User experience of YouTube
How YouTube video recommendation works
-
YouTube’s homepage recommends videos mostly based on viewers’ watch history, so if the person watches one video, even just for fun or by accident, they might suddenly see lots of similar videos on their homepage.
-
The “Up Next” recommendations (the videos shown next to the one the viewer is watching) are based mainly on the current video, so they react quickly to what the viewer just watched.
Problems
Sometimes, people watch a video for a random reason (maybe a friend sent it, or they were just curious), but then YouTube starts recommending lots more of that type of video, even if the user isn’t really interested in seeing more.
In this video ( https://www.youtube.com/watch?v=_geqgdIfsaM ), Cristos Goodrow, the VP of Engineering at Google, mentions that the YouTube recommendations happen right away when a new video is uploaded. He mentioned that when the team interviewed users and asked if they liked a video that much, why didn’t they subscribe to that channel, and the viewer replied that they don’t want to see more like that, they just wanted to encourage the creator for that amazing video. So, he mentions that subscribing can be a starting point to know that the particular video is suitable for a viewer. Still, it doesn’t say that the viewer is interested in all of the videos from that particular channel, but the views, likes are also important when considering a video.
Now we all know how the homepage recommendations are mostly dependent on the watch history of the viewer
Here's a comment I found on Reddit, and People were explaining how YouTube recommendations change right away when they watch.
One user stated that YouTube has
"No Control: It feels like user have lost all control over the viewing experience. Everything is now dictated by an algorithm, and the once customizable interface has been reduced to a series of topic buttons that seem be randomly generated."
While not interested, Reduces similar recommendations, but the feature does not reduce similar videos, it only targets the specific video, and takes feedback on why the user felt that way
A simple solution for this can be
Instead of letting a single recently watched video have a big impact, looking at the user’s overall watch history and giving more weight to videos or topics they watched often or for a long time can help find which content is engaging the user the most.
Not letting one video change everything
If someone watches a video only briefly and doesn’t interact with it, like not giving any likes or comments, treat it as a “casual” watch, and don’t let it affect their recommendations much.
Mixing the content
Including a mix of content on the home page
-
Some videos based on long-term interests (things the user watches regularly).
-
Some new or trending videos to help users discover new topics.
-
Only a few videos based on the most recent watches, especially if there’s no strong engagement.
Taking User Feedback
Improving the “not interested” feature by asking the users if they want to stop seeing similar content, or if they just don’t want to see the particular video, will improve the personalized recommendations.
Why didn’t YouTube provide a feature or make their not interested, don’t recommend channel, dislike features capable of removing similar videos that users were not interested in?
It can be for many reasons why YouTube has not provided a feature to remove all similar videos from recommendations
-
Generating a feature that allows users to block all similar videos (not just one video or channel) would require an extremely detailed and accurate understanding of what counts as "similar."
-
YouTube’s business model relies on keeping users engaged and find new content. Overly filtering might reduce user engagement, which can impact the overall experience.
-
While some say these behaviors are intentional because the platform doesn’t try to block all content related to a topic. ( https://www.theverge.com/2022/9/20/23356434/youtube-dislike-not-interested-buttons-bad-recommendations-mozilla-report )
But seeing how users are experiencing this, it's really time for them to deeply work on this issue
Figma File
Recourses
https://www.tubebuddy.com/blog/how-youtube-recommends-videos/
https://www.youtube.com/watch?v=_geqgdIfsaM
https://www.reddit.com/r/youtube/comments/1dcf1pk/youtube_recommendations_have_never_been_good_and/
https://support.google.com/youtube/answer/6125535?hl=en&co=GENIE.Platform%3DDesktop#:~:text=Click Not interested to remove the video from your feed.
https://www.nngroup.com/articles/instructional-video-guidelines/