Close Menu
TrenderHQTrenderHQ
    Sign Up / Log In

    Subscribe to Updates

    Get the latest creative news and more from TrenderHQ.

    What's Hot

    How does TikTok know what videos you like?

    November 19, 2026

    How does YouTube decide which videos to recommend? The algorithm explained

    November 19, 2026

    How does Google Search decide which results to show you?

    November 19, 2026
    Facebook X (Twitter) Instagram
    TrenderHQTrenderHQ
    • Home
    • Buzz
    • Lifestyle
    • Stories
    • World News
    • Sports
    • Health
    • Culture
    • Videos
    • HQ
    Sign In Register
    TrenderHQTrenderHQ
    Home » How does YouTube decide which videos to recommend? The algorithm explained
    Technology

    How does YouTube decide which videos to recommend? The algorithm explained

    Digital ReporterBy Digital ReporterNovember 19, 202616 Views
    WhatsApp Facebook Twitter Copy Link LinkedIn Telegram Email
    How does YouTube decide which videos to recommend

    YouTube is home to millions of videos covering almost every subject imaginable, from music and comedy to education, news, sports, and entertainment.

    Yet, when two people open YouTube, they rarely see exactly the same collection of videos. One person might see music performances, another might see football highlights, while someone else is presented with cooking tutorials and technology reviews.

    This is not a coincidence. YouTube uses a sophisticated recommendation system designed to predict which videos individual viewers are most likely to watch and enjoy. [1, 2]

    The short answer: How YouTube chooses recommended videos

    • Your watch history: YouTube examines the videos you have watched and the topics you regularly explore.
    • Your search history: The things you search for help the platform understand your interests.
    • Your interactions: Likes, dislikes, subscriptions, comments and feedback help indicate what you enjoy or want to avoid.
    • Video performance: YouTube considers whether people click on a video, how long they watch and whether they appear satisfied.
    • Similar viewers: The system identifies patterns in the interests of people with similar viewing habits.
    • Your circumstances: Factors such as location, language, device and time of day can influence recommendations.
    • Content responsibility: YouTube has systems designed to reduce recommendations of certain harmful or misleading content. [1, 2, 3, 4]

    1. YouTube has a recommendation algorithm

    An algorithm is a set of instructions that a computer follows to perform a task or solve a problem.

    YouTube uses complex computer systems, including artificial intelligence and machine learning, to decide which videos should be recommended to different viewers.

    Machine learning is a type of artificial intelligence that allows computer systems to identify patterns in information and use what they learn to make predictions.

    In YouTube’s case, these systems analyse viewing behaviour and other information to estimate which videos a person might find interesting.

    The recommendation system has two main goals:

    1. Help viewers discover videos they want to watch.
    2. Maximise long-term viewer satisfaction, meaning helping people find content they consider valuable and enjoyable over time. [1]

    This second goal is important.

    YouTube does not want its recommendations to depend entirely on which videos attract the most clicks. A video may attract thousands of clicks because its title and thumbnail look exciting, but viewers may quickly discover that it does not deliver what was promised.

    The platform therefore considers more than whether someone clicks on a video.

    It also tries to understand whether the person actually watches and enjoys it. [5]

    2. Your watch history helps YouTube understand your interests

    One of the most important sources of information for YouTube’s recommendation system is your watch history.

    Watch history is a record of videos you have watched while using your account, subject to your history settings.

    For example, imagine someone regularly watches:

    • Football analysis.
    • Music videos.
    • Interviews with musicians.
    • African entertainment news.
    • Documentaries about the music industry.

    Over time, YouTube can identify patterns in these viewing activities.

    It may recognise that the person has an interest in music, entertainment and football.

    Consequently, when that person opens YouTube, the platform may recommend new music releases, artist interviews, football discussions and related documentaries. [2]

    Now imagine another person who mostly watches cooking tutorials, home improvement videos and travel documentaries.

    That person’s recommendations may look completely different.

    The reason is that YouTube attempts to build an understanding of each viewer’s interests rather than presenting everyone with the same collection of popular videos.

    Does watching one video affect your future recommendations?

    It can.

    If someone who normally watches football suddenly watches several videos about photography, YouTube may begin testing photography-related recommendations.

    However, one unusual viewing session does not necessarily mean that the person’s entire recommendation experience will change permanently.

    The system considers patterns of activity and other signals when estimating interests. [1, 2]

    3. Your searches also influence what YouTube recommends

    YouTube does not only learn from what you watch. It also considers what you search for.

    A search provides information about something you are actively trying to find.

    For example, imagine searching YouTube for:

    “How to start a small business.”

    You may subsequently encounter recommendations about entrepreneurship, marketing, business planning, financial management and interviews with business owners.

    This happens because the search gives the recommendation system additional information about your possible interests. [2]

    The same principle applies to other subjects.

    Someone searching for:

    • How to play the guitar may receive music lessons and instrument tutorials.
    • African history may receive historical documentaries and educational discussions.
    • Smartphone reviews may receive technology comparisons and product demonstrations.
    • Healthy cooking may receive recipes and nutrition-related videos.

    Search activity does not mean YouTube will recommend only those subjects.

    It simply provides another source of information that helps the system estimate what you might enjoy watching. [2]

    4. YouTube pays attention to what you like and dislike

    Your interactions with videos also help shape recommendations.

    YouTube considers several forms of feedback, including likes, dislikes and options that allow viewers to indicate that they do not want particular content.

    Likes

    When you like a video, YouTube receives a signal that you appreciated or approved of that content.

    For example, regularly liking educational videos about science may help the system recognise your interest in science-related subjects.

    Dislikes

    Disliking a video provides information about content you may not want to see again.

    However, a dislike does not necessarily mean that YouTube will completely remove every similar video from your recommendations.

    It is one signal among many. [2, 5]

    Not interested

    This option allows viewers to tell YouTube that they do not want a particular recommended video.

    For example, if your homepage repeatedly displays celebrity gossip and you are no longer interested in that subject, selecting “Not interested” provides direct feedback.

    Don’t recommend channel

    This option goes further by telling YouTube that you do not want recommendations from a particular channel.

    These controls help the platform adjust future recommendations according to your preferences. [2]

    5. YouTube learns from people who watch similar videos

    This is another interesting part of the recommendation system.

    YouTube does not rely exclusively on your personal viewing history.

    It also examines patterns among viewers with similar interests.

    This is related to a recommendation technique called collaborative filtering.

    Collaborative filtering is a method of predicting what someone might enjoy by examining the preferences and behaviour of other people with similar interests.

    Imagine 10,000 people regularly watch videos about African music.

    Among those viewers, a large number also watch documentaries about music production and interviews with African musicians.

    YouTube may identify this relationship and recommend music production documentaries to other people who regularly watch African music content.

    The system can discover these relationships even when viewers have not searched for the recommended videos themselves. [1, 2]

    Another example would be people who watch videos about smartphone photography.

    If many of them also enjoy tutorials about video editing, YouTube may begin recommending video-editing tutorials to other viewers with similar interests.

    This helps people discover content they might not have found through ordinary searches.

    However, it does not mean everyone who watches one particular video will receive exactly the same recommendations.

    Individual viewing habits still matter. [1]

    6. Why does YouTube recommend videos with millions of views?

    Popularity can influence recommendations, but it is not the only consideration.

    YouTube considers how well a video performs when it is presented to viewers who may be interested in it.

    A video that attracts strong interest and satisfies its intended audience may become a candidate for recommendations to more people with similar interests. [3, 5]

    For example, imagine two musicians release new music videos.

    The first video receives 500,000 views, while the second receives 20,000.

    It would be incorrect to conclude that the first video must automatically be recommended to everyone.

    The second video may be particularly relevant to a smaller audience that regularly watches a specific music genre.

    YouTube can therefore recommend it to people who are likely to appreciate that particular style.

    This is one reason why videos from relatively unknown creators can sometimes appear alongside content from established entertainers.

    The recommendation system attempts to match videos with interested viewers rather than simply displaying the biggest channels. [1, 3]

    7. Click-through rate: Does the viewer choose to watch?

    When YouTube displays a recommended video, it observes whether people choose to watch it.

    One measurement used to understand this behaviour is click-through rate, commonly abbreviated as CTR.

    Click-through rate is the percentage of people who click on a video after seeing its thumbnail and title in an eligible impression.

    An impression occurs when YouTube displays a video’s thumbnail in certain places where impressions are counted.

    For example, imagine YouTube displays a video thumbnail 1,000 times and 100 people click on it.

    The click-through rate would be 10%.

    [
    CTR = \frac{100}{1000} \times 100 = 10%
    ]

    A strong click-through rate can indicate that a video’s presentation appeals to viewers.

    However, it does not automatically mean that the video is satisfying or that YouTube will recommend it widely.

    Someone might click because the thumbnail looks exciting, only to leave after a few seconds.

    That is why YouTube also considers what happens after the click. [3, 5]

    8. Watch time: How long do people actually watch?

    Watch time is the amount of time people spend watching a video.

    It is another important measurement used by YouTube to understand viewer engagement.

    Engagement refers to the level of attention and interaction people give to content.

    Consider two videos:

    • Video A is 10 minutes long but most viewers leave after 30 seconds.
    • Video B is 5 minutes long and many viewers watch nearly the entire video.

    Although Video A is longer, Video B may provide a more satisfying viewing experience for its audience.

    YouTube therefore considers more than the total length of a video. It examines how people respond to it. [3, 6]

    Two related measurements are particularly useful to creators:

    Average view duration

    This is the average amount of time viewers spend watching a video.

    For example, if viewers collectively watch a video for 1,000 minutes across 200 views, its average view duration is 5 minutes.

    Average percentage viewed

    This measures the average proportion of a video that viewers watch.

    For example, if a 10-minute video has an average view duration of 5 minutes, the average percentage viewed is 50%.

    Both measurements help creators understand how audiences respond to their content.

    YouTube explains that its discovery systems consider both absolute and relative watch time. Broadly, relative watch time can be more important for shorter videos, while absolute watch time can be more important for longer videos. [6]

    This is why artificially making every video longer is not necessarily a good strategy.

    The length should suit the subject and the value being delivered.

    9. Viewer satisfaction: Does someone actually enjoy the video?

    YouTube has explained that watch time alone is not enough to understand whether a viewer has had a positive experience.

    Someone could spend 20 minutes watching a video without finding it useful, entertaining or satisfying.

    To understand this difference, YouTube also uses viewer satisfaction information.

    One source is surveys that ask viewers to rate videos they have watched.

    These surveys help YouTube understand whether viewers found particular videos valuable rather than simply measuring how long they remained on the platform. [5]

    Other signals, including likes, dislikes and sharing behaviour, can also help the system estimate satisfaction.

    The importance of each signal can differ depending on the individual viewer.

    For example, someone may frequently share videos without necessarily enjoying all of them. YouTube’s systems attempt to interpret signals in context rather than assuming that every action has the same meaning for everyone. [5]

    This is important because the platform is trying to identify content people genuinely appreciate.

    10. Why does YouTube recommend different videos on your homepage and while watching?

    You may notice that the videos displayed on your YouTube homepage are different from the recommendations appearing beside or below a video you are currently watching.

    This happens because YouTube uses different signals for different parts of the platform.

    The homepage

    Your homepage is the collection of videos displayed when you open YouTube.

    It is highly personalised and relies primarily on your watch history, together with other information about your interests and viewing behaviour.

    It may include videos from channels you follow, creators you have never watched before and subjects related to your interests. [2, 3]

    Up Next

    Up Next is the recommendation section shown while you are watching a video.

    Here, the video you are currently watching becomes an important signal.

    For example, if you are watching an interview with a filmmaker, YouTube may recommend other filmmaking interviews, discussions about film production or related documentaries.

    However, the suggestions can also reflect your wider viewing interests. [2, 3]

    The Shorts feed

    YouTube Shorts is designed for short-form vertical videos.

    Its recommendation feed is personalised, using information about the subjects and formats a viewer appears to enjoy.

    The system also considers the different viewing patterns associated with Shorts. [2, 4]

    The Subscriptions tab

    This is different from personalised recommendations.

    The Subscriptions tab primarily displays recent uploads from channels you have subscribed to, generally arranged by publication time.

    It helps viewers keep up with creators they have chosen to follow. [4]

    11. Does the time of day or your location affect recommendations?

    Yes.

    YouTube considers contextual information when determining which videos may be relevant to a viewer.

    Context refers to the circumstances surrounding someone’s use of the platform.

    According to YouTube’s published explanations, factors can include location, language and time of day. Its newer guidance also identifies the device being used as relevant context. [1, 4]

    For example, imagine someone who watches news on a smartphone in the morning and comedy videos on a television in the evening.

    YouTube may learn these viewing patterns and adjust recommendations to match the person’s routines.

    Similarly, someone in Nigeria may receive different locally relevant recommendations from someone in Japan, even if they share an interest in music.

    Language can also influence which videos are useful to a viewer. [4]

    However, location does not mean that YouTube recommends only videos from someone’s country.

    International content can still appear whenever the system considers it relevant.

    12. Does YouTube promote videos that contain misleading or harmful information?

    Not all videos are treated equally when it comes to recommendations.

    YouTube has policies and systems designed to limit the distribution of certain harmful content, including borderline content and videos containing harmful misinformation. [7]

    Borderline content refers to material that may come close to violating YouTube’s Community Guidelines without necessarily crossing the threshold for removal.

    For example, a video may contain misleading claims about a sensitive subject without meeting the platform’s criteria for removal.

    YouTube has explained that its recommendation systems are designed to reduce the visibility of certain borderline content, particularly where it could cause harm. [7]

    This is different from removing a video.

    A video may remain available on the platform while being less likely to appear in recommendations.

    YouTube also applies different recommendation and eligibility rules to certain categories of content.

    Consequently, a video receiving views does not automatically mean it will be recommended widely. [3, 7]

    13. Why do some videos suddenly receive thousands or millions of views?

    A video can experience a sudden increase in views when it begins reaching audiences beyond its existing subscribers.

    This can happen when YouTube identifies groups of viewers who may be interested in the subject and starts recommending the video to them.

    For example, imagine a small channel publishes an interview with an emerging musician.

    Initially, most views may come from the musician’s followers and the channel’s existing audience.

    If viewers respond positively, YouTube may identify other people who regularly watch similar musicians or entertainment interviews.

    The video may then become relevant to a much larger audience. [1, 3]

    However, this does not mean every video that performs well with a small audience will automatically become viral.

    Several external factors influence how widely content can be recommended.

    Topic interest

    Some subjects attract larger audiences than others.

    A video about a major international sporting event may have broader appeal than a specialised tutorial about an uncommon technical subject.

    Competition

    YouTube is comparing videos that could be relevant to the same viewer.

    Even when a video performs well, competing videos may attract stronger interest from the same audience.

    Seasonality

    Seasonality refers to changes in interest associated with particular periods of the year.

    For example, searches and video viewing about Christmas decorations may increase towards the end of the year.

    Interest in certain sporting events may also rise during major competitions.

    These factors can affect the number of opportunities a video receives to appear in recommendations. [6]

    14. Does YouTube favour big channels over small channels?

    YouTube’s published guidance does not describe its recommendation system as simply favouring channels with the most subscribers.

    Instead, it explains that recommendations are intended to match videos with viewers who are likely to find them relevant and satisfying. [1, 3]

    A smaller channel can therefore reach people who have never watched it before.

    For example, a new educational channel publishing a useful explanation of solar energy may attract recommendations among viewers who regularly watch science and renewable energy content.

    However, established channels may already have audiences with strong viewing habits, giving them an existing group of people who are interested in their uploads.

    This can create an advantage in reaching viewers who already know and follow them.

    The important distinction is that subscriber numbers alone do not guarantee recommendations.

    YouTube evaluates how videos perform with audiences and how relevant they are to individual viewers. [3]

    15. Does uploading every day make YouTube recommend your videos more?

    Not necessarily.

    There is a widespread belief that creators must upload every day or several times a week to remain visible in recommendations.

    YouTube has explained that its analysis has found no correlation between growth in views across uploads and the amount of time between uploads.

    It also states that taking a break does not automatically damage a channel’s performance. [6]

    This does not mean publishing schedules are irrelevant to building an audience.

    A consistent schedule can help viewers know when to expect new content and develop a routine around a channel.

    However, uploading frequently at the expense of quality may not produce the results a creator expects.

    YouTube recommends focusing on understanding the audience and delivering content that viewers value. [1, 6]

    16. Can creators influence whether YouTube recommends their videos?

    Creators cannot directly instruct YouTube to recommend a video to a particular number of people.

    However, they can improve the chances of reaching interested audiences by making content that responds to viewers’ needs and expectations.

    Some important considerations include:

    Choose subjects people are interested in

    Understanding the interests of an intended audience helps creators produce videos that people are more likely to discover and watch.

    Use accurate titles

    A title should clearly communicate what the video contains.

    For example, a video explaining how GPS works should have a title that reflects the actual explanation rather than promising unrelated information.

    Create useful thumbnails

    A thumbnail is the image representing a video before someone opens it.

    It should help viewers understand what the video is about and give them a reason to watch.

    Misleading thumbnails and titles can create disappointment when the actual video does not match the promise. [3]

    Deliver what the title promises

    If a video promises to explain how something works, viewers should receive a meaningful explanation rather than lengthy introductions and unrelated discussions.

    Understand audience retention

    Audience retention measures how long viewers continue watching a video and where they stop.

    YouTube Studio, the platform’s creator management and analytics service, provides reports that help creators understand viewing patterns.

    These reports can reveal whether people leave early, skip certain sections or continue watching until the end. [6]

    Study the audience, not just the algorithm

    YouTube provides an Audience tab in YouTube Analytics.

    Creators can use it to understand their viewers’ interests, preferred formats and other channels or videos their audience watches.

    This information can help guide future content decisions. [1]

    17. How can viewers control what YouTube recommends?

    YouTube gives users several controls for managing their recommendation experience.

    These controls are useful when the homepage repeatedly displays content that no longer interests someone.

    Some practical options include:

    1. Use Not interested

    When an unwanted video appears, select the three-dot menu beside it and choose “Not interested”.

    This tells YouTube that you do not want that particular video recommended.

    2. Use Don’t recommend channel

    If you do not want recommendations from a particular channel, select “Don’t recommend channel” from the available feedback options.

    3. Manage your watch history

    You can pause or delete watch history through your YouTube or Google account settings.

    This can affect future recommendations because the system uses watch history to understand your interests.

    However, removing history does not necessarily remove every other source of personalisation. [2]

    4. Remove specific videos from your history

    If you watched something unrelated to your usual interests, you can remove that video from your watch history.

    For example, if you borrowed someone’s phone and watched several videos about a subject you do not normally follow, removing those entries may help prevent them from influencing future recommendations.

    5. Use the Subscriptions tab

    If you want to see recent uploads from channels you have deliberately followed, the Subscriptions tab provides a more direct way to find them without relying entirely on personalised homepage suggestions. [2, 4]

    18. Myths vs facts about YouTube recommendations

    ClaimVerdictExplanation
    YouTube recommends only videos from channels with millions of subscribers.MythSmaller channels can also reach new audiences through recommendations.
    A video with the most views always gets recommended first.MythRecommendations depend on viewer interests, relevance, performance and other signals.
    YouTube learns from your watch history.TrueViewing habits help the system estimate your interests.
    Liking a video can influence future recommendations.TrueLikes are among the signals used to predict what you may enjoy.
    Disliking a video automatically removes all similar content.MythA dislike is one signal, not a guarantee that every similar video will disappear.
    YouTube considers how long people watch a video.TrueWatch time and average view duration help measure viewer engagement.
    A video must be a particular length to go viral.MythYouTube states that there is no universal ideal video length.
    Uploading every day guarantees more recommendations.MythPublishing frequency alone does not guarantee better recommendation performance.
    YouTube can recommend videos from channels you have never watched.TrueIt can identify new content based on your interests and the viewing patterns of similar audiences.
    YouTube considers viewer satisfaction, not just clicks.TrueSurveys and other feedback help the system estimate whether viewers value a video.
    Every video on YouTube is eligible for recommendations.MythCertain content may be ineligible or receive reduced recommendations under platform policies.
    Two people can receive different recommendations after watching the same video.TrueTheir wider viewing histories, interests and contexts may differ.
    Paying for YouTube advertising automatically improves organic recommendations.MythAdvertising and organic recommendation systems are separate.

    Bottom line

    YouTube decides which videos to recommend by analysing what individual viewers watch, search for, like, dislike and ignore. It combines this information with video performance, viewing satisfaction, the interests of similar audiences and contextual factors such as location, language, device and time of day. Its goal is not simply to promote the most popular videos but to match viewers with content they are likely to find valuable and enjoyable. For creators, this means that gaining visibility is not just about uploading frequently, collecting subscribers or chasing views. It is about understanding an audience, producing relevant content, presenting it accurately and delivering the experience viewers expect. Ultimately, YouTube’s recommendation system is designed to find the right videos for the right viewers, rather than expecting every viewer to watch the same popular content.

    Disclaimer: This article is for general educational purposes only and does not replace professional advice relevant to digital publishing, content creation or platform management. If you notice an error or would like to send feedback or corrections, please contact us.

    References

    • [1] YouTube’s Recommendation System, YouTube Help
    • [2] How YouTube Recommendations Work, YouTube Help
    • [3] YouTube Performance FAQ and Troubleshooting, YouTube Help
    • [4] Good to Know About Recommendations for YouTube’s Recommendation System, YouTube Help
    • [5] On YouTube’s Recommendation System, YouTube Official Blog
    • [6] Good to Know About Recommendations for YouTube’s Recommendation System, YouTube Help
    • [7] How YouTube Recommends Videos, YouTube and Google

    Share this:

    • Share on Facebook (Opens in new window) Facebook
    • Share on X (Opens in new window) X
    YouTube
    Share. WhatsApp Facebook Twitter LinkedIn Email Telegram Copy Link
    Digital Reporter
    Digital Reporter

    Digital Reporter is a section under TrenderHQ's editorial team. We make every effort to ensure our information is accurate, though occasional errors may occur. Readers are encouraged to verify details with other trusted sources. For corrections, updates, or feedback, please email: info@trenderhq.com. To advertise with us or support our cause through a donation, reach out via our contact page or email: hello@trenderhq.com

    Related Posts

    How does TikTok know what videos you like?

    November 19, 2026

    How does Google Search decide which results to show you?

    November 19, 2026

    How does your phone know when you are moving?

    November 19, 2026
    Leave A Reply Cancel Reply

    You must be logged in to post a comment.

    Ads
    Advertise with us
    Advertise with us
    Top Posts

    Odo Broni: Daddy Lumba’s Wife Priscilla Ofori

    July 30, 20252,838

    10 Simple Steps to Achieve Greatness in Your Life

    August 13, 20232,369

    Easter dates for the next 20 years (2026 – 2045)

    March 14, 20262,143

    British Man who posted luxury on social media gets drugged and murdered for his watches!

    January 13, 20232,136
    IntekHost
    Don't Miss
    Technology

    How does TikTok know what videos you like?

    November 19, 20260

    TikTok’s ability to recommend videos can sometimes feel almost unbelievable. You might watch a few…

    Share this:

    • Share on Facebook (Opens in new window) Facebook
    • Share on X (Opens in new window) X

    How does YouTube decide which videos to recommend? The algorithm explained

    November 19, 2026

    How does Google Search decide which results to show you?

    November 19, 2026

    How does your phone know when you are moving?

    November 19, 2026

    What actually happens when you turn on airplane mode?

    November 19, 2026

    How does Wi-Fi actually work?

    November 19, 2026

    How does Bluetooth work?

    November 19, 2026
    Stay In Touch
    • Facebook
    • Twitter
    • Pinterest
    • Instagram
    • YouTube
    • LinkedIn
    • TikTok
    • Telegram
    • WhatsApp
    • Tumblr
    • Threads
    • Bluesky

    Subscribe to Updates

    Get the latest creative news and more from TrenderHQ.

    TrenderHQ © 2026 | powered by Intek Multimedia.

    • About Us
    • Contact
    • Privacy
    • Disclaimer
    • Services
    • Advertise

    Type above and press Enter to search. Press Esc to cancel.

    TrenderHQ

    Create a TrenderHQ Account for free

    Join TrenderHQ to stay connected and get more from our platform.

    Sign Up

    The close button will appear in 15 seconds.

    Already a member? Log in