Abstract

The South African business banking sector has seen a high number of new to the bank, business banking customers closing their accounts within three years of having their account and operating their business, which then leads to them ending their relationship with the bank, either due to lack of funding, business insight or support from various stakeholders. Through the use of Predictive Modelling and Analytical Techniques, variables significant to business failure can be identified and a customer retention strategy can be created, to help business banks to successfully retain business banking customers and provide them with knowledge to help grow and groom their business. A logistic regression model is used with the objective to explain the relationship between the dependent binary variable and the independent predictive variables. For the interpretation of the logistic regression, the Gini cofficient as well as the information value is looked at. The logistic model is evaluted using the confusion matrix as well as the receiver operating characteristic chart. Then conclusions are drawn from this study.