Pemanfaatan Teknik Clustering untuk Segmentasi Pelanggan pada Industri Perbankan Digital

Kevin Pratama (Program Studi Informatika, Fakultas Ilmu Komputer, Universitas Indonesia, Depok, Indonesia)
Nabila Aulia (Program Studi Teknologi Pangan, Fakultas Teknologi Pertanian, Universitas Andalas, Padang, Indonesia)

Primary Journal of Multidisciplinary Research
Vol. 2 No. 3 (2026): PRIMARY: Journal of Multidisciplinary Research, June 2026

The rapid development of the digital banking industry in recent years has led to a significant increase in the volume and complexity of customer data. This condition requires more effective analytical approaches to gain a deeper understanding of customer behavior. One of the approaches widely used in data science is clustering, a technique that aims to group data based on similarities in specific characteristics without prior labeling. This study aims to develop a clustering model for customer segmentation in the digital banking industry in order to improve the effectiveness of marketing strategies, service quality, and data-driven business decision-making. The data used in this study consist of various customer-related variables, including demographic information, transaction history, frequency of digital service usage, transaction value, and interaction patterns with digital banking platforms. The research process began with data collection, followed by data cleaning, normalization, and feature selection to ensure optimal data quality. Subsequently, clustering methods such as K-Means were employed to classify customers into several segments based on the similarity of their characteristics. Model evaluation was conducted using the silhouette score and elbow method to determine the optimal number of clusters. The results indicate that customers can be grouped into several major segments, including highly active customers, passive customers, and potential customers who have opportunities to increase their use of digital banking services. Each segment exhibits distinct behavioral characteristics, enabling banks to design more targeted marketing strategies and provide more effective personalized services. Furthermore, the segmentation results offer valuable insights for the development of digital banking products that better meet the needs of each customer group.Therefore, the implementation of clustering models in customer segmentation has proven to enhance the understanding of customer behavior more comprehensively. This study is expected to serve as a reference for the development of data science-based analytical systems in the digital banking sector and to support digital transformation in the financial industry toward becoming more adaptive, efficient, and customer-oriented.