Implementation of K-Means Clustering in Grouping Sales Data at Zura Mart

The efficiency of inventory management and targeted marketing strategies relies on understanding sales patterns and stock levels dynamically. This study proposes a K-Means Clustering-based approach combined with a real-time stock monitoring system to classify products adaptively. The dataset consists of 87 products with variables including total sales, average sales, and remaining stock. The analysis process begins with data normalization to standardize parameter scales, followed by the application of the Elbow Method, which determines the optimal number of clusters as three. The clustering results indicate that Cluster C0 (21 products) has high sales but low stock, Cluster C1 (59 products) has stable sales with moderate stock, and Cluster C2 (7 products) has low sales but abundant stock. These findings not only provide strategic insights for inventory optimization but also serve as the foundation for developing an automated recommendation system that links clustering results with adaptive promotional strategies and restock prediction. Thus, this study contributes to enhancing Zura Mart's business efficiency through the integration of data-driven decision-making in inventory management and marketing.