Shall brands create their own virtual influencers? A comprehensive study of 33 virtual influencers on Instagram

Increasing customer-brand engagement on social media is crucial for brand success. Prior research highlights the role of social media in building online communities and its impact on brand loyalty, perceptions, and purchase intentions. The rise of virtual influencers presents a new avenue for brands to enhance engagement, leading to increased partnerships and even the creation of branded virtual influencers. While some research exists on virtual influencers from an influencer marketing perspective and their perceived authenticity, there's a gap in understanding their role in customer-brand engagement, particularly comparing branded versus non-branded influencers. This study addresses this gap by investigating whether brands should create or collaborate with virtual influencers to maximize engagement, proposing that collaboration with non-branded influencers is a more effective strategy.

The literature review examines existing research on customer-brand engagement on social media, emphasizing the importance of cognitive, emotional, and behavioral activities positively evaluated by customers. Studies show a strong relationship between brand image, customer perception, and brand loyalty. However, the literature lacks a comprehensive analysis of customer engagement with virtual influencers, especially concerning the differences between branded and non-branded virtual influencers. Prior studies mainly focus on virtual influencer authenticity and credibility, often using case studies with limited generalizability, neglecting the nuances of customer-brand engagement driven by different virtual influencer types and communication strategies. This study aims to bridge this gap.

This study employs a quantitative and qualitative mixed-methods approach. Data was collected from 33 Instagram-verified virtual influencers (14 branded, 19 non-branded) by analyzing their posts up to December 2022 (23,260 posts total). Descriptive statistics on likes, comments, and followers were gathered using SPSS, and customer-brand engagement was calculated as the average number of likes and comments relative to followers. Qualitative content analysis, aided by DiVoMiner software, was used to categorize virtual influencers based on four variables from Kozinets et al. (2010): character narratives, forums, promotional characteristics, and community reactions. This enabled the creation of a typology of virtual influencers and analysis of their engagement.

The study reveals that non-branded virtual influencers exhibit significantly higher customer-brand engagement rates than branded virtual influencers. Nobody Sausage, a non-branded virtual influencer, had the highest engagement rate (30.74%), while Lu of Magalu, a branded virtual influencer, had the lowest (0.07%). The top six highest-engagement influencers were all non-branded and doll-like. The bottom six were all branded. The analysis did not show a consistent relationship between engagement and the degree of human-likeness or authenticity of virtual influencers, contradicting the Uncanny Valley hypothesis. A typology of virtual influencers was developed, categorizing them into four types: virtual storytellers, social connectors, product demonstrators, and brand assistants. Virtual storytellers and product demonstrators showed lower engagement than social connectors and brand assistants. The study found that directly showcasing marketing intentions in posts negatively impacts engagement.

The findings support the hypothesis that collaboration with non-branded virtual influencers is more effective for brands than creating their own. The high engagement rates of non-branded influencers indicate that audiences value authenticity and genuine connection, which is often compromised when a brand directly controls the influencer's messaging. The typology reveals that engagement strategies focusing on storytelling, fostering social connections, and implicitly integrating brand messages are more successful than direct product demonstrations. This highlights the importance of aligning brand messaging with the influencer's established personality and audience connection.

This study contributes to the literature by providing a comprehensive analysis of customer-brand engagement with virtual influencers on Instagram. It demonstrates the superior engagement of non-branded influencers, challenging previous research suggesting all virtual influencers yield positive brand benefits. The typology of virtual influencers provides valuable insights into effective engagement strategies. Future research could explore engagement across different social media platforms and investigate the role of additional factors (product characteristics, demographics) in influencing engagement.

The study's limitations include focusing solely on Instagram, limiting the generalizability to other platforms. The typology, while grounded in existing literature, may benefit from further refinement by incorporating other relevant factors. Future work could address these limitations to enhance the understanding of virtual influencer marketing.