Ali Abdaal: What All 1,458 Videos Reveal

Ali Abdaal has 6.68 million subscribers and 558,911,935 lifetime views on the channel. The oldest two videos are music covers from 2016 and the channel proper starts in 2017. The public library now runs to 1,458 videos: 683 long-form uploads at a median length of 15 minutes, 761 Shorts at a median of 52 seconds, and 14 livestreams that are held out of every number below.

I put all 1,458 through a full channel export from the YouTube API, with every view, like and comment count. A 2018 video has had eight years to collect views, so comparing raw totals would make every old video look like a winner. Instead each video is compared only with the ones he published in the same six months, in the same format. I also read the full caption tracks of two videos and pulled Google Ads search volumes for the UK and the US.

The analysis in this article comes from the same process Stewart runs on client channels every month inside Channel Intelligence.

What comes out is a channel that has been about five different subjects in nine years while keeping the same person, the same category and the same credential. Each change moved him to a bigger audience for six years. Money and business subjects are now the biggest share of his uploads they've ever been, those titles come in slightly below his typical video, and the typical video has been falling since late 2023.

The 5 moves

Five moves came out of the analysis. The evidence for each one sits further down the article.

The Takeout: These are the moves that show up in Ali Abdaal's content. They're what his audience rewards on his channel, at his size, in his subject. Nothing here transfers automatically to yours, and the honest way to use any of this is as a hypothesis: pick one, have your channel manager run it against your own library, and keep it only if your numbers agree.

1. He kept the credential and dropped the subject. 114 of the 1,458 titles carry medicine or doctor, worth 5.2% of the channel's views. His last three long-form videos about medicine are all about why he left, and the most recent was December 2022. The credential moved into the spoken introduction instead.
2. The business videos are filed as Education. 1,322 of the 1,458 uploads, 91% of the library, sit in YouTube's Education category, including the ones about making money.
3. A checkable claim beats a big one. Long-form titles carrying "evidence-based" get 3.10x his typical video of the same six months. Titles carrying a money figure or the word "million" get 0.93x.
4. He's cut the Shorts hard. Shorts are 53% of the library and 20% of the views. He published about 23 a month from September 2025 to June 2026, then 7 in July 2026 and 6 in August.
5. Every change of subject grew the channel until 2023. The typical long-form video peaked at 740,329 views in the second half of 2023. The most recent complete half-year is 214,129, a fall of 71%.

1. He kept the credential and dropped the subject

Ali Abdaal describes himself as a doctor turned entrepreneur, in his channel description and in the opening seconds of his videos. Medicine is close to absent from his titles.

Across all 1,458 videos, 114 carry medicine or doctor in the title. Between them they hold 28,748,124 views out of the 552,531,848 the library has collected, which is 5.2% of the channel's views. Nearly all of that comes from early in the channel: the study technique and medical admissions videos he made as a Cambridge medical student.

His last three long-form videos on the subject are all about leaving it:

• Why I Left Medicine (And Why I'm Going Back), March 2021, 486,129 views
• Why I Left Medicine... Forever, November 2021, 896,925 views
• My Regrets about Quitting Medicine, December 2022, 308,505 views

Since then medicine hasn't come back to a title. It has stayed in the videos. This is how a video published in February 2026 introduces him, taken from its caption track:

"hello, my name is Ali, I'm a doctor turned entrepreneur and author of the New York Times bestseller Feel Good Productivity"

The channel's own description opens the same way. So the credential does its work in the first thirty seconds, where it costs him nothing, and the title is left free to sell whatever the video is actually about. Whether viewers trust him more because of it isn't something outside data can measure, so treat that as a hypothesis rather than a finding.

The Takeout: If you've got a credential that isn't what you sell, say it out loud in the first thirty seconds of every video and keep it out of your titles. The title has one job, which is to earn the click on the subject the viewer came for, and a credential sitting in it competes for space with the promise.

2. His videos are filed as Education, not business

1,322 of the 1,458 uploads sit in YouTube's Education category, which is 91% of the library. The next biggest groups are People and Blogs at 86 videos, Howto and Style at 18, Music at 16, and Science and Technology at 14.

The category is one of the few things you tell YouTube directly about a video, and on this channel it stays the same whether the video is about revision technique or about making your first thousand dollars online. The money videos are filed as Education too.

What that does to how YouTube distributes a video isn't visible from outside an account, so the effect on distribution is a hypothesis and I've labelled it as one. What is measurable is the consistency: nine uploads in ten describe the channel the same way, across nine years and five changes of subject.

A channel gets two places to tell YouTube what subject it covers. One is the category on each video. The other is the channel keywords box, and when we read what 20 UK business channels had put in theirs, five had filled in nothing at all and the fifteen that had used it claimed 247 different subjects between them.

The Takeout: Open Studio, sort your library by category, and count how many different ones you've used. If your channel has one subject and three categories, you're describing it three different ways in the one field where you get to describe it at all, and picking one takes a few minutes per video.

3. A checkable claim beats a big one

Two title elements move in opposite directions on this channel, and they're the two that set what kind of promise the video is making.

Thirteen long-form videos have "evidence-based" in the title. Against the videos he published in the same six months, their median is 3.10x. Nine of the thirteen beat his typical video and four didn't, and the spread inside the group is very wide, so this is a strong hint rather than a rule. The weakest of them, How to Actually Get in Shape in 2026 - My Evidence-Based Guide, did 0.53x.

Thirty-five long-form videos carry a money figure or the word "million". They get 0.93x, with sixteen above his typical video and nineteen below.

Across the six elements the pattern is that titles promising a method do better than titles promising a number. "How to" openers get 1.27x on 107 videos and "My" openers 1.23x on 70, while a number anywhere in the title gets 0.97x on 322 videos. A number on its own leaves him 3% below his typical video, which is close to no effect at all.

There's a search-side version of the same point. In the US, "productivity" gets 60,500 Google searches a month at a $15.12 cost per click, with advertisers bidding $6.79 to $51.27 for the top of the page. In the UK it's 8,100 searches a month at $9.50. His own name gets 4,400 a month in the US and 2,900 in the UK. All of those figures come from DataForSEO's Google Ads data and are in US dollars. So in his largest market the subject he teaches is searched for around fourteen times more often than the person teaching it, which is the opposite of the position Steven Bartlett has built, where his own name out-searches the guests on his biggest episodes.

The Takeout: Split your own titles into the ones promising a method and the ones promising a number, then compare each group against the videos you published around the same time. If method titles win on your channel too, you've learned which kind of promise your audience believes, and writing more of them costs you nothing.

4. He's making far fewer Shorts

Once livestreams are held out, Shorts are 761 of 1,444 uploads, so 53% of everything he's published. They've returned 110,893,460 views out of 551,312,896, which is 20%. Half the library, a fifth of the result.

The per-video number has been falling for years. The typical Short in the second half of 2020 got 416,912 views. In the first half of 2026 it got 28,849, a fall of 93%.

He published between 11 and 33 Shorts a month from September 2025 to June 2026, an average of about 23. Then 7 in July 2026 and 6 in August. Long-form output didn't move across the same two months: 2 in July and 2 in August, against 3 to 5 a month for the year before.

Two months is a short read, and a summer gap can be a holiday rather than a decision. What the data supports is narrower and still useful: the cut is Shorts-only, and the long-form schedule held while it happened. Daniel Priestley's channel shows the same shape, long-form first with Shorts demoted, while Simon Squibb does the opposite, roughly eight Shorts a week alongside about one long video a week, across a 1,723-video library and 2.42 million subscribers.

The Takeout: Work out what share of your views your Shorts return, then compare it with the share of your production time they take. If a format is half your output and a fifth of your result, it needs to be doing a job you can name, like feeding people into a specific long-form video, or it's the cheapest thing on the channel to cut.

5. Five eras, and where the growth stopped

The channel has been about five different things. Each era below shows the typical long-form video published inside it.

The sharpest boundary is between the two halves of 2021, and it happens in one step. In the first half of 2021, 20% of his uploads had a tech word in the title: iPads, keyboards, desk setups, Mac apps. In the second half, none did. Money and business words went from 9% of uploads to 18% across the same line.

Money and business words have kept climbing since. They're 20% of his uploads in the first half of 2026, a bigger share than in any earlier period, and 33% of the 18 uploads since July carry them.

Every video in this article is scored against the median of its own six months, so those medians are worth looking at on their own.

The typical long-form video peaked in the second half of 2023 at 740,329 views. The most recent complete half-year, the first half of 2026, is 214,129, a fall of 71%. The low point was the second half of 2025 at 188,204, so the most recent period is up on the one before it.

The money era's card above shows a lower typical video than the tech era's card, and that's because the era covers both the climb and the fall. At its best the money era was well clear of the tech era, whose strongest half-year was 505,426.

Volume went the other way. He published 95 videos in 2021 and 267 in 2024. Videos published from the start of 2021 onwards account for 72% of the channel's lifetime views, so the recent library is doing most of the work in total even while each individual video does less.

Put those together and the mechanism is visible. Medical admissions to study technique to tech each moved him to a bigger audience, and the money era kept climbing for two more years after that before it turned. The subject he's publishing most of now is the one whose titles come in below his typical video, at 0.93x, and its share of his uploads is still rising.

Several things changed at the same time, so outside data can't separate them. He roughly tripled his output, Shorts became half the library, and the subject moved to money. Any of those would pull a per-video median down on its own. What the data does show is that the decline and the subject change happen together, and that a channel is a live test rather than a finished broadcast. The most recent half-year is already up on the one before it.

The Takeout: Plot your own typical video's views by six-month period instead of looking at videos one at a time. That's the view that separates a channel getting better from a library getting bigger, and it's usually the number a change of subject moves first.

Put simply

The five moves work as one machine. A credential that lives in the introduction and stays out of the titles, one category declared 1,322 times, titles that promise a method rather than a number, a Shorts habit he's willing to cut, and a subject that changes whenever the audience moves.

The part most people would copy is the subject, because that's the visible part. The part doing the work is the discipline underneath it. He stopped making study videos when he stopped being a student, dropped tech titles from 20% of a half-year to zero in one step, and has cut his monthly Shorts output by roughly two thirds in two months. Nine years, five subjects, one person, one category.

The open question on the channel is whether money and online business reaches the audience the earlier subjects did. So far it hasn't, and he's making more of it than ever.

Method and limitsSources, dates and what this can't see

Public data only, pulled 9 September 2026: the full 1,458-video library from the YouTube Data API with every view, like and comment count, the automatic caption tracks of two videos, and Google Ads search volumes and costs per click for the UK and the US. The channel's own page reports 1,461 videos, so three of them sit outside the public uploads playlist and aren't in this analysis. They're worth about 1% of the views.

Fourteen livestreams were identified and held out of every sample and every baseline. Their own median is 32,689 views. YouTube's public data carries no Shorts flag, so format is assigned by duration and a video of 60 seconds or less is counted as a Short.

Every per-video comparison scores a video against the median of the videos he published in the same six-month period in the same format, and only videos over 90 days old are scored. Periods with fewer than five videos aren't used as baselines. Search figures are labelled with their market and are in US dollars, because that's the currency the source reports in. One of them, the US cost per click for his own name, sits outside its own top-of-page bid range, so the bid range is the better guide there.

YouTube changed how it counts a view on 24 August 2026. Two of the 1,458 videos were published after that date, which is 0.1% of the library, and the 90-day rule excludes both from every score. So 99.9% of the figures here sit on one side of that boundary, and nothing in this article compares views across it.

A paid-traffic screen was run on the channel's six biggest videos and on every video in the evidence-based group, comparing like rate and comment rate against this channel's own long-form norms of 3.09% and 0.130%. Every one of them came back with a like rate between 2.42% and 4.70%, so none of them looks bought.

There's no Studio access here, so nothing in this article estimates impressions, click-through, retention, traffic sources or revenue. Where a cause is suggested for something, it's labelled as a hypothesis. Subscriber history was left out on purpose, because the daily series available for this channel isn't reliable enough to carry a conclusion.

Analysis and write-up: Stewart Read.