One upload does five times your channel's usual views. Does that change anything for the videos after it?
I scored 63,375 uploads from 472 sampled channels with the free checker's formula. 466 of them had judged videos (the method explains), and 27,985 uploads had ten earlier ones to compare against.
Quick answer
- After 24% of outliers, the channel's next five uploads ran 3x or more its old normal, against 5% after an ordinary video. 53% went back to normal (80% after an ordinary video).
- About 1 in 12 long videos is an outlier: 8.3% of 16,331 long videos reached 5x their channel's normal (95% range 7.6% to 9.1%). 4.1% reached 10x, a very big outlier.
- Shorts break out slightly less often: 7.1% of 10,386 Shorts reached 5x.
- Smaller channels see more outliers: 11.2% of long videos hit 5x under 10,000 subscribers, against 4.9% over 1 million.
- Title patterns are modest: a number, a list or a tutorial made a 5x video 1.28x to 1.38x as likely within the same channel.
Check your own video: paste the link in the free outlier checker. It shows the maths and takes a second.
- How Often YouTube Videos Become Outliers (About 1 In 12)
- What Happens After A Video Goes Viral On YouTube (Against A Placebo)
- Small YouTube Channels Get More Outliers (Partly Through Arithmetic)
- Outlier Rates By Niche
- What YouTube Outlier Titles Have In Common (Within The Same Channel)
- How This YouTube Outlier Study Was Run (And Where It Can Be Wrong)
- How To Measure YouTube Outliers On Your Own Channel
- Conclusion (What I Would Do)
- FAQ: YouTube Outliers
How Often YouTube Videos Become Outliers (About 1 In 12)
About 1 in 12 long YouTube videos is an outlier: 8.3% of 16,331 long videos from 388 channels beat their channel's median (the sorted middle) by 5x or more. Only channels with 20 judged videos count.
| Format | 5x or more | 10x or more | Channels judged |
|---|---|---|---|
| Long videos | 8.3% (7.6% to 9.1%) | 4.1% (3.6% to 4.7%) | 388 |
| Shorts (3 min or under) | 7.1% (6.3% to 7.8%) | 3.3% (2.8% to 3.8%) | 252 |
The brackets are the 95% range: the true rate almost surely sits between them.
The 5x Line And How A Video Gets Scored
An outlier score is a video's views divided by its baseline, the median of ten same-format uploads before it.
Harry Uppal's outlier video had 329,431 views on 25 September 2026 against a 64,793 median: 329,431 / 64,793 = 5.1x, just over the 5x line our explainer uses.
Most Outliers Land Just Over The Line
Of the 8.3% that clear 5x, most sit close to the line, not deep in the tail.
- 4.2% score 5x to 10x, the typical outlier.
- 2.0% score 10x to 20x.
- 2.1% score 20x or more, the extreme tail everyone quotes.
Most Channels Get At Least One Outlier
So per channel, outliers are common: 73.1% of 338 channels with 30 judged videos had at least one 5x video in their newest 30.
OverseerOS reports 92.0% of 625 channels had one, counting a year of uploads against our newest 30 videos, so the numbers don't match.
What Happens After A Video Goes Viral On YouTube (Against A Placebo)
The after-outlier result is the finding I think matters most. Just over half of outliers don't create a new normal: after 53.4% of them, a channel's next five uploads were back under 1.5x its old, 20-video normal.
We tested this against a placebo, the same measure after an ordinary video.
- Back to normal (under 1.5x): 53.4% after an outlier, 80.5% after an ordinary video.
- Lifted (1.5x to 3x): 22.2% after an outlier, 14.6% after an ordinary video.
- Moved the channel (3x or more): 24.4% after an outlier, 5.0% after an ordinary video.
The 24% is the number to watch, since ordinary videos rarely manage it. 1of10's small-versus-big study does not look at what comes after a hit.
The Placebo Test In Plain Words
A placebo is the dummy pill a trial gives patients, to see what happens anyway.
Our dummy is an ordinary video, one scoring 0.8x to 1.25x. Channels drift up on their own, so we ran the same test after 3,649 ordinary videos and 1,122 outliers. Only the gap belongs to the outlier.
Why The Normal Uses 20 Videos, Not 10
Outliers often land after a quiet patch. Before an outlier, the ten-video normal sat at 88.3% of the ten videos before that; before an ordinary video, 95.6%.
That dip means any recovery looks like a lift, so we use 20.
One Channel That Moved And One That Went Back
Two study channels, as numbers only:
| Channel | Score | 20-video normal | Next-5 median | After |
|---|---|---|---|---|
| Dine Joy | 12.5x | 2,180 | 8,053 | 3.69x, moved |
| Josie Woah | 14.6x | 3,492 | 3,377 | 0.97x, back to normal |
Worth testing after a hit: OverseerOS found same-topic follow-ups became another 3x hit 37.8% of the time, against 15.9% for other follow-ups, with no ordinary-video comparison.
Small YouTube Channels Get More Outliers (Partly Through Arithmetic)
Small channels post more outliers than big ones, and I think part of that is plain arithmetic.
| Subscribers | 5x rate | Channels |
|---|---|---|
| Under 10,000 | 11.2% (9.4% to 13.0%) | 85 |
| 10,000 to 100,000 | 9.9% (8.4% to 11.5%) | 104 |
| 100,000 to 1,000,000 | 7.7% (6.3% to 9.1%) | 101 |
| Over 1,000,000 | 4.9% (3.7% to 6.2%) | 98 |
Why Small Channels Swing Further
To hit 5x, a 300-view normal needs 1,200 extra views; a 300,000-view one needs 1.2 million. Same multiple, far fewer people, and small counts jump more.
1of10 reported the same direction on 17 September 2026: 68% of 1 to 5 minute videos from 1K to 10K-subscriber channels reached 2x their normal, against 17% over 1M. Their bar is 2x, not our 5x, so only the direction matches.
Tiny channels don't drive the number: drop the 566 videos with under-100-view normals and the 5x rate only falls from 8.3% to 8.1%.
The Tail Behind The Gradient
These two show how far the tail runs; just over half of all outliers score 5x to 10x.
999 Things To Cook, 5,750 subscribers: a 757-view normal, 656,375 views on 25 Sept, 867x.
Pap, 22,800 subscribers: a 3,495-view normal, 1,530,122 views on 25 Sept, 438x.
Outlier Rates By Niche
Also, science had the highest rate (10.3%), fitness lowest (6.2%); the rest are too close to rank.
What YouTube Outlier Titles Have In Common (Within The Same Channel)
Tutorial, number and list titles raised the odds of a 5x outlier; story and gameplay titles lowered them. I'd test them, not trust them.
We read 16,849 titles with an AI classifier, each compared only against the same channel's others.
Five Title Patterns That Held Up
"Held up" means it showed in small and big channels and passed the usual 95% fluke test. So 1.38x means a channel's tutorials hit 5x 1.38 times as often as its other videos.
- A tutorial title: 1.38x as likely.
- Has a number: 1.28x.
- A list-style title: 1.28x.
- A personal-story title: 0.74x, less likely.
- Gameplay, vlog or highlights: 0.68x, less likely.
With 16 patterns tested, about one could pass by luck. Also, most list titles carry a number, so those two overlap.
Title Tricks With No Reliable Signal
- Holding back the answer: 1.01x.
- Promising something extreme: 1.00x.
- A question mark (0.93x), capitals (0.89x) or brackets (1.13x).
- A longer-than-usual title (0.96x) or "I", "my", "we" (1.16x).
MrBeast's production guide estimates an extreme idea is easily 100 times more viral. Our test reads only titles: within a channel, titles promising something extreme scored 1.00x.
Why Pooled Numbers Mislead
Two patterns flip once you compare within one channel:
| Title | Pooled across channels | Within one channel |
|---|---|---|
| First person ("I", "my", "we") | 0.83x | 1.16x |
| Explainer | 1.17x | 0.87x |
So a "titles that work" list built across channels describes those channels, not what to change on yours.
A title's other half is the thumbnail: see our thumbnail study.
How This YouTube Outlier Study Was Run (And Where It Can Be Wrong)
The study sampled 472 channels from 60 YouTube searches in six niches.
- 60 searches found 2,125 channels; 1,939 eligible (a subscriber count, 40+ uploads).
- 472 were sampled across six niches and four size bands.
- 63,375 uploads were scored; 27,985 judged, with a full ten-video baseline and 14 days' age.
- A channel needed 20 judged videos in a format to count: 388 for long videos, 252 for Shorts.
The Same Formula As The Free Tools
Our outlier checker and channel analyzer run this same formula.
The full method, videos.csv and channels.csv reproduce every number here, including the after-effect and placebo shares.
What Public View Counts Cannot Show
Search favours channels that rank, and public counts include paid promotion. Impressions (how many times a thumbnail showed) and click-through rate (the share who clicked) stay invisible, and views climbed after we read them.
How To Measure YouTube Outliers On Your Own Channel
You can run this study on your channel today, free, in under a minute.
- Paste your channel into the channel analyzer, reading your newest 50 videos and scoring Shorts separately.
- Compare your rate with your size band above: rekrap1's 1 of 45 is 2.2%, against 7.7% for 100k to 1M.
- Check any video with the outlier checker for one exact score.
Where your rate sits among our 388 channels:
- Median long-video channel: a 5x outlier in 6.4% of its videos.
- A channel at the top-quarter mark: 13.9%.
- At the top-tenth mark: 20.1%.
- 19.8% had none among their judged videos.
Reading The After Number On Your Own Outliers
The analyzer's "after" divides your next five uploads' median views by the old normal: near 1.0x is unchanged, 3x or more means it moved, like rekrap1's 3.4x there.
The stricter 20-video version reads 2.95x for that video, just under the moved line.
Conclusion (What I Would Do)
I would run my channel through the analyzer, note my 5x rate, then test one title pattern on my next three uploads against my normal there.
Our own YouTube outlier breakdown on Toby shows the far end: 115,963 views on a 228-view normal, 508x.
Every Sunday, we take one apart to separate luck from strategy.
FAQ: YouTube Outliers
Do Shorts get outliers as often as long videos?
Slightly less often. 7.1% of 10,386 Shorts reached 5x their channel's Shorts normal, against 8.3% of long videos. 3.3% of Shorts reached 10x.
How many YouTube views are considered good?
Good means above your median, not a fixed number. The median long video scored exactly 1.00x its normal, 22.9% reached 2x, and 17.7% landed under 0.5x.
Is 10,000 views considered viral?
10,000 views is viral only against a channel's normal: on a 757-view normal, it is roughly 13x.
Across YouTube, Pex found only 0.64% of videos, as of 2018, ever pass 100,000 views.
Is 5x or 10x the outlier line?
5x. It is the only bar we use. 10x or more is a very big outlier, not a second definition. Our outlier score explainer has the maths.
Can I reuse the data?
Yes, with credit to Breakdown Theory. The method, videos.csv and channels.csv are public, and the free-method guide shows how to score a video by hand.