Consumers’ Responses to Virtual and Human Social Media Influencers and Strategic and Ethical Implications for Advertising

Abstract

Virtual influencers (VIs) are digital non-human social media influencers that have become increasingly popular over the past few years. While VIs have been frequently employed by advertisers to promote brands and social issues, little is known about their effectiveness as endorsers and how they compete with human influencers (HIs). This dissertation examines (1) how consumers perceive and respond to VIs as compared to HIs, (2) how their perceptions and responses influence advertising outcomes for VIs and HIs across branding advertising and social issue advertising, and (3) how consumers’ knowledge about the source influences their perceptions, responses, and advertising outcomes. Drawing on the Persuasion Knowledge Model, the concepts of parasocial interaction and relationship, and prior research on the effectiveness of influencers as advertising endorsers, this study examined VIs and HIs using an interdisciplinary multi-method approach, including a novel computational research approach and an online experiment. The key finding of both studies was that VIs were overall less effective social media influencers than HIs. Specifically, the computational study revealed that social media followers had a less positive attitude toward the ad for VIs than HIs. Followers also tended to be more skeptical toward VIs than HIs for branding advertising, which in turn led to a less positive attitude toward the ad. The results also showed that social media followers engaged in fewer parasocial interactions with VIs than HIs, but when they did, it mitigated the negative effects of skepticism for branding advertising. The computational study also demonstrated that the differences between VIs and HIs were not greater for social issue advertising than branding advertising. The online experiment also revealed a less positive attitude toward the influencer and ad for the VI than the HI for branding advertising, and a less positive attitude toward the VI than the HI for social issue advertising. Regarding consumers’ persuasion knowledge, the study demonstrated that the VI did not increase consumers’ ad recognition or skepticism in comparison to the HI. There were also no serial mediation effects, however, the results showed that skepticism had a negative influence on consumers’ attitude toward the influencer, ad, brand, and behavioral intention for branding advertising. For social issue advertising, ad recognition and skepticism negatively influenced consumers’ attitude toward the influencer and brand, while skepticism also negatively influenced consumers’ attitude toward the ad. Moreover, the findings provided more insights into the role of agent knowledge. Consumers were more likely to believe that someone else than the influencer created the VI posts than the HI posts. In a similar vein, VIs were perceived to have less agency than HIs, which had a negative impact on consumers’ attitude toward the influencer, ad, brand, and behavioral intention for branding advertising, and a negative impact on attitude toward the influencer and brand for social issue advertising. Lastly, the online experiment showed that the predicted differences between VIs and HIs were not greater for social issue advertising than branding advertising. This study advances research on VIs and is the first to apply the PKM to investigate the effectiveness of VIs in comparison to HIs. It also provides new insights into the role of agent knowledge and parasocial interaction with regard to VIs. Moreover, it provides a novel computational approach that allows to identify and examine consumers’ skepticism toward social media influencers. Lastly, this study also has important practical implications to advertisers for developing effective influencer campaigns with VIs.