Pengembangan Sistem Customer Follow-up Management Menggunakan Metode Analisis Recency Frequency Monetary

Abstract

Penentuan pelanggan prioritas untuk tindak lanjut pada PT XYZ masih subjektif dan tidak konsisten antarcabang karena data transaksi pada Accurate Online belum terolah otomatis. Penelitian ini bertujuan mengembangkan sistem customer follow-up management dengan kalkulasi dan segmentasi Recency, Frequency, Monetary (RFM) otomatis terjadwal. Sistem dibangun menggunakan framework Laravel dan metode Extreme Programming (XP) dalam dua iterasi, dengan skor RFM ditetapkan melalui pembagian kuintil per cabang dan kategori serta dijalankan otomatis melalui cron job harian. Pengujian mencakup uji unit dan fitur otomatis, black box testing dengan teknik equivalence partitioning, serta User Acceptance Testing (UAT) terhadap 29 admin cabang, dengan seluruh 65 kasus uji black box berstatus pass dan rata-rata UAT 94,21% (sangat baik). Segmentasi yang dihasilkan memperlihatkan pemusatan nilai yang tajam, yaitu segmen tertinggi menyumbang porsi pendapatan jauh melampaui proporsi jumlah pelanggannya, sehingga urutan tindak lanjut dapat ditetapkan dari satu aturan seragam yang dapat dihitung ulang.

Determining priority customers for follow-up at PT XYZ remained subjective and inconsistent across branches because Accurate Online transaction data was not processed automatically. This study aims to develop a customer follow-up management system with automated, scheduled Recency, Frequency, Monetary (RFM) calculation and segmentation. The system was built using the Laravel framework and the Extreme Programming method in two iterations, with RFM scores assigned through quintile partitioning per branch and category and executed automatically via a daily cron job. Testing covered automated unit and feature tests, black box testing with the equivalence partitioning technique, and User Acceptance Testing (UAT) involving 29 branch administrators, in which all 65 black box test cases passed and UAT obtained an average of 94,21% (very good). The resulting segmentation shows a sharp concentration of value, in which the highest segment contributes a revenue share far exceeding its proportion of customers, so that follow-up priority can be determined by a single uniform and recomputable rule.