Banks and service companies in general are facing problems with multi-channel management, especially in the context of outbound communication. Problems are the high costs of the multi-channel systems, unsatisfied cus-tomers, and few customers who interact with companies. Personalization of the selection of communication channel to reach a customer is seen a solution to these problems. However, a complication is that currently no insight exists in what factors can explain channel preferences of customers. These factors are required for estimating the channel preferences of customer. In order to identify these factors a survey has been used to collect channel preferences of banking customers in the context of outbound contact. Furthermore, hypotheses about what factors are expected to explain channel preferences were constructed. These hypotheses have been tested through multinomial logistic regression models. Multiple relations between channel preference and predictors were identified. To assess the impact of the findings on the presented problems, it is recommended to start pilots in which the selection of a channel to reach a customer is based on the identified predictors.