Advances in information technology produces wide range of choices in accessing information including reading books. The increase in the number of readers who turning to electronic books making sales of printed books has decreased in the recent years. PT Gramedia Asri Media is one of book retail company in Indonesia. Gramedia implement CRM by launching a member card named Kompas Gramedia Value Card (KGVC). Promotion given has not been able to increase book transaction of KGVC members.This research focus on make customer segmentation in CRM at PT Gramedia Asri Media. Data mining process is done by clustering using K-means algorithm for segmenting customers based on RFM, as well as hierarchical clustering algorithms for segmentation of customers based on the number of books type. Evaluation is done on cluster result using elbow method, silhouette method, and Calinski-Harabasz index. Customer segmentation based on the RFM produce two optimal clusters, occasional customers and dormant customers. Customer segmentation based on the number of types of books purchased produce 3 optimal cluster, namely low, medium, and high. With these results, it is expected to help the company classifying KGVC members to determine the appropriate strategies, so company can increase the number of books transactions.