<![CDATA[The rapid growth of e-commerce has generated massive customer transaction data that can support business decision-making. However, many companies have yet to fully leverage this data for understanding customer behavior and developing targeted strategies. This study aims to perform customer segmentation on the Online Retail II dataset using the Recency, Frequency, Monetary (RFM) approach combined with the K-Means Clustering algorithm in order to generate customer profiles that can serve as the basis for business strategy development. The dataset used in this study consists of 1,067,371 customer transaction records with eight main attributes. The preprocessing stage included the removal of missing values in Customer ID, duplicate records, invalid transactions, and return transactions. Subsequently, RFM values were calculated for 5,878 unique customers. To improve data quality, outlier removal was performed using the Interquartile Range (IQR) method, resulting in 5,182 customers being retained for the clustering process. Prior to modeling, the data were normalized using the Min-Max Scaling technique. The optimal number of clusters was determined using the Elbow Method and Silhouette Score, which indicated that k = 3 was the best configuration with a Silhouette Score of 0.4780. The clustering results identified three customer segments with distinct characteristics. Cluster 0 had an average Recency of 91.53 days, Frequency of 2.88 transactions, and Monetary value of 776.59, representing potential customers. Cluster 1 exhibited the highest Recency value of 479.40 days along with the lowest Frequency and Monetary values, indicating at-risk customers. Cluster 2 showed the lowest Recency value of 69.31 days and the highest Frequency and Monetary values of 8.80 transactions and 2,943.38 respectively, representing loyal and high-value customers (Champions). Based on the characteristics of each segment, different marketing strategies can be implemented, including loyalty programs, upselling strategies, and customer reactivation campaigns. The findings demonstrate that the combination of the RFM method and K-Means Clustering is effective in generating customer segmentation and supporting data-driven customer intelligence to improve marketing strategy effectiveness in the retail industry.]]>
