What is an outlier video? How to find outliers on Instagram, TikTok and YouTube
An outlier is a post that beats its own author's typical views by a wide margin — not a post with a big number. We measured how often that happens across 64,471 Reels and TikToks, and here is how to find outliers in your niche, with or without a tool.
SFOM.AI
·7 min read
An outlier video is a post that performs far above its own author's normal — usually measured as its views divided by that creator's median views. A video with 40,000 views is an outlier for an account that typically gets 3,000, and an ordinary day for an account that typically gets 400,000. The number alone tells you almost nothing. The ratio tells you something worked.
Key takeaways
- Outlier score = views ÷ the author's own median views. 2× is notable, 5× is a clear outlier, 10× is rare.
- Across 64,471 Reels and TikToks we measured (March–September 2026), 27% of posts reached 2× their author's median, about 15% reached 5× and about 9% reached 10×.
- 82.5% of the Instagram creators in that sample had at least one 5× post in roughly six months — outliers are common per creator and rare per post.
- Use the median, not the average: one viral post inflates an average and hides the next outlier.
- Copy the format of an outlier from a creator your size, not the view count of a celebrity.
Why raw views mislead
Absolute view counts rank accounts, not ideas. Sort any niche by views and the top of the list is the same large accounts having ordinary days — their audience is big enough that a mediocre post still clears half a million. The post that did something genuinely new, from an account with 9,000 followers, sits halfway down that list where nobody scrolls.
Dividing by the author's own baseline takes account size out of the comparison. What is left is the part you can actually copy: the hook, the format and the angle that made people watch longer and share more than this creator's audience normally does.
How to calculate an outlier score
- Take the creator's last 15–30 posts of the same type — Reels with Reels, Shorts with Shorts. Never mix formats; long-form videos and Shorts have completely different baselines.
- Drop anything posted in the last three days. Its views are still climbing and would drag the baseline down.
- Find the median views of that set: sort the numbers and take the middle one.
- Divide the post's views by that median. The result is the outlier score.
Why the median and not the average: averages are pulled up by the very outliers you are trying to find. A creator with nine posts at 5,000 views and one at 500,000 has an average of 54,500, and against that number their next breakout looks ordinary. The median stays at 5,000, which is what a normal post of theirs actually gets.
How often do outliers actually happen?
We measured it on our own index: 21,813 Instagram Reels from 1,044 creators and 42,658 TikToks from 1,489 creators, posted between 1 March and 17 September 2026. Every creator had at least eight posts and 1,000 followers, and every post was at least three days old when we read its views. The two platforms came out almost identical.
- 2× the author's median or more: 27% of posts on both platforms.
- 5× or more: 14.6% on Instagram, 14.8% on TikTok.
- 10× or more: 9.4% on Instagram, 9.5% on TikTok.
- 50× or more: 3.4% on Instagram, 3.6% on TikTok.
- Below half the median: 21% on Instagram, 20.5% on TikTok — a flop is almost as common as a 2× hit.
One caveat matters. Part of our corpus enters through posts that were already spreading — that is how a trend tool finds creators in the first place — so these shares are probably higher than you would see on a random account. Treat them as a ceiling, not a target.
The per-creator view is the useful one: 82.5% of the Instagram creators in the sample had at least one post at 5× their median within about six months, and 70.9% had one at 10×. Almost every consistent creator produces outliers. The skill is noticing which post it was, and why.
How to find outliers by hand
On Instagram
Open a creator's Reels tab, which shows a view count on every tile. Scan 20–30 tiles, note the typical number, and look for tiles that are clearly an order of magnitude above it. Do this for 10–15 accounts of your size in your niche and the same two or three formats usually start repeating among the outliers.
On TikTok
The profile grid also shows views per video. Pinned videos sit first and are often old hits, so skip them when you estimate what normal looks like. Judge each video against the creator's usual number, not against what looks big.
On YouTube Shorts
A channel's Shorts tab shows views per Short. Sort by Popular to see the head of the distribution, then check the Latest view to see what an ordinary Short gets. Keep Shorts and long-form videos apart — they live on different baselines.
Doing this by hand works, but it costs about a minute per account, and an outlier is only useful for a couple of weeks before the format gets crowded. That is the part worth automating.
What to do with an outlier once you find one
- Name the format, not the topic. “Before/after with the price revealed at the end” transfers to your niche; “a burger” does not.
- Check that it repeats. One outlier can be luck; the same format producing outliers for three different creators is a pattern.
- Study the first two seconds. That is usually where an outlier separates from the creator's normal posts — the rest of the video only has to keep the viewer it already earned.
- Shoot your own version this week, then judge it against YOUR median, not against the original's views.
What is an outlier video on YouTube?
A video or Short whose views are far above the channel's own typical views — usually 2× or more of the channel's median for the same format. Outlier tools compare each video with its channel's baseline instead of ranking by raw views.
What counts as a good outlier score?
2× the creator's median is notable, 5× is a clear outlier and 10× is rare. In our 2026 sample of 64,471 Reels and TikToks about 15% of posts reached 5× and about 9% reached 10× — and those shares are probably on the high side, because our index over-represents posts that were already spreading.
Should I use average or median views as the baseline?
Median. The average is pulled up by the very hits you are trying to detect, so the next outlier looks ordinary against it. The median reflects what a typical post actually gets.
How many posts do I need to calculate a creator's baseline?
At least 8–10 posts of the same format, ideally 15–30, leaving out anything posted in the last three days. With fewer, a single viral or flopped post moves the median too much.
See the outliers in your niche, already scored
SFOM.AI scores Instagram, TikTok and YouTube posts against their own author's median, so a small account having its best week shows up before a celebrity having an ordinary one — with a breakdown of why it worked.
Find outliers in my nicheRead next
Related articles
What is a good engagement rate on Instagram Reels? 2026 benchmarks from 23,596 reels
The median creator in our index gets 3.0% engagement on Reels by views and 1.35% by followers. Here are the benchmarks by follower tier, the two formulas behind them, and why the one most people use makes big accounts look broken.
How to plan a week of Reels and TikToks: a 7-day shot list that doesn't burn you out
Most creators don't run out of ideas — they run out of decisions. A fixed weekly rhythm (three proven formats, two light posts, one conversation, one day off) turns “what do I post today?” into a 20-minute Monday job. Here is the template and how to fill it.
Google Trends alternatives, grouped by the question you're actually asking
Most roundups list the same twelve tools with the same vendor copy. The reason people search for a Google Trends alternative is simpler than that: it measures search demand, and they need to see what is happening on social. Here is which tool answers which question — including where we do not fit.