<![CDATA[PT. XYZ is a telecommunications operator company in Indonesia that provides prepaid GSM (global system for cellular communication) services. PT. XYZ support to serve the whole city until remote areas of Indonesia are no exception Malang Regency by continuing to increase the number and increase the number of customers when providing the services provided. Companies that provide different services based on the characteristics or each customer in the form that provides services and build loyal and long-lasting relationships. The data used in this study is the transaction data of PT. XYZ customers in Malang Regency operational area with a period of time January 1 2019 - March 31 2019 that made 35,868 transactions. Customer segmentation is a strategy to divide customers into different, different, and homogeneous subgroups according to their characteristics. The main purpose of customer segmentation is to determine the customer base and gain customer insights that will enable the design and development of different marketing strategies. Customer characteristics can be seen through customer transaction data modeling using RFM (recency, frequency, and monetary) models, namely the time span of the last transaction (recency), number of transactions (frequency), and money spent (monetary). Fuzzy C-Means Clustering can be an option in solving customer segmentation problems. The Elbow method is used to support the clustering process by determining the number of clusters in the application of Fuzzy C-Means Clustering. Modification of Partition Coefficients (MPC) and Euclidean Distance (EU) is a validation method used to obtain the accuracy of cluster results using data to the nearest central point. In this study clustering with 2 clusters is the best result. The results of customer segmentation are visualized into a dashboard page consisting of a combination of integrated tables, diagrams and graphs and information needed by the company. The results of dashboard visualization are sponsored by users to obtain the level of user acceptance of using the System Usage Scale (SUS) analysis. Dashboard test results show an average value of 80 which is included in the acceptable category.]]>
Copyrights © 2020
