The challenge of analyzing influencer communities in the algorithmic age: a methodological framework for audiovisual social media platforms with algorithmic recommendation feeds

Abstract

The evolution of social media platforms toward algorithmic recommendation systems calls for a rethinking of how communities built around influencers are identified and analyzed. Public metrics, such as follower counts and engagement rates, do not provide a clear picture of participation structures: following an account affects content exposure but does not determine it, and interaction does not require users to already follow the account. This study therefore proposes a methodological model based on publicly visible actions to identify and compare communities of observable interaction through public comments. The model combines social network analysis (SNA) with comment analysis. It maps relationships between influencers and participants, distinguishes communication directed at creators from exchanges among users, identifies recurring participation, and estimates the overlap of commenters across accounts. The protocol is applied on an exploratory basis to TikTok posts by ten Spanish food influencers. The findings reveal that participation is predominantly occasional and centralized, alongside a smaller share of conversations among individual users. The communities identified are organized primarily around each influencer’s ecosystem rather than as dense, topic-based subgroups. The model enables communication professionals to compare participation structures and assess whether different accounts or strategies generate one-off responses or sustained relationships. It is operationalized as a replicable protocol that can be adapted to other audiovisual social media platforms or platforms with algorithmically curated feeds, while incorporating ethical and regulatory criteria for data collection and processing.