Product clustering is one of the determinants of product development in sales activities. Zeeida Herbal products are engaged in health and beauty, of all the products sold, not all of them are sold, some are less well sold. Sales at the distributor of zeeida Medan products are still not computerized, namely by using manual recording. Every buyer who purchases either an agent, sub-agent, reseller or general customer who makes purchases through social media such as WhatsApp, Facebook, marketplace, and other E-Commerce is recorded in the manual bookkeeping, so there is often stock accumulation and even stock shortages at distributors. In this study, the authors apply the k-means clustering algorithm to classify products that do not sell (C0), sell very well (C1) and sell (C2). Clustering is a technique of one of the data mining functionality, the Clustering Algorithm is an algorithm for grouping a number of data into a certain data group (cluster). From this study, the output generated from the last 4 months, namely January-April 2022, shows that from 47 Zeeida products, sales of Zeeida products did not sell well in cluster 0, there were 39 products, while sales were very good in cluster (C1), there were 4 products and sales were sold in cluster. (c2) there are 4 products.