Abstract:
"Companies are aiming to improve their market share and profitability in the current
competitive business environment. A crucial aspect of this endeavor involves dividing their
customer base into ideal segments to enable targeted marketing campaigns. In the retail
industry, traditional data mining methods are commonly used to understand the distinct
characteristics of customer profiles. Among these methods, the K-Means clustering algorithm holds a significant position due to its usefulness in data mining and machine learning, especially for customer segmentation. However, K-Means clustering has a notable limitation -it struggles to accurately identify customer segments based on their unique transactional behaviors. Combining the RFM (Recency, Frequency, Monetary) model with K-Means clustering appears to be a possible solution to this constraint. This integration offers the possibility for a more thorough and accurate understanding of consumer behavior, which can result in more exact customer segmentation and marketing tactics. The RFM scoring can be applied through various methods, including Percentile Ranking, Equal Interval, Standard Deviation, and Weighted RFM. While numerous segmentation approaches have been developed, utilizing various RFM scoring methods alongside the K-Means algorithm, none have definitively determined which RFM scoring approach optimally increases the K-Means algorithm’s efficiency in achieving superior customer segmentation."
