Follow the algorithm: An exploratory investigation of music on YouTube

Abstract

This article presents an exploratory study of the network of associations among 22,141 YouTube music videos retrieved by ‘following’ the platform’s recommender algorithm, which automatically suggests a list of ‘related videos’ to the user in response to the video currently being viewed. As YouTube’s recommendations are predominantly based on users’ aggregated practices of sequential viewing, this study aims to inductively reconstruct the resulting associations between the musical content in order to investigate their underlying meanings. Network analysis detects 50 clusters of tightly connected videos characterised by a strong internal homogeneity across different axes of similarity. We discuss these findings with reference to the literature on music genres and classification, arguing that the emerging clusters can be considered as ‘crowd-generated music categories’. That is, sets of musical content that derive from the repeated, crowd-based actions of sequential viewing by users on YouTube in combination with the platform’s algorithm. Interestingly, 7 out of 50 clusters are characterised by what may be seen as a ‘situational’ culture of music reception by digital audiences. Such culture is not so much founded on music genres as traditionally conceived, but rather on the purposes of reception which are rooted in the context where this takes place.