<![CDATA[Inaccuracies in sales strategies are often caused by the lack of optimal utilization of sales data. Large amounts of available transaction data are generally only used as archives, without further analysis to explore consumer purchasing behavior patterns. This condition makes it difficult to determine product segmentation, stock priorities, and appropriate marketing strategies. This study aims to assist the sales decision-making process by applying sales data clustering techniques using the K-Medoids algorithm. The method used is a data mining approach with the stages of collecting historical sales data, data preprocessing to handle empty data and outliers, data normalization, and the clustering process using the K-Medoids algorithm. The dataset used comes from sales data from a store with attributes such as sales volume, item price, and transaction frequency. The K-Medoids algorithm was chosen because of its ability to produce clusters that are more stable against outliers than other clustering algorithms. The results show that the K-Medoids algorithm is able to group sales data into several clusters that represent product sales levels, such as products with high, medium, and low sales. The information obtained from this clustering can be used as a basis for decision support in determining sales strategies, managing inventory, and planning promotions. Thus, the application of the K-Medoids algorithm has proven effective in supporting data-driven sales decision-making.]]>
