YouTube · Data study
Why YouTube Videos Go Viral: 1,019 Outliers vs Normal Videos From the Same Channel
of viral videos made a different promise
By Elijah Westbrook · 2026-10-19 · 2 min read
We compared 1,019 viral videos (5× or more their channel’s usual views) with a normal video from the same channel. In 74% of the pairs we checked by hand, the viral video made a different promise. In 26% there was no visible difference at all. Some of it is luck.
share of pairs
Broader beats deeper
A sleep channel’s video on the entire history of Area 51 got 177× its normal views. The video next to it, the entire history of the British Empire, had the same format, the same music and the same voice.
More people are curious about a secret base than about a history class they half remember.
Ask whether your next video can be about something more people already care about.
Names, stakes and the complete version
An animated summary of Atomic Habits got 16.7× its channel’s usual views, made exactly like that channel’s other summaries. A creator who uploaded once a week for a year and shared the results got 50.8×. A compilation of old Halloween commercials got 14.8×.
Name something people know, put something on the line, or make the version that means nobody needs another video after yours.
How-tos win on search
A video on the right way to upload a YouTube video got 85.7× its channel’s usual views, more than the same creator’s video about making money on YouTube.
If you make a how-to, make it about a problem people type into the search bar today.
The titles barely differed
Viral and normal titles were the same length, 61 characters. Viral titles used a few more numbers and words like “most” and “best”, but only by a few points.
The promise matters more than the title formula. Write the promise in the plainest words you can.
From my channel
How I did it
We found 1,019 videos with at least 5× their channel’s median views and paired each with a normal video from the same channel. 150 pairs were checked by hand (145 after removing noise) for what the viral video promised differently.
The limits
Hand-tagging is a judgment call, and 145 pairs is a sample. The categories describe the pairs; they don’t prove the promise caused the views.
Questions
What makes a YouTube video go viral?
In this data, mostly the promise: a broader topic, higher stakes, a bigger name, the complete version or a practical how-to. A quarter of viral videos had no visible reason.
Is going viral on YouTube luck?
Partly. In 26% of hand-checked pairs, the viral and normal video looked the same. You can’t control that part, so give yourself more chances.