Hybrid Intelligence for Customer Targeting: The HM2CA Framework for Optimizing E-Commerce Marketing and Acquisition

This study introduces a novel framework, the Hybrid Modeller for Maximizing Customer Acquisition (HM2CA), which uses machine learning techniques to enhance marketing efficiency in e-commerce by identifying potential customers with high purchase intent. Using the “Online Purchase Intention” dataset, consisting of 12,330 sessions from e-commerce transactions, the model combines classification and clustering methodologies to optimize marketing strategies. Initially, the model applies the Synthetic Minority Oversampling Technique (SMOTE) with Gradient Boosting, achieving a prediction accuracy of 93.37%, surpassing the baseline by 3%. This is followed by the application of K-Median clustering to segment customers into three distinct groups: High Intention and Revenue, High Intention and Non-Revenue, and Low Intention and Non-Revenue. These clusters enable more targeted retargeting campaigns, reducing cart abandonment rates and maximizing marketing return on investment (ROI). The hybrid approach offers a strategic advancement in pinpointing high-propensity customers, thereby enhancing the overall effectiveness of e-commerce marketing efforts.