<![CDATA[Customer churn, or the loss of banking clients, represents a major challenge in the banking industry due to its potential to cause significant financial losses. This study aims to segment customers based on characteristics that influence their churn risk using the K-Means algorithm. The data used in this research is secondary data consisting of 9,763 customer records from a bank customer churn dataset obtained via the Kaggle platform. The data processing follows the CRISP-DM framework. Clustering was conducted using RapidMiner, and performance was evaluated using the Davies-Bouldin Index to determine the optimal number of clusters (K). The results indicate that the optimal number of clusters is K = 4. Centroid analysis revealed that balance and estimated salary are the primary variables contributing to cluster formation. Cluster 1 and Cluster 3 had the highest number of churned customers. Cluster 1 consisted of customers with high balances but low salaries, while Cluster 3 included customers with both high balances and high salaries. These findings suggest that a high balance does not necessarily guarantee customer loyalty, and that income level plays an important role in preventing churn. Based on the analysis, recommended strategies include providing financial education and loyalty programs for customers in Cluster 1, and offering exclusive services and personalized approaches for those in Cluster 3. This study demonstrates that the K-Means algorithm is effective in producing relevant customer segmentation, serving as a valuable foundation for developing more targeted and efficient marketing strategies.]]>
Copyrights © 2025
