Abstract:
Product segmentation is one of the marketing strategies used by business units to achieve
goals, which include decisions regarding target markets, product placement, and product
promotion. To could perform product segmentation more effectively and efficiently, it is
necessary to process sales data, by grouping product data from the level of sales. Er
Coffee as the object of research, has a problem with how to group products based on the
level of sales, to determine the right strategy. The sales transaction data used in this
research are 27330-row records consisting of 186 product data with attributes used are
the number of items and unit prices. This study uses the K-Medoids using the Minkowski
Distance to group product data into 3 clusters based on high, medium, and low sales
levels with the aim that each group will produce different promotional strategies for each
level of sales. The product grouping analysis system in this research is web-based using
the PHP programming language. The cluster results from this study were 76 items with
low sales levels, 106 medium sales products, and 4 high sales products, then an
evaluation of the cluster results was carried out to determine how well the accuracy level
of the cluster results was. Testing the results of clustering using Silhouette Coefficient,
based on this process, the k-medoids are good at grouping products with an accuracy
rate of 0.570 and are included in the medium clustering
