3.7Million: script patterns by public-view group
@3.7million
3.7Million'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 23.4× 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 3.7Million'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 underperforming videos shift from high-concept, long-term personal challenges (like building a business over months or testing AI) to low-stakes, short-term tasks (like 5-hour delivery shifts, interviewing others, or working out of a small car) which lack a compelling narrative hook or significant financial stakes.
3.7Million has 121K subscribers and a median of 72.2K views across its recent long-form uploads. Its highest-viewed video in this sample did about 23.4× 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 272.1K 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
Testing popular internet claims or guru advice with a high-stakes financial target
23.4× channel median“for the next one month I follow the advice of some of the biggest Financial Youtubers with a go to make 10,000 from all the side hustles combined”
Delegating business decisions entirely to artificial intelligence to test its profitability against humans
10.9× channel median“today I'm following charge's advice on building a service business from scratch all to find out will this be more profitable than if a human did it.”
Documenting a long-term, high-effort business build from scratch with zero prior experience
8.3× channel median“This is me building an e-commerce business from scratch for six months with zero experience.”
Patterns in the lower-public-view group
Focusing on low-barrier, highly saturated side hustles that are framed as bad or low-yield
0.4× channel median“in this video I am going to try three of the Wild Side hustles to see how much money I can make”
Conducting passive interviews with other business owners rather than executing a personal challenge
0.3× channel median“today we are going to find out how much money these businesses make how long how long have you been here”
Imposing restrictive physical limitations that make the execution of a basic side hustle highly inefficient
0.1× channel median“and it's almost impossible to do this High hustle if you don't have a but today I'm going to ATT doing it in a small car for the next 5 days”
A writing hypothesis to test
Frame the next video as a high-concept, multi-week challenge where you build a business or test a system using an external authority (like an AI model or a specific industry expert) with a clear, high-stakes financial benchmark to beat.
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 3.7Million breakdown actually measure?
It compares the transcripts of 3.7Million's higher-viewed and lower-viewed recent long-form uploads, grouped against the channel's own median of 72.2K 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 July 7, 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 come with a Creator plan 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.