This article will look to YouTube’s algorithm to assess how such forms of mechanical decision-making can create a discriminatory visibility hierarchy of vloggers, favouring middle class social actors who make highly gendered content aligned with advertisers’ demands and needs. I have selected entrepreneurial beauty vloggers as a case study for this article; beauty vlogging is defined as the demonstration and discussion of cosmetic use, often from a vIoggers’ own bedroom (Banet Weiser, 2017; Nathanson, 2014). This is a deeply entrenched genre on the site; beauty vlogging is a full-time job for some successful participants, and a source of pocket money for many more. Moreover, beauty vlogging is an effective illustration of how the YouTube algorithm causes the polarization of identity markers such as gender. Indeed, for female participants, I hypothesize that YouTube actively promotes hegemonic, feminized cultural outputs, created by beauty vloggers with significant embodied social and cultural capital. That is to say, for women on YouTube, the algorithm privileges and rewards feminized content deeply entwined with consumption, beauty, fashion, baking, friendships and boyfriends in the vein of the historical bedroom culture of the teenage magazine. A secondary hypothesis is that beauty vloggers’ own understandings of YouTube’s algorithmic processes are learned and embodied within their own practices, influencing modes of self-presentation, tone of voice, choice of content covered, words and sentence structures used. I argue that it is essential to situate all beauty vloggers’ experience and content as specific to the platform of YouTube; it is their continued success on the YouTube platform that underwrites the value of their brands. In other words, even highly successful vloggers remain beholden to YouTube’s technologies of visibility, they are not safe from the sovereignty of the algorithm.

Beauty vloggers’ feminised outputs often position them outside of traditional spheres of technical expertise, however, their strategic management of algorithmic visibility makes them an illuminating source of algorithmic knowledge. I draw from an ethnography of beauty vloggers and industry stakeholders to study the collaborative and directive processes used to formulate and sustain algorithmic expertise – algorithmic gossip. Algorithmic gossip is defined as communally and socially informed theories and strategies pertaining to recommender algorithms, shared and implemented to engender financial consistency and visibility on algorithmically structured social media platforms. Gossip is productive: community communication and talk informs and supports practices such as uploading frequently and producing feminised beauty content to perform more effectively on YouTube. Taking gossip seriously can present a valuable resource for revealing information about how algorithms work and have worked, in addition to revealing how perceptions of algorithms inform content production.