Viewed vs. swiped away: why viewers skip your Short in the first second (2026)

In the analytics of each Short, YouTube Studio shows a split called viewed vs. swiped away. It answers one narrow question: of the times the Short was shown in the Shorts feed, in what percentage did the viewer choose to stay and watch, and in what percentage did they swipe to the next video. When the line reads 40% viewed, 6 in 10 people who were served your Short moved on. When it reads 75%, only 1 in 4 did.

It is a feed metric and nothing else. Views that come from search, from your channel page, from a link in a description or from a playlist do not enter the calculation. Those viewers asked for your video, so their decision says little about the opening. The feed viewer did not ask for anything. They were served your Short between two other videos, and the metric records whether you earned the next few seconds of their attention.

That is why the first second carries the whole metric. In the Shorts feed there is no thumbnail to click and no title to read first. The Short starts playing the moment it lands on the screen. The opening frame is the thumbnail, the first line of audio or on screen text is the title, and the decision happens while the thumb is still resting on the glass, usually in under a second. Anything you put before the subject, a logo, a fade, a greeting, is spent from that single second.

Because of this, the metric deserves its own reading instead of being one line in a general checklist. The broader diagnosis of why a Short stalls, including the second number that tells you whether the middle held, is in why Shorts are not getting views. This guide stays on the first second, because when viewed is low nothing after it gets a fair chance.

• Base of the metric: only impressions in the Shorts feed.
• Out of the metric: search, channel page, links and playlists.
• In the feed, the opening frame works as the thumbnail and the first line works as the title.
• 40% viewed means 6 in 10 people served your Short skipped it.

YouTube does not publish a target for this metric. Every number you read online, including the ones in this guide, comes from creators comparing their own Studio screens and sharing what separates Shorts that expanded from Shorts that stalled. Treat them as ranges reported by the community, not as official thresholds. With that honesty in place, the ranges are consistent enough to be useful.

The most common target creators report is 70% to 80% viewed or higher. Below roughly 60%, the feed tends to stop testing the Short on new audiences, and some creators place that line closer to 50%. The Shorts that break out often sit between 80% and 90% viewed. Niche moves these numbers: a calm recipe or a satisfying process video can hold a different baseline than a fast opinion Short, so the same 65% can be weak in one niche and normal in another.

The math shows why a few points matter so much. Take 10,000 impressions in the feed. At 55% viewed you get 5,500 starts. At 75% viewed you get 7,500 starts. That is 2,000 more people, 36% more starts, from exactly the same impressions, before retention, loops or likes have any say. And since the feed tends to widen distribution for the Shorts that hold the test batch, the winner usually earns more impressions in the next round too, so the gap tends to grow instead of staying at 36%. Tends, not certain: nothing in distribution is promised.

The reverse is just as important. Average percentage viewed only counts the people who stayed. A Short with 95% average viewed on top of 45% viewed still lost more than half the room before it started. Looking only at retention makes that Short look excellent, when the feed saw a video most people rejected.

• Community target: 70% to 80% viewed or more.
• Below about 60%, sometimes cited as 50%, expansion tends to stop.
• Breakout Shorts often read 80% to 90%.
• 10,000 impressions: 5,500 starts at 55% against 7,500 at 75%, which is 36% more.

When viewed is low, the cause is almost always visible in the first frame or audible in the first half second. These are the eight patterns that show up again and again when creators take apart the Shorts that were swiped:

• A black or fade in first frame: the feed shows darkness for the exact moment the decision is made.
• A logo, an intro sting or a greeting before the subject: every frame of branding is a frame without a reason to stay.
• Text too small, or placed in the bottom 20% of the screen, where the title, the channel name and the buttons cover it.
• A first sentence without the subject and the promise: the viewer hears words but does not know what the video is about or why it pays to stay.
• Audio that starts after 0.5 to 1 second of silence, so the Short feels broken or slow next to the video before it.
• A horizontal frame recycled into 9:16 with bars: a 16:9 image occupies 607 pixels of the 1,920 vertical pixels, 31.6% of the height, and the rest is padding.
• A topic that does not match the audience the feed is testing, so even a good opening is shown to people who never cared.
• The same template posted 30 times: identical openings teach returning viewers to swipe before the first word.

Reverse those eight mistakes and you get the shape of an opening that holds. Frame zero is already the most interesting image of the Short, with the subject on screen and something in motion, never a fade from black. Audio starts at 0.0 seconds, not after a breath. The first words carry the subject and the promise together, as in a sentence that names what the viewer will learn and hints at the payoff, so the brain has a reason to wait for the second sentence before the thumb moves.

Text belongs in the middle band of the screen, large enough to read at arm length on a phone, and out of the bottom 20% where the interface sits. Captions that appear word by word keep the eye on the video while the audio catches up, which is important because a large share of feed viewers watch with the sound low or off. Which caption style holds the thumb best, and where each style breaks, is covered in the guide to Shorts subtitle styles.

Contrast with the previous video matters too. The viewer arrives from something loud or bright, and an opening that looks like the last three Shorts they skipped gets skipped by reflex. A different color, a close object, a number on screen or a question in the first line breaks that reflex without any trick.

In FalconVid this is how the Short is built from the start. The script writer writes a hook for the first second before writing the rest, the narration begins on the first frame, the opening scene is chosen to match the first sentence, and the karaoke style captions, in more than 15 styles, are placed out of the interface zone. The Short is born in 9:16, so there is no letterbox to fight and no 607 pixel window to fit the subject into.

The first reading lies. In the first hours the Short is shown to small batches, and a few hundred impressions swing the percentage by ten points in either direction. A practical rule many creators use is to read viewed vs. swiped away only after about 48 hours and at least 1,000 impressions in the feed. Before that you are reading noise and reacting to it.

Then compare the Short against your own channel, not against a number from a forum. Take the median viewed of your last 10 to 20 Shorts and ask whether the new one sits above or below it. A channel whose median is 58% and whose new Short reads 66% has found something worth repeating, even if 66% sounds weak next to the 80% someone posted online. A channel whose median is 78% and whose new Short reads 66% has a problem in that opening.

Viewed never works alone. Read it with average percentage viewed and, for Shorts designed to loop, with whether that average passes 100%. Length plays into the second number, not the first: a longer Short does not change the first second, but it changes how much of the video people finish, and the ideal length of a Short in seconds walks through that trade. The four combinations below tell you which part of the Short to fix.

• Low viewed, high average viewed: the Short is good and the opening hides it. Rebuild the first second only.
• High viewed, low average viewed: the opening works and the middle or the length fails. Keep the hook, cut or restructure.
• Both low: the topic or the audience is wrong. A new opening will not save it.
• Both high: repeat the idea with a different opening so the template does not wear out.

The manual fix for a low viewed rate is clear, and it is slow: rewrite the first line, find a stronger opening image, move the text, re-record the voice so it starts on frame one, export, upload, wait 48 hours, compare. FalconVid, the platform that runs channels on autopilot, turns that loop into a production line where the first second is a setting of the channel, not a manual edit on every Short.

For each Short the AI specialists work in parallel. The script writer starts from the hook for the first second and then writes the rest in the length you set. The narrator reads in one of the premium voices or in your cloned voice, in 63 languages, starting on the first frame. The editor uses generated scenes or a real media bank with movement, sound design adds music and effects, and the SEO is written with the Short. The AI Shorts generator produces native 9:16 Shorts and Reels automatically, with the captions out of the interface zone, and schedules them on YouTube, Instagram, TikTok, Rumble and Facebook.

The Studio is where the first second gets adjusted without starting over. You watch V1, and if the opening is slow you shorten the intro, swap the opening media or change the music, or open the timeline for a finer edit. Testing a new first second does not mean regenerating the whole Short. The channel DNA keeps the voice, the caption style, the palette and the structure the same from one Short to the next, so what changes between two versions is exactly the opening you are testing.

Choosing what to open with is also not guesswork. The Spy, the niche modeler, shows which Shorts are already working in your niche and how they open, so you do not start from zero, you start from what already monetizes. On the Pro plan and up, the Senior AI Analyst, a fixed person with a name, a face and a voice, reads your account and writes to you every 2 days with the next move inside the product, such as which Shorts sit below your median viewed and which opening to test next.

The most useful test for a low viewed rate is narrow: keep the idea, the middle and the ending, and change only the opening. Two or three versions of the same Short with different first seconds, published on different days to comparable audiences, tell you more than ten unrelated Shorts ever will. One version opens on the result, one on a question, one on a number on screen. Whatever wins becomes the opening pattern for the next batch.

By hand, this is where channels stop. Each version means rewriting the hook, finding new opening footage, re-recording or re-timing the voice and exporting again, and even a fast editor spends a meaningful slice of an hour per version. Three openings for each of eight ideas is 24 Shorts a month, which is most of a part time job before a single one has been read. So most creators test nothing, keep the opening they like and guess.

On the line, the ceiling moves. A Short of up to 1 minute costs around 580 credits in economy mode, and the exact credit cost appears on screen before you generate. With that, the Starter's 15,000 credits cover about 25 Shorts a month in economy, enough for 8 ideas with 3 openings each. The Pro's 30,000 credits cover about 51, which is 17 ideas tested three ways. The Shorts are generated side by side, 2 at a time on the Starter, 5 on the Pro and up to 50 on the Scale, and you approve the calendar so each version goes out on its own day.

What stays yours is the reading: which opening won against your median, and whether the win repeats on the next idea. That part should never be automated away, because it is where the channel learns. What disappears is the month of editing between the question and the answer.

• One variable per test: the first second. Idea, middle and ending stay the same.
• Three opening types to start: the result first, a question, a number on screen.
• Around 580 credits per Short of up to 1 minute in economy.
• Starter: about 25 Shorts a month. Pro: about 51.

Since 31 March 2025, the public view count of a Short counts every start and every replay. That makes the counter bigger, but it is not the number that pays. Monetization and YouTube Partner Program eligibility use engaged views, the stricter count. A feed viewer who swipes away is not the audience that builds either one, so the first second is still where the qualified numbers are won or lost.

The bar is rising. For new entrants to the Partner Program through Shorts, from 1 February 2027 the rule becomes 1,000 subscribers plus 20 million qualified Shorts views in 90 days, up from 10 million. That is about 222,000 qualified views a day for three months. At that scale the 36% more starts from the same impressions in the example above is not a detail, it is the difference between reaching the line and spending another quarter below it.

Longer Shorts do not buy time for a weak opening. Shorts can run up to 3 minutes since 15 October 2024, but the first second is still one second, and the length only changes how much of the video the viewers who stayed will finish. There is also a trap tied to openings that lean on a trending song: a Short over 1 minute with an active Content ID claim is blocked worldwide, so a borrowed track that holds the thumb can take the whole video off the feed.

Shorts also feed the rest of the channel. A Short that holds the feed introduces the channel to people who never searched for it, and how that traffic moves into long videos and watch hours is covered in the comparison of Shorts and long videos. Either way, the path starts with the same decision: whether the first frame earned the second one.