E-Commerce Customer Segmentation: A Clustering Approach in A Web-Based Platform

This study develops a K-means clustering model to segment e-commerce customers into distinct personality groups (e.g., Platinum, Gold, Silver, Bronze). The model utilizes a dataset encompassing customer demographics (income, age, family size), spending behavior (total expenditure, product preferences), customer tenure, and engagement with marketing campaigns (campaign responses). Model accuracy is evaluated through comparison of predicted cluster assignments to established customer segment characteristics. A web application, built with the Flask framework, provides an interactive interface allowing users to input new customer data for personalized predictions and detailed cluster-specific insights regarding product preferences, campaign responsiveness, and suggested marketing strategies. The application outputs cluster assignments, key spending/purchase tendencies, typical campaign response profiles within a respective segment, and prioritized product recommendations. Findings demonstrate the model's ability to effectively group customers with theoretical implications and suggests potential for improving targeted marketing campaigns. This work highlights the application of K-Means clustering with a practical online platform through an implemented web app for data visualization. Acknowledged limitations in the generalizability of the dataset to the entire customer base are addressed.