Abstract
e-retailing and database marketing are two emerging industries that require strong support from the CRM system. It is important for a website to keep its customers interested and come back frequently to visit. As web data and direct marketing data are available in huge volumes, data mining is an important and popular tool for both industries to develop good CRM systems to target loyal customers. Since most of this data is primary purchasing data, one could even go one step further to develop models to describe and predict behaviour of customers.In this study two statistical models from the theory of repeat purchase behaviour were used to analyse customer loyalty. The models were able to predict the percentage of repeat-customers, and were able to identify marketing variables which affect the repeat-rate.