Consumers’ Willingness to Accept Product Recommendations from Virtual Influencers: The Roles of Persona Cues, Relational Closeness, and Influencer–Brand Congruence

Abstract

Virtual influencers (VIs) have become visible actors in digital marketing, yet prior research has often examined anthropomorphism, relational distance, and influencer-brand congruence separately and has frequently treated recommendation acceptance as equivalent to purchase intention. This study examines how three perceived VI persona cues—coolness, cuteness, and warmth—are associated with close influencer-customer distance, close brand-customer distance, influencer-brand congruence, and consumers' willingness to accept product recommendations. A cross-sectional, single-stimulus survey was conducted with 237 undergraduate students recruited through convenience sampling at one university in Xi'an, China. All participants evaluated the same McDonald's advertisement featuring the virtual influencer Xi Jiajia, and the proposed relationships were assessed using PLS-SEM. Perceived coolness was positively associated with both forms of relational closeness and with influencer-brand congruence. Perceived warmth showed the same pattern. Perceived cuteness was positively associated only with close brand-customer distance; its relationships with close influencer-customer distance and influencer-brand congruence were not significant. Influencer-brand congruence was positively associated with close influencer-customer distance and recommendation acceptance but was not significantly associated with close brand-customer distance. Both forms of relational closeness were positively associated with recommendation acceptance. The measurement model showed acceptable reliability, convergent validity, and discriminant validity at the construct level. The findings identify cue-specific and relationship-specific associations rather than a uniform anthropomorphism effect. Because the study used a cross-sectional convenience sample, one university, one VI, and one advertisement, the results should be interpreted as associative and context-bound; they do not establish causality or mediation. These results offer context-specific guidance for virtual influencer marketing practice.