Veritasium: script patterns by public-view group
@veritasium
Veritasium's higher-viewed and lower-viewed uploads open differently, and this page shows the difference in their own words. Its widest-travelling upload here reached about 3.8× the channel median. We compared 3 recurring patterns from the higher-view group against 3 from the lower-view group, quoted verbatim. This is an observational read of the scripts, not a measurement of retention.
What this page is
A risk map derived from the words in Veritasium's public scripts. The uploads are split into groups by public view count against the channel's own median, and the analyzer reports the script patterns that recur in each group, quoted verbatim. It is not a measurement of this channel's audience retention, which is private to the channel owner, and it does not establish that any pattern caused any view count.
Script difference to review
The difference in performance may relate to how quickly the script establishes a high-stakes, real-world threat or engineering marvel. Winner videos immediately present a massive physical or digital system on the brink of failure or operating at the limits of physics, whereas loser videos tend to open with academic tools, historical anecdotes, or personal channel updates.
Veritasium has 21.1M subscribers and a median of 9.4M views across its recent long-form uploads. Its highest-viewed video in this sample did about 3.8× that median. This breakdown compares recurring transcript patterns in the channel-relative public-view groups, each backed by a verbatim quote. The comparison is observational and does not reveal private retention or establish what caused the view counts. The lower-view group has a cumulative gap of 26.2M public views below the channel median. That is descriptive arithmetic, not an estimate of views lost because of a script pattern.
Patterns in the higher-public-view group
Establishes massive scale with specific numbers early
3.8× channel median“transistors, billions of them. They are the ones and zeros of our computer.”
Builds tension using a precise timeline sequence
1.7× channel median“At 7:03 pm, around 50 people were riding the long escalator down to the platform”
Introduces a high stakes global vulnerability immediately
1.6× channel median“In 2021, a hacker uncovered a fatal weakness in the world's most important operating system.”
Patterns in the lower-public-view group
Begins with historical anecdotes from centuries ago
0.3× channel median“This is from a 1247 book, "The Washing Away of Wrongs" by Song Ci.”
Focuses on personal channel updates and history
0.5× channel median“You know, I started this YouTube channel in 2011 after quitting my full-time job.”
Relies on academic textbook diagrams as hooks
0.5× channel median“This is the scariest chart in electrical engineering. - Puts the fear of God into all undergraduate students.”
A writing hypothesis to test
In the next script, test opening with a high-stakes, real-world system failure or physical limit within the first 30 seconds, rather than starting with historical context or academic diagrams.
What this page cannot tell you
It cannot tell you whether these patterns would do anything for your channel, because it never saw your script. The patterns above are what recurs in one channel's published uploads; whether your own opening holds attention is a question about your draft, and it is answerable before you record.
Questions about this breakdown
What does this Veritasium breakdown actually measure?
It compares the transcripts of Veritasium's higher-viewed and lower-viewed recent long-form uploads, grouped against the channel's own median of 9.4M views, and reports the script patterns that recur in each group with a verbatim quote for each. It is a read of the scripts. It does not use audience retention data, which is private to the channel owner, and it does not establish that any pattern caused any view count.
Is this a prediction of how the videos performed?
No. The analyzer reads the words in a script and flags where attention is most at risk on the page. Published view counts are used only to split the uploads into groups to compare. Nothing here is a measurement or a forecast of what real viewers did, and the same script can travel very differently depending on topic, packaging, timing, and distribution.
How current is this analysis?
This page was analyzed on August 8, 2026 and the result is frozen at that point, with the analyzer version stamped on the page. YouTube changes, channels change, and a scan run today could group different uploads. Re-read the quotes against the channel's current videos before relying on any pattern.
Can I get this for my own script?
Yes, and that is the more useful direction. This page shows what the analyzer found in someone else's published scripts. Running your own draft through the free check gives you the same read on the script you are about to record, while you can still change it: hook, pacing, and structure scores, and the passages most at risk of losing attention.
This page reads someone else's script. Read your own.
Paste the draft you are about to record and get the same kind of read on it: hook, pacing, and structure scores, and the passages most at risk of losing attention, while you can still change them. Free, no login. The line-by-line rewrites are a $9 unlock if you want them.
Check my script freeRelated reading
- All channel breakdowns : the same read on other channels.
- Hook examples that open on a claim : the pattern most often separating the two groups above.
- The retention guide : what the curve shapes mean and which are script problems.
- Our 349-hook study : first-party research with the row-level data published.
- Run this scan on your own channel : free, needs captioned long-form uploads.