Write Better YouTube Titles In 54 Minutes (Jake Thomas | Creator Hooks)

Jay Clouse

Abstract

Fifty-four minutes on titles as a modelling problem rather than a creative one. Three emotions do the clicking — curiosity, fear and desire, usually curiosity plus one of the other two — and the working method is finding what already produced them and moving it into your own subject.

• Model creators a few steps ahead rather than the biggest names, because their conditions still resemble yours.
• Patterns get found by spotting outliers, and good patterns do not care what niche they came from.
• The idea comes first: a clever title cannot save a bad one, and is only lipstick on a pig.
• One subject per title — if you confuse, you lose, and no focus means no audience.
• A test waits five days, because a fresh video's first day goes to the people who already subscribe.
• A hook score normalises for channel size, which is what makes outliers comparable at all.

Everything below follows the conversation in order, from the three emotions through modelling and swipe files, into click triggers and the ethics of fear, out through guests, testing, curation and the search-versus-browse split, and finally to the creator behind the packaging. For anyone building a personal brand the last beat is the one that generalises: most people never answer why anybody should listen to them, and audiences can feel it.

Chapter summaries

00:13 - The best titles create an emotional response, and there are three emotions

The best titles create an emotional response that makes you click; the three click-worthy emotions are curiosity, fear and desire, usually curiosity plus fear or curiosity plus desire.

"emotional response that cause you to click the three I'm calling them click-worthy emotions"

— 00:13

Curiosity, fear and desire — usually curiosity plus one of the other two. Three emotions is a small enough set to check any draft title against, which is most of what makes the taxonomy worth having.

01:51 - The models worth copying are a few steps ahead rather than at the top

Do not model the biggest names; model creators a few steps ahead of you (roughly five thousand to five hundred thousand).

"I don't think it's the the right way to do it so if I"

— 01:51

Not the biggest names. Creators a few steps ahead — roughly five thousand to five hundred thousand — because their conditions still resemble yours and their moves are still copyable.

03:07 - Proven models get copied before anybody tries to be original

Copy proven models first; don't try to be super original.

"the best way to do it is to kind of copy first like people"

— 03:07

Hand-copying old sales letters has been recommended for training copywriters for some seventy years and still works, which makes the practice older than any platform it is being applied to. Copy work comes before invention because the pattern has to be in the hand before it can be varied.

04:24 - Analysis paralysis is what stops most people

Don't let analysis paralysis stop you; you can't steer a parked car — just move, then figure it out.

"yeah definitely but also like if you're just getting started like just just do"

— 04:24

Analysis paralysis is the thing that stops most people, and a parked car cannot be steered. Move first and figure it out afterwards, which is a claim about sequencing rather than about carelessness.

05:11 - A title is the gate on everything already spent

A good title is the gate on all the money and effort already spent making the content.

"then figure it out yes how did you get to YouTube titles of all"

— 05:11

Everything already spent making the content passes through the title or does not pass at all. Framing it as a gate rather than a label is what justifies the disproportionate attention.

06:50 - Originality is not required, and modelling into a chosen niche is

You don't have to be original — find something that works and model it into your own niche.

"but but somehow I kind of figured it out and the thing that like"

— 06:48

Following a proven structure from an adjacent niche carries the working pattern across rather than duplicating the content. A hunting channel's newbie-gear episode reworked as newbie fishing gear became the channel's best episode, and revenue then multiplied several times over — which is why it is described as not like copying.

08:05 - Patterns are found by spotting outliers

Patterns are found by spotting outliers: a video with far more views than its neighbours signals something worth noting.

"content how did you go about finding these templates of breakout sort of successes"

— 08:05

Publish date and view count are both exposed, so an outlier against its neighbours is visible without any tool. Hours every week for nearly two years produce hundreds of notes, and the notes turn into patterns: a video at 100k where the neighbours sit at 10k is a pattern waiting to be named.

10:43 - Emotions are what make a good title

Emotions are what make a good title; curiosity plus fear or curiosity plus desire is the simplest explanation, with nuances below that.

"because I I've heard you talk about some of these emotions that you think"

— 10:43

Bland statements carry no emotion, and a channel can improve thumbnails, storytelling and editing while its titles stay emotionless statements and it stays flat. The emotional pair is described as the high-level version, with the nuances sitting below it.

11:36 - The emotional principles do not go stale even as phrasings do

The emotional principles don't go stale, even as specific templated phrasings do.

"good title I like that you're focusing on kind of emotional terms and ideas"

— 11:36

It is psychology that gets attention and clicks, which is why copywriting books from the 1930s to the 1960s say the same thing as a 2014-era listicle promising you will never guess number seventeen. The phrasings expire; curiosity and a little fear do not — and the claim is hedged to a couple of years of watching.

13:05 - Opening a loop in a title builds the curiosity

Opening a loop in a title builds curiosity: you start a story but don't finish it.

"super nerdy here I've got a database of 408 titles right now and 61"

— 13:05

Start a story and do not finish it. An open loop in a title is the most reliable curiosity device and the easiest to over-apply, which is why it appears everywhere at once.

13:49 - One particular word signals insider access

The word "actually" (or "really") signals insider access — you feel you're getting a secret, which is curiosity, plus authority you can trust.

"how a pro YouTuber actually makes thumbnails like just the word actually is like"

— 13:49

Pulling back the curtain implies the reader already knew the surface but not the real method, and naming the pros adds instant credibility on top. A title about how a pro actually makes thumbnails stacks three things at once: desire, curiosity by secret, and authority.

14:36 - The idea comes first, because a clever title cannot save a bad one and is only

The idea comes first, because a clever title cannot save a bad one and is only lipstick on a pig.

"the pros do it when you are writing titles for yourself or for other"

— 14:35

No amount of wordsmithing, graphic design or thumbnail skill rescues a weak idea. It is named as the most important thing, ahead even of angle, which puts every technique in the session downstream of a decision made before any of them apply.

15:05 - The same thing gets said a hundred different ways, and calling out a specific audience by

The same thing gets said a hundred different ways, and calling out a specific audience by name is a strong click driver.

"the most important thing and then just coming from there it's like you know"

— 15:05

Named beginners see themselves in the title and think this is for me, and beginners are the hungriest and most insecure viewers on any topic. Picking stocks as a beginner is the same video as picking stocks; the guest binged beginner piano videos precisely because he felt too insecure for the advanced ones.

16:48 - There are around thirty recurring click triggers

There are ~31 recurring "click triggers" — reveal a secret, open a loop, regrets, timeliness, call-outs — treated as moves in your arsenal.

"if you're you know making your stocks video how to pick stocks for beginners"

— 16:48

Around thirty-one recurring triggers — revealing a secret, opening a loop, regrets, timeliness, call-outs. Treated as moves in an arsenal rather than as a formula to follow.

18:55 - A creator's own clicking behaviour is the study material

Study your own behaviour: when you click or buy something, ask why.

"were talking about his research process and something I shared with Jake is that"

— 18:55

Your own recorded reactions surface the patterns before you can name them. A personal document kept since around 2015 collects the titles that caught the eye — and the first five entries in it were all fear-based, which is a finding about the collector as much as about the titles.

19:44 - Fear works, and leaning on it feels antithetical to the work

Fear works, and titling everything to prey on fears and insecurities feels antithetical to what you want to do.

"here fear is working very well on me I hate that about our psychology"

— 19:42

The conflict is left unresolved rather than argued away, which is the honest position when a tactic works and the person using it would rather it did not. Naming the discomfort rather than resolving it is the honest move.

19:59 - There is a standing tension between what gets rewarded and what the show is

There's a standing tension between what the audience/platform rewards and what the show is about; you play the game "a little bit" while watching how far is too far.

"works we have this little bit of tension with with our Channel where we"

— 19:59

A standing tension between what the audience and the platform reward and what the show is actually about. The stated resolution is to play the game a little while watching how far is too far.

22:04 - Fear gets framed as a lived experience rather than as an attack

To use fear but turn the ickiness dial down, frame it as your own experience, not an attack on "you".

"that is not what you do one one way to uh to like make"

— 22:04

Frame it as your own experience rather than as an attack on the reader. It uses the same emotional pull with the ickiness dial turned down, which is a craft answer to an ethical problem.

22:50 - Aggression gets dialled up or down by the state of the business

How aggressive to be depends on business state: growing, dial it down and serve; growth stalling, run what you know works.

"do it but for me it kind of depends on like how is my"

— 22:50

Growing, dial it down and serve. Growth stalling, run what is known to work. The answer depends on the state of the business rather than on principle, which is more honest than a fixed rule.

23:25 - Competing title ideas get settled by searching them

To choose between competing title ideas, search the title (or a broader adjacent topic) and read whether it has worked before.

"when you have like competing ideas for a title let's say that the idea"

— 23:25

Visible publish dates and view counts answer did this work for a similar title without running the test yourself. An obscure niche gets checked against a broader adjacent topic — steam mops against vacuums — because the trend is easier to see where the volume is.

24:34 - An unfamiliar guest's name comes out of the title

Unless the guest is a heavy hitter, skip the name or give yourself/them a label rather than leaning on the name.

"not worked let's say that I wanted to lean into this strategy and as"

— 24:33

Unless the guest is a heavy hitter, the name comes out and a label goes in. The name is only an asset if the audience already recognises it, which is a smaller set than anybody assumes.

25:30 - A rising guest gets the common element across their best work

For a rising-profile guest, find the common creative element across the top videos and include it; if they're all different, pick the single most interesting thing.

"into Justin Welsh let's say he has a growing profile more people are searching"

— 25:30

Find the common creative element across their top videos and include it. If those are all different, pick the single most interesting thing — which is the fallback rather than the method.

26:57 - One subject per title, and the commas come out

One subject per title — drop the commas and "ands"; if you confuse, you lose, and no focus means no audience.

"you think about brevity because as you're listing off some of these they they"

— 26:57

Drop the commas and the ands. If you confuse, you lose, and no focus means no audience — and cutting from three listed things to one concise idea lifted an email open rate materially. One subject is the whole rule, and it is the most frequently broken one. Descriptions probably do not matter much either, because the platform is watching the video itself rather than reading around it.

30:22 - Thumbnails matter more in entertainment and titles more in education

Thumbnails matter more in entertainment; titles matter more in education.

"personal opinion but I think that thumbnails are more important in entertainment and titles"

— 30:22

It is offered as a personal read rather than as a finding, and it is useful anyway because knowing which side you are on decides where the effort goes. An entertainment viewer is choosing a feeling from an image; an education viewer is checking whether the words name their problem.

31:37 - A test waits at least five days before it starts

Wait at least five days before running an A/B test, because a fresh video's first day is shown to loyal subscribers who click most.

"b testing is I wait five days days because if you test too early"

— 31:37

Day-one click-through is inflated by loyal followers, so the early data is bad data and the test measures the wrong audience. Videos have blown up after a day and a half with no changes at all and then dropped off, and tools that switch the variant every twenty-four hours are testing that noise.

32:45 - A test result is a compass rather than a verdict

Use A/B test results as a compass showing which themes work, not as precise verdicts.

"it out I know a lot of people test it out like within an"

— 32:45

A compass showing which themes work rather than a precise verdict. One variable at a time — title or thumbnail — unless a whole new idea is being tested.

34:32 - Higher click-through with lower duration reads as too clickbaity

A higher CTR with lower average view duration reads as too clickbaity; often, though, higher CTR correlates with higher watch time.

"sometimes you might have a much higher click-through rate but maybe your average view"

— 34:32

A higher click-through with a lower average view duration reads as too clickbaity. Often, though, higher click-through correlates with higher watch time, which means the warning sign is the divergence rather than the number.

38:25 - Curation runs on a hook score, which is how many views a video got above its

Curation runs on a hook score, which is how many views a video got above its own channel's recent average.

"how do you curate an uh an issue of Creator hooks how do you"

— 38:25

Comparing a video to its ten neighbours — five before, five after — isolates what beat the channel's own baseline, so there is something specific to learn from it. A channel averaging around 100k that posts one video at a million is the clear case; the measure fails on channels that post banger after banger with no deviation to detect.

39:35 - The feed gets trained to surface outliers

Train the feed to surface outliers: keep a list of channels and hunt one niche until the home page fills with that niche's outliers.

"a list of like 300 channels that I I'll scroll through those 300 channels"

— 39:35

Watching and clicking inside one niche steers the recommendations until the home page becomes an outlier-finder. Twenty or thirty minutes poking around travel channels is enough to turn a feed into travel outliers, which is how a list of three hundred channels gets built without searching for any of them.

40:33 - The same hooks work across industries

The same hooks work across industries, so you can model a title from an unrelated niche for your own subject line.

"how often do you use the Curiosity plus desire or fear framework for the"

— 40:33

A hook is portable: the emotional structure carries even when the topic changes completely. An email subject line modelled on a photography video title opened at around half its list — and it read slightly clickbaity where it did not literally deliver, which a few readers said before the explanation landed.

41:44 - Email subject lines fight a word ceiling

Email subject lines fight a word ceiling — staying ~six to eight words so they don't truncate on mobile makes the emotional model harder to apply.

"well for like 52 open rate I might try this model more closely aligned"

— 41:44

Six to eight words so nothing truncates on a phone. The ceiling makes the emotional model harder to apply, which is a real constraint rather than a reason to abandon it.

44:46 - Writing for the human or for the machine is decided by where the content is meant

Writing for the human or for the machine is decided by where the content is meant to show up.

"so you're saying right for the human don't write for the robot um I"

— 44:46

It is a routing decision made before any words get written, and getting it wrong cannot be repaired by writing the other version better.

45:53 - For search the keyword goes first and the trigger goes last

For search, put the main subject/keyword up front and a click trigger at the end.

"and stuff so like going back to SEO one of the things that you"

— 45:53

Leading with the keyword is what ranks, and a trailing trigger — a desire, a year — still earns the click once the ranking has done its job. A solopreneur revenue claim with the year at the end is the shape: keyword front, timeliness and desire trailing. It is offered as one of the better ways rather than the way.

46:34 - The year goes at the end of an evergreen tutorial

Put the current year at the end of evergreen tutorials, heavily from about November through January.

"you can use like a big desire so like made a million dollars you"

— 46:34

It is a seasonal mechanic rather than a permanent one. Outside that window it adds nothing and inside it the date is doing real work. An objection the audience actually holds can go in alongside it, with the outcome promised without that objection attached.

48:22 - Every title gets narrowed to one single subject

A hunch: drop the word "and"; make every title about one single subject.

"so we talked about one earlier is I think that people should not use"

— 48:22

The rule is borrowed from story structure — but and therefore, never and then — because and signals more than one topic and dilutes the title. Beats joined by but or therefore carry conflict, and and-then is boring. The borrowing is flagged as uncertain: I don't know if that's true or not.

49:39 - Titles have a science and an art

Titles have a science (modelling proven patterns) and an art (cleverness plus knowing your audience).

"single subject another thing we could talk about is like the art versus the"

— 49:39

Neither half substitutes for the other: modelling without judgement produces competent imitations, and judgement without models produces nothing repeatable. The science half can be taught in an afternoon, which is why almost everybody stops there.

51:16 - The art is knowing the audience's desire and wording it better

The art is knowing the audience's deepest desire and wording it better than they can.

"knowing your audience too so you could say how to make money online or"

— 51:16

That is the art half stated precisely enough to work toward, and it is why the research into what an audience believes matters more than any technique applied afterwards.

51:49 - Beyond the idea and the packaging there is the creator

Beyond idea, title and thumbnail there's the creator: most people never answer "why should I listen to you?".

"topic uh so the idea there's the title there's a thumbnail but there's also"

— 51:49

Most people never answer why anybody should listen to them. Beyond the idea, the title and the image there is the person behind them — and the question goes unanswered because answering it feels immodest, not because there is no answer.

52:18 - Audiences can feel whether somebody has been in the trenches

Audiences can feel whether you've been in the trenches; failing to prove it holds a lot of creators back.

"yes exactly I don't know like what it is about that but like people"

— 52:18

People sense first-hand experience against repackaged reading and allocate their attention accordingly. The failure case named is creators who never prove they have accomplished the thing their audience wants to accomplish — which no title can compensate for.

Personal Branding Lessons

A conversation about writing lines that get clicked. The moves below are the method.

Model creators a few steps ahead rather than the biggest names

Somewhere between five thousand and five hundred thousand followers is where the useful examples live, because those creators are still operating under conditions that resemble yours. What the largest accounts do works because of who they already are. 01:51

Keep a swipe file with the emotional pattern tagged

Title, publish date, URL — and the emotional pattern noted as you spot it. The tagging is what turns a collection into a resource, because the thing being reused is the emotion rather than the phrasing. 09:49

Study your own clicking and buying

When you click something or buy something, stop and ask why. Your own behaviour is the only sample anybody has unrestricted access to, and it is the one most people never examine while studying everybody else's. 18:55

Frame the fear as your own experience

Fear works and leaning on it feels wrong. Framing it as something you went through rather than as an accusation aimed at the reader uses the same emotional pull with the discomfort dial turned down — a craft answer to what is really an ethical problem. 22:04

Search a title idea before committing to it

Search the title, or a broader adjacent topic, and read whether it has worked for anybody before. Competing ideas get settled by evidence rather than by argument, and the evidence costs nothing to gather. 23:25

Keep one subject per title

Drop the commas and the ands. If you confuse, you lose, and no focus means no audience. Cutting from three listed things down to one concise idea lifted an email open rate materially, which is about as clean a test as this gets. 26:57

Wait five days before starting a test

A fresh video's first day is shown to loyal subscribers, who click at a rate nobody else will match. Testing inside that window measures the audience you already had rather than the one the title is supposed to reach. 31:37

Answer why anybody should listen to you

Beyond the idea, the title and the image there is the person behind them, and most people never answer this. Audiences can feel whether somebody has been in the trenches, and failing to show it is what holds a lot of creators in place. 51:49

Questions

Each answer ends at the moment in the recording where it is given.

What actually makes someone click a title?

An emotional response, not a piece of information. A plain, descriptive line gives a scroller nothing to react to, so nothing prompts the click; a line that carries a charge does. Three emotions do almost all of the pulling — curiosity, fear and desire — and they tend to work in pairs, usually curiosity with a fear of missing out or with a specific desire. The simplest working rule is to check whether a title makes anyone feel something before worrying about anything else. 10:43

Should I copy the titles of the biggest creators in my niche?

No. A celebrity can post a one-word title and win on reach alone, and copying that teaches nothing because the reach, not the craft, is doing the work. The better target is someone a few steps ahead — roughly five thousand to five hundred thousand followers — close enough that their moves are skill you can borrow rather than fame you cannot. The test that you have the right ladder is that their structures reproduce for you when the celebrity's would sink. 01:51

How do I research titles that actually work?

By hunting outliers and logging them. The platform shows when something was published and how many views it drew, so a piece that far outperforms its neighbours stands out as a signal worth studying. Scan a broad spread of industries, because the strong patterns are portable and turn up on unrelated channels, and record each one — title, date, the emotion it pulled — in a simple spreadsheet. A few hundred notes turn into a pattern you can name and reuse. 08:05

What is an open loop in a title, and how do I use one?

It is a title that starts a story without finishing it, so the reader needs the resolution and feels the pull of missing out if they scroll past. Something like "seven things experts do that you probably don't" opens the loop by naming a gap without filling it. It works by stacking curiosity with the fear of missing out, the two charges an unresolved loop pulls at once. Use it by withholding the one detail the title exists to promise. 13:05

How do I use fear in a title without feeling clickbaity?

Reframe it from an accusation into your own experience. Aimed at the reader — you will regret not doing this — a warning feels like an attack; turned inward — I regret doing this — the same emotion becomes a story being shared, and the pull survives without the discomfort. How hard to lean on it is a question of need rather than taste: a growing project can afford to serve gently, while a stalling one leans on what is known to work, still short of outright fear-mongering. 22:04

How long should a title or email subject line be?

Short enough to read at a glance and to survive truncation. One subject beats a list, because multiple topics, commas and conjunctions slow the read and split the focus until the line loses its force. Subject lines fight a harder ceiling than titles, since emotional openers eat words before the substance arrives, so staying within roughly six to eight words keeps the line from being cut off on a phone. Cutting a three-item line to one idea lifted a real open rate by about ten points. 26:57

When and how should I A/B test my titles?

Wait at least five days before starting, and read the results as a direction rather than a verdict. A fresh piece is shown first to loyal followers who click at the highest rate, so testing into that early spike produces inflated, unrepresentative data. Once it settles, treat the numbers as a compass: a theme that wins across ten tests is signal, an exact percentage gap is not. Test one variable at a time unless you are trying a wholly new idea. 31:37

How do I title something to rank in search versus the browse feed?

Match the line to the door. An interruption context — an inbox or a browse feed — rewards emotion, so a reframed, feeling-first line can double a click rate against a plain keyword version. A ranking context rewards the plain keyword instead. For search you can serve both at once: lead with the main subject or keyword so it ranks, then trail a trigger — a desire, or the current year — at the end to earn the click once it is found. 45:53

Does putting the year in a title still work?

Yes, within a season. A current year at the end of an evergreen tutorial is newsjacking a number already on everyone's mind, and it reads as fresh rather than dated; people also like clicking numbers. The effect is strongest from about November into January and decays every month after, because an out-of-date year signals old news. Plan evergreen tutorials to carry the coming year early in the calendar, and expect the pull to fade as the months pass. 46:34

Sources

Write Better YouTube Titles In 54 Minutes (Jake Thomas | Creator Hooks)

The sections above follow the episode's own order. Each timestamp is the point in the recording where the idea is discussed, so the post reads as a map of the conversation rather than a rearrangement of it.

Quotations come from the episode transcript with spoken filler ("okay", "right", "you know", "um", "actually", "kind of", "sort of") and false starts removed. No words are added: square brackets mark anything inserted for sense, and an ellipsis marks any cut. British spellings are restored, and speech-to-text errors were repaired against context only where the intended word is unambiguous. Where a line appears both in a cold-open teaser and in its place in the conversation, the timestamp cites the second — the point at which it is actually said.

Quotes are stamped with the moment they are said rather than with a speaker's name.