Application of The RFM Model and K-Means Clustering for Customer Segmentation in E-Wallet Top-Up Services
DOI:
https://doi.org/10.23969/infomatek.v28i1.42246Keywords:
customer segmentation, recency, monetary, frequency, K-Means ClusteringAbstract
The implementation of digital payment technology through e-wallet top-up services requires financial institutions to understand user characteristics and behavior comprehensively The objective of this study is to segment customers based on their e-wallet top-up behavior by analyzing 143,836 bill payment transaction records using the RFM (Recency, Frequency, Monetary) model combined with the K-Means clustering algorithm. The dataset contains more than one hundred thousand transaction entries, with RFM parameters representing the time since the last transaction, the frequency of top-ups, and the monetary value spent by users. The RFM scoring process is applied to quantify user activity levels before entering the clustering stage. The K-Means clustering model successfully grouped customers into three distinct segments. The first segment represents low-activity users, the second consists of moderately active customers with stable transaction behavior, while the third segment captures highly engaged users with the highest transaction frequency and value. Evaluation metrics, including a silhouette score of 0.64, a Calinski-Harabasz index of 21690.50, and a Davies-Bouldin score of 0.70, demonstrate strong clustering performance and reliable separation between groups. The findings provide valuable insights for designing service strategies, improving mobile banking system performance, and developing targeted marketing approaches tailored to each customer segment. This research highlights the potential of RFM based clustering as a decision-support tool for enhancing digital payment service optimization and customer engagement.
Downloads
References
Al Janabi, M. A. (2022). Optimization algorithms and investment portfolio analytics with machine learning techniques under time-varying liquidity constraints. Journal of Modelling in Management, 17(3), 864–895
Anitha, P., & Patil, M. M. (2022). RFM Model for Customer Purchase Behavior Using K-Means Algorithm. Journal Of King Saud University - Computer And Information Sciences, 34(5), 1785–1792.
Das, S., & Nayak, J. (2022). Customer Segmentation via Data Mining Techniques: State-of-the-Art Review. Smart Innovation, Systems and Technologies, 281(July), 489–507.
Egorova, E., Glukhov, G., & Shikov, E. (2022). Customer Transactional Behaviour Analysis Through Embedding Interpretation. Procedia Computer Science, 212(C), 284–294.
Firdaus, U., & Utama, D. N. (2021). Development of Bank’s Customer Segmentation Model Based on Rfm+B Approach. ICIC Express Letters, Part B: Applications, 12(1), 17–26.
Haddadi, S. J., Mohammadi, M. O., Bahrami, M., Khoeini, E., Beygi, M., & Khoshkar, M. H. (2022). Customer churn prediction in the Iranian banking sector. The 2022 International conference on applied artificial intelligence (pp. 1–6). IEEE.
Lestari, D., Fauzan, A. C., & Harliana, H. (2022). Penerapan Algoritma Pillar Untuk Optimasi Penentuan Titik Awal Centroid Pada Algoritma K-Means Clustering. JOISIE (Journal Of Information Systems And Informatics Engineering), 6(1), 15-24.
Machado, M. R., & Karray, S. (2022). Assessing credit risk of commercial customers using hybrid machine learning algorithms. Expert Systems with Applications, 200, Article 116889
Puspitasari, N., Widians, J. A., & Setiawan, N. B. (2020). Customer Segmentation Using Bisecting K-Means Algorithm Based on Recency, Frequency, and Monetary (RFM) Model. Jurnal Teknologi dan Sistem Komputer, 8(2), 78–83.
Sundari, A., Lydia, M. S., & Muchtar, M. A. (2024, September). Customer Segmentation Based on Recency, Frequency, Monetary, Variety and Duration (RFMVD). In 2024 2nd International Conference on Technology Innovation and Its Applications (ICTIIA) (pp. 1-5). IEEE.
Syahputra, S., Ramadani, S., & Pardede, A. M. H. (2020). Menentukan Strategi Promosi Menggunakan Algoritma Clustering K-Means. JOISIE (Journal Of Information Systems And Informatics Engineering), 4(1), 7-14.
