Product returns represent a growing operational and financial challenge for retailers, particularly in the e-commerce and omnichannel sectors. This research investigates the key factors that drive customer product returns and proposes data-driven strategies for minimizing return rates while maintaining customer satisfaction. The study uses a mixed-method approach, incorporating exploratory data analysis, statistical inference, and supervised machine learning to model and predict return behavior. Using publicly available and simulated retail datasets, we analyze product-level, customer-level, and transaction-level variables to understand return patterns across product categories and consumer segments. Predictive models, including logistic regression and random forest are applied to identify the most influential factors contributing to product returns. The study also segments return behavior by product type and customer demographic to offer targeted managerial insights. Preliminary findings suggest that product category, price, prior return behavior, and inadequate sizing information are strong predictors of return likelihood. The machine learning models achieve high predictive accuracy, with a reasonable AUC score in multiple scenarios. Based on these insights, the research outlines actionable strategies for retail managers, such as implementing AI-powered sizing tools, refining product images and descriptions, and deploying real-time return-risk scoring at the point of sale. This work contributes to the literature on reverse logistics, customer behavior analytics, and decision support in retail operations. It also provides a roadmap for industrial engineers and operations managers seeking to leverage predictive analytics for return reduction and profit optimization