CUSTOMER CHURN PREDICTION ON OTT PLATFORMS

Churn prediction is crucial for organi-zational growth across various sectors worldwide. Customer churn can significantly harm a company’s revenue and profits. The key to preventing churn lies in identifying its causes and implementing appropriate measures. Today, machine learning is essential in addressing this issue. Churn can result from several factors, such as switching to a competitor, canceling subscriptions due to poor customer service, or losing touch with a brand because of inadequate engagement. Maintaining long-term connections with customers is often more effective than acquiring new ones. Addressing underlying issues can lead to increased customer satisfaction. This project aims to identify the primary factors influencing customer churn and analyze them using machine learning algorithms. By understanding churn, businesses can gauge how many existing customers are likely to leave, which can have a significant positive effect on revenue. Churn rates help track lost customers, while growth rates monitor new customer acquisition; comparing these metrics provides a clear picture of business growth over time. The analysis involves examining data such as customer data, viewing habits, and subscription details to identify patterns and reasons for churn. The methodology includes comprehensive data cleaning, feature selection, and training various models like logistic regression, decision trees, random forests, and neural networks.[2][4] Each model’s effectiveness is evaluated using metrics such as accuracy, precision, recall, and F1-score to ensure reliability[2].