<![CDATA[One of a company's most important assets is its customers. Companies use various methods to retain customers by implementing various strategies, such as PT Bintang Multi Sarana which uses various methods to establish good relationships with customers, such as holding promos and discounts. Companies offer promotional items or discounts by contacting customers by telephone one by one or offering them directly when the customer has finished purchasing the previous item. Offering discounts and promos to all customers is less efficient because it costs a lot of operational costs and time. Therefore, customers at PT Bintang Multi Sarana need to be grouped first to make it easier for the company to determine appropriate services. Apart from that, sales transaction data at PT Bintang Multi Sarana is only processed into monthly reports. In fact, if processed properly, sales transaction data can be used to help determine strategies for retaining customers. The aim of this research is to segment customers to determine customer characteristics. One solution to this problem is to process sales transaction data by utilizing the role of data mining with clustering techniques using the k-means algorithm. The k-means algorithm is a distance-based algorithm that divides data into separate clusters. Based on calculations using k-means, customers are clustered into 3 groups, namely cluster 1 with 742 customers, cluster 2 with 795 customers and cluster 3 with 223 customers. From the results of the customer segmentation analysis, the results show that cluster 1 is a silver segment customer, cluster 2 is a gold segment customer, and cluster 3 is a platinum segment customer. The marketing strategy for the silver segment is by providing price discounts, promotions and advertising on social media, for the gold segment services are provided in the form of reward points or special discounts, while for the platinum segment, this is the provision of attractive prizes or exclusive access to certain products.]]>
