YouTube Shorts Recommendations: How Videos Get Views

What Are YouTube Shorts Recommendations

YouTube Shorts recommendations are a selection of videos that the platform shows to a specific viewer based on their interests and behavior. The main source of these views is the Shorts feed, but videos can also be discovered through the home page, subscriptions, search, and channel pages.

YouTube does not disclose the full recommendation formula. That's why there's no single action that will automatically push a video to a large audience. The system evaluates several groups of signals: viewer reaction, how well the video matches their interests, viewing quality, and overall satisfaction with the content.

Subscriber count affects the initial audience but doesn't determine the entire reach. A new channel can also get recommendations if the video is interesting to the viewers it's shown to. At the same time, no format, hashtag, or sound guarantees a large number of views.

How a Video Enters the Feed

After publishing, YouTube determines what the video is about and finds potentially interested viewers for it. This uses the video's content, title, on-screen text, description, selected audience, and data on what videos users watch.

The first impressions don't always mean a full-scale promotion launch. A video may gradually find its audience, get views in waves, or barely spread in the first hours. This is a normal feature of recommendations: the system matches the video with different groups of viewers rather than making a final decision only at the moment of publication.

If people often swipe past the video, close it quickly, or rarely watch it to the end, further impressions may slow down. If the video holds attention and gets a positive reaction, YouTube has more reason to offer it to similar viewers.

It's important to distinguish a recommendation from a subscription. A user can watch a Short without subscribing to the channel. So the video's goal is not only to bring a new subscriber but also to give the viewer a clear reason to watch to the end and move on to other content.

What Signals YouTube Considers

Recommendations don't have one main metric for all videos. The system looks at a combination of metrics and compares viewer behavior with how they interact with other videos.

• View or swipe. A viewer can stay on the video or swipe past it right away. The clearer the opening, the higher the chance the user will keep watching.
• Retention. What matters is how much of the video people watch and at what moments they leave. For a short video, sharp drops at the beginning are especially noticeable.
• Completions and rewatches. If a video makes people want to watch it to the end or rewatch it, that shows the content holds attention.
• Reactions. Likes, comments, shares, and subscriptions help gauge interest, but on their own they don't make a video recommended.
• Negative signals. Hiding recommendations, frequent skips, and reports show that the video doesn't suit part of the audience.
• Interest match. The same Short can perform well for one group of viewers and fail to interest another.

You shouldn't try to artificially improve a single metric. For example, a large number of likes with weak retention doesn't replace an interesting story. What matters is the viewer's overall reaction: they stopped, understood the topic, watched to the end, and wanted more or another video from the channel.

Why the First Seconds Matter

In the Shorts feed, a viewer makes a quick decision: keep watching or swipe past. So the opening should immediately show the topic, result, or conflict. A long greeting, intro, or generic phrase without substance takes up the most valuable spot in the video.

The hook doesn't have to be loud or provocative. For an educational video, it can be a specific mistake and how to fix it. For a gaming stream, it can be an unusual moment with a short explanation. For an expert channel, it can be a conclusion that's usually only revealed at the end of a long piece.

The content of the first seconds should match what follows. If the title and opening promise one thing but the video quickly moves to another topic, viewers will leave. That disconnect hurts retention and reduces trust in the channel.

How to Hold Attention Without Deception

• Cut pauses, repetitions, and phrases that don't move the story forward.
• Change the visual shot only when it helps comprehension, not for the sake of constant motion.
• Add subtitles if the meaning matters even without sound.
• Build the video around one clear idea rather than several unrelated topics.
• End the video with a conclusion, a result, or a natural transition to the next piece of content.

The Role of Topic, Presentation, and Quality

Recommendations take into account more than just the reaction after publishing. YouTube needs to understand who might be interested in the video. So it helps to clearly state the topic in speech, on-screen text, title, and description. This doesn't mean you should fill the description with a long list of keywords.

The title should explain the content, not promise an impossible result. Hashtags can clarify the topic but don't replace the script and clear delivery. Add them in moderation and only relevant to the video.

Image and sound quality also affect perception. A blurry picture, quiet speech, noise overpowering the voice, or small text make viewing harder. If the video uses someone else's music, a stream fragment, or an image, check the rights and platform rules.

A consistent thematic focus helps the channel build a clear audience. If one video is about games, the next about makeup, and the third about repairs, it's harder for the system to determine who to offer new posts to. Different topics are possible, but each one is better off having a clear format and an expected audience.

How to Read Shorts Analytics

Analytics aren't for finding one "correct" number but for understanding viewer behavior. First, look at how many users chose to watch instead of swiping. Then study retention and the moments where the audience leaves.

If viewers swipe past the video at the beginning, check the first frame, how the topic is phrased, and how quickly you get to the point. If they watch the beginning but leave in the middle, there's likely a drawn-out explanation, weak structure, or a mismatch with the promise. If the video gets watched to the end but gets almost no reactions, that's not necessarily a problem: the video can be useful but not spark a desire to like or comment.

Compare videos of the same type with each other. Don't compare a short funny clip, a detailed tutorial, and a stream highlight based on views alone. They have different goals, audiences, and script structures.

Check the stats not right after publishing but after enough data has accumulated. Early numbers can change while the video is being shown to different viewers. One video doesn't always tell you enough to draw conclusions about the entire channel's format.

What Prevents Videos From Getting Recommendations

• A long intro without a clear topic.
• A title that doesn't match the content.
• Overloaded editing that makes it hard to follow the thought.
• Vertical video with small text and important details at the edges of the frame.
• Posting on random topics without understanding the target audience.
• Trying to get reactions through spam, deceptive promises, or artificial activity.

Don't delete a video just because of a weak start, and don't re-upload it multiple times without changes. First figure out the cause: is it the topic, delivery, quality, first seconds, or a mismatch with the audience. Re-publishing only makes sense with noticeably reworked content and in line with YouTube's rules.

In Short: What to Do Step by Step

1. Choose one topic and determine which viewer the video is for.
2. Formulate the main idea and start with it, without a long intro.
3. Make a clear title, on-screen text, and description.
4. Check the sound, subtitle readability, vertical frame, and rights to materials.
5. After publishing, study swipes, retention, and the moments where viewers leave.
6. Compare the result with similar videos and change one or two elements in the next video.