Publicação

Using Machine Learning in E-commerce to Predict Customer Churn and Identify Supply Chain Inefficiencies

Ver documento

Detalhes bibliográficos
Resumo:In the highly competitive landscape of retail e-commerce, customer churn poses a significant challenge, undermining profitability and growth. Understanding the drivers of customer churn within supply chain operations can illuminate pathways to enhance operational efficiencies. This study aims to identify critical factors contributing to customer churn in retail e-commerce, leveraging advanced predictive analytics to uncover underlying supply chain inefficiencies. The study followed the CRISP-DM methodology. Four classification models were tested: two single classification methods (Neural Networks and K-Nearest Neighbors) and two ensemble methods (Random Forest and XGBoost). The results revealed that the XGBoost model outperformed the others, demonstrating superior performance across all evaluation metrics. Utilizing the SHAP method, as predicted, it was determined that supply chain variables had the most significant impact on the model's predictions. This finding underscores the critical role of supply chain factors in influencing customer churn rates. This research contributes to the academic literature by integrating predictive analytics of customer churn with supply chain operations. For business practitioners, the implications are insightful. Companies can leverage this insight to optimize their operations and enhance customer retention strategies by identifying specific supply chain inefficiencies that drive customer churn.
Autores principais:Ferreira, Katia Florinda Narcy
Assunto:Customer Churn Machine Learning Supply Chain Ensemble Methods E-commerce Churn Prediction SDG 4 - Quality education SDG 8 - Decent work and economic growth SDG 9 - Industry, innovation and infrastructure
Ano:2024
País:Portugal
Tipo de documento:dissertação de mestrado
Tipo de acesso:acesso embargado
Instituição associada:Universidade Nova de Lisboa
Idioma:inglês
Origem:Repositório Institucional da UNL
Descrição
Resumo:In the highly competitive landscape of retail e-commerce, customer churn poses a significant challenge, undermining profitability and growth. Understanding the drivers of customer churn within supply chain operations can illuminate pathways to enhance operational efficiencies. This study aims to identify critical factors contributing to customer churn in retail e-commerce, leveraging advanced predictive analytics to uncover underlying supply chain inefficiencies. The study followed the CRISP-DM methodology. Four classification models were tested: two single classification methods (Neural Networks and K-Nearest Neighbors) and two ensemble methods (Random Forest and XGBoost). The results revealed that the XGBoost model outperformed the others, demonstrating superior performance across all evaluation metrics. Utilizing the SHAP method, as predicted, it was determined that supply chain variables had the most significant impact on the model's predictions. This finding underscores the critical role of supply chain factors in influencing customer churn rates. This research contributes to the academic literature by integrating predictive analytics of customer churn with supply chain operations. For business practitioners, the implications are insightful. Companies can leverage this insight to optimize their operations and enhance customer retention strategies by identifying specific supply chain inefficiencies that drive customer churn.