Organizations always take part in competition and as the competition grows; they are more concern about their customers rather than products. Organization always focuses on customer's behavior to retain in market competition. Churn prediction models are developed to manage and control customer churn in order to retain existing customers. Churn prediction aims to predict profitable customers. To predict the customer's behavior, data mining techniques are used. Various algorithms are used for the prediction of customer's behaviors. This paper proposes the enhancement in the performance of existing boosted tree algorithm by removing repeated scanning for scoring of the classifiers.