Learn Copywriting in 76 Minutes – Harry Dry

Harry Dry | David Perell

Abstract

Seventy-six minutes of copywriting worked through line by line, with a host pulling up real advertisements and asking what makes each one function. The spine is three questions run on every sentence — can I visualise it, can I falsify it, can nobody else say this — and three noes means the line is probably rubbish. Around that sit the two halves the craft is usually confused about: having something to say, and saying it well.

• If you cannot visualise it you will not remember it, and the route from abstract to concrete is rewriting downward until an object remains.
• A sentence that can be proved false puts the writer at risk, and that risk is why a reader sits up.
• The adjective is the failure and the fact is the fix — and a fact does not have to be numeric.
• The strength of an idea is inversely proportional to its scope, which is why the word and on a landing page is a warning.
• Simple writing is not written, it is rewritten: versions rather than drafts, four or five side by side.
• Your standards are your work — nine out of ten and above goes out, and it will cost you.

Everything below follows the conversation in order, from the three rules through positioning and the two-second test, into the writing process itself — the dirty water out of the tap, the two-line paragraph ceiling, the burrito test for a paragraph — and out to what a predictive tool cannot supply. For anyone building a personal brand the useful part is how mechanical most of it is: almost every rule here is a test you can run on one sentence, alone, in seconds.

Chapter summaries

00:00 - Three questions run on every sentence, and three noes mean trouble

Three questions run on every sentence: can I visualise it, can I falsify it, can nobody else say this — three no's and you are probably writing a lot of rubbish; three yes's and you are onto something.

"I know that you have these three rules that you apply to every sentence."

— 01:28

Three noes means the line is probably rubbish and three yeses means it is probably good, which is a crude instrument by design — it is meant to be run on every sentence rather than deliberated over. The same three questions separate a slogan any competitor could run from a line only one brand owns, which is why they are worth more than a style guide.

02:51 - Copywriting's reputation is deceptive tricks, and the craft described here is the opposite

If you cannot visualise it, you will not remember it.

"Rule one, can I visualise it? So let's start with a game."

— 02:51

The reputation is hucksters gaming psychology with deceptive tricks, and what follows is the opposite: a set of tests that make a claim checkable rather than persuasive. The first rule is visualisation, and the reason is memory — an unpicturable phrase has nothing to attach to. The abstract/concrete line is physical: an abstract thing cannot be dropped on your foot.

03:54 - The route from abstract to concrete is rewriting downward until an object remains

You get from abstract to concrete by writing the word at the top of a sheet and rewriting downward, asking what you actually mean, until you reach an object.

"So how do you go from abstract to concrete?"

— 03:54

The word goes at the top of a sheet and gets rewritten downward, each pass asking what the writer actually means, until a physical object is left at the bottom. It is a mechanical procedure rather than a knack, which is what makes it teachable — and the output is not a better adjective but a thing.

05:46 - A sentence that can be proved false puts the writer at risk

A sentence that can be proved true or false puts the writer's head on the chopping block, and that risk is what makes people sit up.

"Rule two, can I falsify it?"

— 05:46

A claim that can be proved false costs the writer something, and that risk is what makes a reader sit up. A claim nobody could disprove costs nothing and is therefore worth nothing. The checkable version is the one that carries weight.

06:38 - Pointing replaces every subjective descriptor with something checkable

You cannot talk, you can only point — replace every subjective descriptor with something checkable.

"Deep in a Reddit thread by a now deleted account, so I can't."

— 06:38

Pointing requires an object, so every subjective descriptor has to be swapped for something checkable. The instruction came from a forum post by an account since deleted, which is a fair description of where most usable craft advice actually lives.

08:21 - Pointing gets a line off the adjective trail, because it takes work

Pointing gets you off the adjective trail — and the reason it is better is that it takes work.

"When I say don't talk, only point, let's say I'm trying to."

— 08:21

The reason pointing beats describing is that it takes work: a fact has to be found, and an adjective only has to be chosen. That asymmetry is why the adjective trail is the default and why leaving it is a decision rather than a talent.

09:24 - An ad a competitor could sign forces no look at what is being sold

Never write an ad a competitor can sign; the test forces you to look deeper at what you are actually selling.

"Finally, final test: can nobody else say it?"

— 09:24

A line any rival could put their name to transfers preference to nobody, so the test forces a look at what is actually being sold rather than what is being claimed. It is the last gate rather than the first, and it fails a great many lines that pass the other two.

10:22 - Two identical products diverge on the wrapper and the words around them

Two products that look and taste the same can end as the best-selling bar in the world and a discontinued one; strip away the wrapper, the copy and the branding and businesses do roughly the same thing.

"Why should people learn copywriting? In this hand I got a Snickers."

— 10:22

Two products that look and taste the same ended as the best-selling bar in the world and a discontinued one, which strips the explanation down to the wrapper and the words. The foundational question follows from it: what can be done that nobody else is doing and that people care about. Positioning is the input the copy converts.

13:49 - Six words, white space and an enemy land instantly

Six words, a lot of white space, an enemy, and something only this brand could claim — and you get it instantly.

"They don't write songs about Volvos. What's going on here?"

— 13:49

You get it instantly, which is the whole test: no second pass, no explanation, nothing held back for a caption. Everything the three rules ask for is present in one line, which is why it works as a demonstration rather than an argument.

14:26 - An intuitive two-second read knows more than two hours of analysis

An intuitive test governs: sometimes you know more in two seconds than you know in two hours.

"There's an intuitive test with copywriting. Sometimes more in 2 seconds."

— 14:26

Sometimes you know more in two seconds than in two hours, and the two-second read is the one the audience actually performs. Analysis after the fact can justify a line that failed the first test, which is precisely the risk the intuitive read guards against.

15:44 - An ad pitched at the in-crowd risks losing the out-crowd entirely

If everyone gets it, you are possibly doing something wrong; not getting it is fine if you talk about it — what is worse is a line already said.

"I think if everyone gets it, you're possibly doing something wrong."

— 15:44

An ad pitched at the in-crowd risks losing the out-crowd entirely, and if everyone gets it you are possibly doing something wrong. Not getting it is survivable if people talk about it; the failure that is not survivable is a line nobody notices either way.

16:38 - Feature-led copy loses to a line about what the buyer gets

Feature-led copy loses to a line about what the buyer gets.

"These Mac and PC ads — this is the problem with PC ads."

— 16:38

It is the oldest rule in the trade and still the most broken, because a feature answers a question nobody asked while feeling like diligence to write. The benefit is what the buyer is counting.

17:07 - Metonymy substitutes a visual word for the literal one

Metonymy: substitute the literal word for a visual one, even where the sentence stops being literally true.

"It's also a metaphor. It's an example of metonymy. What's metonymy?"

— 17:07

A visual stand-in is more surprising and more punchy than the accurate term, and the reader supplies the correction for free. The literal version names the device; the famous version names where you carry it — and he grants the sentence does not literally make sense.

18:15 - A weak product becomes the intelligent choice when the buyer feels smart

Make the customer feel they are making a smart decision, and a weak product becomes the intelligent choice.

"I love this, because who said this? I'm paraphrasing someone here."

— 18:15

The frame moves the judgement from the object to the buyer's self-image. A plain saloon was sold to people who inherited brains instead of wealth, with the alternative reframed as a quarter of a million dollars you could crash. He flags that he is paraphrasing someone and cannot place the source.

19:11 - An ad built on one element beats an ad built on four

An ad built on one element rather than the assumed four — header, subheader, picture, logo — is the winning version, because an ad that does not look like an ad gets read.

"I've never read The Economist. Management trainee, aged 42. What's going on here?"

— 19:11

Most people think an ad needs four elements — header, subheader, picture, logo — and doing it with one is the winning version. The single-element ad forces the idea to carry the whole thing, which is also why it exposes an idea that cannot.

20:19 - Juxtaposition creates the contrast that makes a line work

Juxtaposition creates the contrast that makes a line work, and demand can be manufactured by welding a product to an occasion.

"How can you make two months' salary last forever? The diamond engagement ring."

— 20:19

Two incompatible terms in one sentence create a tension the reader resolves by reading on. A 1948 line set an amount of salary against forever, for a diamond company, at a time he says wedding rings were other stones — with the writer's name and the year both hedged, and the resulting share of engagement rings given as I think eighty-five per cent.

21:28 - Writing starts from the current attitude and the desired one

Writing starts from the current attitude of the consumer and the desired attitude — point A and point B; you cannot start a race in the middle.

"I start with, what am I trying to do?"

— 21:28

Naming both ends converts a vague brief into a distance to be crossed. Two telephone poles first, then the wire strung between them. The attribution is hedged: he apologises in advance for possibly getting the name wrong.

22:02 - The first piece is who the writing is talking to

Piece one is who you are talking to, and getting it wrong wastes the whole spend.

"Well, there's three pieces here. First one, piece one: who you're talking to."

— 22:02

Getting the audience wrong wastes the whole spend, which is why it is the first piece rather than a refinement. The audience question is answered before the writing begins, not discovered during it.

22:53 - Form follows function, so the winning sign is read from the corner of an eye

Form follows function: the sign that wins is the one a distracted passer-by reads out of the corner of their eye.

"One more example. I'm in St Ives last year and doing a coffee run."

— 22:53

The context of reading sets the constraints, and a line optimised for a document fails on a street. Length, contrast and word choice are all downstream of where the thing will be read and how much attention is going spare there.

24:11 - The writer spends hours in the mode and the reader spends none

The writer spends hours in the mode of thinking and the reader spends none, so the work has to be judged in reality — mocked up among the competing ads with people walking past.

"So when you're writing copy, how do that something is going to instantly."

— 24:11

The writer spends hours in the mode of thinking and the reader spends none, which is the single most common source of copy that works in a deck and dies in the world. The fix is to judge it in reality: mocked up, at size, in place. Compose in the tool the reader will read it in, never in a document first.

25:15 - Having something to say means building for more than the money

Piece two is having something to say, which means building for more than the money — a belief, a vision, or something you think is wrong with the world.

"Piece two. You got to have something to say."

— 25:15

The practical form is counter-positioning: find what nobody is doing and build the opposite, which is how this particular body of work started.

27:06 - Saying it well is a separate job, and the brief is never the line

Piece three is saying it well — quickly, visually, persuasively; the brief is never the line, and turning it into something memorable is the writer's job.

"Peace free. Peace free to peace free said it well. So this is what."

— 27:06

Saying it well is a separate job from having something to say, and the brief is never the line. Turning the brief into something memorable is the work, and conflating the two is why so much copy reads as a restated requirement.

29:19 - A seed kept because it was true and achievable starts the work

Work starts from a seed: a line seen years earlier and kept because it was true and achievable.

"So I want to show you this for real. I want to show you what."

— 29:19

The archive is the raw material, and the discipline is keeping things long before there is a use for them. A line kept because it was true and achievable is one that can be tested against later work rather than merely admired.

30:07 - Three types of enemy sit behind any conflict worth writing

A line down the middle of the page and any two conflicts that come to mind is the whole exercise, and the three types of enemy behind it are different approaches, beliefs and competitors.

"Great copywriting tip. We were talking about conflict earlier."

— 30:07

The exercise is a line down the middle of the page and any two conflicts that come to mind. Different approaches, different beliefs and direct competitors are the three sources, and any of them will generate the contrast a line needs.

31:20 - What a writer takes in is what they have to build with

What you take in as a writer is what you have to build with.

"Now at this point, this is why I feel like what you take."

— 31:20

What a writer takes in is what they have to build with, which makes the reading list an input to the output rather than a hobby. The ceiling is set by the diet.

32:39 - Liking a draft is not the same as signing it off

You can like a draft and it can still get far better; liking something is not signing it off.

"At the time I wrote this, I was like, you."

— 32:39

Stopping at there's something here is the common failure, and the gain past that point is large — he says it got about two hundred per cent better after the version he already liked, and that he kept walking down what might have been a blind alley. The standard is refusal rather than enthusiasm: I wouldn't sign this off.

33:55 - A line that wraps to two is a defect worth restructuring for

A line that wraps to two lines is a defect worth restructuring the whole piece to fix.

"At this point something was really irritating me. I could not put this out."

— 33:55

It sounds disproportionate until you consider that the wrap is evidence the idea did not fit in the first place. The fix is upstream of the typography, in the thought the line was trying to carry.

34:31 - Solving one problem creates another, and that trade is the work

Solving one problem creates another — one forward, one back — and that trade is the work.

"1% to 2%. This creates a problem."

— 34:31

One forward and one back is the normal rhythm rather than an interruption to it. Every fix moves the defect somewhere else, and the craft is in choosing where.

34:52 - Spitballing runs the dirty water out of the tap without judging it

Spitball without judging: you are running the dirty water out of the tap, and you have to trust that it gets clean.

"Forward, one back. So then I think, all right, I need to write something that's."

— 34:52

Judgement and generation compete for the same attention, so suspending judgement raises volume, and volume is what reaches the usable line. Five discarded attempts written in a row, none of them good, arrive at the phrase he keeps — and he calls them not good while writing them, which makes the bad versions the method rather than the failure.

35:45 - Copy and design are one act, composed in the tool it will live in

Copy and design are one act; he never works anywhere but in the tool the thing will live in.

"One thing that's really revealing. You just said I've been writing this for two to."

— 35:45

He never works anywhere but in the tool the thing will live in, so the words are judged in the setting that will carry them. Feedback comes from people whose taste is respected, and an idea you liked can be tried and reverted — which requires the versions to be cheap.

37:06 - Design is the last pass, and it pulls out rhythm already in the lines

Design is the last pass, and it pulls out rhythm that is already in the lines.

"But the last thing I do in most ads is I try and add."

— 37:06

Whatever rhythm it finds was already in the lines. Illustrations are iterated to serve the argument rather than to fill the space, which is a distinction most decks do not survive.

39:46 - The claim stays small enough to be believed

The claim stays small enough to be believed — a one-point lift, not a fortune.

"Also, what I like about this is it's not a big claim."

— 39:46

A modest number is checkable and a large one invites disbelief, so the smaller claim does more work. The ad took two days and twenty-five rewrites, built piece by piece.

41:19 - Expanding the time frame makes a small number matter

Expand the time frame to make a small number matter.

"I like this ad from you. Here's what it says. Marketers, you'll."

— 41:19

Expanding the time frame makes a small number matter — hours across a career rather than hours this week. The same line flips the frame from a cost to a saving without changing a single fact in it. The framing is the argument.

43:07 - An unfamiliar product makes sense against products already known

Make an unfamiliar product make sense by comparing it to products already known.

"Tell me about this ad that you made. Better truck than an F-150, faster."

— 43:07

An unfamiliar product makes sense by comparison to products already known, which is how a new category borrows the reader's existing furniture rather than asking them to build some.

43:45 - A concrete comparative replaces the abstract one and buys parallelism

Replace the abstract comparative with a concrete one; the shorter version also buys parallelism.

"Based on what I've learned from you, there's two things I would do to improve."

— 43:45

Concreteness and compression turn out to be the same edit: cutting the vague word removes a noun and lands the symmetry. A comparative rewritten to a single physical adjective drops a word and makes both halves scan. Shorter is better in copy most of the time — hedged, not absolute.

44:29 - The first line's only job is to make the second unavoidable

First line, second line: the first line's only job is to make the second unavoidable.

"Read that one. This from the Ad Professor. It takes 3.1 seconds."

— 44:29

An unresolved statement is a question the reader has to answer by reading on. An ad states how long it takes to read itself, and the second line resolves it by naming what else takes that long; the precision of the figure makes it feel exact, and it breaks the fourth wall. The attribution passes through two names, both hedged.

45:26 - A fact guarantees that something was said rather than word-shaped air

A fact guarantees that you have said something; most people open their mouths and say nothing — wallpaper, word-shaped air.

"What do you think of facts? Do you like facts?"

— 45:26

A fact is checkable, so it cannot be empty, where an assertion can be — a stated annual sales-and-marketing spend, a writer who lived on the road seven years and wrote the book in three weeks. He offers the counter-case himself: the population of a country is a fact and is not interesting. The boundary is explicit — you have to work with a fact.

47:16 - A fact does not have to be numeric

A fact does not have to be numeric.

"To be numeric. I'll give you a fact."

— 47:16

A place, a name, an act, a date and a count all qualify, which widens the rule considerably. The requirement is checkability rather than arithmetic, and that distinction matters because the numeric reading of the rule is what makes writers conclude they have nothing to work with. Almost every subject has facts available to it that were never counted.

47:54 - The adjective is the failure and the fact is the fix

The adjective is the failure; the fact is the fix — give me a fact instead of a judgement.

"So inside of this, though, is an implicit critique of how most people communicate."

— 47:54

The adjective is the failure and the fact is the fix, which is the whole critique in six words. A judgement asks the reader to agree; a fact gives them something to do.

48:23 - Facts give a commentator weight, and their absence reads as imprecision

Facts give a commentator weight and credibility, and their absence is a lack of precision.

"Now this is a great example. I'll use a golf analogy."

— 48:23

Specific quantities can be checked and compared; vague effort claims cannot. The golf commentary constructed on the spot runs averages per round, a percentage difference and a likelihood, against a claim that a player did not want it enough — and it is a constructed example rather than a report of one. Quoted alongside it: modern writing does not consist in picking out words for the sake of their meaning.

49:21 - A writer keeps facts in their arsenal about whatever they write on

A writer should keep facts in their arsenal about whatever they are writing on; turn the adjective into a reality.

"So the point is, as a writer, as a communicator, you should have some facts."

— 49:21

Each adjective gets turned into a reality as it appears on the page. It is a preparation habit rather than an editing one. On a newsletter the number worth talking about is engagement rather than subscriber count.

50:45 - Every issue is a struggle, and the form is a letter rather than news

Every issue is a struggle, and the form is a letter, not news — which is why it opens and closes rather than just listing.

"I mean, it's tricky. And honestly, every newsletter, genuinely, is."

— 50:45

The distinction determines the voice: a letter has a sender, and news does not. Each intro opens on time, place, and what is going on in the writer's life.

52:05 - Density is five or six examples in a single issue

Density: five or six examples in a single issue.

"Firstly, you get five things in a newsletter, six things. You'll get five examples."

— 52:05

Density is five or six examples in a single issue, which is a lot by any normal standard and is the reason to open it. The examples are the product; the commentary is the wrapper.

52:15 - Conflict appears in nearly every example, because memory works relatively

Conflict appears in pretty much every example, because human beings remember things relatively.

"Secondly, we talked about this a lot today, but there's conflict in pretty much every."

— 52:15

Nothing is memorable on its own — it is memorable against the thing beside it, which is why the contrast is engineered rather than found. Human beings remember relatively, so the writer's job is to choose what the thing is standing next to.

52:45 - Naming a new category beats competing as an easier version of a commodity

Open on how everyone else does it, then on how this one does.

"I did one recently about positioning."

— 52:45

Opening on how everyone else does it and then on how this one does is the compressed form of the same move. Naming a new category beats competing as an easier version of a commodity, because the comparison the reader would otherwise make is removed.

53:44 - No paragraph runs longer than two lines, because short paragraphs are monkey bars

No paragraph longer than two lines — short paragraphs are monkey bars, a third line usually means the explanation failed, and lines need room to breathe.

"Third thing, I only noticed this the other day. I was reading back previous issues."

— 53:44

Three mechanisms run at once — swing-ability for a distracted reader, a diagnostic for unclear thinking, and emphasis through white space. He says he noticed it reading back issues and has not broken it since a stated month and year, and offers it as idiosyncratic to himself rather than as good or bad.

54:37 - The intro is a live report from the desk

The intro is a live report from the desk: the hour, the room, the drink, the unfinished item, the doubt, the send.

"Walk me through this intro. Hey, it's 3:47 a.m. I'm in the big."

— 54:37

The hour, the room, the drink, the unfinished item, the doubt, the send: it is specific enough to be unfakeable, which is what makes it read as a letter rather than as a format.

56:32 - Simple writing is not written, it is rewritten

You cannot write simply; you can only rewrite simply.

"Writing simply — what we're really talking about here is rewriting."

— 56:32

Simplicity is a subtraction performed on an existing draft, so demanding it at the first pass blocks the draft from existing. It is named as the first thing he learned about writing simply.

56:53 - Any words that are not working for a piece are working against it

Any words that aren't working for you are working against you — and the common failure is not taking that seriously enough.

"First one, my favourite: Kaplan's Law of Words."

— 56:53

The common failure is not taking that seriously enough. There is no neutral word: the ones that do nothing are costing the reader something.

57:16 - The strength of an idea is inversely proportional to its scope

The law holds for ideas as well as words: the strength of an idea is inversely proportional to its scope.

"Kaplan's Law of Words is true for words, it's also true for ideas."

— 57:16

The law holds for ideas as well as for words. A claim about everything is a claim about nothing, and narrowing it is what makes it land. The narrow version is the strong one.

57:27 - The word and on a landing page is normally never a good thing

The word "and" on a landing page is normally never a good thing; do one thing well.

"This is why the word and on a landing page is normally never a good."

— 57:27

And is the visible symptom of a scope that has widened past the point of force. A jeans maker's tagline is that it makes jeans; an album cover that is one image sits against one he calls among the worst, made of seven disparate ones. The rule is hedged as normally never.

58:18 - A good paragraph can be thrown and caught without coming apart

A good paragraph is like a burrito — you should be able to throw it and have it caught without it coming apart.

"We talked about simple sentences there. So I'd talk about simple paragraphs."

— 58:18

A good paragraph can be thrown and caught without coming apart, which is a test rather than a metaphor. Pull one sentence out: if the paragraph still works, that sentence should not have been there.

58:58 - Versions rather than drafts buy the freedom to be terrible

Write versions, not drafts — four or five of the same paragraph side by side; the real payoff is the freedom to be terrible.

"Third thing I talk about when I'm talking about writing simply."

— 58:58

Four or five of the same paragraph side by side buys the freedom to be terrible in some of them. Versions also change the feedback: one paragraph returns a verdict, and three return a specific edit.

1:00:29 - Structure is dividing lines and parallelism

Structure is just dividing lines and parallelism.

"How do I structure things?"

— 1:00:29

It is a deflationary account of something usually treated as architecture. A line can be the structure: the first step to ten thousand readers is a hundred readers. The dividing line sets up the comparison and the parallelism makes both halves audible in one pass.

1:01:22 - A 1971 paper gives a formula for framing an idea as interesting

A 1971 paper offers a formula for framing an idea as interesting: take two things thought different and show they are one, or take one thing and show it is two.

"And parallelism. There's a paper that was published in 1971. It's called That's Interesting."

— 1:01:22

Take two things thought different and show they are one, or the reverse. A book without chapters is long and heavy, and structure is something for the reader to hold onto.

1:02:14 - Two rules: a high-quality bar, and never letting it fall

Two rules: learn how to have a high-quality bar, and never let it fall — and it will cost you.

"Another rule, and this is something that I value too. Many years ago, I."

— 1:02:14

The cost is named in the same breath as the rule. Your standards are your work, and your work is nothing but your standards. Nine out of ten and above goes out, and otherwise nobody sees it. The bar is the whole of the reputation.

1:03:34 - Leaving and coming back is how a piece is known to be done

You know it is done by leaving and coming back: if nothing can be taken away, it works.

"I think it's when you go for a walk around the garden."

— 1:03:34

Distance restores the reader's eye, and subtraction rather than addition is the test that ends the work. A walk around the garden and then the return — and the trainer-brand line he says you can neither add a word to nor take one from.

1:04:15 - Conflict is created by hinging the story on but without writing the word

Create conflict by hinging the story on "but" without writing the word.

"I'm basically trying to hinge something around the word but."

— 1:04:15

Conflict is created by hinging the story on but without ever writing the word, which is the difference between structure and signposting. The reader feels the turn rather than being told about it.

1:05:16 - A tool is only as good as the taste behind it

A tool has never made the work better; a tool is only as good as the taste behind it.

"What went wrong? That's conflict. I got to ask, how does AI."

— 1:05:16

Capability applied without taste destroys the thing it is applied to, which is a claim about amplification rather than about the technology.

1:06:21 - Taste, conviction and experience are unavailable to a predictive tool

Taste, conviction and experience are the three things named as unavailable to a predictive tool.

"At 60 mph the loudest noise coming from the."

— 1:06:21

The pattern in good writing is lived time — the work comes out of the years spent in it — and two different kinds of experience are distinguished: the noticing kind and the accumulated kind.

1:08:29 - A machine that sees nothing and believes nothing cannot do the work

Writing is: you see something, you believe something, you write about it — and a machine that sees nothing and believes nothing cannot tell you something surprising.

"What's writing? You see something, you believe something, you write about it."

— 1:08:29

Seeing something and believing something come before writing about it, and a machine does neither — which is the argument, rather than an objection to the output. Collecting excellence and letting the inspirations set the standard is offered as the working alternative.

1:09:23 - Silence and action give a better start than a feed

Silence and action: a feed first thing makes the day drudge; silence, no talking, and a walk with no earphones give you a better shot.

"It's in the everyday. You get up, you."

— 1:09:23

Silence, no talking, and a walk with no earphones give a better start. It is the last thing said and it is consistent with everything before it: the input sets the ceiling, and attention is the input.

Personal Branding Lessons

Seventy-six minutes of craft, almost all of it reducible to tests you can run on a single sentence without asking anyone's permission or opinion.

Run the three questions on every sentence

Can I visualise it. Can I falsify it. Can nobody else say this. Three independent failure modes — unpicturable, unprovable, unowned — and a line can clear one while failing the others, which is why they are asked separately rather than as a general sense of quality. Three yeses is probably something good; three noes is probably rubbish. The same three questions also separate a slogan any competitor could run from a line only one brand owns. 00:00

Rewrite downward until an object remains

Put the abstract word at the top of a sheet and rewrite it downward, each pass asking what you actually mean, until what is left at the bottom is something you could drop on your foot. The procedure is mechanical rather than a knack, which is what makes it teachable — and the output is not a better adjective, it is a thing. If you cannot visualise it, you will not remember it. 03:54

Replace every subjective descriptor with something checkable

Imagine you cannot talk and can only point. Pointing requires an object, so every judgement has to be swapped for something that can be verified. The reason it produces better copy is unglamorous: it takes work. A fact has to be found and an adjective only has to be chosen, which is exactly why the adjective trail is the default and leaving it is a decision rather than a talent. 06:38

Put a competitor's name under the line

A line any rival could put their name to is transferring preference to nobody, so write the line, put a competitor's logo underneath it, and see whether anything breaks. The test is the last gate rather than the first, and it fails a great many lines that pass the visualise and falsify tests cleanly. What it forces is a look at what is actually being sold rather than at what is being claimed. 09:24

Compose in the tool the reader will read it in

The writer spends hours in the mode of thinking and the reader spends none, which is the most common reason copy works in a document and dies in the world. Judge it in reality — mocked up, at size, in the place it will actually appear. Never write it in a document first and move it afterwards. Copy and design are one act, and the sign that wins is the one a distracted passer-by reads out of the corner of their eye. 24:11

Give a fact instead of a judgement

The adjective is the failure and the fact is the fix. A judgement asks the reader to agree with you; a fact gives them something to do. A fact also guarantees you have said something at all — most communication is wallpaper, word-shaped air — and it grounds the claim, makes it real, and hands you something to tell a story from. It does not have to be numeric: a place, a name, an act and a count all qualify. 47:54

Keep no paragraph longer than two lines

Short paragraphs are monkey bars: they carry the reader forward by giving them something to reach. A third line usually means the explanation failed upstream, so the overrun is a diagnostic rather than a formatting problem. A good paragraph should survive being thrown and caught without coming apart — and pull one sentence out of it: if it still works, that sentence should not have been there. 53:44

Write versions rather than drafts

Four or five of the same paragraph side by side, not one paragraph revised five times. The real payoff is the freedom to be terrible in some of them, which is what unlocks the range that a single precious draft suppresses. Versions also change the feedback you get: one paragraph returns a verdict, and three return a specific edit. Simple writing is not written, it is rewritten. 58:58

Set a quality bar and never let it fall

Two rules: learn how to have a high-quality bar, and never let it fall — and it will cost you, which is said plainly rather than softened. Your standards are your work, and your work is nothing but your standards. Nine out of ten and above goes out; otherwise nobody sees it. Know it is done by leaving and coming back — if nothing can be taken away, it works. 1:02:14

Questions

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

How do I know whether a line I have written is any good?

Run three tests on it. Can you picture it, can it be proved true or false, and could a competitor sign it unchanged? Three no's means you are probably writing a lot of rubbish; three yes's means you are onto something. The value of the filter is that each question catches a different defect, so a line can be vivid and still borrowable, or exclusive and still unprovable. It takes seconds and it is brutal. 01:28

How do I make vague writing concrete?

Put the vague word at the top of a sheet, write concrete at the bottom, and rewrite downward, asking at each rung what you actually mean, until you arrive at an object you could drop on your foot. Regaining fitness descends through getting off the couch and running five kilometres to a two-noun name. The common failure is stopping one rung early, at something like pretty people in big cities, which he grants can kind of be seen, where a supermodel is unmistakable. 03:54

What makes a claim believable instead of just bold?

Falsifiability. A sentence that can be proved true or false puts the writer at risk, and it makes readers sit up a little in their seat, ears pricked. The practical route is to stop talking and start pointing: instead of calling a thing a good investment, show fifty years of its chart. That takes research, and the research is precisely why it beats the adjective — the research is what a reader's trust is actually built on. 06:38

How do I write something a competitor could not also say?

Apply the exclusivity test after the line already works: could a rival run it tomorrow, unchanged, having done none of the work? If so, it is describing the category rather than you. The escape is to look deeper at what is actually being sold, which usually means going and finding something only you have seen. Whatever is unique to you is where the unsignable lines are. 09:24

Why should I learn copywriting if I am not a copywriter?

Because two things that look and taste the same can end as the best-seller and the discontinued one. Strip away the wrapper, the copy and the branding and most offers do roughly the same thing, which puts the difference in the layer above the product — and in which one is already in the reader's mind. The foundational question is what you can do that nobody else is doing that people care about — a positioning question that a weaker asset can win on with a little storytelling. 10:22

How long should it take someone to get what I have written?

Two seconds, tested by counting one Mississippi, two Mississippi against the finished thing. The reader's verdict is made at glance speed, so the writer's verdict should be made there too — sometimes you know more in two seconds than in two hours. For a reader who is walking past rather than sitting down, the threshold is tighter still: not two seconds, but instant. Attention is decided before comprehension. 14:26

Where do good lines actually come from?

From things collected long before they were needed. The advertisement built in the episode starts from a line seen roughly four years earlier and written down because it was true and achievable, and its layout comes from an older advertisement he says he admired a couple of years before. What you take in is what you have to build with, which makes reading published work in your own field a raw material rather than a hobby. 29:19

How do I make a small number sound like it matters?

Expand the time frame instead of changing the number. Four hours a day is unremarkable; multiplied across three hundred days and a thirty-year career it becomes tens of thousands of hours, and the same quantity now carries weight. Note the honesty attached: the sum produced twenty thousand and the poster says twenty-two thousand, rounded on purpose to rhyme with the second half of the line. 41:19

How do I write simply?

You do not write simply; you rewrite simply. Simplicity is a subtraction performed on something that already exists, so demanding it at the first pass only blocks the pass. Then apply the law: any words that are not working for you are working against you, and the same holds for ideas, whose strength falls as their scope widens. The word and on a landing page is the visible symptom. 56:32

How do I know when a piece is finished?

Leave it, walk around outside, come back and ask whether anything can be taken away — not whether anything is missing. When nothing can go, it works. The threshold above that is explicit: nine out of ten or better goes out, and anything below it nobody sees. Holding that line costs you something socially, which is the part usually left out of advice about holding your own standard. 1:03:08

Sources

Learn Copywriting in 76 Minutes – Harry Dry

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.