Persuasive Differences between Human and Virtual Influencers in Health Supplement Advertising: Evidence from Eye-Tracking

Mengqi DuKyung Han You (Ryu)^*

• Jeonbuk National University, Jeonju-si, Republic of Korea

Introduction: Attention is scarce in fast-paced, visually saturated feeds—especially for trust-sensitive categories such as health supplements, where credibility relies on source cues. Eye-tracking shows how viewers engage with advertising elements, yet authenticity and perceived human-likeness vary between human and virtual endorsers. We posit that source type moderates whether attention translates into evaluation. To examine these effects, we investigate whether source type influences effectiveness, whether attention drives outcomes, and whether these relationships depend on source type. Method: In a laboratory experiment with an East Asian university student sample (N = 120), an Instagram-style vitamin C advertisement was displayed using a 2 × 2 design (human vs. virtual influencer; gender). Gaze was tracked with Tobii Pro Nano for three areas: the endorser's face, the product, and the text. After viewing, participants rated their attitude toward the advertisement and purchase intention. Hierarchical regressions with interactions were used to test main and moderating effects. Results: There are persuasive differences between human and virtual influencers. Virtual influencers tended to attract more visual attention, but this did not consistently correspond to more favorable outcomes. By contrast, human influencers were generally associated with more positive advertising attitudes and stronger purchase intentions. Attention to text and face appeared to be predictors of advertising effectiveness. Importantly, source type appeared to moderate the link between attention and outcomes: for human endorsers, greater face attention was associated with more favorable evaluations, whereas for virtual endorsers, similar levels of attention were not accompanied by comparable gains. Discussion: The findings shift focus from "who performs better" toward understanding "when and why" attention persuades. Consistent with source credibility accounts in credence contexts, its persuasive yield hinges on source human-likeness processing fluency—how fluently the source is resolved as human, visually and mentally. Eye-tracking complements self-reports by revealing processing before conscious awareness. The patterns yield preliminary hypotheses that human endorsers may better support credibility-intensive claims and suggest a testable framework for how attention could translate into persuasion, offering a starting point for future targeted evaluations.