One video gets thousands of views, while another gets only dozens, even though the authors spent equally long on both. The difference is often not a "secret algorithm" but how well the video fits a specific viewer and holds their attention. Let's look at how YouTube and TikTok describe their recommendations and what practical conclusions authors can draw from this.
YouTube Matches Videos to Viewers, Not Viewers to Videos
A common belief about the algorithm goes like this: the platform "promotes" a video by showing it to an ever-larger audience. YouTube describes its recommendation system differently: it selects videos for each user, taking into account their interests, viewing history, and reactions to similar content.
That is why a video does not have a single chance of success for all viewers. YouTube may offer one person a detailed review, another a short tutorial, and a third neither. The system tries to predict which video will be most relevant to a specific user.
This explains why the number of subscribers alone does not guarantee a large reach. Subscribers may give the first viewers a reason to open a video, but further recommendations depend on how the audience responds. If viewers leave quickly or do not find what they expected, a large channel gets no automatic protection from low views.
Practical takeaway: when creating a video, think not only about the channel's topic but also about a specific group of viewers. The more precisely you define the video's purpose, for example "show how to choose a microphone for home recording," the easier it is to make the content clear and useful for your audience.
Why Watch-Through Matters for Shorts
For short videos, what matters is not only how many people started watching but also how much time they spent with the video. A high watch-through rate can signal that the video holds attention. Repeat views can also indicate interest, especially if the video is short and the viewer rewatches it to catch a detail or understand a sequence of actions.
However, watch-through should not be treated as the only measure of success. Other behavioral signals matter too: for example, a viewer might move on to the next video, open the channel, or interact with the video. In addition, the meaning of the same metric depends on format and purpose. A short entertainment clip and a long explanation have different expected audience behavior.
Shorts have no familiar scenario in which a viewer first chooses a video from a thumbnail on a recommendations page. A person scrolls through a feed, and the video starts playing immediately. That is why the first seconds become especially important: this is when the viewer decides whether to keep watching or scroll past.
New videos first reach a limited audience. Roughly speaking, the first impressions can be seen as an early test of reaction: the platform collects signals about who might be interested in the video and how people interact with it. This does not mean there is a fixed "100-view test" or a guaranteed threshold after which a video will definitely start gaining reach. Recommendations are personalized, and results depend on many factors.
To increase the likelihood of watch-through:
- start with the point, not a long greeting;
- immediately show the situation, question, or result that made the viewer open the video;
- remove pauses and repetitions if they do not help understanding;
- end the video logically, without dragging it on after the main conclusion;
- do not promise at the start something that is not in the video itself.
How the TikTok "For You" Feed Works
TikTok also describes its recommendations as personalized content selection. What appears in the "For You" feed is influenced by the user's reactions, information about the video, and the viewing context. Important signals include, for example, watch-through, skips, likes, comments, follows, and information about the video's content.
At the same time, the number of followers is not a direct condition for a video to be recommended. An account with a small audience can get views if the video turns out to be interesting to suitable viewers. Conversely, a large number of followers does not guarantee that every new video will be actively shown.
TikTok's description of recommendations also notes the role of content freshness. In particular, videos older than 90 days are generally not considered for recommendations in the "For You" feed. This does not mean that a video posted today will necessarily appear in the feed: freshness is just one factor, not a guarantee of reach. The video's relevance to a specific audience and viewers' reactions still matter.
Authors should also consider the nature of the feed itself. Users switch between videos quickly, so the meaning of a video should be clear without a long introduction. The caption, visuals, and sound should help convey the topic rather than distract from it. If a video targets a narrow audience, that is not necessarily a drawback: what matters is that it reaches exactly the people for whom the topic may be useful.
Myths About Promoting Videos
"Just re-upload the video"
Republishing the same video does not guarantee new reach. Moreover, multiple identical uploads may look like spam and do not solve the original problem: viewers may keep scrolling past or not understand the video's value.
If the first version did not work, it is more useful to figure out where attention is being lost. Perhaps an overly long introduction can be shortened, an unclear topic can be formulated more precisely, and a weak start can be replaced with a more concrete one. A substantially reworked version is no longer just a re-upload but a new version of the content.
"The #FYP hashtag leads to recommendations"
#FYP is an ordinary hashtag, not a command to the algorithm. On its own, it does not get a video into the "For You" feed and does not replace a clear topic, an engaging opening, and interesting content. Hashtags can help label the context of a video, but you should not fill captions with them hoping for guaranteed reach.
It is better to choose a few tags that genuinely describe the topic and audience. For example, for a video about setting up home lighting, topical tags are more useful than a set of the most popular tags unrelated to the content.
"The main thing is to post at the right time"
Posting time can influence initial views, especially if an active audience follows the channel. But it does not determine the long-term result: recommendations can keep finding viewers later, and posting at peak hours will not save a video that fails to hold attention.
If your audience regularly watches content at certain times, the schedule can be considered as an organizational factor. But do not draw conclusions about a video's quality based only on the time it was posted. It is more useful to compare several videos with a similar topic and format rather than explaining every result with a single external circumstance.
How to Evaluate Results and Improve Future Videos
Views show the scale of reach but on their own explain little about why a video succeeded or failed. To understand what happened, compare them with other metrics: retention, watch-through, reactions, and click-throughs. Study analytics separately for each format: short vertical videos and long videos solve different tasks.
For your next post, choose one hypothesis and test it. For example: "If I show the finished result in the first two seconds, more viewers will watch the video to the end." Then change exactly that element while keeping the topic and other parameters as constant as possible. This approach yields more information than changing the topic, length, presentation, and posting time all at once.
It is useful to keep a simple table:
| What to compare | Question it answers |
|---|---|
| Video opening | Does the viewer understand the topic immediately? |
| Retention and watch-through | At what point does the audience lose interest? |
| Topic and format | Which subjects suit your audience best? |
| Reactions and click-throughs | Does the viewer want to keep exploring your content? |
Practical Takeaway
YouTube and TikTok do not promise every video the same reach: they select content for users and assess audience reactions. Therefore, growth in views depends not on a single hashtag, posting time, or number of subscribers, but on a combination of topic, presentation, and viewer interest.
Start your video with a clear idea, deliver on the promise made in the first seconds, and analyze audience behavior rather than just the final view count. If a video got few views, this is not always a reason to delete it or re-upload it: first find the specific place that can be improved in the next version.
