting our final results. Transactions on Information and Communications Technologies vol 29, © 2003 WIT Press, www.witpress.com, ISSN 1743-3517