Big Data in logistics marketing for customer churn prediction and dynamic pricing

Анотація

The relevance of the study is determined by the transformation of the logistics market under the conditions of digitalization, which requires a shift from traditional approaches to customer base management and pricing toward analytically driven models. The purpose of the article is to identify the possibilities of implementing Big Data technologies in logistics marketing to improve the accuracy of customer churn prediction, optimize demand management systems, and develop an effective mechanism of dynamic pricing aimed at maximizing the profitability of enterprises. The research methodology is based on a systemic approach and employs methods of comparative analysis, predictive analytics, and machine learning (XGBoost, LightGBM, LSTM). The study generalizes international experience of major logistics operators (DHL, UPS, FedEx, Maersk, DPD) and analyzes models of dynamic pricing based on streaming data, telematics, and customer behavior patterns. Results. The findings demonstrate that the integration of Big Data into logistics marketing enables more accurate prediction of customer behavior, personalized pricing, and flexible demand management. It was found that the use of machine learning algorithms reduces customer churn by an average of 10–15%, increases demand forecast accuracy by 20%, and optimizes logistics profitability. The effectiveness of combining dynamic pricing with behavioral analytics was proven, forming the foundation for adaptive business models in delivery and warehousing logistics. Conclusions.Therefore, the full realization of Big Dataʼs potential in enhancing customer orientation and profitability of logistics companies is possible only under conditions of information flow standardization, CRM and ERP system integration, and the development of analytical competencies and human capital. Future studies should focus on developing intelligent dynamic pricing systems based on reinforcement learning, improving explainable analytics methods (SHAP, LIME), and creating a unified logistics data management architecture to build fully adaptive models of marketing management.