What’s “Up Next”? Investigating Algorithmic Recommendations on YouTube Across Issues and Over Time

Abstract

YouTube’s “up next” feature algorithmically selects, suggests, and displays videos to watch after the one that is currently playing. This feature has been criticized for limiting users’ exposure to a range of diverse media content and information sources; meanwhile, YouTube has reported that they have implemented various technical and policy changes to address these concerns. However, there is little publicly available data to support either the existing concerns or YouTube’s claims of having addressed them. Drawing on the idea of “platform observability,” this article combines computational and qualitative methods to investigate the types of content that the algorithms underpinning YouTube’s “up next” feature amplify over time, using three keyword search terms associated with sociocultural issues where concerns have been raised about YouTube’s role: “coronavirus,” “feminism,” and “beauty.” Over six weeks, we collected the videos (and their metadata, including channel IDs) that were highly ranked in the search results for each keyword, as well as the highly ranked recommendations associated with the videos. We repeated this exercise for three steps in the recommendation chain and then examined patterns in the recommended videos (and the channels that uploaded the videos) for each query and their variation over time. We found evidence of YouTube’s stated efforts to boost “authoritative” media outlets, but at the same time, misleading and controversial content continues to be recommended. We also found that while algorithmic recommendations offer diversity in videos over time, there are clear “winners” at the channel level that are given a visibility boost in YouTube’s “up next” feature. However, these impacts are attenuated differently depending on the nature of the issue.

A method for time-varying analysis of YouTube search results and related videos: The case of the war in Ukraine

“…There are differences between clusters participating in the PageRank and search results top five before and after the EU Council added the violation of restrictive measures to the list of 'EU crimes'. Both YouTube search and related videos algorithms favour mainstream channels (RQ3) with a high number of subscribers (higher than a million subscribers each), confirming that despite being considered a user-generated platform, YouTube favours mainstream media when it comes to newsy events such as the war in Ukraine (Matamoros-Fernández et al, 2021;Glaesener, 2022).…”

“…Research on time-varying analysis of YouTube algorithms for ranking of videos, though, is more focused on the search results – with interesting findings (Rieder et al, 2018) – or recommendations algorithm (Matamoros-Fernández et al, 2021), considered isolated. Once a set of keywords is used for collecting data from the API, it is possible to obtain a list of videos shown as a result and can use them as seeds (starting point) for the related video data collection.…”

YouTube vloggers set the stage: How public (non)compliance with COVID‐19 regulations affects adolescents

“…For one, we should study the cumulative effects of being exposed to multiple types of influencer content over time. On YouTube, adolescents easily end up in algorithm loops; if they watch certain types of content, they will likely be directed to similar content in the future (Matamoros‐Fernandez et al, 2021). Such algorithm biases may lead to a false consensus effect; when multiple sources share similar messages, people tend to believe there is consensus on what the majority of people believe or do, even when that consensus is based upon misinformation or inappropriate experts (Höttecke & Allchin, 2020; Yousif et al, 2019).…”

Freedom and Constraint in Digital Environments: Implications for the Self

“…Specifically, algorithms widen inequality by recommending popular accounts and content at higher rates than less popular accounts and content, creating a “rich get richer” effect (Fabbri et al, 2020; Fleder & Hosanagar, 2009). For instance, YouTube’s “up next” algorithm tends to recommend channels that already have more than 100,000 followers and videos that already have more than 1 million views (Matamoros-Fernandez et al, 2021). Another example of the ways in which algorithms differentially benefit those already at the top comes from Twitter.…”