- What is YouTube Algorithm
- Uses of YouTube Algorithm
- History of YouTube Algorithms
- How does the YouTube Algorithm work?
- Negatives of the Algorithm
- Algorithm in Trending Page
- Algorithm & Ad Companies
What is YouTube Algorithm
Have you ever noticed that when you watch multiple videos of same genre or same channel you will find more of those in the recommended section or in the homepage of your YouTube. If this ever happened to you, then you need to know about why its happening. Its because the machine learning technology of YouTube that try to get you to keep on watching the videos. That is YouTube Algorithm.
Uses of YouTube Algorithm
As i stated before YouTube algorithm is machine learning technology that tries to guide the user behavior in YouTube. Overtime the algorithm will understand your interests and tastes in videos and it will keep recommend them. It also affects the persons who uploads the videos in YouTube. The AI will check what is the video is about, how long is it, time of posting, relevance of the video, the keywords used in the metadata, appropriation of the content etc. It affects the overall marketing of the “creators” and for YouTube. Because of giving the people what they like, they can increase watch time.
History of YouTube Algorithms
The current algorithm was introduced by YouTube in 2016. Before they used View Counts and View Duration.
At first YouTube recommended the videos to the users through only one metric, view counts. YouTube highly recommended those videos which have most views, which were, in fact, helped in increasing the watchtime of YouTube but also helped in creating the infamous clickbaits. These clickbaits mislead the user in clicking in the video. Due to this the viewers left the videos early.
After the “failure” of the first metric YouTube started to recommend the videos which have longer watchtime. In theory, this was a good tactic. It would help the users to identify what they were looking for instead of ending up in some cliackbaits. But, instead of that users got long and confusing videos which did provides the titled information with a large amount of unnecessary things.
To be fair, the creators were only trying to improve their channel, but this time it was in favor of YouTube. YouTube doesn’t even cared about the quality if the video and they got watchtime which helped the ad companies.
How Does The Algorithm Works?
In 2016 YouTube released the algorithm which is described as deep neural networks and machine learning in its recommendation system. The AI introduced a ranking system for the videos uploaded to YouTube. A higher ranking video is most likely to get recommended at higher rates. In the paper published in 2016 the Google engineers break down the signals that were used in ranking the videos for YouTube recommendations:
- The likelihood of someone clicking on your video (Views)
- The combined amount of time users spend watching the video (Watchtime)
- Number if videos a user watched from a particular channel
- User Topic preferences
- Search preferences
- Watch history
- User location
YouTube doesn’t completely removed the View Counts and View Duration. They added some additional metrics to it and called it a different name. So the click-baits and the pointless long videos are sill exists.
Negatives of the Algorithm
The AI is a sophisticated and highly integrated system but it still have flaws. The tech reviewers referred the AI as a misinformation engine due to its tendency to recommend conspiracy theories, fake news, and disturbing contents to the users. After all its a machine, it can be manipulated into thinking that a particular news or an incident should be highly recommended and it is relevant to the users.
The AI want you stay in the site for a long period, so video descriptions containing another site addresses are less likely to get recommended.
Algorithm in Trending Page
The YouTube trending page is a complete mess because of the confused algorithm. If you check the trending tab the probability of finding a video from your taste or your subscribed channels are very low. Half of the time it will be some music videos or weird video from a channel we never even heard of. For the trending page, the algorithm look for the consistency of the channel for uploading videos and number of people watching or clicking the video in a given time. That’s why smaller channels and music videos are very likely to end up in the Trending page.
Algorithm & Ad Companies
Ad companies are the main source of income for YouTube and the video creators, so YouTube will always in favor of the ad companies. The algorithm is a main part in deciding the eligibility of monetizing a persons video. The factors deciding the eligibility are:
- Copyright Claims: Copyright claims are the most sensitive issues in YouTube. In a nutshell, Copyright law prevent from anyone from using another persons content without their prior permission
- Breaking Community Guidelines: Community Guidelines are the censor boards for the YouTube, it allows YouTube to censor the sensitive contents like sexual, violence etc from users that are not in age.
YouTube algorithm will check the above mentioned parameters closely and tell YouTube about the monetizing eligibility of the video. This helps the ad companies to invest their ads in safe and family friendly content.
In conclusion, YouTube algorithm is not only for the smooth running of YouTube but also for the user satisfaction and its a way to increase the revenue of the YouTube.