Aplicação de análise estatística na previsão de perda de clientes: estudo em uma empresa atacadista.

Abstract

Winning the customer represents a high cost for a company. Generally, it is required investments in marketing, relationships, discounts, working hours in negotiations and several other actions that demand financial resources. In order to recover those expenses, the company needs to hold the client for as long as possible. In this context, the aim of this work is to propose a statistical analysis model able to foresee which clients are about to leave and stop being clients, through a study in a wholesale company. It was used the client’s registration data and relationship with the company to develop a predictive model of Logistic Regression. The proposed model achieved an accuracy of 94.6% and specificity of 48%, showing that it is effective for the prediction of loss of customers.