Customer Segmentation using Data Science Techniques

Many businesses send the same marketing message to every customer, which wastes budget on customers who are unlikely to respond and under-serves the most valuable ones. This paper applies data science techniques to segment the customers of an online retailer and to link each segment to a practical marketing strategy. Using a sample of 150,000 transaction records from the public Online Retail dataset, each customer’s Recency, Frequency and Monetary (RFM) values are calculated, log-transformed and standardised, and customers are grouped with K-Means clustering. The Elbow method, supported by the silhouette score, indicates four clusters. After cleaning, 2,225 customers were segmented and labelled in business terms as Loyal, Big Spenders, New and At Risk. Loyal customers form only 12.2% of the customer base but generate 51.6% of revenue, while At Risk customers make up 42.0% of customers but only 9.9% of revenue. A targeted strategy is attached to each segment, such as rewards for Loyal customers and win-back discounts for At Risk customers, and the segments are presented in a dashboard of charts. The study shows that RFM-based K-Means segmentation, when interpreted in business terms, can turn transaction data into actionable marketing decisions.