PREDICTIVE ANALYSIS OF CUSTOMER CHURN IN THE TELECOM INDUSTRY USING BUSINESS ANALYTICS TECHNIQUES

The telecom industry is one of the fastest-growing and most competitive sectors in the global economy. Due to intense competition and changing customer expectations, customer churn has become a major challenge for telecom companies. Customer churn refers to the situation in which customers discontinue telecom services and switch to competing service providers. High churn rates negatively affect organizational profitability, customer loyalty, and long-term business sustainability. The main objective of this research is to analyze customer churn behavior in the telecom industry using business analytics and machine learning techniques. The study focuses on identifying important customer-related factors influencing churn and developing predictive models capable of classifying customers into churn and non-churn categories. The research uses secondary telecom customer datasets containing customer demographics, billing information, internet service usage, contract details, customer complaints, payment methods, and customer support interactions. Data preprocessing techniques such as cleaning, normalization, transformation, and feature selection were applied to improve dataset quality. Exploratory Data Analysis (EDA) and statistical analysis were conducted to understand customer behavior patterns and identify relationships among variables. Machine learning algorithms such as Logistic Regression and Decision Tree models were implemented for customer churn prediction. The findings of the study revealed that factors such as monthly charges, customer tenure, customer complaints, service quality, and contract type significantly influence customer churn behavior. Customers with higher monthly charges, frequent complaints, shorter tenure, and monthly contracts were more likely to discontinue telecom services. The Decision Tree model achieved higher prediction accuracy compared to Logistic Regression and successfully classified customers into churn and non-churn categories. The research demonstrates that predictive analytics and machine learning techniques are effective tools for customer retention management and business decision-making in the telecom industry. The study concludes that telecom organizations can significantly improve customer retention, operational efficiency, and profitability by adopting data-driven customer relationship management strategies and predictive analytics systems.