Recommended for You Exploring user perceptions of the YouTube recommender system

YouTube is the second-most used social media platform worldwide and its popularity is partly due to its refined recommender system. This system provides users with specifically tailored video suggestions and it is based on an algorithm that is responsible for over 70% of the daily watchtime on the platform. The best-watched genre on YouTube concerns entertainment videos and the objective of this research is to inquire about user perceptions of the recommender system and how this affects entertainment content consumption. The study at hand provides a literature review on recommender systems, algorithmic imaginary, user consumption behavior and the uses and gratifications theory in light of the YouTube recommendation algorithm. However, literature that tied together user perceptions of the recommendation algorithm to YouTube entertainment content consumption was still lacking. Therefore, the research question of this study is as follows: To what extent do perceptions of YouTube's recommendation algorithm shape user entertainment content consumption? More specifically, it was studied whether user perceptions of the recommendation algorithm affect perceived entertainment content diversity, watchtime, recommendation satisfaction and perceived user agency. Moreover, it was inquired about whether content, social, process and/or technology gratifications influence recommendation satisfaction. A quantitative approach was implemented to measure these concepts using a survey (N = 161). The data was gathered among adult YouTube users who consume entertainment content, whereafter the data was statistically analyzed using SPSS. The results demonstrated that positive perceptions of the recommender system do not influence perceived entertainment content diversity, nor affect watchtime and neither influence the perceived amount of agency over the user's entertainment content consumption. However, more positive perceptions of the recommendation algorithm were found to lead to higher recommendation satisfaction. Also, content and technology gratifications were positively related to recommendation satisfaction, whilst social and process gratifications were not related. There are three key takeaways that can be derived from this study. Firstly, the attitude that the user holds of the algorithm does not influence the amount of time users spend watching YouTube videos. So, regardless of whether a user has a negative perception of the recommendation algorithm, the user does not necessarily watch less YouTube which might be related to the privacy paradox theory. Secondly, an interesting outcome from this study that substantiates the algorithm appreciation theory is that positive perceptions regarding the recommender system lead to more recommendation satisfaction. Thirdly, users who experience convenience while using YouTube were also more satisfied with their entertainment recommendations. The second and third key takeaway justify the reasoning for digital media companies to optimize and refine their platform’s algorithm, user interface and user experience as much as possible since this leads to higher satisfaction and therefore possibly to higher platform usage. The findings of this research contribute to the conclusions of previous studies and they might also be relevant to digital media companies, governmental entities and social actors.