Testing the predictive power of PLS through cross-validation in banking

This paper aims to predict the relationship quality (satisfaction, trust, and commitment) of a customer with a Spanish saving bank by using both the quality of service operations and the usage of e-banking, on the relationship quality (satisfaction, trust, and commitment) of a customer with a Spanish saving bank. A competitive strategy follows testing three alternative models employing PLS path modeling with a primary dataset obtained through both an online survey and internal bank databases of almost one thousand customers. This study includes the predictive validation of models using hold out samples and testing for casual asymmetry. The results indicate that both the quality of service operations and the customer's use of e-banking successfully and independently contribute to predict the customer's relationship quality. The face-to-face service encounter is still the major contributor of relationship quality, but the use of e-banking tools also improves the relationship quality of customers in the banking industry. This model is consistent across a multi-group of customers.