Product recommendation using typicality based collaborative filtering and churn analysis

Product recommendation is a widely used technique to attract customers and improve sales of an organization. This research proposes a novel methodology for retaining customers by combining the ideas of customer segmentation, product recommendation, and churn analysis using customer lifetime value, RFM analysis and survival rates. This methodology highlights the potential customers of an organization for targeting. In this research a clustering algorithm segments customers on the basis of weighted RFM values and marks heavy weights customer to be loyal. Product recommendations to loyal customers are given by using the Typicality based collaborative filtering. In this research, experiments are performed that show the superiority of TB-CF algorithm over other CF algorithm. The research also identifies effective ways and strategies for retaining loyal customers. The churn analysis determines the rates of customer attrition and the attributes which effect the customer attrition the most. The research discusses the organization's future dependance on the churn rate and the importance of customer retention. The churn analysis in due time is highly effective to prevent a catastrophic downfall of an organization if used properly.