The YouTube Shorts algorithm shows each new Short to a small group of viewers it predicts will like it, measures how they react, and keeps widening the audience while the reactions stay strong. YouTube's own Shorts guidance names three groups of signals: whether viewers choose to watch or swipe away (the “stayed to watch” rate), how long they watch (average view duration and percentage viewed), and whether they enjoyed it (likes, post-watch surveys and “not interested” clicks). Each viewer's history with topics, channels and sounds decides who sees it first. Hashtags, upload time and subscriber count matter far less than most advice claims, and a Short stuck at 0 views usually has not been shown in the feed yet.
This explainer is for anyone posting Shorts in 2026, from first-time faceless creators to marketers whose Shorts stall at a few hundred views. It is built on what YouTube itself has published: its Help Center pages on Shorts discovery and analytics, its official blog, and interviews with the Shorts product team. Where we add our own channel data, we say so and give the sample size. You'll get the test-and-expand loop, every published signal with the lever that moves it, the view-count changes of 2025 and 2026, the common myths, a 0-views checklist, and how each signal maps to a setting in Vidpal Studio. Facts are current as of October 2026.
Related guides: how to get more views on YouTube Shorts, what a good Shorts retention rate looks like and YouTube Shorts length and size requirements.
What Is the YouTube Shorts Algorithm?
The Shorts algorithm is the set of recommendation systems that decides which Short each viewer sees next in the Shorts feed, on the Home tab and in search. YouTube's Shorts search and discovery tips put it in one line: “Shorts are ranked based on performance and viewer personalization.” Performance is how viewers react when the Short is shown to them; personalization is how well it matches a particular viewer's history.
The system has no taste of its own. The same help page says “our system has no opinion about what type of Shorts you make,” and YouTube's Shorts product lead Todd Sherman summed up the philosophy in a 2023 Creator Insider interview as “the audience is the algorithm,” as Social Media Today reported. It follows what viewers do, one swipe at a time, which is why the same Short can stall on Monday and take off weeks later when it reaches a different set of viewers.
A few terms you'll see throughout. The Shorts feed is the vertical, swipe-to-advance player where most Shorts views happen. Stayed to watch is the percentage of times viewers kept watching past the first seconds of a Short instead of swiping away. An engaged view is a view where the viewer stayed past those initial seconds, not counting loops. Average view duration is the mean watch time per engaged view, and average percentage viewed is that watch time as a share of the Short's length.
How Does YouTube Decide Which Shorts to Push?
YouTube pushes a Short by repeating one loop: show it to viewers likely to enjoy it, measure the response, and show it to more people while the response holds up. YouTube has never published a fixed “test audience” size or schedule, so treat the steps below as a description of how distribution behaves, consistent with what the Shorts team has said, rather than a documented pipeline.
Step 1, classification: after upload, YouTube treats a vertical or square video of up to three minutes as a Short. That has applied to uploads since October 15, 2024, according to its three-minute Shorts help page. A horizontal video, or anything longer, goes to the long-form system instead.
Step 2, first impressions: the Short is placed into the feeds of viewers whose watch history, favorite topics and sounds suggest interest. These are mostly people who have never heard of your channel, which is why a brand-new channel can get its first thousand views in a day and why subscriber count matters less on Shorts than on long-form.
Step 3, measurement: for each impression, YouTube records whether the viewer watched or swiped, how much they watched, whether they liked it, and any survey answer or “not interested” click.
Step 4, expansion or plateau: if early viewers respond better than the alternatives competing for the same feed slots, the Short is shown to a wider and more varied audience; if not, impressions taper off. Search Engine Journal's summary of the Sherman interview describes those early views as exploratory: a Short can collect a few hundred or a few thousand views and then drop off while the system searches for its audience.
Step 5, later rounds: because ranking is relative and personal, a Short that plateaued can get a second wave when its topic trends or when it starts reaching better-matched viewers.
One detail is easy to miss. YouTube's help page warns that “even if you have good metrics on your Short, you may get fewer impressions if Shorts from other channels are performing even better.” The loop is a competition for limited feed slots on a topic, not a pass-or-fail exam against a fixed score, so the same numbers can earn different reach in different weeks and niches.
What Signals Does the Shorts Algorithm Use?
YouTube lists the signals in three plain groups, plus personalization. When a Short is recommended, the system checks whether the viewer chooses to watch it, ignores it or clicks “not interested”; it uses the “% of viewers who chose to view, avg. view duration and avg. % viewed” to inform ranking; and it looks at “whether or not viewers enjoyed the Short using likes and post-watch survey results.” For personalization, it considers the Shorts and channels a viewer enjoyed before, the topics they watch, and trending songs or sampled audio they engaged with.
YouTube's general page on how recommendations work explains why surveys are on the list: they help the system understand satisfaction, “not just watch time.” Two signals creators talk about constantly, shares and comments, do not appear in YouTube's published Shorts list. They may matter indirectly, because a shared Short reaches new viewers who then generate the measured signals, but no YouTube page calls them ranking inputs, so treat any claim about their exact weight as a guess.
The table maps each signal to what it tells the system, how to improve it, and the Vidpal Studio setting that moves it.
| Signal | What it tells YouTube | How to improve it | Studio lever |
|---|---|---|---|
| Chose to view (stayed to watch vs swiped away) | The opening earned attention in the first seconds | Open on the payoff, a number or a question; never on a greeting, logo or setup | Hook Intro: Viral Clip (Library, Text → Video, Image → Video) or Text Reveal |
| Average view duration | Viewers kept watching once they stayed | Cut lines that don't move the story; change the visual every few seconds; keep captions readable | Per-section AI rewrite, 20 visual styles, 75 caption presets with word-by-word animations |
| Average percentage viewed | Watch time relative to length; shorter Shorts reach high percentages more easily | Match length to the idea and end as soon as the payoff lands | Duration picker: 30, 45, 60, 75 or 90 seconds |
| Replays and loops | Viewers watched again; each replay counts as a view, but engaged views exclude loops | End on a line that connects back to the opening | Script ending; shorter 30 to 45-second durations |
| Likes | Viewers enjoyed it enough to tap | Deliver a clear payoff before asking for anything | Script ending and call to action |
| Post-watch surveys | Direct satisfaction ratings from a sample of viewers | Make the hook, title and content keep the same promise | One topic or one full script per Short |
| Not interested clicks | The Short was a poor match or felt low-value | Stay in one niche so the system can find the right viewers | Brand Kit for a consistent look across Shorts |
| Personal relevance (history, topics, sounds) | Which viewers see it first | Give each Short one clear topic and keep the channel consistent | Topic, Full Script or From URL input focused on one subject |
| Shares and comments | Not on YouTube's published list; may help indirectly | Give people a reason to send it: a surprising fact or a useful tip | Script angle; From URL for timely stories |
What Does “Stayed to Watch” Mean, and What Is a Good Rate?
Stayed to watch is YouTube's measure of how many feed impressions turned into real viewing. Its content performance help page defines it as “the percentage of times viewers stayed to watch past the initial seconds of a Short.” In the Content tab, the same idea appears as “How many chose to view,” which compares viewed with swiped away, next to “Shown in feed,” the number of times your Short appeared in the Shorts feed, according to YouTube's Shorts analytics tips.
YouTube publishes no target. In our own October 2026 review of about a dozen Shorts on our test channels, the ones that dropped below roughly 25 percent stayed to watch stopped getting new impressions within a day or two, while the ones that kept being pushed sat around 30 percent or higher. That is a small sample in a handful of niches, so read it as our observation rather than a rule, and compare each Short with your own channel's typical numbers.
The same review showed where the swipe happens. Our finished Studio stories held 60 to 70 percent average view duration once people stayed, but versions that opened with a few seconds of context before the payoff lost most viewers at the start. The fix was to move the result into the first second. If your stayed-to-watch rate is low but average view duration is healthy, the problem is the opening, not the video. For tiers and benchmarks, see our guide to good retention rates for YouTube Shorts.
How Are Shorts Views Counted in 2026?
Since August 24, 2026, YouTube counts a view the moment any video starts to play, Shorts included, and calls the older, stricter measure an engaged view. Shorts got this change first: from March 31, 2025, a Short's view counted each time it started or replayed, with no minimum watch time, as Entrepreneur reported at the time. YouTube's August 2026 post on engaged views extended first-frame counting to long-form videos, podcasts and live streams.
Two things did not change, according to that post. There is “no change to recommendations,” and engaged views remain how you qualify to earn in the YouTube Partner Program. Most analytics, including average view duration and retention, are still calculated from engaged views.
The practical effect is that the public view count now includes every quick swipe-past, so it flatters weak openings. A Short with 10,000 views and a 20 percent stayed-to-watch rate had roughly 2,000 engaged views; a Short with 5,000 views at 60 percent had roughly 3,000. Judge Shorts by engaged views and stayed to watch, and use the headline count only to compare reach.
Which Shorts Algorithm Myths Should You Ignore?
Most Shorts advice treats hashtags, posting time and a 48-hour window as levers. YouTube's own statements say otherwise on all three, and on a few other popular tactics.
Myth 1, hashtags push Shorts into the feed. YouTube's hashtag help page describes hashtags as a way to connect your content with other videos that share the same tag, shows up to three of them above the title, and ignores every hashtag on a video that has more than 60. Sherman said hashtags aren't required and help mainly when they relate to a real-world event or a specific topic. A relevant tag or two is harmless; a block of 30 does nothing for distribution. Our guide to the best hashtags for YouTube covers when they help search.
Myth 2, there is a best time to post Shorts. In the same interview, as MediaPost summarized it, Sherman said that outside breaking news, timing matters less than the content itself. Publishing when your audience is active can help early viewership on long-form, where subscribers are the first wave, but the Shorts feed mostly serves people who don't follow you. Pick a time you can keep; our guide to the best time to post has benchmarks if you want a starting point.
Myth 3, a Short is dead after 48 hours. No YouTube page mentions a 48-hour window. YouTube's recommendations guidance does say the Shorts feed “may tune up on the recency of content,” so most Shorts get the bulk of their views early; one of ours with 70 percent average view duration went flat about eight hours after posting. But recency is a weighting, not a cutoff, and a Short can resurface weeks later when it finds a better-matched audience or its topic trends.
Myth 4, more Shorts automatically means more reach. YouTube's Shorts page says “there is no minimum posting cadence required,” and Sherman said volume alone doesn't expand reach; a flood of weak Shorts mostly produces weak signals. Cadence helps because each Short is another test, not because the system rewards frequency. Our guide on how often to post Shorts has sensible starting rates.
Myth 5, delete and re-upload a Short that flopped. Sherman advised against it and said re-uploads risk being treated as spam. Make a new Short with a different opening instead.
Myth 6, one bad Short, or posting Shorts at all, hurts your channel. YouTube says one underperforming video doesn't penalize the whole channel, that experimenting with formats such as Shorts won't confuse the algorithm, and that “Shorts performance doesn't negatively impact long form Video recommendations.”
Why Do Some Shorts Get 0 Views?
A Short at 0 views has almost always not been served in the feed yet, because under the current definition a view counts the moment a Short starts to play. Check “Shown in feed” in the Short's analytics first: if it is zero or close to it, the problem is eligibility or matching, not your hook. These are the usual causes.
- It isn't a Short: horizontal videos, and anything over three minutes, are treated as long-form.
- It isn't public yet: private, unlisted and still-scheduled videos don't enter the feed.
- It has a copyright claim: YouTube's three-minute Shorts page says Shorts over one minute with an active claim are blocked and not recommended, while its newer Create Shorts page says they are no longer automatically blocked, so check the Copyright column in Studio either way.
- It is age-restricted: signed-out viewers and viewers under 18 can't watch it, which shrinks the pool the system can show it to.
- The topic has little demand: YouTube names topic interest, how many people are watching Shorts about a subject, as a reason impressions vary.
- Stronger Shorts are competing for the same viewers on the same topic.
- It reuses other people's footage: unedited clips from other people's movies or TV shows are ineligible for Shorts revenue sharing and run into YouTube's reused-content rules.
If “Shown in feed” is healthy but views stay tiny, the Short was served and swiped. That is a stayed-to-watch problem: rewrite the first line, change the opening visual, and post a new version rather than re-uploading the same file.
New channels are not locked out of the feed, but YouTube notes that for new channels “gaining this credibility and building a community takes time,” and that a library of quality videos helps new viewers explore. Early Shorts with a few dozen views are normal; the number to watch is whether stayed to watch improves from one upload to the next.
How Do You Work With the Shorts Algorithm? A Checklist
You work with the algorithm by giving it a clear topic to match and a strong opening to test; everything else is detail. These steps follow directly from the signals YouTube names.
Step 1: Choose one promise per Short. One question, one payoff. Mixed topics confuse personalization and invite “not interested” clicks from viewers who came for the first half.
Step 2: Put the payoff in the first second. Lead with the result, the surprising number or the question, and move context after it. Our scroll-stopping hook examples and the free YouTube hooks generator give you openings to test.
Step 3: Write to length. At about 150 words a minute, a 30-second Short holds roughly 75 words and a 60-second Short about 150. Cut until every line earns its seconds; a tight 40-second Short beats a padded 60-second one on percentage viewed.
Step 4: Keep the screen moving and the captions readable. Change the shot every few seconds and let captions carry the key words. Our list of retention techniques for faceless videos covers pacing in detail.
Step 5: End cleanly. Stop when the payoff lands, ideally on a line that makes the opening make sense again, which invites a replay.
Step 6: Read the right numbers two or three days after posting: shown in feed, stayed to watch, average view duration and average percentage viewed. Compare them with your channel's typical Short, not with viral outliers.
Step 7: Iterate on openings. When a Short stalls, keep the topic and test a new first line. When one takes off, make three more on the same angle while the topic is warm.
How Does Vidpal Studio Map to Each Signal?
Vidpal Studio is a faceless video creator: you give it a topic, a full script or a URL, and it writes the script, voices it, picks visuals and captions, and renders a 9:16 Short. Each setting lines up with one of the signals above.
For stayed to watch, the Hook Intro option adds a Viral Clip before the narration, chosen from a curated library or generated with Wan 2.2 from text (Text → Video) or from an image (Image → Video, a 5-second vertical clip), or a Text Reveal title. A hook clip costs one extra credit and is charged only when it generates successfully, so you can test a hooked and an unhooked version of the same idea.
For average view duration, scripts are calibrated to about 150 words per minute, any section can be rewritten with AI, and visuals come from stock footage, AI-generated images in 20 styles from cinematic to stickman, or a looping gameplay or oddly satisfying clip behind the narration. AI images are stills with camera motion rather than animation, so pacing comes from the script and the cuts. Captions come in 75 presets such as Mozi, Beasty and Karaoke, with about 18 animations including Word Punch and Karaoke Pop and a Lines control that shows one word or three at a time.
For percentage viewed, you pick the duration up front: 30, 45, 60, 75 or 90 seconds. Brand Kit keeps your logo, colors, caption preset and font identical across Shorts, so viewers who liked one recognize the next. On Pro and Business, each finished video's Publish panel posts to YouTube, Instagram, TikTok and LinkedIn immediately or on a schedule at no credit cost, and the optional Social Scheduler can write per-platform copy for 1 credit.
What Studio cannot do is pick a winning topic for you or guarantee distribution; no tool can. It is not an AI avatar tool, and narration captions are English. A Short costs 1 credit, and the Free plan's 4 one-time credits let you try Studio but not export Studio videos, so testing hooks at volume starts on Starter at $19 a month for 15 credits or Pro at $39 a month for 40. See Vidpal Studio and pricing for the details.
Start free on Vidpal and build your next Short with the payoff in the opening second.
What Are the Limits of What Anyone Knows About the Algorithm?
YouTube publishes which signals it uses but not how much each one weighs, and it changes the system constantly. Everything above combines YouTube's statements with our own observations; where they conflict, trust YouTube's current Help Center over any blog, including this one.
Thresholds also vary by niche and format. A 30 percent stayed-to-watch rate might be strong for a dense explainer and weak for a meme format. And the feed finds whoever responds: on one of our test channels, the Shorts feed delivered a mostly teenage audience, the wrong one for a product aimed at creators, even while views were healthy. Good signals and the right audience are separate goals, and only your topic and framing control the second.
Finally, the system optimizes for viewer satisfaction, so tactics that lift one metric at the cost of satisfaction, such as bait openings that never pay off, tend to come back as survey answers and “not interested” clicks. For the AI-specific side of this question, read our post on whether YouTube suppresses AI-generated videos.
Frequently Asked Questions
How does the YouTube Shorts algorithm work? It shows each Short to viewers whose history suggests they will enjoy it, measures whether they watch or swipe, how long they watch and whether they liked it, then widens the audience while those signals beat competing Shorts. YouTube says ranking combines a Short's performance with each viewer's personal history of topics, channels and sounds.
Why are my YouTube Shorts getting 0 views? Usually the Short has not been served in the feed yet, because views now count the moment a Short starts to play. Check Shown in feed in its analytics, confirm it is public, vertical or square and under three minutes, and look for copyright claims or age restrictions. Low topic demand and stronger competing Shorts also limit impressions.
Do hashtags help YouTube Shorts get views? Only a little. YouTube treats hashtags as a way to connect videos on the same topic, shows up to three above the title, and ignores all of them if a video has more than 60. The Shorts team has said they are optional and most useful when tied to a real-world event or a specific topic.
Does the time you post a YouTube Short matter? Not much. YouTube's Shorts product lead has said that, outside breaking news, what you post matters more than when. The Shorts feed favors recent content, so most views arrive in the first days, but there is no fixed best window. Choose a consistent time you can keep and spend the effort on the opening instead.
Is a Short dead if it gets no views in the first 48 hours? No. YouTube has never published a 48-hour rule. The Shorts feed weighs recency, so most Shorts get the bulk of their views early, but ranking is relative and personal, and a Short can get a second wave weeks later when its topic trends or it reaches better-matched viewers.
What is a good stayed-to-watch rate for YouTube Shorts? YouTube publishes no target. In our own small sample, Shorts that fell below about 25 percent stayed to watch stalled within days, while those around 30 percent or higher kept getting impressions. Compare each Short with your channel's typical results and treat any outside benchmark as a rough guide only.
Do Shorts hurt long-form videos on the same channel? No, according to YouTube. Its Shorts help page says Shorts performance doesn't negatively impact long-form recommendations, and its recommendations guidance says experimenting with formats won't confuse the algorithm. Each video is judged on its own, although Shorts viewers do not automatically go on to watch your long videos.
