The Impact of AI-Driven Personalized Recommendation on College Students’ Purchase Decisions on Small E-Commerce Platforms

Abstract

This study investigates how AI-driven personalized recommendations influence college students’ purchase decisions on small e-commerce platforms. As even smaller retailers increasingly adopt affordable AI tools and SaaS-based recommender solutions, it becomes crucial to understand whether and how personalization works for young consumers beyond big platforms. We adopt a 2×2 between-subjects experiment (explanation provided vs. not provided; short vs. long recommendation list) combined with a structured survey. Using real data (N = 391), we find that adding an explanatory rationale to recommendations does not significantly change subjective evaluations or purchase intention, whereas shorter lists (3–4 items) outperform longer lists (10–12 items) on purchase intention. Mediation analyses reveal that perceived usefulness and enjoyment are significant predictors of purchase intention; however, the paths related to “explanation usefulness/enjoyment” are not significant, indicating that the overall mediation does not hold. Additionally, moderation tests demonstrate that trust and privacy concerns do not significantly influence the effect of explanation on purchase intention. The results suggest that, for small platforms targeting college students, streamlined presentation matters more than adding generic justifications; strengthening actual usefulness and enjoyment is key to improving responses. We discuss practical implications for balancing personalization with user control and transparency.