Interpretable prediction of private brand purchases by pet type in e-commerce for consumer behavior analysis using real-world transaction data

Abstract

Background: The global pet care market is rapidly expanding, and private brand (PB) products are becoming increasingly important for e-commerce retailers. However, how PB purchasing behavior differs between dog and cat owners remains underexplored. Methods: This study analyzed PB purchasing behavior in pet e-commerce using real-world transaction data and machine-learning techniques. We developed separate predictive models for dog and cat owner segments, with extreme gradient boosting (XGBoost) demonstrating superior performance (F1-scores: 0.7806 and 0.7876, respectively). SHapley Additive exPlanations (SHAP)-based interpretability analysis identified key drivers of PB purchasing behavior for each segment. Results: Pet supplies and snacks emerged as universal predictors across both segments; however, their relative importance and underlying mechanisms differed significantly. Dog owners showed stronger associations with delivery convenience features, whereas cat owners demonstrated greater sensitivity to price and product quality factors. Implications: These findings suggest segment-specific marketing strategies: convenience-focused approaches for dog owners and value-oriented trust-building strategies for cat owners. This work contributes to the limited literature on PB behavior in pet e-commerce and demonstrates the practical applicability of explainable artificial intelligence (XAI) for customer segmentation in digital retail.