Persaingan industri barbershop yang semakin kompetitif menuntut Juragan Cukur Ciawigebang untuk memiliki strategi retensi pelanggan yang tepat. Selama ini, identifikasi pelanggan yang akan berhenti berlangganan (churn) masih dilakukan secara konvensional dan subjektif, sehingga rentan terhadap bias. Penelitian ini bertujuan untuk membangun sistem pendukung keputusan pemodelan risiko churn menggunakan integrasi model LRFM (Length, Recency, Frequency, Monetary) dan algoritma Fuzzy C-Means (FCM). Data yang digunakan mencakup 15.500 riwayat transaksi yang disintesis melalui tahapan preprocessing menjadi 3.780 data pelanggan unik. Hasil penelitian menunjukkan bahwa model FCM memiliki performa yang sangat valid dengan nilai Fuzzy Partition Coefficient (FPC) sebesar 0,9445 dan Silhouette Score sebesar 0,6046. Berdasarkan hasil klasterisasi, populasi pelanggan terbagi menjadi tiga segmen risiko: Low Risk sebanyak 698 pelanggan (18,4%), Medium Risk sebanyak 1.098 pelanggan (29,1%), dan High Risk sebanyak 1.984 pelanggan (52,5%). Penelitian ini diimplementasikan ke dalam sebuah Web Dashboard yang dilengkapi fitur Email Blast untuk mempermudah admin dalam mengeksekusi strategi promosi yang personal dan terukur. Hasil penelitian ini memberikan rekomendasi strategis bagi Juragan Cukur untuk melakukan pengambilan keputusan berbasis data guna menekan angka churn pelanggan secara efisien.

Kata Kunci : Customer Churn, LRFM, Fuzzy C-Means, Data Mining, Barbershop

The increasingly competitive barbershop industry requires Juragan Cukur Ciawigebang to have an appropriate customer retention strategy. Traditionally, identifying customers likely to leave (churn) has been conventional and subjective, making prone to bias. This study aims to develop a churn risk modeling decision support system by integrating the LRFM (Length, Recency, Frequency, Monetary) model and the Fuzzy C-Means (FCM) algorithm. The data uses comprises 15,500 transaction records, synthesized through preprocessing into 3,780 unique customer records. The research results show that the FCM model performs exceptionally well with a Fuzzy Partition Coefficient (FPC) of 0.9445 and a Silhouette Score of 0.6046. Based on the clustering results, the customer population is divided into three risk segments: Low Risk with 698 customers (18.4%), Medium Risk with 1,098 customers (29.1%), and High Risk with 1,984 customers (52.5%). This research is implemented into a Web Dashboard equipped with an Email Blast feature to assist administrators in executing personalized and measurable promotional strategies. The results of this study provide strategic recommendations for Juragan Cukur to engage in data-driven decision-making to efficiently reduce customer churn rates.

Keywords : Customer Churn, LRFM, Fuzzy C-Means, Data Mining, Barbershop.