Most of the service providers and product based companies while launching brand new products, services or releasing new versions of existent products need to campaign to reach at the potential customers. While doing so they target their already existing customers who are the ambassadors of their company. To address the existing customers, they maintain the detailed customer data at all levels as customer maser data [9]. In this paper, we have built a prediction model to identify the customers who would most likely respond to the prospective offerings of the company basing on their past purchasing trends. Experiments have been conducted using the well known classifiers, viz., Naïve Bayes, KNN and SVM to classify a bank customer data. Subsequently, we have compared the effectiveness of these techniques and found out which one produces the maximum accuracy for the existing data set. © 2015 IEEE.