A prediction model of retention in a Spanish fitness centre

Abstract

[Abstract]: Studies on customer loyalty commonly use future behavioural intentions, even though these do not always correspond to actual behaviour. This study aimed to create a model for predicting dropouts at a sport centre based solely on actual customer behaviour. Monthly data corresponding to the actual behaviour of 1977 members at a sport centre in Spain were recorded over the course of one year; these data represented the historical behaviour of the participants. The variables selected to develop the model have been used by other authors to study behavioural loyalty, although in this study users were not surveyed. A dropout prediction model was created by means of logistic regression based on real and common records in the management of users at sports centres. The predictive model was formed by: age of the subjects, number of attendances to the sport centre and their duration, their link to the centre (length of membership and ex-members) and economic aspects. The model’s goodness of fit was considerably high showing more than 70% of effectiveness. These results suggest that the recording of historical behaviour from sport centres costumers can be a useful tool to predict the chance of dropping out.