Subscription box companies routinely lose 40-70 percent of their subscribers within the first year, yet the causal effects of personalized marketing interventions on repeat purchase behavior across heterogeneous product categories remain poorly understood. This study addresses that gap by analyzing a 24-month panel of 47,318 subscribers from 12 subscription box firms in the food, beauty, and fashion sectors, employing a quasi-experimental design that combines difference-in-differences estimation with propensity score matching. Integrated personalization strategies increased repeat purchase rates by 23.1 percent (Cohen's d = 0.47), reduced quarterly churn by 34.0 percent, and raised average customer lifetime value by an estimated $87 per subscriber. Email personalization, algorithmic product recommendations, and customized promotional offers exhibited complementary rather than substitutive effects, with product recommendations generating the largest individual contribution. Significant category-level heterogeneity emerged: food boxes captured the greatest repeat purchase gains (28.7%), followed by beauty (22.4%) and fashion (17.9%), patterns attributable to differences in preference stability and feedback immediacy. Personalization benefits compounded with subscriber tenure and were attenuated by privacy concerns, though transparency measures substantially mitigated the latter effect. These findings offer subscription commerce managers an evidence-based framework for allocating personalization investments across channels and product categories to maximize retention and profitability.