YouTube gives you at least three numbers worth tracking: audience retention, watch time, and (for Shorts) viewed vs. swiped away (VVSA). Retention is how much of a video people watch. Watch time is total minutes across your channel. VVSA is whether Shorts viewers stop scrolling or swipe away. Fix the metric that matches your format.
Retention, Watch Time, and VVSA
What is audience retention?
Audience retention is the percentage of your video that viewers watch on average. YouTube divides average view duration by total video length, so a 10-minute video watched for 5 minutes on average shows 50% retention. The retention graph shows where viewers leave moment to moment, which makes it the best diagnostic for hooks, pacing, and drop-off points on long-form uploads.
What is watch time?
Watch time is the total number of minutes viewers spend watching your content. Unlike retention, it is an absolute number that adds up across every video on your channel. YouTube uses it to measure channel authority, session depth, and monetization eligibility (4,000 hours in the past 12 months for the Partner Program).
What is VVSA?
VVSA (viewed vs. swiped away) is a Shorts-only metric. It shows the percentage of people who chose to watch your Short after it appeared in their feed, rather than swiping past immediately. Think of it as the Shorts equivalent of click-through rate: it measures whether your opening stopped the scroll before retention even starts.
| Metric | Audience Retention | Watch Time | VVSA (Shorts) |
|---|---|---|---|
| Measures | Avg view duration as % of video length | Total minutes watched (absolute) | % who watched instead of swiping past |
| Where | Analytics > Engagement > Audience retention | Analytics > Content > Watch time | Per-Short analytics in YouTube Studio |
| Best for | Where viewers drop off in one long-form video | Channel growth, session time, YPP (4,000 hrs / 12 mo) | Whether the opening stops the scroll |
| Algorithm | Stronger signal for recommending this video | Stronger signal for channel authority | Gate before retention matters on Shorts |
| Edit lever | Cut dead air, remove tangents, visual resets | More videos, longer videos (if retention holds) | First second hits like a thumbnail |
For long-form, retention is the lever and watch time is the outcome. For Shorts, VVSA comes first: if viewers swipe before second one, retention never gets a chance.
How to read the retention graph
Three patterns map to fixes on long-form:
- Steep drop at 0:00-0:30: hook failed. The opening did not match the title/thumbnail promise.
- Steady slide from start to finish: pacing is too slow. Cut dead air, reorder sections, add visual resets.
- Flat middle, drop at the end: payoff or CTA arrived too late. Move the key result earlier, or shorten the outro.
Compare each upload to your last 3 videos in the same length range. A 45% average on 8-minute videos tells you more than comparing an 8-minute video to a 90-second Short.
VVSA benchmarks for Shorts
In a 2023 study of roughly 5,400 Shorts across 33 channels (~3.3 billion views), analyst Paddy Galloway found a clear floor: Shorts under 60% VVSA almost never performed well. Strong uploads usually landed in the 70-90% range. High VVSA did not guarantee viral reach (similar to CTR on long-form), but dropping below 60% was a reliable warning that the opening failed.
Practical rule: treat the first second of a Short like a thumbnail. Punchy motion, clear subject, immediate payoff. Pure likes, comments, and shares did not show a strong correlation with reach in that sample. The feed still rewards attention held, especially on longer Shorts with solid average view duration.
Before You Publish
Run this once against your analytics and timeline:
- Open Audience retention in YouTube Studio (Analytics > Engagement) for long-form
- Compare to a past video in the same length band: relative drop matters more than one number
- First 30 seconds: sharp cliff means weak hook. Fix the opening before polishing the middle
- Mid-video cliff: cut the tangent, or add a pattern interrupt or visual reset
- End retention: viewers stay for the payoff, or leave before your CTA?
- Title promise lands in the first 15-30 seconds
- Dead air, filler, and off-topic segments are cut (trim in browser if you need a quick pass)
- Something changes on screen or in pacing roughly every 30 seconds
- Re-watch at 1.25x speed. Still drags? Tighten again
- Shorts only: check VVSA on the last upload. Under 60%? Re-cut the first second before the next post
When Watch Time Comes First
- Near YPP threshold: publish more videos. Volume helps; filler does not.
- Series format: each episode needs solid retention so viewers click the next one.
- Shorts from long-form: extends reach. Does not replace long-form watch-time hours.
Mistakes that tank all three metrics
- Chasing retention on a 2-minute video: 90% retention still yields under 2 minutes of watch time per viewer. Length matters for total hours.
- Padding length for watch time: a 15-minute video full of filler drops retention and hurts recommendations.
- Fixing the middle when the hook fails: no amount of B-roll rescues a weak first 30 seconds on long-form, or a weak first second on Shorts.
- Ignoring VVSA on Shorts: sub-60% VVSA means viewers never gave the video a chance. Retention edits cannot fix a swipe-away opening.
FAQ
Is 50% retention good?
Yes for most niches. Ballpark average is 30-50%. A 2-minute video at 50% is not the same as a 20-minute video at 50%.
What is a good VVSA score?
Aim for 70%+ on Shorts. Under 60% is a red flag that the opening failed to stop the scroll.
Does watch time still matter for monetization?
Yes. You need 4,000 hours in 12 months. Retention gets you there faster; total hours are what count.
Can I fix retention after publishing?
No. The curve is fixed. Use it to edit the next upload.
Which metric do I check first?
Shorts: VVSA (did the first second stop the scroll?). Long-form: retention (where did viewers leave?). Channel level: watch time (growth, session depth, YPP).
Try GrepCut
Ready to run the checklist on your next upload? Open GrepCut Studio and tighten pacing in the browser. New here? See grepcut.com for how on-device editing works.
GrepCut
Author of this article