Does the shape of the title move the number?
Every structural feature a title can carry, scored by how uploads with it performed against uploads without it, on the WatchMojo main channel and on the competitor set separately.
26 structural title features, matched on long-form titles (1,590 of the 1,595 uploads collected). The number beside each is the median index of titles with the feature over the median index of titles without it, where a video’s index is its views over its own channel’s median for the same type, on videos at least three days old. WatchMojo’s main channel and the 15-channel competitor set are scored separately and never pooled. This is correlation on a single collection window, not causation: a feature can travel with a subject or a release moment rather than cause the views. Where n is small it is marked, and a lift built on a handful of titles should be read as a handful of titles.
What moves the needle on WatchMojo
Features carried by at least five main-channel titles, 20 of the 26, ranked by lift. Green is 1.5x or better; red is 0.6x or worse.
| Feature | n | With | Without | Lift |
|---|---|---|---|---|
| The number is 20 or more | 89 | 1.94x | 0.77x | 2.5x |
| Says 'ever' | 9small | 1.37x | 0.99x | 1.4x |
| Says 'of all time' | 10 | 1.35x | 0.97x | 1.4x |
| Starts with 'Top N' | 143 | 1.09x | 0.79x | 1.4x |
| Has an all-caps word | 69 | 1.08x | 0.86x | 1.3x |
| Speaks to 'you' | 24 | 1.07x | 0.98x | 1.1x |
| 'You missed / didn't know' | 8small | 1.07x | 0.98x | 1.1x |
| Carries a year | 18 | 1.03x | 0.98x | 1.0x |
| Carries 2026 | 17 | 1.00x | 0.99x | 1.0x |
| Hyperbole (insane, shocking, crazy) | 12 | 1.00x | 1.00x | 1.0x |
| The number is under 10 | 7small | 0.98x | 1.00x | 1.0x |
| The number is 10 | 154 | 0.87x | 1.02x | 0.9x |
| Says 'times' (moments list) | 11 | 0.83x | 1.00x | 0.8x |
| Says 'secret / hidden / truth' | 8small | 0.76x | 1.00x | 0.8x |
| Says 'best' | 17 | 0.74x | 1.01x | 0.7x |
| Over 60 characters | 29 | 0.64x | 1.01x | 0.6x |
| No number at all | 39 | 0.25x | 1.06x | 0.2x |
| Has a colon, dash or pipe break | 33 | 0.23x | 1.05x | 0.2x |
| Is a question | 26 | 0.20x | 1.05x | 0.2x |
| Versus | 18 | 0.18x | 1.03x | 0.2x |
A lift of 1.0x means titles with the feature did no better and no worse than titles without it. The table beside carries the medians the ratio was made from.
The same features on the field
Features carried by at least eight competitor titles, 24 of the 26, scored against each channel's own median so a large channel cannot drown a small one.
| Feature | n | With | Without | Lift |
|---|---|---|---|---|
| Says 'of all time' | 10 | 3.37x | 1.00x | 3.4x |
| 'You missed / didn't know' | 13 | 1.73x | 1.00x | 1.7x |
| Versus | 17 | 1.60x | 1.00x | 1.6x |
| Says 'ever' | 21 | 1.57x | 1.00x | 1.6x |
| Says 'explained' | 34 | 1.45x | 1.00x | 1.4x |
| Starts with 'Top N' | 57 | 1.39x | 0.97x | 1.4x |
| Hyperbole (insane, shocking, crazy) | 21 | 1.36x | 1.00x | 1.4x |
| The number is 20 or more | 21 | 1.31x | 1.00x | 1.3x |
| Has a parenthetical | 50 | 1.26x | 0.99x | 1.3x |
| The number is 10 | 104 | 1.08x | 0.98x | 1.1x |
| Is a question | 98 | 1.07x | 0.99x | 1.1x |
| Speaks to 'you' | 119 | 1.04x | 0.99x | 1.0x |
| Says 'secret / hidden / truth' | 36 | 1.03x | 1.00x | 1.0x |
| Says 'actually / really' | 17 | 1.01x | 1.00x | 1.0x |
| Over 60 characters | 317 | 1.00x | 1.00x | 1.0x |
| Says 'times' (moments list) | 8small | 1.00x | 1.00x | 1.0x |
| Has a colon, dash or pipe break | 281 | 0.99x | 1.00x | 1.0x |
| Carries a year | 26 | 0.99x | 1.00x | 1.0x |
| Has an all-caps word | 128 | 0.97x | 1.00x | 1.0x |
| Says 'every' (completist) | 22 | 0.97x | 1.00x | 1.0x |
| No number at all | 701 | 0.99x | 1.03x | 1.0x |
| Says 'worst' | 8small | 0.93x | 1.00x | 0.9x |
| Says 'best' | 23 | 0.84x | 1.00x | 0.8x |
| Says 'ranked' | 14 | 0.56x | 1.00x | 0.6x |
A lift of 1.0x means titles with the feature did no better and no worse than titles without it. The table beside carries the medians the ratio was made from.
The widest gaps between the main channel and the field are on Versus (0.2x on WatchMojo against 1.6x on the field), Is a question (0.2x on WatchMojo against 1.1x on the field), Has a colon, dash or pipe break (0.2x on WatchMojo against 1.0x on the field) and No number at all (0.2x on WatchMojo against 1.0x on the field). The field is ahead on every one of them. The shapes the field wins with are the ones its explainer channels use; on this channel the same shapes read as not-a-list to an audience that came for lists.
Supply and performance on the field are measured across the 15 channels in the competitor set only, not across YouTube. A feature that looks unused on the field may be common elsewhere; the comparison holds between these channels and WatchMojo, nowhere wider.
The number is the format
Four ways a main-channel title can carry a count, and what each did against the titles that did not carry it.
On the main channel, titles numbered 20 or more run at 2.5x the median index of titles without that number, across 89 uploads. Titles numbered 10, the house default, run at 0.9x across 154. Titles with no number at all run at 0.2x across 39. A number under 10 appears on 7 titles at 1.0x, which is too few to read either way. On the field the same shapes sit at 1.3x for 20 or more, 1.1x for 10 and 1.0x for no number, and a number under 10 appears on 1 competitor title, too few to score: the count matters far less to the competitor set than it does here.
Titles numbered 20 or more run at 2.5x the channel median; titles numbered 10 run at 0.86x. Of the summer's twenty biggest breakouts, fourteen carried a 20, 30, 45, 50, 75 or 100.
Thumbnails: not in this run
The brief asks for faces, composition and text correlated with performance. This collection did not measure it, and the honest thing is to say what was seen and what it would take.
No image analysis was performed
Thumbnails were not downloaded or scored. Nothing on this page joins a visual feature to a view count.
- A single high-contrast face or character filling the frame.
- The round cyan mojo badge, bottom-left, on the main channel.
- Rarely any headline text laid over the image.
- No large Top 10 numeral on recent uploads.
- MsMojo carries the same layout with a crimson badge.
Observations of the house style, not measurements. None of them has been tested against the index.
- Face count per thumbnail, and how much of the frame the largest face takes.
- Text presence: whether any headline is laid over the image, and how much.
- Brightness and contrast per thumbnail, as simple scalar features.
- Each joined to the video’s index, so a visual feature gets the same with and without read the title features get above.
Listed on the inputs ledger as number 14, Thumbnail patterns, status not in this run.
How the audience talks about packaging
175 comments in the corpus carry the craft complaints tag (padding, recycled lists, voiceover, thumbnails, misleading titles.). 80 survived screening for display, 29 of them under WatchMojo, MsMojo or MojoPlays uploads. The 8 most liked are below, verbatim.
Tiny Toon Adventures - this game took a lot of inspiration from Super Mario 3 and is still fun to play, downside it was way too short. A remake could give us a longer game and more characters with different abilities.
The narrator's pronunciations are cringe. Foyer is pronounced foy-ay and Je Suis Charlie is pronounced Zhe Swee Charlie. A quick google could have fixed that before the upload.
I'm Chilean, my blood ran cold when I saw the thumbnail.
This narrator is from the compilation videos... he does a great job! 👍
This is way too long. I got tired of it after 20 minutes. Probably should've broken it up into a two or three parter.
Do you think the show's gone on too long? Let us know in the comments.
Yeah.. I clicked.. I really loved his villains and voice over work.
they do this same video like 3 times a year
Comments are the top-sorted sample YouTube returns for each video, capped by our own collector, so likes rank a comment within that sample rather than across the video. A class this size is a handful of voices, useful for the words they use and the things they notice, not as a survey of the audience.