Data-Driven Product Segmentation for Sustainable Microbusiness Optimization

The rapid growth of retail sales data offers both opportunities and challenges for businesses in understanding customer behavior and optimizing operational strategies. This study addresses the issue of manual and intuition-based product categorization, which often leads to suboptimal pricing, inventory control, and marketing efforts. To overcome this, we implement and compare two unsupervised machine learning algorithms—K Means and Hierarchical Clustering—for segmenting products based on transaction patterns. Using a dataset consisting of one month of sales from a retail microbusiness, the study includes data preprocessing, algorithm implementation, and performance evaluation using three widely accepted metrics: Silhouette Score, Davies-Bouldin Index, and Calinski-Harabasz Index. The results indicate that K-Means with five clusters (K=5) yields more optimal segmentation in terms of compactness, separation, and overall cluster quality. These findings provide valuable insights for data-driven decision-making in microbusinesses, particularly in optimizing pricing strategies, stock management, and targeted marketing initiatives. This study shows that using machine learning is important not just for sorting data, but also for making smart plans in small retail stores. The findings can help build better tools for predicting customer behavior and creating tailored marketing approaches.