Implementation of K-Medoids Clustering Method for B2B Service Price Segmentation at PT Telkom Tasikmalaya

Elsa Amalinda (Department of Information Systems Universitas Siliwangi, Tasikmalaya, Indonesia)
Andi Nur Rachman (Department of Information Systems Universitas Siliwangi, Tasikmalaya, Indonesia)
Cecep Muhammad Sidik Ramdani (Department of Information Systems Universitas Siliwangi, Tasikmalaya, Indonesia)

Journal of Computer Networks, Architecture and High Performance Computing
Vol. 8 No. 3 (2026): Research Paper July 2026

PT Telkom Indonesia Witel Tasikmalaya faces strategic challenges in monitoring customer behavior in real-time and determining optimal Business-to-Business (B2B) service pricing strategies due to conventional data management. This study aims to design a Web-based Decision Support System to analyze B2B service pricing using the K-Medoids clustering algorithm with Manhattan Distance metrics. The research and system development are fully structured using the Agile methodology. The model evaluation is measured using the Davies-Bouldin Index. The data mining modeling indicates that the formation of three clusters is the most optimal partition with the smallest Davies-Bouldin Index value of 0.638. The clusters successfully map customer profiles from small enterprises to large corporations as a baseline price recommendation. The system is built using an implementation of vanilla JavaScript and Google Firebase serverless architecture, achieving success in functional black-box testing and a 90% score in user acceptance testing. This research provides a strategic contribution by accelerating the issuance of quotation documents and optimizing company revenue.