X-Means Clustering Algorithm in Property Customer Payment Pattern

Understanding customer behavior is essential for ensuring the sustainability and competitiveness of property businesses. This study aims to segment customers of PT X based on installment payment patterns using the X-Means clustering algorithm, which automatically determines the optimal number of clusters. From 9,615 transaction records, 386 customer profiles were analyzed using four features: number of transactions, number of late payments, payment differences, and payment status. The analysis produced five customer clusters with a silhouette score of 0.571, reflecting good cluster separation and internal consistency. The results reveal distinct payment behaviors, such as customers who consistently pay on time, those frequently late, and those who have fully completed their payments. These clusters provide practical insights that can support targeted communication, billing, and retention strategies. Furthermore, the study highlights the effectiveness of adaptive clustering techniques in improving segmentation accuracy. The findings contribute to data-driven decision-making in customer management, offering valuable guidance for enhancing operational efficiency and supporting long-term business performance.