<![CDATA[Background: Identifying high-conversion customers is a key challenge in retail CRM. Traditional segmentation methods often fail to accurately predict customer behavior, leading to suboptimal targeting strategies. This study proposed a predictive segmentation framework based on Engagement, Intent, and Value (EIV) to enhance conversion rates in the Indonesian retail sector. Objective: This study aimed to evaluate the effectiveness of the EIV-based segmentation framework through offline benchmarking and online causal experimentation, particularly in terms of conversion uplift and CRM performance. Methods: Five algorithmic models—Logistic Regression, Random Forest, Gradient Boosting Machines (GBM), Artificial Neural Networks (ANN), and Deep Neural Networks (DNN)—were trained using omnichannel behavioral data. Model performance was evaluated using Precision, Recall, F1-score, and Matthews Correlation Coefficient (MCC). The framework was further validated through a large-scale field experiment involving 59 email campaigns across 15 e-commerce platforms. Results: The GBM model demonstrated the most robust predictive performance and was selected for deployment. Experimental results showed that the EIV-based predictive segmentation consistently produced statistically significant conversion uplift compared to traditional operational segmentation. Conclusion: This study highlighted the effectiveness of a unified, propensity-based segmentation approach in improving CRM strategies. By integrating offline predictive modeling with online causal validation, the framework demonstrated how predictive segmentation can enhance CRM outcomes in real-world omnichannel environments.]]>
Copyrights © 2026
