Теоретичні та прикладні питання економіки. Збірник наукових праць.
Випуск 2 (51)
UDC 339.1:004.9
JEL M31, D83
ORCID ID: 0000-0003-4889-1247
ORCID ID: 0009-0009-5725-6975
DOI https://doi.org/10.17721/tppe.2025.51.7
Galyna Chornous,
doctor of economics, professor,
Taras Shevchenko National University of Kyiv, Kyiv
Anastasiia Khyzhniak,
student, Taras Shevchenko National University of Kyiv, Kyiv
CUSTOMER SEGMENTATION MODELING VIA ADVANCED RFM ANALYSIS FOR DECISION-MAKING IN MARKETING
In today’s e-commerce landscape, one of the key challenges is the need for effective customer segmentation to enhance the performance of marketing strategies and ensure stable customer relationships. Traditional RFM analysis - based on recency, frequency, and monetary value of purchases - is widely used in practice, but it has certain limitations, particularly the lack of consideration for customers’ demographic and behavioural characteristics, which reduces segmentation accuracy.
The aim of the study is to develop and evaluate the effectiveness of the improved RFMP-DOV+AIC customer segmentation model, which combines classical RFM parameters with additional features: purchase diversity, online behaviour characteristics, and long-term customer value assessment.
The methodological basis of the study includes methods of economic and mathematical modelling, statistical analysis, clustering, and machine learning algorithms implemented in the Python environment. For empirical validation, an open dataset from the Kaggle platform was used, containing information on online store transactions.
During the research process, data preprocessing, normalization of indicators, and feature engineering for the model were carried out. Based on the K-means algorithm, customer segments were constructed and evaluated using the following metrics: SSE, Silhouette index, Calinski-Harabasz index, and Davies-Bouldin index. The consistent results of these metrics allow for a well-grounded selection of the number of clusters, ensuring internal homogeneity and sufficient distance between clusters.
The results confirmed that the extended model enables clearer customer grouping and allows for the identification of segments that consider behavioural and demographic characteristics, which are not captured by classical RFM analysis.
The practical significance of the obtained results lies in the possibility of integrating the RFMP-DOV+AIC model into CRM systems of e-commerce enterprises to optimize marketing communications, reduce customer churn, and increase customer loyalty. The proposed approach can be applied both in academic research in the field of marketing analytics and in the practical activities of companies operating in the e-commerce market.
Keywords: RFM analysis, e-commerce, clustering, marketing analytics, personalization.
Full Text: PDF
DOI: https://doi.org/10.17721/tppe.2025.51.7