PREDICTING CUSTOMER CHURN IN INDONESIAN ISPs: AN MLP AND MARKETING INTELLIGENCE APPROACH - A CASE STUDY ON PT XYZ

Abstract

To forecast customer churn in the Indonesian Internet service provider (ISP) industry, this research proposes a unique combination of Multilayer Perceptron (MLP) neural networks with marketing intelligence and strategy. This research examines various aspects of customer retention issues using a comprehensive dataset from a top Indonesian ISP that includes user demographics, subscription information, geographic information, and customer interaction. The research focuses on using MLP, a sophisticated artificial neural network, to predict churn potential, providing a deeper understanding of churn dynamics by examining key factors influencing customer preferences. With the expected results for above 90% accuracy rate in predicting customer churn, this research will provide a valuable marketing insights for the company. This will enable PT XYZ to tailor customer engagement and loyalty campaigns by combining machine learning (MLP) with targeted marketing approaches. This combination can lead to reduced churn rates, improved customer satisfaction, and a long-term competitive advantage. The research highlights the importance of advanced data analytics in enhancing telecom marketing techniques. It emphasizes the need for machine learning to understand consumer behaviour and develop proactive marketing strategies to maintain customer loyalty in the competitive Indonesian ISP industry. The study aims to provide ISPs with valuable insights into customer behaviour, enabling them to develop effective retention strategies and maintain a competitive edge in a constantly evolving industry.