A weird YouTube thing?
2020-02-25 · letter vlog · John · 3:44 · watch on YouTube
Topic: Internet & YouTube · mode: Personal essay · tags: self promotion, civic or action cta
John reports that for over a year Vlogbrothers videos have drawn almost exactly one like per ten views in their first 24 hours, explains why the brothers stopped asking for likes and subscriptions (they want viewers who care, not the most viewers), and nervously asks viewers to like this video to test the ratio.
What is said
The parts of the video and what is said in each (★ — how much it matters to the video).
00:00:00 — Tour plug; the one-like-per-ten-views ratio on Vlogbrothers videos, with examples
- ★★ John says he and Hank are coming on tour to Indianapolis, Columbus and Ann Arbor, hopes to see viewers there, and points to ticket links in the doobly-doo. — 00:00:00 · Indianapolis, Hank Green
- ★★★ John says that for over a year almost every Vlogbrothers video has drawn one like for every 10 views in its first 24 hours, which must be coincidental but is very strange. — 00:00:00 · Vlogbrothers
- ★★ John gives examples (75,700 views and 7,600 likes; 99,000 views and 9,900 likes) and says it holds for Hank's and John's videos and for high- and low-view videos alike. — 00:00:29 · Vlogbrothers
- ★★ John says that when the ratio seems not to hold, as with a video showing 30,000 views and 3.3 thousand likes, analytics reveal exactly 33,000 views: the watch-page count just hasn't updated. — 00:00:29
00:00:58 — The ratio decays after 24 hours; the old five-star rating days; not knowing what the algorithm values now
- ★★ John says that after the first 24 hours fewer viewers click like, so likes decouple from views, but for that first day the ratio has held for over a hundred videos. — 00:00:58
- ★★ John recalls that before likes, when videos were rated on a five-star scale, the brothers often asked viewers to rate videos five stars because 'it makes new nerdfighters'. — 00:00:58 · Hank Green, Nerdfighteria
- ★★ John says the five-star asks really did make new nerdfighters, because the algorithm then valued a video's average rating, so people found videos they otherwise wouldn't have. — 00:01:27 · recommender system
- ★★ John says one of the chief joys of his professional life is no longer having to chase the whims of YouTube's recommendation engine. — 00:01:27 · YouTube, recommender system
- ★★ John has no idea what YouTube's algorithm values now but suspects it isn't as simple as showing a video to 10 people per like; maybe it is and he's cracked the code. (playful, hedged) — 00:01:27 · YouTube, recommender system
00:01:56 — Why the brothers don't ask for likes or subscriptions: they want viewers who care; the ratio reflects a consistent audience
- ★★★ John says that for several years the brothers have made a point of not asking viewers to like and subscribe, first because YouTube users hear that enough every day. — 00:01:56 · YouTube
- ★★★ Second, John says they don't want the most possible viewers (he doesn't want to live in trending comment sections) but to reach people who find something useful or helpful connections. — 00:01:56
- ★★★ Echoing what Hank has often told him, John says what matters is not how many people watch their videos but how many care about them. — 00:02:26 · Hank Green
- ★★ John says that partly because they've been around so long, asking for likes or subscriptions or hunting the clickbaitiest thumbnail feels uncomfortable or disingenuous. — 00:02:26
- ★★ John says almost every video he has made for several years uses whatever thumbnail YouTube chose, not as the right way in general but as right for this channel and viewership. — 00:02:26 · YouTube
- ★★ John suspects the real reason for the astonishingly consistent like-to-view ratio is that their audience is really consistent. (hedged) — 00:02:55 · Nerdfighteria
- ★★ John thanks viewers, new and longtime, for sharing their attention and says he doesn't want to do anything to mess that up. — 00:02:55
00:03:20 — John nervously asks viewers to like the video to test the ratio; sign-off
- ★★★ Feeling as nervous as if asking someone to prom, John asks viewers to click like to break the cycle or find the ratio is real; probably not, but he asks them to like it so he can disprove the theory. — 00:03:20
In the description and comments
- ★★ In a pinned comment John asks viewers to like this video, saying he hasn't asked that in years and won't ask again for years, but is very curious. — creator comment
- ★ In a comment John says the video has a bad cut because he rushed the edit; he already re-uploaded once, removing a version with a long blank stretch, and won't again. — creator comment
- ★★★ In comment updates John reports the 1:10 ratio is dead: at 17 hours the video had 107,000 views and 47,000 likes, about 1:2.2, with views about 35% above normal. — creator comment
- ★★ John concludes the 1:10 ratio was not necessary or inevitable and probably partly human behavior, and that online, human and algorithmic behavior aren't really separable. — creator comment · recommender system
- ★★ Asked whether other Complexly channels will ask for likes, John says maybe if the effect is huge, but asking for likes pulls learners out of the learning experience. (unsure) — creator comment
Threads in this video
- reminder · Hank invites viewers to the March 2020 Dear Hank and John and Anthropocene Reviewed live shows (2020) — John says he and Hank are coming on tour to Indianapolis, Columbus and Ann Arbor, hopes to see viewers there, and points to ticket links in the doobly-doo. (00:00:00)
- restate · Does commercial success or reach measure how good or valuable a work is? — Second, John says they don't want the most possible viewers (he doesn't want to live in trending comment sections) but to reach people who find something useful or helpful connections. (00:01:56)
- asked · John asks viewers to like his video to test the one-like-per-ten-views ratio (2020) — Feeling as nervous as if asking someone to prom, John asks viewers to click like to break the cycle or find the ratio is real; probably not, but he asks them to like it so he can disprove the theory. (00:03:20)
- result reported · John asks viewers to like his video to test the one-like-per-ten-views ratio (2020) — In comment updates John reports the 1:10 ratio is dead: at 17 hours the video had 107,000 views and 47,000 likes, about 1:2.2, with views about 35% above normal.
- promised · John: not ask viewers to like a video again for years (2020) — In a pinned comment John asks viewers to like this video, saying he hasn't asked that in years and won't ask again for years, but is very curious.
- predicted · John: the one-like-per-ten-views ratio will probably not hold for this video (2020) — Feeling as nervous as if asking someone to prom, John asks viewers to click like to break the cycle or find the ratio is real; probably not, but he asks them to like it so he can disprove the theory. (00:03:20)
- came true · John: the one-like-per-ten-views ratio will probably not hold for this video (2020) — In comment updates John reports the 1:10 ratio is dead: at 17 hours the video had 107,000 views and 47,000 likes, about 1:2.2, with views about 35% above normal.
Where the stories that moved here went next
- John asks viewers to like his video to test the one-like-per-ten-views ratio (2020) — reported back; last: result reported, 2026-09-15
- John: the one-like-per-ten-views ratio will probably not hold for this video (2020) — came true; last: recalled, 2026-09-15
- John: not ask viewers to like a video again for years (2020) — standing commitment (last affirmed 2020); last: promised, 2020-02-25
Mentioned
Ann Arbor, Michigan · clickbait · Columbus, Ohio · Indianapolis · recommender system · YouTube
Replied to by
← Previous: The Muck of Millennia · Next → I'M FINE EVERYTHING IS FINE
Later videos replying to this one
- 2026-09-15 · YouTube Says This Video Has 683,183 Views, It Actually Has 312,584 (John) — continues — “many years ago, I made a video about a phenomenon I'd noticed on YouTube called the ratio” (00:00:29)