Customer Lifetime Value Assessment via K-Means Clustering
· Analytics · Sabanci University
K-MeansRFMClustering Supervisor: Dr. Kemal Kılıç
RFM (Recency, Frequency, Monetary value) is the canonical CLV starting point but treats each dimension as independent. I extended it with a weighted K-Means clustering step that lets the algorithm discover segments where dimensions interact — e.g. customers whose frequency and monetary value are co-elevated but recency is low (lapsed VIPs). Implemented on an e-commerce dataset and validated against holdout spend.