Customer segmentation is a tool to communicate effectively with customers. Companies cannot satisfy all potential buyers of their products or services. They should choose a group of customers that they can serve better than their competitors. To do this, customers must be segmented. Analyzing customer perceptions to distinguish between diverse customer groups and devising tailored strategies for each group poses a significant challenge for customer-centric organizations. Utilizing machine learning techniques for customer segmentation serves as a method to assess and categorize customers. In addition, customer clustering is crucial for any enterprise, especially agricultural goods retailers. In this article, we leverage transactional records from an online retail company, extracting pertinent features to construct RFM model indicators including Recency, Frequency, and Monetary from sales invoices. Following data preparation and processing, the optimal number of customer clusters is determined through both hierarchical methods and the k-means algorithm and also the silhouette score for model evaluation. Performance evaluation reveals a slightly better preference for the k-means algorithm. Finally, a correlation analysis between RFM indicators identified the optimal customer group, referred to as the “golden customers.”