Pemetaan Pelanggan dengan LRFM dan Two Stage Clustering untuk Memenuhi Strategi Pengelolaan

Abstract

Maibus is a company of transportation services located in Bali. Transaction data that is owned has not been managed properly. This results in data accumulation and only as a turnover calculation, so LRFM and clustering methods are needed to assist the calculation and processing data in fulfilling customer management strategies. The research was conducted by collecting and understanding data, preprocessing, applying LRFM (Length,Recency,Frequency,Monetary), normalizing LRFM, evaluating the number of clusters with Davies Bouldin Index (DBI), clustering with K-Means, and analyzing cluster results. The data used is transaction data from January 2017 to December 2018 with a total of 14.292 data. The clustering method with the K-means algorithm helps in mapping customers based on transaction data. DBI was used to determine the optimal number of clusters and LRFM used to test the determination of variables in determining customer behavior and loyalty. The results of testing 7.193 invoice using 5 clusters with DBI value is 0.135. The result of customers in cluster 0,2,4 are new customer groups with the proposed strategy is enforced strategy, while the customer in cluster 1 and 3 are lost customers with the proposed strategy is let-go strategy that refers to the customer value and customer loyalty matrix.