Related Experiment Videos
Integrating Business Intelligence and CRM Systems With a Machine Learning Approach for Predictive Customer Retention
Mohammad Zeinali1, Leila Ramezani Asli2, Mohammad Amin Khalili3,4
1Faculty of Engineering, University of Isfahan, Isfahan, Iran, ui.ac.ir.
Thescientificworldjournal
|May 28, 2026
Summary
This study introduces a data-driven framework for predictive customer retention in e-commerce, integrating business intelligence (BI) and machine learning (ML) with customer relationship management (CRM). The framework effectively segments customers and predicts churn, enabling targeted retention strategies.
Area of Science:
- E-commerce Analytics
- Machine Learning Applications
- Customer Relationship Management
Background:
- Customer retention is more cost-effective than acquisition in e-commerce.
- Existing data-driven approaches often analyze BI, ML, or CRM in isolation.
Purpose of the Study:
- To propose and validate an integrated, end-to-end analytical framework for predictive customer retention.
- To enhance e-commerce customer retention strategies through data integration and advanced analytics.
Main Methods:
- Utilized the Brazilian E-Commerce Public Dataset (Olist) for data integration and feature engineering.
- Applied K-means clustering on RFM variables for customer segmentation (loyal, at-risk, occasional).
- Trained Random Forest and XGBoost models for churn prediction using behavioral, satisfaction, and service features.
Main Results:
- XGBoost outperformed Random Forest in churn prediction, achieving accuracy of 0.81 and AUC of 0.85.
- Customer segmentation identified distinct behavioral groups for targeted interventions.
- The framework successfully integrated data preparation, predictive modeling, and CRM decision support via Power BI dashboards.
Conclusions:
- The proposed framework offers a reproducible and practical approach to e-commerce retention analytics.
- Integrating BI, ML, and CRM provides a holistic solution for improving predictive customer retention.
- The study lays the groundwork for live deployment and A/B testing of CRM strategies.