Customer Segmentation on Returned Product Customers Using Time Series Clustering Analysis

Returning product is one of the supply chain strategies the company implemented. Managing the returned product is essential, the return policy is one of the strategies. Regardless, the implementation of the return policy is mostly applied in one procedure to all customers instead of a dynamic return policy based on customer segmentation. This study will analyze the return customer segmentation using data mining analysis. Time-series clustering is used in this study considering the dynamics of the data based on time series with Dynamic Time Wrapping (DTW). The RFM model is used as the attribute to cluster the customers' return transactions. The research shows the clusters based on the frequency and monetary attribute of customers in the pharmaceutical industry in Indonesia as the case to be analyzed with Recency, Frequency, and Monetary (RFM) model consists of three models for frequency attribute and two models for monetary attribute. The result on this study is clustering with time series could classified return customer which resulting three clusters for frequency attribute and two clusters for monetary attribute.