Customer Churn Analysis Using Machine Learning to Improve Customer Retention on Vissie Net

Customer Retention Service Science Fishbone Diagram Service Improvement

Authors

Vol. 12 No. 09 (2024)
Economics and Management
September 10, 2024

Downloads

The industrial transformation that positions the internet as essential for all social aspects has intensified competition among internet service providers in Indonesia, including efforts to retain customers. This study focuses on predicting customer churn at Vissie Net using the CRISP-DM method combined with supervised learning algorithms, particularly Random Forest, and utilizing a fishbone diagram to identify areas for improvement. The research analyzes a dataset of 1,119 customers, considering variables like subscription length, contract terms, bills, packages, demographics, and churn data. Results indicate that the Random Forest algorithm excels in predicting churn, with 21.8% of customers unsubscribing. The fishbone diagram highlights factors influencing churn and offers suggestions for service variety, AI integration, customer satisfaction, and staff training.