Abstract

Customer Lifetime Value (CLV) prediction is crucial for effective marketing resource allocation and strategic decision-making. Traditional CLV models often rely on simplifying assumptions and aggregated historical data, limiting their predictive accuracy. This research proposes a hybrid approach that integrates machine learning algorithms with traditional marketing metrics to enhance CLV prediction. We develop a model incorporating both transactional data and customer behavioral features extracted from CRM systems. The methodology involves feature engineering, model selection (comparing algorithms such as linear regression, decision trees, random forests, and gradient boosting), and rigorous model evaluation using metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared. Our results demonstrate that the hybrid model, particularly the gradient boosting algorithm, significantly outperforms traditional CLV models in predicting future customer value. The findings offer actionable insights for marketers to optimize customer acquisition, retention, and engagement strategies. The paper concludes with a discussion of limitations and potential avenues for future research, including the incorporation of real-time data and advanced deep learning techniques.