Application of clustering methods in machine learning to segment bank customer retention groups

Abstract

This research examines customer retention strategies in banking, focusing on identifying patterns among loyal customers using machine learning clustering techniques. The study segments customers based on features like age, credit score, income, and product usage, providing insights to help banks develop targeted retention strategies to reduce churn. Data was collected from the bank's customer database and preprocessed to handle missing values and normalize features. Clustering algorithms, including K-Means grouped customers with similar characteristics. Key factors influencing retention, such as credit scores, age, location, and banking relationships, were analyzed in-depth. Results show that Cluster C2, comprising customers aged 34–38 with credit scores around 650 and multiple products, represents high-value, loyal customers. In contrast, younger customers with lower credit scores and fewer products have higher churn rates. The study recommends personalized retention strategies, such as offering premium services for Cluster C2 customers and educational resources or incentives for younger segments. Enhancing digital platforms and customer service availability is also crucial for improving loyalty. This research bridges traditional banking theories and modern machine learning techniques, providing actionable insights for practitioners and opening pathways for further studies using enriched data sources. By leveraging data-driven approaches, banks can optimize customer retention strategies, reduce churn, and boost profitability