In this paper, a predictive model to determine customer desertion in Peruvian banking entities is proposed. The purpose of the model is the early identification of customers that reflect a behavior tending towards desertion based on financial movements, transactions, product acquisition, etc. The model is based on the analysis of a customer dataset to identify common traits through the use of SAP Predictive Analytics, and then comparing these traits to a different customer dataset, identifying those that are more likely to leave the entity. The commercial use of this model is the immediate application of loyalty initiatives that would enable the entity to retain the customer. The model was tested in order to identify the most efficient and precise one, being the R-K Means algorithm the best performing one, with a 93.20% accuracy and a better false positive/negative relation (8 and 3 respectively).