To widen their chances of receiving attention on the platform, creators must work with the YouTube algorithm. In fact, Pew Research Center found that over 80% of users watch recommended videos, which are determined by the algorithm.

This must be kept in mind when creating videos for YouTube. However, there is a certain mystery surrounding what is recommended to you on YouTube. Finding a way to work efficiently with the algorithm, then, is a major key to success as a YouTube creator. To better understand YouTube’s algorithm, we must look at where the platform’s recommendations began to where they are today.

A brief timeline of YouTube recommendations

The process by which people watch YouTube videos has changed since its dawn in 2005. Initially, YouTube recommended videos that were being frequently clicked. The issue with this method is that it presented a ton of clickbait. Users would click, feel misled and then stop watching. YouTube eventually noticed and adjusted their system of choosing recommended videos.

In 2012, YouTube updated their discovery system to recommend longer watch times. The update was put in place to identify videos people actually wanted to watch. It suggested videos that held viewers attention throughout. This method inevitably kept users on the platform longer. At the time, YouTube encouraged creators to not try and solve the new update in order for their video to be recommended. Creators would try this by posting shorter videos to get a high retention rate, or by posting longer videos to have more watch time.

In 2016, YouTube released a model of a machine-learning diagram. The diagram helps creators visualize how videos are delivered to viewers.


Deep Neural Networks for YoutTube Recommendations. Recommendation system architecture demonstrating the “funnel” where candidate videos are retrieved and ranked before presenting only a few to the user. 

Although the model looks simple, it still uses language that is unfamiliar to the average person. We do know that millions of videos go through this system to give viewers a highly-personalized selection. Many of the selections are based proportionately on your search and watch history.

The algorithm today

In 2021, creators on YouTube are highly aware the algorithm exists. So much so that it seems to be an entity of its own. An entity, it should be noted, that needs to be satisfied. YouTube defines the algorithm to be “real-time feedback loop that tailors videos to each viewer’s different interests.” The algorithm tracks your activity YouTube, learns your habits and recommends accordingly.

In its modern form, the YouTube algorithm is a complicated entity that creators must do their best to adhere to in order to boost views and engagement with their content.

The goal of the sophisticated algorithm is to keep users on the platform longer with videos they want to watch. Focusing on user behavior is just as important as a singular video’s success. YouTube explains the best way for the algorithm to like creators videos is to make the audience like said videos. The platform says simply that “the algorithm follows the audience”. There are undoubtedly more factors to how it works. But to put it simply, creators should strive to make videos that people want to watch.

The search box on YouTube is a major aspect to how the algorithm functions. The search results yield the most popular videos for the combination of words a user searches. Search results factor in titles, description and how the actual video matches the wording of the search. Aside from the initial factors for results, YouTube considers watch time and engagement between results.

For creators to optimize their video in search they must be thorough with multiple aspects of video. The wording in your titles and descriptions are critical. Be relevant and clear. Take your time on descriptions and be thorough with information pertaining to the video. YouTube recommends writing 1-2 paragraphs. Creators should also consider creating videos around trends from their own personal perspective.

YouTube recommendations

Recommended videos on YouTube combine multiple factors to show users what they may like. The factors include what viewers were watching previously, related topics and watch history. The suggested videos are accessible on the right side of the desktop or under the video users watch on mobile.

Most creators strive to get their videos suggested to as many users as possible. YouTube gives some tips on how creators may do this. A main tip to consider is making a strong call-to-action to the viewer to watch more of your videos. Suggestions should be like short and sweet sales pitches for your videos. Another way to optimize your chances for being recommended is to provide resources to your channel. Creators provide resources by using links, playlists and organized end screens to suggest a video to watch next.

YouTube homepage

The homepage of YouTube is another space of where the algorithm is hard at work. Users are given an array of video options from videos watched by similar viewers, new videos or from their subscriptions. Home is YouTube’s strongest guess to what viewers want to watch. YouTube bases the videos on the homepage on performance and personalization.

Creators should note that to appear on the home page, there are tips to help improve your chances. These tips include keeping your audience engaged, uploading frequently and keep creating videos that receive positive feedback.

Working with the algorithm

Creators should consider YouTube search, recommendations and the homepage when producing content for the platform. Taking time to make sure video postings are optimized will be integral to creators’ success with the algorithm.

Image courtesy: Deep Neural Networks for YouTube Recommendations