Grow Creator Field Notes

Good AVD but bad swipe-away rate on Shorts? Fix it

Good AVD next to a bad swipe-away rate is not a contradiction. The two numbers score different gates. How to read the pair and fix the right one.

A good AVD next to a bad swipe-away rate is not a contradiction. The two numbers score different moments. Swipe-away is decided in the first frames, on your cover and your opening line. Your watch numbers only describe the people who stayed. So the number to fix is the packaging, not the edit.

That is the whole answer. The rest of this page is how to confirm it against your own Studio numbers, because a high swipe-away share is not always the binding problem, and there is one reading error that sends creators after the wrong fix entirely.

First, check which number you are reading

Two mix-ups show up constantly, and both change the diagnosis.

AVD is in seconds. It cannot be a percentage. Average view duration is a clock reading, something like 0:22. If the number you are looking at has a percent sign on it, you are reading average percentage viewed. YouTube defines it as the average share of a video watched among those who stayed to watch, and for Shorts it is calculated from engaged views and their watch time. Loops add watch time without adding another engaged view, which is why the figure can pass 100%. Creators post things like "my AVD is 120 to 140%" all the time. That is APV, and 120% on a 20-second Short means the average person who stayed watched about 24 seconds of it, which is a good result for a different gate than the one they are worried about.

Viewed and swiped away are two halves of one bar. YouTube introduced the Shorts split as viewed vs swiped away, and the documented metric underneath it is stayed to watch, the percentage of times viewers stayed to watch past the initial seconds of a Short. Some creators quote the viewed half, some quote the swiped half, and both call it "swipe rate". A creator saying "my swipe rate is 85%" usually means 85% stayed. A creator saying "my swipe-away rate is 80%" means 20% stayed. Before you diagnose anything, open the card and read the label on the side you are quoting.

The four gates, and which metric scores each one

A Short is not judged once. It passes through a sequence, and each metric only sees the people who cleared the gate before it. This is why the pair can look contradictory when it is not.

Gate 2 is decided before anyone has heard your point. Gates 3 and 4 are scored only on the survivors of gate 2. So a Short can be genuinely excellent and still be swiped past by most of the feed, and the analytics will show exactly that: a mediocre viewed share sitting next to an APV over 100%.

The denominators are the part you can rely on, because YouTube publishes them. Average percentage viewed is an average taken among the people who stayed to watch. The people who swiped were never counted in it. That alone means a strong showing at gates 3 and 4 tells you nothing about gate 2. On ranking, YouTube's search and discovery guidance says only that its systems use both average view duration and average percentage viewed as signals to inform ranking. Where the viewed-versus-swiped split sits in that stack is not something YouTube publishes, so treat it as a measurement of what viewers did rather than a lever you know the weight of.

What YouTube also does not publish is how many people the first test batch contains, or what viewed share triggers a wider push. Anyone quoting you a specific number for that is guessing. Treat the sizes as unknown and work the signals you can see.

Where your pair sits

We are not going to hand you a benchmark for viewed-versus-swiped. YouTube publishes no target for it, and the numbers floating around creator forums are individual screenshots, not a sample you can position yourself against. The comparison that carries real information is your own back catalogue, because it holds your niche, your face, your format, and your audience roughly constant.

Open Content, then the Shorts tab, and write down two figures for each of your last ten uploads: the viewed share and the average percentage viewed. Sort by the first column.

The pattern creators describe over and over when they do this, and the one you will probably see, is that viewed share behaves like a gate you clear rather than a dial you maximise. Below your own usual band it is the thing holding an upload back. Inside or above that band, pushing it higher stops tracking with views, and what separates the Shorts that travel is something else entirely: the subject, the audience they get served to, and how far outside your subscribers they go. This is why the pair can look scrambled. A Short with a middling viewed share can travel and a Short with an excellent one can stall, and neither outcome means a metric is broken.

Read your sorted list against these four shapes:

• Viewed share well below your own band, watch numbers normal or better. Packaging is the binding constraint on that upload. Nothing downstream is worth touching yet.
• Viewed share inside your band, APV near or over 100%. Both gates are behaving. If views are still flat, the question has moved to subject and audience match, not to the first three frames.
• Viewed share at the top of your range and views still flat. Look at who the Short reached rather than at the Short. Check the non-subscriber share in the Audience tab: a Short loved by a small, well-matched pocket and never served outside it produces exactly this shape.
• Every upload low, with no spread at all. That is a channel-level pattern rather than a video-level one. Packaging is still the first thing to test, because it is the only gate on the list you fully control.

Three readings

These are the three shapes people bring to this question. Find the one closest to your own card.

APV 120% to 140% (often reported as "AVD"), viewed share around half. The content works on everyone who stays, and half the feed never gives it that chance. Verdict: packaging. The fix is the cover frame and the opening clause, and nothing in the body of the video. Editing this Short tighter moves APV from great to slightly greater and changes nothing about views.

Viewed share around 20%, so 80% swiped away, APV just under 100%. The rare people who watch it watch nearly all of it, and four in five never start. Verdict: the same problem, much worse, and most often an opening frame that is either illegible at feed size or reads as an ad. Re-shoot the first two seconds and re-upload rather than trying to rescue the existing one.

Viewed share at the top of your range, APV around 100%, views flat. Verdict: not a packaging problem at all. Both gates are clear. When both numbers are strong and views stop, the usual explanations are a subject with a small ceiling or an audience pocket that is already saturated. That is a channel-level question, and reading the traffic and audience side of your Shorts analytics tells you more than any per-video metric will.

The fix, when it is packaging

The swipe decision happens on the first frames and the first sound. Five checks, in the order they matter:

• Legibility at thumbnail size. Shrink your first frame until it is about the size of a fingernail. If you cannot tell what it is, neither can a viewer mid-scroll. Busy backgrounds and thin fonts fail this test more than anything else. You can run a frame through the free reel cover analyzer to see how it holds up small before you commit to it.
• Contrast on the opening text. White text on a bright or moving background disappears. A solid stroke, a shadow, or a flat block behind the words fixes it.
• No logo, no intro, no title card. Any frame that is about you rather than about the viewer is a free swipe. The first frame should be the middle of something.
• A claim, not a greeting, in the first spoken clause. "Hey guys, so today I wanted to talk about" spends the entire gate on nothing. Lead with the thing you are going to prove.
• Cover matches the first frame. If the cover you picked and the actual opening frame are different images, the viewer gets a small jolt of mismatch at exactly the wrong moment.

If you want the per-video version of this split, Reel IQ scores hook and retention separately for a single upload, which is the same distinction this whole page is built on.

Two things that do not fix a swipe-away problem, despite being the most common responses to it: tightening the edit, which improves gate 3 when gate 3 is already fine, and rewriting hashtags and the description, neither of which changes the frame the viewer is deciding on. If your retention curve is genuinely the weak side, the Shorts retention benchmarks are a better starting point than this page.

The same split on Instagram Reels

Instagram does not give you a swipe-away metric, so the gate is invisible and you have to infer it. In Insights, the pair to look at is reach against average watch time. Wide reach with a short average watch time points at the hook. High average watch time relative to the Reel's length sitting next to low reach points the other way: it cleared the hook for the few who saw it and never got distributed, which is the Instagram version of the paradox on this page.

The practical difference is that Instagram lets you set the cover frame independently of the first frame of the video, which makes the mismatch problem above much easier to create by accident. Pick a cover that is the first frame, or something within the first half second of it. The Reels completion rate benchmarks cover the downstream half of the same split.

If you only do one thing after reading this: open your last ten Shorts, write down the viewed share and the average percentage viewed as a pair, and sort by the first column. Where your weak uploads sit against your own band tells you which of the two fixes is worth your next week.

Canonical: https://growcreator.pro/blog/good-avd-high-swipe-away-rate-shorts

Leumos Private Limited · hello@growcreator.pro