YouTube Shorts data · a Virlo study

How the YouTube Shorts algorithm works

Updated

The YouTube Shorts algorithm shows each Short to viewers in the Shorts feed and keeps showing it while people choose to watch rather than swipe away. Subscribers are not the main audience: in Virlo’s scan of 10,827 Shorts posted October 13, 2026 to October 31, 2026, the median Short got 4.94× its channel’s subscriber count in views. The format and the length moved views far more than the posting time (7% between the best and worst time).

  • What YouTube itself says, quoted and dated
  • Why subscribers matter less on Shorts
  • The formats, hooks and lengths that got the most views
  • What a brand should change in its next Short
YouTube Shorts
Views per subscriber
4.94×
median Short: most views come from non-subscribers
Top format
Clip compilation
1.88× the views of similar Shorts
Under 15 seconds
1.08×
vs 0.96× for 30-60 seconds
Posting time
7%
best vs worst 3-hour window

As of November 20, 2026 · YouTube Shorts posted October 13, 2026 to October 31, 2026

What does YouTube say about how Shorts are recommended?

YouTube publishes the signals in general terms. These are its own words, with the source and date, so you can check them.

YouTube’s own statements about Shorts recommendations
What YouTube saysSourceWhat it means for a brand
“The Shorts Feed is personalized to reflect what we think you want to see next.”YouTube Help: How YouTube recommendations workEach viewer gets their own feed; there is no single chart to get onto
Recommendations use watch history, search history, subscriptions, likes, dislikes, “Not interested”, “Don’t recommend channel” and satisfaction surveysYouTube HelpSatisfaction counts, not only watch time: clickbait that disappoints can count against you
People swipe through Shorts without knowing what comes next, so performance rests on viewers choosing to watch, not skip (Todd Sherman, Shorts product lead)TechCrunch, Aug 25, 2023The first second decides it: show the product or the payoff at once
Time of day is not a factor to optimise for, except for newsTechCrunch, Aug 25, 2023Post when your team can answer comments
Since March 31, 2025 a Shorts view counts every start or replay; the old metric is “engaged views”YouTube Help: Get started creating ShortsJudge a Short on engaged views and on what viewers do next, not the headline count
Shorts can be up to 3 minutes longYouTube HelpRoom for a full product story when it earns the time

Do subscribers decide who sees a YouTube Short?

No. In Virlo’s scan the median Short got 4.94× its channel’s subscriber count in views, and the smallest channels made up most of the sample. Most views come from people who don’t subscribe, which is why a brand with no Shorts audience can still reach people.

YouTube Shorts by channel size: share of the sample and median views
SubscribersShare of ShortsMedian views75th percentile viewsShorts
Under 1K52%1,3121,7875,631
1K-10K20.5%5,25625,1232,217
10K-100K17%16,81962,6981,840
100K-1M8.3%20,374116,573902
1M+2.2%49,931281,443237

Which Shorts formats does the algorithm reward most?

Clip compilation came out on top at 1.88× the views of similar Shorts; event footage came last at 0.87×. Talking-head Shorts, the most common format in the sample, got 0.95×.

YouTube Shorts views vs similar Shorts by visual format (formats with 100+ Shorts)
FormatShare of ShortsViews vs similar ShortsMedian viewsShorts
Clip compilation6.3%1.88×8,952328
Green screen commentary3.7%1.08×3,038192
Activity demonstration13.7%1.07×2,210710
Native gameplay2.5%1.01×1,461128
Slideshow text3.8%1.00×1,568198
Vlog handheld8%1.00×1,975416
Animation motion graphics15.2%0.99×1,374789
Dance full body6.1%0.99×5,894315
Product closeup3%0.97×1,466156
Talking head19.8%0.95×1,4441,027
Screen recording2.2%0.94×1,357112
Outfit showcase2.8%0.93×1,398143
Event footage4%0.87×3,555205

Format is tagged by Virlo’s AI after watching each Short. “Clip compilation” is a run of short clips cut together, the format clipping channels and highlight accounts use.

Which opening hooks work best on YouTube Shorts?

Shock statement openings did best (1.05×) and statistic openings worst (0.92×). The spread is smaller than for format: the hook type matters less than whether the first second delivers it.

YouTube Shorts views vs similar Shorts by spoken hook type (hook types with 100+ Shorts)
Hook typeShare of ShortsViews vs similar ShortsMedian viewsShorts
Shock statement8%1.05×2,049422
Pov setup2.8%1.04×2,042148
Relatable scenario6.2%1.02×1,702325
Story tease6.2%1.02×2,288328
Negation3.4%1.01×1,731178
Comparison2.5%0.99×1,965133
Question16.5%0.98×1,512874
Bold claim15.8%0.97×1,636834
Direct address2.8%0.97×1,824148
Tutorial promise11.7%0.96×1,568620
Statistic2.1%0.92×1,426111

See the hooks winning in your niche

This is every niche at once. In Virlo, each top video in your niche shows its exact hook, format and length. 7-day free trial.

Start for $0 to see all

Do faces, length and posting time change Shorts reach?

Length did: under 15 seconds got 1.08× against 0.96× for 30-60 seconds. A face on screen did not (0.99× with a face, 1.01× without), and posting time moved views by 7% at most.

What else goes with Shorts views in Virlo’s scan
FactorOptionViews vs similar ShortsShorts
Lengthunder 15s1.08×2,869
Length15-30s1.00×3,262
Length30-60s0.96×2,797
Length60-90s0.95×1,082
Length90s-3 min1.04×817
FaceFace on screen0.99×5,947
FaceNo face1.01×4,387
Posting timebest 3-hour window (12-15 UTC)1.05×1,354
Posting timeworst 3-hour window (21-24 UTC)0.98×1,181

What does this look like in real Shorts?

The most-viewed real Shorts for the top hook types in the sample. Use them as a reference for the first second of your next brand Short, counts as of November 20, 2026.

Hook types, best first
Hook typeViews vs similar ShortsShorts
Shock statement1.05×422
Pov setup1.04×148
Relatable scenario1.02×325
Story tease1.02×328
Negation1.01×178
Comparison0.99×133
Question0.98×874
Bold claim0.97×834

What does a shock statement hook look like on Shorts?

Shock statement openings got 1.05× the views of similar Shorts across 422 sampled Shorts. The top real examples follow, each with its opening line.

Views vs similar
1.05×
same subscriber band
Share of Shorts
8%
in the sample
Median views
2,049
Shock statement
Shorts compared
422
posted October 13, 2026 to October 31, 2026

Top shock statement Shorts · counts as of November 20, 2026

What does a pov setup hook look like on Shorts?

Pov setup openings got 1.04× the views of similar Shorts across 148 sampled Shorts. The top real examples follow, each with its opening line.

Views vs similar
1.04×
same subscriber band
Share of Shorts
2.8%
in the sample
Median views
2,042
Pov setup
Shorts compared
148
posted October 13, 2026 to October 31, 2026

Top pov setup Shorts · counts as of November 20, 2026

What does a relatable scenario hook look like on Shorts?

Relatable scenario openings got 1.02× the views of similar Shorts across 325 sampled Shorts. The top real examples follow, each with its opening line.

Views vs similar
1.02×
same subscriber band
Share of Shorts
6.2%
in the sample
Median views
1,702
Relatable scenario
Shorts compared
325
posted October 13, 2026 to October 31, 2026

Top relatable scenario Shorts · counts as of November 20, 2026

What does a story tease hook look like on Shorts?

Story tease openings got 1.02× the views of similar Shorts across 328 sampled Shorts. The top real examples follow, each with its opening line.

Views vs similar
1.02×
same subscriber band
Share of Shorts
6.2%
in the sample
Median views
2,288
Story tease
Shorts compared
328
posted October 13, 2026 to October 31, 2026

Top story tease Shorts · counts as of November 20, 2026

Hook type and format are tagged by Virlo’s AI after watching each Short. Example Shorts are the most-viewed English, brand-safe Shorts with that hook type, one per creator.

See the winning hooks in your niche

A marketing agent watches your niche and shows the hook, length and format of every breakout video. 7-day free trial.

Start for $0 to see all

To see the hooks and formats winning in your own niche, start a free 7-day Virlo trial at app.virlo.ai.

How was this YouTube Shorts algorithm study done?

YouTube does not publish its ranking model, so this page pairs what YouTube says with what Virlo measured. Go deeper with the Shorts length study and the YouTube Shorts vs TikTok comparison.

How this data was made

Source
Virlo’s database of social videos (23.1M+ TikTok, Instagram and YouTube Shorts videos), read through the Virlo public API. YouTube’s statements are quoted from YouTube Help and TechCrunch (linked in the table).
Window and sample
10,827 YouTube Shorts (3 minutes or shorter) posted October 13, 2026 to October 31, 2026: for each of the 445 hours sampled in that window, the most recent Shorts Virlo’s library holds for that hour (UTC), so every posting hour is equally represented.
Method
“Views vs similar Shorts” = views ÷ the median views of sampled Shorts from channels in the same subscriber band (under 1K, 1K-10K, 10K-100K, 100K-1M, 1M+); tables show the median of that ratio. Views per subscriber = views ÷ the channel’s subscriber count. Views are the latest count Virlo holds for each Short, read on or before November 20, 2026. Limitations: Virlo collects the Shorts its trend scans and marketing agents find, which leans toward Shorts already getting views; the study shows what goes with more views, not proof of cause.
As of
November 20, 2026. Re-run every month.
How this page is made
Numbers are computed automatically from Virlo’s data; the text is written and checked by the Virlo team.

Source: Virlo, How the YouTube Shorts algorithm works, November 2026, virlo.ai/data/youtube-shorts-algorithm

Want the full list?

To see the formats and hooks winning in your own niche this week, start a free 7-day Virlo trial at app.virlo.ai.

Frequently asked questions

How does the YouTube Shorts algorithm work?

The YouTube Shorts algorithm shows each Short to viewers in the Shorts feed and keeps showing it while people choose to watch rather than swipe away. Subscribers are not the main audience: in Virlo’s scan of 10,827 Shorts posted October 13, 2026 to October 31, 2026, the median Short got 4.94× its channel’s subscriber count in views. The format and the length moved views far more than the posting time (7% between the best and worst time).

What does YouTube say about the Shorts algorithm?

YouTube’s Help Center says “The Shorts Feed is personalized to reflect what we think you want to see next.” Its recommendation signals include watch history, search history, subscriptions, likes, dislikes, “Not interested” feedback and satisfaction surveys. In 2023, Shorts product lead Todd Sherman explained that on Shorts people swipe without knowing what comes next, so a Short’s reach depends on viewers choosing to watch it rather than skip it (TechCrunch, August 25, 2023).

Do subscribers matter on YouTube Shorts?

Less than on long videos. The median Short in Virlo’s sample got 4.94× its channel’s subscriber count in views, so most views came from people who did not subscribe. Channels under 1,000 subscribers made up 52% of the sample, with a median of 1,312 views per Short.

Does posting time affect the Shorts algorithm?

Barely. The best and worst 3-hour posting windows in Virlo’s data were 7% apart, and Todd Sherman said time of day isn’t something to optimise for, except for news. See the best time to post YouTube Shorts for the full table.

How are YouTube Shorts views counted?

Since March 31, 2025, a Shorts view counts each time a Short starts to play or replay, with no minimum watch time (YouTube Help). The older metric is now called “engaged views” in YouTube Analytics, and Partner Program eligibility and Shorts revenue sharing still use engaged views.

What kind of Shorts does the algorithm push most?

In Virlo’s scan, clip compilation Shorts got the most views compared with similar Shorts (1.88×), while talking-head Shorts got 0.95×. Under-15-second Shorts got 1.08× against 0.96× for 30-60 seconds.

Why did my YouTube Short stop getting views?

A Short gets shown to new viewers in rounds; when the people it reaches start swiping away, the feed shows it less. Todd Sherman put it simply: a Short “may get a lot more traffic or it may taper off.” A new Short with a sharper first second is usually a better bet than waiting.

Do YouTube Shorts hurt a brand’s long-form videos?

YouTube says it doesn’t think of Shorts as cannibalising long-form (TechCrunch, 2023). For a brand, Shorts are a reach channel: most views come from people who have never seen your channel.

From the signal newsletter

Every issue is in the signal archive.

See what the Shorts feed is rewarding in your niche this week

Virlo ranks every video in your niche by Virality Score and shows the hook, length and format behind the ones that took off.

7-day free trial. Cancel any time.