PAN, MEIZHI (2025) Influencer Marketing Effectiveness. Doctoral thesis, Durham University.
Full text not available from this repository. Author-imposed embargo until 14 January 2026. |
Abstract
Influencer marketing plays a critical role in shaping consumer behavior, yet strategies for improving its effectiveness across both traditional (human) and emerging (AI) forms remain underexplored. This thesis addresses these gaps by conducting a meta-analysis and a two-part experimental study.
This thesis begins by examining the existing literature on human influencers through a meta-analysis of 1,531 effect sizes, guided by the persuasion knowledge model (PKM). This analysis identifies the antecedents, mediators, and moderators that shape influencer marketing effectiveness, emphasizing the importance of trait-based mechanisms such as source credibility. However, it also reveals a lack of research on the influence of influencer type (virtual vs. real). This highlights the need to test traditional mechanisms in AI influencer contexts and explore alternative explanations for non-human influencers.
In response, the second experimental study tests trait-based mechanisms in AI influencer contexts using two separate experiments, based on the theory of simulacra and simulation. Study 1 finds that AI clone influencers (AI-generated replicas modeled after real individuals) outperform pure AI (entirely computer-generated characters with no link to real people) on traditional traits but do not produce significantly higher engagement. To address this disconnect, Study 2 introduces experiential mechanisms, staged authenticity and immersion, and finds that AI clones are more effective for symbolic products, while pure AI performs better for functional products, with effects mediated by the experiential mechanisms.
This thesis develops a comprehensive framework for influencer marketing effectiveness across human and AI contexts. It advances the influencer marketing literature by identifying the key mechanisms that shape consumer responses to different influencer types and by clarifying when trait-based or experiential mechanisms are more effective. It offers practical implications for improving influencer marketing strategies in an evolving digital landscape.
Item Type: | Thesis (Doctoral) |
---|---|
Award: | Doctor of Philosophy |
Faculty and Department: | Faculty of Business > Management and Marketing, Department of |
Thesis Date: | 2025 |
Copyright: | Copyright of this thesis is held by the author |
Deposited On: | 15 Jul 2025 13:24 |