Implementation of BCBimax algorithm to determine customer segmentation based on customer market and behavior

Abstract

Customer loyalty and long term profitability are organizational goals to maintain their existence in uncertain business environment. In order to achieve the goal, knowing and understanding customers play a crucial part in the phase of product arrangement and development. By appropriately segmenting markets as well as applying different relationship management strategies, business organization can manage potential value of each identified segments. This research aims to cluster customer based on product beneficial expectancy using biclustering method named BCBimax algorithm. While hierarchical clustering and k-means clustering can only cope with one data source and demand similar data behavior over all phase of experiment, Bimax algorithm offers a new approach in processing both internal and external data source associated as united rows and columns. The approach does not require to reduce variables, instead it creates matrix and divide the matrix into sub-matrices. As a result, customer characteristics with high number of matched criteria are effectively obtained.