A New Method of B2B Customer Segmentation Based on Firmographic Data, and RFM and Graph Models

Customer segmentation is a crucial aspect of mar-keting strategies. It enables companies to tailor products and services to specific customer groups, thereby enhancing enterprise competitiveness. The RFM (Recency, Frequency, and Monetary) model has been widely used to identify valuable customers in various marketing areas. However, the simplicity of this model limits its effectiveness in the B2B context, as it does not consider firmographic characteristics. Additionally, clustering approaches are commonly used to address customer segmentation. Most of these approaches require transforming mixed variables (i.e., numerical and categorical data) into a unified format for use as input. However, such transformations can result in a loss of information and reduced segmentation accuracy. In this paper, we propose an RFMSG model that incorporates firmographic data (i.e., activity Sector and Geolocation) in addition to the standard RFM model. The selection of these firmographic features is driven by their significant influence on customer behavior and needs in a B2B context. We design a comprehensive data collection process to help companies enrich their datasets with these B2B customer features. Furthermore, we propose a graph-based method for handling mixed data to achieve effective clustering. Through a real-world case study, we demonstrate the importance of the enriched dataset and our proposed method in enhancing the B2B customer segmentation process.