Live "Office Hours" with Harry Dry and Sabba Keynejad ✨

Harry Dry | VEED STUDIO

Abstract

An hour and a half of public critique: submitted adverts and social posts played and judged live by a copywriter and a video-platform operator, with the makers watching. The two perspectives pull in useful directions — one asks what the line is doing, the other asks what the numbers do — and they converge on the same diagnosis more often than not. Almost every verdict comes back to the opening, the stakes, or the absence of anything to say.

• There are two approaches to a hook, and which one applies is decided by the material rather than by taste: promise it when the substance is strong, open a loop when it is thin.
• Naming the product in the hook ends the transaction; delete it and the same line becomes a situation the viewer wants finished.
• Tension is carried by but, never by and — put imaginary buts between your points and see whether they connect.
• Production quality is not a hook, and nothing said well is still nothing.
• If any other subject could be swapped into the advert without changing it, it is not an advert for this product.
• Show the client rather than yourself, and film the reaction rather than the claim.

Everything below follows the session in order, through the anatomy of hooks, the stakes that hold a viewer, and then submission by submission — a van insurer, a guitar course, an explainer video, a gardening company, a local trades business — with the verdicts attached to the specific defect each time. For anyone building a personal brand the useful part is the register: these are the notes you would get if somebody watched your work honestly and said where they stopped.

Chapter summaries

01:06 - The session's form is public critique, with submitted work judged live

Public critique is the session's form: submitted adverts and social posts are played and judged live, with more taken from the chat as it runs.

"One thing is, we've had video."

— 01:06

Work is played and judged in front of the people who made it, which sets the register for everything that follows: the verdicts are specific, occasionally blunt, and always attached to a named defect rather than to a rating.

02:16 - The media keep changing and the people do not

The media keep changing and the people do not, so the question that mattered when readers flicked through a magazine is the question that matters when they scroll.

"I probably learned more about advertising from him."

— 02:16

The question that mattered when readers flicked through a magazine is the question that matters on a feed, because the substrate changed and the psychology did not. The old tests still apply, which is why examples from decades ago are still doing work in a session about short video.

03:10 - A promise is a handshake the content has to honour

A promise is a handshake: if the content does not deliver what the line said, the viewer feels betrayed.

"When I say promise, what I mean is."

— 03:10

The promise creates an obligation the piece has to discharge, which is what makes it a strong claim on attention rather than a description. A title that names the thing the viewer will actually see, with the thumbnail proving it up front, has already paid part of the debt.

03:46 - The device is decided by the material rather than by taste

Which device to use is decided by the material: when the substance is good, promise it; when it is thin, withhold it and open a loop.

"One is an open loop. Now you decide as a creator."

— 03:46

There are broadly two approaches, and which one applies is decided by the material rather than by preference. When the substance is genuinely strong, promise it; when it is thin, open a loop instead. Choosing the wrong one commits you to a debt you cannot pay.

04:39 - A creator who visibly delivers builds trust in the next promise

A creator who promises something extreme and then visibly does it builds the audience's trust in the promise itself.

"We'll start off by talking about a promise. Videos which make."

— 04:39

What accumulates is not trust in the creator but trust in the promise: the next extreme claim is believed because the last one was kept on camera. The device is not a trick in that case — it is a record of delivery, and the record is what the audience is actually responding to.

05:02 - Where there is something to work with, the lead is not buried

Where there is genuinely something to work with, there is no need to bury the lead.

"This guy runs a body armour company, and there's no."

— 05:02

The strongest version states the matchup and withholds the outcome, so a promise and a loop run at the same time: the viewer knows what they are getting and still does not know how it ends.

06:53 - Juxtaposition states a point A and a point Z and leaves the middle empty

Juxtaposition works because it states a point A and a point Z and leaves the space between them empty; that space is the loop.

"And a point zed. A being this person's lazy, and then zed."

— 06:53

Juxtaposition puts two things together that should not go together, stating a point A and a point Z and leaving the space between them empty. The brain wants the middle filled. The gap is the whole mechanism, and neither end is interesting by itself.

07:14 - Doing the opposite of the category is the second way to open a loop

The second way is to do the opposite of whatever the category does.

"To click for. Another idea, this is actually what I."

— 07:14

Inverting the obvious noun in a category's stock line produces a usable hook immediately, because the audience's filing shortcut fails and they have to look. It costs nothing to try, which is rare among the moves in this session.

07:55 - Storytelling is the art of withholding information

The third way is withholding, and storytelling is basically the art of withholding information.

"Now finally, I want to talk about withholding information."

— 07:55

Storytelling is basically the art of withholding information, which is why the third route is the most transferable. Fit the product into a story somebody already cares about rather than showing the product, and the withholding does the work the pitch cannot.

08:52 - Deleting the product turns the same line into a situation

Delete the product from the hook and the same line becomes a situation the viewer wants finished.

"So what if we just delete the product?"

— 08:52

Naming the product in the hook ends the transaction: the viewer classifies it as an advert and stops. Delete the product and the same line becomes a situation the viewer wants finished, which is the entire difference between something skipped and something watched.

09:32 - Adverts open loops rather than make promises, because nobody cares about the product

Adverts should open loops rather than make promises, because the audience does not care about the product enough for a promise to be worth anything.

"Don't start with the product. In advertisements it's more likely."

— 09:32

A promise trades on desire that already exists; where there is none, the attention has to be taken before anything can be promised. It is stated as what you are more likely to want — a lean rather than a rule.

09:43 - A good hook tells about eighty per cent and leaves twenty

A good hook tells about eighty per cent of the story and leaves twenty per cent for the viewer to work out.

"Attention. My wife's boyfriend left these special chocolates for me."

— 09:43

The twenty per cent left over is what the viewer stays to resolve; withhold much more and there is nothing to be curious about yet. The proportion matters: too little and there is nothing to follow, too much and there is nothing left to want.

10:10 - Raising the stakes is universal to hooks and to storytelling

Raising the stakes is universal to hooks and storytelling: something has to be at risk.

"Two points which are more universal in hooks and storytelling."

— 10:10

Something has to be at risk for anyone to stay, which is as true of a thirty-second advert as of a novel. Without a gamble there is no reason to care how it ends, and attention decays back to where it was.

10:46 - A public challenge is interesting because of the chance of failure

A public challenge with a real chance of failure is interesting because of the chance of failure.

"Raised. It's a real person. He has started this challenge recently."

— 10:46

Remove the possibility of losing and there is nothing to watch: the audience is there for the outcome, not the attempt. The outcome is genuinely undetermined, which is what holds attention rather than the content of the activity. The risk is the content.

11:24 - Vulnerability at the opening outperforms competence at the opening

Vulnerability at the opening outperforms competence at the opening.

"You will probably have seen this. It went incredibly viral."

— 11:24

Opening on your own failure raises the stakes further than opening on your work, and vulnerability at the opening outperforms competence at the opening. Real exposure cannot be simulated cheaply, so it reads as a genuine risk being taken in public.

12:09 - Tension is carried by but, and never by and

Every story runs on tension, and tension is carried by "but", never by "and".

"Finally, every story has stakes. You got to raise them."

— 12:09

A commercial becomes a story at the point where the direction changes, and the same sentence can be rewritten with tension without changing a single fact in it. The conjunction is the structure.

13:24 - Imaginary buts and sos test whether the points connect at all

Put imaginary "buts" and "sos" between your points and see whether they connect; if it all reads as "and", it is a shopping list.

"Tension. There's a page which thrives off this."

— 13:24

If every join reads as and, there is no story in the piece — only an inventory. The A-to-Z structure is what makes a career-change advert clickable, because the gap between the two states is the whole draw.

15:51 - The first three seconds lead with something emotional, visual or unexpected

The first three seconds are critical, and they must lead with something emotional, visual or unexpected.

"So let's jump into them right now. The first one is hook them."

— 15:51

Emotional, visual or unexpected: any of the three will do, and nothing else will. It is the operator's version of the same claim the copywriter has been making, arrived at from the other end — from what the numbers do rather than from what the craft says.

16:16 - A dry topic becomes interesting restated as the saving it produces

A dry instructional topic becomes interesting when the same content is restated as the saving it produces.

"And when we think about the tools that we."

— 16:16

Nothing about the material changes; what changes is which end of it the first line starts from, which is the cheapest edit available. The saving is the reason anybody sits through the instruction, so it belongs at the front rather than in the summary at the end.

16:41 - Hooks are collected rather than invented

Hooks are collected, not invented: write down good ones when you hear them and adapt them to your own industry.

"And you can get good at this by literally just."

— 16:41

A swipe file is the working memory of the craft, and it removes the blank page from the job. Adapting an existing hook to another industry is also a test of whether it was structural or merely topical.

17:03 - A short video has one hook and one payoff, and a long one needs several

A short video has one hook and one payoff; a long one needs several of each.

"Quite long, you might need to remind them what the hook is, because they've got all."

— 17:43

The reason is memory rather than pacing: the audience forgets the hook, so it has to be restated. Length changes the structure of a piece rather than merely stretching it.

17:56 - One clear message produces higher retention

One clear message produces higher retention, and everything not serving it should be cut.

"Of engagement content. And the final tip, and I think this is."

— 17:56

Everything not serving the one message should be cut, which is easy to state and hard to apply, because the material that gets cut is usually the material somebody is fond of.

20:16 - A swap test exposes an advert that is not about this product

If any other subject could be swapped into the advert without changing it, it is not an advert for this product.

"The problem with this setup is this whole video is not bespoke to van."

— 20:16

The test is mechanical and brutal, and it fails a great deal of work that looks finished. What it exposes is that the piece was built around a category rather than around this product, so nothing in it is unavailable to a competitor.

20:37 - Provocation is legitimate only when it stems from the product

Provocation is legitimate only when the provocation stems from the product.

"Insurance. You've created a tension you have, which I like."

— 20:37

Provocation attached to nothing is noise that costs trust, and the audience can tell the difference between an argument and an attempt to be noticed. Where it does stem from the product, the provocation and the proof are the same sentence, which is why it survives the reaction it causes.

21:03 - The method is to brainstorm conflicts true of the product and dramatise one

The method is to brainstorm conflicts that are true of the product, then dramatise one of them.

"For example, if I was brainstorming a conflict for this, it would be."

— 21:03

The constraint is in the first half: the conflict has to be true of this product, or the dramatisation is a story about something else. Starting from the conflict rather than from the message means the tension is inherent rather than applied, which is why it survives contact with an audience.

21:31 - Production quality is not a hook

Production quality is not a hook; a beautifully shot opening with no question in it loses the viewer anyway.

"What I thought when I first saw it? I was like."

— 21:31

The viewer's decision is made on interest, and interest is not a function of the budget. An opening in which a cheerful man begins cutting wood gives nobody a reason to wait — and the video gets called really well shot before it is rejected.

22:54 - Nothing said well is still nothing

Nothing said well is still nothing: craft cannot rescue an empty idea.

"Quote in advertising. This is really respectful and I appreciate you."

— 22:54

If the maker cannot say what the piece is meant to communicate, the piece cannot be fixed. Nothing said well is still nothing, and the session keeps returning to this: craft cannot rescue an empty idea, it can only make the emptiness more expensive.

23:31 - What is missing under a weak advert is an unwritten selling point

What is missing underneath a weak advert is usually a unique selling point nobody has written down.

"Insurance. So I think maybe the heart of the problem is."

— 23:31

Two questions come before the hook — where are you starting, and what are you competing against — and both are positioning questions rather than writing ones.

25:02 - An outcome is a scene the buyer can picture rather than a skill

"Get better at [a skill]" is not an outcome; the outcome is a scene the buyer can picture.

"Get better at guitar. Pause there. Pause. What is."

— 25:02

The difference is between a category description and a reason, and only the second can be evaluated before buying. A scene can be pictured and therefore wanted; a skill can only be nodded at, which is why it converts nobody.

25:29 - Positioning starts from what the buyer would otherwise do

Positioning starts from what the buyer would otherwise do, not from the category.

"Point, what are you competing against?"

— 25:29

The alternative defines the objection, and the objection is the hook. A card-based learning product competes against free video tutorials, books, a private teacher and passive learning — four alternatives, four objections. It is named as positioning 101.

25:53 - The competitor becomes the but in the hook

The competitor becomes the "but" in the hook.

"Trying to learn. So a hook might be, you're trying to learn guitar."

— 25:53

A positioning decision turns into a sentence at this point, and the two halves of the work meet: the enemy identified upstream is the tension deployed downstream. Naming what the thing is not is faster than describing what it is, because the comparison does the explaining.

26:26 - No length fixes a bad opening

Very few videos need to be a minute long, and no length fixes a bad opening.

"I think that's really good feedback. I'm just going to."

— 26:26

Very few pieces need the length they are given, and adding time to one that lost the viewer in the first seconds only extends the part nobody watched.

27:56 - The competing product won by explaining the problem first

The competing product won by explaining the problem first.

"That's enough. But what did you just — they've done a better job of."

— 27:56

It is the same ordering rule stated from the outcome end. An audience made to feel the problem goes looking for the solution; a passive one ignores it. The competitor was not better funded or better made, it simply arrived in the right order.

29:20 - Credibility is the second unanswered question

A tip becomes worth watching when it is attached to a result it produced for somebody.

"I agree. I think it's really good that, again, what I would."

— 29:20

Credibility is the second unanswered question: why this seller rather than any other. The comparison being drawn is not like-for-like either — the model video was a very expensive production, which is worth naming before drawing a lesson from it.

30:01 - A single tip demonstrated cleanly is enough to build an audience on

Demonstrate the technique while explaining it rather than describing it.

"Doing that, I would actually exhibit. So I would do the walking."

— 30:01

A single tip, demonstrated cleanly in about twenty seconds, is enough to build an audience on. Nothing about the format is expensive and nothing about it requires an idea nobody has had — only that the tip be genuinely useful and genuinely shown rather than described.

30:57 - A tip becomes worth watching when a result is attached to it

Given a real photograph of a person or an illustration of one, buyers want the real person.

"No. Well, the thing with this is, I feel strongly about."

— 30:57

A tip becomes worth watching when it is attached to a result it produced for somebody. The result converts an instruction into evidence: without it the viewer has to decide whether to trust the advice, and with it they have already watched it work once.

31:25 - Demonstrating the technique beats describing it

Illustrated explainer video is drifting toward being a commodity as generated art gets cheaper.

"Content of not real people. Just contrast this to a video where it's a real guy walking."

— 31:25

Demonstrate the technique while explaining it rather than describing it. The demonstration is an event the viewer witnesses rather than a claim they have to accept, which removes the step where they decide whether to believe you.

33:00 - Buyers want the real person rather than an illustration of one

"Saves you time and money" is what every company claims, so it carries no information.

"I feel like time and money is an outcome which is often banded."

— 33:00

Given a real photograph of a person or an illustration of one, buyers want the real person. The illustrated explainer format is drifting toward commodity as generated art gets cheaper, and it was popularised as a way to give spoken audio something to look at. People are what people attend to.

33:16 - A time-and-money saving is what every company in the category claims

Agitate the pain with arithmetic the buyer can do: a per-session price multiplied out into an annual total.

"Good to use that as an outcome. If I was doing this, I'd."

— 33:16

Saves you time and money is what every company in the category claims, which is exactly why it carries no information. A claim available to everyone transfers preference to nobody, and the reader has learned to skip it.

34:00 - Agitating the pain with arithmetic the buyer can do

Customers do not want to be talked to as if they were slightly stupid, and a voiceover can do that without saying anything condescending.

"Minutes you could do. I just feel like the voiceover was talking."

— 34:00

Agitate the pain with arithmetic the buyer can do: a per-session price multiplied out into an annual total. The number does the work the adjective cannot, and the buyer arrives at the conclusion themselves.

34:10 - Customers do not want to be talked to as if slightly stupid

A founder talking to camera for twenty or thirty seconds, naming the product and the reason it was made, outperforms the produced version.

"That is true. I do want to give a."

— 34:10

Customers do not want to be talked to as if they were slightly stupid, and a voiceover can do that without anybody intending it. Tone is a decision even when nobody makes it, and the default register of explainer narration is condescension.

35:37 - A founder talking to camera outperforms a produced advert

Human beings are interested in one thing above all, which is people, and objects and abstractions are hard to engage with alone.

"I'll give you one very specific point."

— 35:37

A founder talking to camera for twenty or thirty seconds outperforms the produced version, and an advert whose product cannot be identified cannot even be critiqued. The second usually means the effort went into the making rather than the saying.

36:23 - Human beings are interested in one thing, which is people

Well-shot, well-lit content with no message is filler, and may still have a supporting role somewhere smaller.

"Different. Look, I think it's a nice bit of well produced content."

— 36:23

Human beings are interested in one thing, which is people. Objects and abstractions are hard to engage with alone, which is why the fix in almost every critique here is to put a person in the frame. Attention follows faces.

36:44 - Well-shot content with no message is filler

Before making anything, name the reaction you want: a like, rage, love, or wanting the thing.

"But what am I — what is it I really like."

— 36:44

Well-shot, well-lit content with no message is filler, and may still have a supporting role somewhere small. The verdict is not that it is worthless — it is that it cannot carry the job it was made for.

38:27 - Naming the reaction wanted comes before making anything

Opening a paid advert as somebody's small domestic story is not a bad start.

"Campaign. What do you think? I think there's one thing."

— 38:27

Before making anything, name the reaction you want: a like, rage, love, or wanting the thing. Without that named in advance there is nothing to judge the result against, and a cluttered frame hides the product anyway — if the viewer cannot make it out, nothing else matters.

39:18 - Opening a paid advert as a small domestic story is not a bad start

A lot of story with a vague solution loses the product inside its own advert.

"I also feel like there's not. Do I actually want to."

— 39:18

Opening a paid advert as somebody's small domestic story is not a bad start, but the resolution has to answer the opening. A payoff that drops the setup's characters leaves the story unclosed, and the viewer notices even if they cannot say why.

39:57 - A lot of story with a vague solution loses the product inside it

A paid advert that does not feel like a paid advert has already done something valuable.

"All, I'm going to be a bit fairer here."

— 39:57

A lot of story with a vague solution loses the product inside its own advert. The balance is the whole craft: enough story to hold, enough product to land, and the failure here is a piece that entertained without selling. A paid advert that does not feel like a paid advert has already done something valuable, and in this category the value proposition people actually want is high-quality video produced quickly.

41:01 - In a category of near-identical products the differentiator is the specialism

In a category with many near-identical products, the differentiator is what you specialise in.

"Actually, what I would say? Final thing that I would say."

— 41:01

Search in these categories is transactional — people type the thing they want done — so the specialism is also the query. Narrowing what is offered widens the number of searches it answers exactly, which is the opposite of how it feels to do.

42:09 - A simple product's homepage is one line and a way to act

For a simple product the homepage should be one line and a way to act, with the input box above the fold.

"What it is, one thing I change here is if you go."

— 42:09

The input box belongs above the fold, and stripping an elaborate homepage back to the working thing itself is a move already proven elsewhere, so the risk has been taken by somebody else.

43:06 - Sometimes the best homepage is no homepage at all

Sometimes the best homepage is no homepage: let the visitor use the product and paywall them after a use or two.

"Much of the rest of it. I simply add your videos, choose."

— 43:06

Letting the visitor use the product and paywalling them after a use or two removes the deciding step, and with it the place the decision lapses.

44:40 - Building the thing unasked and aiming it at one buyer is the idea

Building the thing unasked and aiming it at one named buyer is a strong idea in itself.

"Built a live example. I think there's a — it looks like what's."

— 44:40

The idea stands before any of the execution is judged, because the work is the pitch, which is a harder thing to ignore than a message.

44:59 - A pitch aimed at one person is titled as a letter to them

A pitch aimed at one person should be titled as a letter to that person.

"Video, but you've titled it completely wrong."

— 44:59

The title sets the frame, and a generic title on a targeted piece throws away the only advantage the piece had. Addressed as a letter, the piece is unmistakably for one reader, and that reader has no way to file it as marketing aimed at everybody.

45:29 - An intuitive I do not understand this is sometimes the whole feedback

An intuitive response — I do not understand this — is sometimes all the feedback a piece needs.

"And the video completely — it's being honest."

— 45:29

It is worth more than a considered critique, because it is the response the audience will actually have and the only one they will act on. Addressing a pitch to a named buyer also carries stakes: it is a public risk taken to reach them.

46:27 - A founder talking honestly to camera can beat a produced video

A founder talking honestly to camera, without stock footage, can be more engaging than a produced video.

"Think something that I often see that's missing, and I don't want to."

— 46:27

The production adds nothing the viewer wanted and removes the thing they did, which is the sense that somebody is actually there. The cheap version wins on the axis that matters, and it wins by being less finished rather than more.

47:06 - Talking to camera works because it feels exposing

Talking to camera works precisely because it feels exposing and almost nobody wants to do it.

"Simple videos. So I'll tell you one thing, the reason why people do."

— 47:06

The reluctance is the moat: a format that is uncomfortable stays uncrowded, which is the same logic as effort being a differentiator. The discomfort is legible on camera too, and that is what the viewer reads as somebody meaning it.

47:20 - Fast editing and cutaways are a way of hiding

Fast editing and cutaways are a way of hiding; a real idea survives being said plainly to a camera.

"Like normal language, I think it would have resonated more. You can hide."

— 47:20

The same fix runs backwards through the earlier submissions: the founder holding the thing and stating what it is. Nothing in that requires an editor, which is why it is available to everybody being critiqued.

49:05 - Articulating the buyer's pain better than a competitor is close to sufficient

Articulating the buyer's pain better than a competitor does is close to sufficient: he says he is ready to buy without knowing what the product is.

"Well, three things. I'm ready to buy this product, whoever's made it."

— 49:05

The proof offered is his own reaction: ready to buy, without yet knowing what the product is. A strong statistic is a legitimate opening too. Naming the pain precisely is most of the sale.

49:23 - Zooming in on the abstract noun produces something visual

Zoom in on the abstract noun: a named specific is more visual than the large category it belongs to.

"I think everyone in the chat laughed at that."

— 49:23

It is the same ladder from abstract to concrete that runs through the whole session, applied here to a single word in a single line rather than to a structure.

50:07 - Where a video will live decides what the video should be

Where a video will live decides what the video should be, and without that context the critique cannot be given.

"We only saw — I think the only question that I'd have."

— 50:07

Without that context the critique cannot be given at all, which is why the question comes before any judgement of the work. The placement is part of the brief, not a decision made afterwards.

50:26 - A small personal detail changes how the speaker is heard

A small personal detail — the size of the speaker's family, the years spent at the skill — changes how the same advice lands.

"Absolutely, yeah. One thing which the guitar video didn't."

— 50:26

The size of the speaker's family, the years spent at the skill: neither adds information, and both change who is talking. A local service video should say where it operates, and a person talking to camera with clean subtitles is already engaging; what is left is pace.

53:00 - The two presenters disagree about whether the opening is plausible

The two presenters disagree about the opening line: one finds it implausible that people stress about their lawn, the other says that in the market it is aimed at they do.

"I'm a little bit confused by that, because this is a gardening company."

— 53:00

A judgement about a first line is a judgement about an audience the critic may not belong to. The disagreement is recorded rather than resolved: the rebuttal is that a lawn says something about a family's values, which makes the stress plausible after all.

53:39 - Raising something the piece never resolves leaves a question open

Do not raise something you will not resolve: naming packages without describing them leaves a question open.

"I think there's a few little subtle changes. You mentioned free packages."

— 53:39

Naming packages without describing them leaves a question open in the viewer's head, and an open question at the close is a reason to leave rather than a reason to act.

54:07 - A script that goes through the motions reads as machine-made

A script that goes through the motions reads as something a chatbot would have produced.

"I think it's slightly going through the motions with the script."

— 54:07

Generic construction is now attributable, because a machine produces it for free. The personalisation language in the lawn-care script is the tell — said with the hedge if I was being honest.

54:21 - One real customer with a before and an after beats a claim

Pick one real customer and show the before and the after, with the customer in it.

"For example, could you pick a specific customer? Pick a tree."

— 54:21

One specific case, with the customer in it, is more persuasive than a summary of many, because it can be pictured and a summary cannot.

54:47 - Filming the reaction beats filming the claim

Film the reaction rather than the claim.

"Fun with it. I remember my dad once did a."

— 54:47

The reaction is evidence the viewer watches happen; the claim is something they are asked to accept, and only one of the two survives scepticism. The demonstration outranks the description because the viewer draws the conclusion themselves rather than being handed it.

55:22 - Video may not be the right channel for a local business

Video may not be the right channel for a local business at all.

"Advice. I'd zoom out for a minute here."

— 55:22

It is the most useful thing said in that stretch, because the critique zooms out from the piece to the decision behind it and finds the decision unexamined.

56:04 - Existing reviews are hook material waiting to be used

Existing reviews are hook material: pull a line out of one and open with it.

"Just one more. You have the reviews, so if."

— 56:04

A customer's words are proof and copy at the same time, which is what nothing written in-house can be. Opening on a named customer's verdict on the service is the whole move.

56:17 - Promises rather than fluffy outcomes are what can be stated

Deal in promises, not in fluffy outcomes: state the things you will always do.

"Also, I feel like in terms of a USP, I would think."

— 56:17

A promise is checkable and an outcome is aspirational, and only one of the two can be held against you, which is why only one of them is worth anything.

57:26 - A hook that makes the viewer laugh has done its job

A hook that makes the viewer laugh has done its job, even if it takes a moment to land.

"Honestly, I think that's really creative. It took me a while to."

— 57:26

A delay before it lands is not a fault. A piece like that suits re-engaging somebody who has already visited the site, which is a placement judgement rather than a quality one.

58:13 - A joke extends into something longer once it has a stake

To extend a joke into something longer, give it a stake: the relationship the chore is costing.

"I thought it was very good. The only other thing I."

— 58:13

The stake here is the relationship the chore is costing. The joke gets the attention and the stake keeps it, which is the same two-part structure as hook and tension.

59:06 - A good video can be awfully packaged

A good video can be awfully packaged: a title and a placement that tell nobody when to watch it.

"This video is titled May 1st 2025. What do you think about."

— 59:06

The title and the placement here tell nobody when to watch it. For a short-form platform the same piece would need stakes or tension added. The packaging is a separate job from the making, and it is usually nobody's.

1:02:07 - A shopping list is not a structure

Do not present a shopping list; give structure — this is the past, this is what we do today, this is five years out.

"Two things just bloody smart which people don't do. When you're presenting, please."

— 1:02:07

The structure on offer is a simple one: this is the past, this is what we do today, this is five years out. A list has no direction, and direction is what carries a viewer through something longer than a few seconds.

1:02:45 - Half the words and double the examples

Half the words, double the examples — and an example told in full is proof of the skill being sold.

"Thing on this story. Half the words."

— 1:02:45

An example told in full is proof of the skill being sold rather than a claim about it, and the withholding principle explains why that story held: the punchline is still owed. Density is what earns a long piece.

1:03:44 - Production has to match the buyer being courted

Production has to match the buyer being courted: a domestic set under blue-chip client logos is a juxtaposition working against you.

"Generally, I think — things I was like, what's that in your hand?"

— 1:03:44

A domestic set under blue-chip client logos is a juxtaposition working against the pitch, and the mismatch is read as evidence, whatever the words say.

1:05:31 - A problem that does not resolve keeps the audience engaged

A problem that does not resolve keeps the audience engaged.

"What do you think? I like the push and the pull."

— 1:05:31

It is the same rubber band described from the audience's side. The tension is the reason to stay, and releasing it too early ends the piece before it is over. Nothing after the release is watched, whatever is still in the edit.

1:06:04 - What a service business is usually missing is proof of clients

What a service business is usually missing is proof of clients who have used the work.

"It. I've seen it. It's a brand strategy. What I think you're."

— 1:06:04

Making the thing at all — planning it, filming it, editing it — already puts you ahead of almost everyone, which is worth saying to somebody being critiqued.

1:07:33 - Point A without point Z is half a story

Point A without point Z is half a story: state where the work has got to now, however small.

"You are. I just think you need an outcome. Like, where are."

— 1:07:33

The gap needs both ends and a modest second number is better than none. Point Z can be small and still close the arc; what it cannot be is absent.

1:09:11 - Opening on a sentence written a hundred times before wastes it

Never open on a sentence that has already been written a hundred times.

"Idea. I think for me, I think you've buried the lead here."

— 1:09:11

The opening is the one line guaranteed to be read, and spending it on a phrase the reader has already skipped elsewhere wastes the only certainty you have.

1:09:29 - A coined term borrows tension from the term it rewrites

A coined term is the thing to lead with, and it works by borrowing tension from the term it rewrites.

"Before, where I found it interesting, is when you started talking to me."

— 1:09:29

It is also the thing to lead with. The familiar shape carries the unfamiliar idea, which is why the coinage lands rather than confuses.

1:10:02 - Thirty seconds a chatbot could have written are not worth keeping

If a chatbot could have written the first thirty seconds, they are not worth keeping; write what nobody else can.

"Not give me — unfortunately, I thought this had a lot of legs."

— 1:10:02

What a machine produces cheaply cannot differentiate anything. The problem setup in this video sits against its own coined idea — and he says the piece had a lot of legs, so the objection is to the opening rather than to the thinking.

1:10:16 - Examples and evidence do the talking rather than the author's opinion

Let the work do the talking: examples and evidence rather than the author's opinion.

"Start giving me. If you read the market, it's all."

— 1:10:16

An opinion invites a counter-opinion and an example invites a look, and the second is the only one that moves anybody. An example also carries the opinion inside it, so nothing is lost by withholding the verdict and letting the reader arrive at it.

1:13:05 - The honest verdict is where the reviewer would have stopped watching

The honest verdict on a video is where the reviewer would have stopped watching.

"Let me just jump in here. I didn't want to press."

— 1:13:05

It is a far more useful note than a score, because the drop-off point names the defect more precisely than any adjective. A score can be argued with; a moment cannot, because it is the moment the reviewer stopped caring and they can point at it.

1:13:34 - A synthetic voice nobody can place is doing damage

A synthetic voice you cannot place is still doing damage, and a bad one is doing more.

"Look, the great news is, if this business is doing well, you're only."

— 1:13:34

A bad one does more, and the damage is done whether or not the viewer can name it. It is a small thing the audience registers without articulating, and the cost is trust rather than attention.

1:14:12 - The organic-post test asks what would happen if this were posted unpaid

A usable test: if you posted this organically to a professional network, what would happen?.

"A good test for a video is, if you posted this."

— 1:14:12

The question strips away the spend and asks whether the thing is worth watching on its own merits, which is the only version of the question that matters.

1:14:51 - Twenty seconds of setup could have been five

What took twenty seconds could have been done in five: open on the statistic, move straight to the solution.

"I think we might have been a bit unfair. There is."

— 1:14:51

The faster version opens on the statistic and moves straight to the solution. The compression is not stylistic — every second of setup is a second in which the viewer decides whether to continue. Avoid catch-all phrases and zoom down onto the specific object instead.

1:16:04 - On a low-traffic high-value page every visitor matters, so there is one exit

On a low-traffic, high-value page every visitor matters, so the page should have one exit.

"Still good. I would, if you got serious and this is."

— 1:16:04

Every additional option is a place the decision dissolves, and on a page with few visitors each one is expensive. Where traffic is large the leak is averaged away; where it is small, a single lost visitor is a measurable share of the month.

1:16:52 - Social proof belongs above the fold on every homepage

Every homepage needs social proof above the fold.

"What do you think? Help me out. I don't know what it."

— 1:16:52

It answers the second question a visitor has — is this any good — in the same glance as the first, and the count is doing work no sentence can. Below the fold it answers a question the visitor has already given up on asking.

1:18:02 - The objection gets answered at the point of contact

Answer the objection at the point of contact: a face, an availability signal, the hours the phone is answered, a reply time.

"With contact information, I would add a face here."

— 1:18:02

A face, an availability signal, the hours the phone is answered — each answers a worry at the moment it arrives. The objection arrives at a specific moment, and answering it anywhere else is answering it too late.

1:18:51 - Showing the product working means a before and an after of the thing

Show the product working: a before and an after of the thing itself.

"I'd like to see the product, or what the product does."

— 1:18:51

A before and an after of the thing itself is the shortest route to it. The submission's opening is rocky, and there is something behind it — which is a more useful verdict than either praise or dismissal.

1:20:54 - Showing the client is better evidence than showing the seller

Showing the client is better evidence than showing yourself, and more interesting.

"I thought the start was really good. I felt."

— 1:20:54

The seller's account of the work is a claim; the client's is testimony, and the audience discounts the first automatically. It is also cheaper to shoot, because the client already has the story and does not have to be scripted.

1:21:11 - A short trading period set against a large sales figure is a real claim

The juxtaposition principle reappears in a real claim: a short trading period set against a large sales figure.

"Also, we said, all right, in a search for just."

— 1:21:11

Two numbers make an argument that neither makes alone, and both of these are checkable. The short period is what makes the large figure mean something, and the large figure is what makes the short period worth mentioning.

1:21:24 - Production matters when it contradicts the proof being shown

Production usually does not matter much, but it matters when it contradicts the proof you are showing.

"Once you get to your bedroom, I think, or wherever you're filming."

— 1:21:24

The contradiction is the defect rather than the production value, which is why the rule cannot be stated as a standard. A rough piece claiming rough is fine; a rough piece claiming premium is evidence against itself, and the viewer trusts the evidence.

1:21:44 - Rhyming jargon is what people reach for with nothing to say

Rhyming jargon is what people reach for when they have nothing to say.

"You list off stuff like scroll scuffing."

— 1:21:44

Revenue popping and scroll stopping are audible as filler precisely because they are pleasant, and the pleasantness is doing the work of hiding the absence. The phrases survive because they sound like a claim while committing to nothing that could be checked.

1:22:11 - A revision pass highlights every place the script could be more specific

A concrete revision pass: highlight every place in the script where you could be more specific.

"Brutally honest. I would rewrite your whole script and I would highlight every."

— 1:22:11

It is a mechanical instruction rather than an aesthetic one, which is what makes it runnable by somebody who does not yet have taste. Specificity is the fix with the widest application. Cut to the hook as well: the strongest line is often already in the video, sitting behind something weaker, so the fix is reordering rather than writing.

1:23:07 - Length has to be earned by the storyteller

Length has to be earned; a minute and a half needs a phenomenal storyteller.

"Get into the frame of it. I think, to that point, I."

— 1:23:07

Runtime is a cost the audience pays and the work has to repay. The comparison is with the storytelling submission earlier in the session, which earned the time it took.

1:23:39 - The artefact nobody else holds is the evidence worth showing

Show the artefact only you have: the campaign dashboard, the spend and the return, the thank-you email — with permission.

"Ads. Yeah, but how do you."

— 1:23:39

The campaign dashboard, the spend and the return, the thank-you email: it is evidence nobody else can produce, which makes it the one thing a competitor cannot copy into their own advert.

1:24:02 - Anything sayable on the first day in business is not worth saying

Anything you could say on your first day in business is not worth saying.

"Get anywhere else. And revenue popping, scroll stopping is just, honestly."

— 1:24:02

The test is sharp because it cuts exactly the claims that feel safest — the ones about care, quality and service, all of which were available before any work was done.

1:24:26 - The close is the client rather than the seller

End on the client, not on yourself: a short unpolished clip of the customer saying thanks is the close.

"Then you being like, imagine a video — like you show her."

— 1:24:26

A short unpolished clip of the customer saying thanks does the job. A scarcity line at the close — a stated small number of places left — gives the ask a shape. The last thing said is the thing remembered.

1:25:58 - Filming it and pressing upload is half the battle

Filming it, editing it and pressing upload is half the battle, and nobody's first videos are any good.

"Once again, thanks so much for coming, and more importantly, everyone who."

— 1:25:58

Nobody's first videos are any good, and the point is that they exist. It is the closing note of an hour of hard critique, and it is the only one that addresses whether any of it gets done at all.

Personal Branding Lessons

An hour and a half of live critique, and the notes that recur are the ones worth stealing. Each move below was applied to a real submission in front of the person who made it.

Decide the hook by the material rather than by taste

There are broadly two openings: promise what the viewer gets, or open a loop and withhold it. Which one applies is decided by what you actually have. When the substance is genuinely strong, promise it — the line is a handshake, and delivering on it builds trust in the next one. When the substance is thin, a promise is a debt you cannot pay, and the loop is the only honest option. Choosing by preference rather than by material is how a piece ends up betraying its own opening. 03:46

Delete the product from the hook

Naming the product in the first line ends the transaction: the viewer files it as an advert and stops. Delete the product and the same line becomes a situation they want finished. Fit the thing into a story somebody already cares about rather than showing the thing, and introduce it once the tension has done its work. The audience does not care about your product and does care about the situation it belongs to, which is the whole basis of the move. 08:52

Put imaginary buts and sos between the points

Every story runs on tension, and tension is carried by but, never by and. Write the points out and put imaginary buts and sos between them: if it all reads as and, there is no story in it — only an inventory. The same sentence can usually be rewritten with tension without changing a single fact in it, which means the fix costs nothing but attention. A commercial becomes a story at the point where the direction changes. 13:24

Collect hooks rather than inventing them

Write down the good ones when you hear them, and adapt them to your own subject. A swipe file is the working memory of the craft and it removes the blank page from the job entirely. Nobody in the critiques invented a hook shape from nothing: the strong submissions were running known structures on new material, and the weak ones were reaching for a phrase. 16:41

Run the swap test on a finished advert

If any other subject could be swapped into the advert without changing it, it is not an advert for this product. The test is mechanical and brutal, and it fails a great deal of work that looks finished. Provocation is legitimate only when it stems from the product — a conflict borrowed from somewhere else is noise that costs trust. The method is to brainstorm conflicts that are true of the product, then dramatise one. 20:16

Name the reaction wanted before making anything

Decide in advance what the viewer should do: like it, rage at it, love it, want the thing. Without that named up front there is nothing to judge the result against, and a critique becomes a matter of taste. The same discipline catches the cluttered frame — if the viewer cannot make the product out, nothing else in the piece matters, and no amount of lighting fixes it. 36:44

Put social proof above the fold

Every homepage needs it, because it answers the second question a visitor has — is this any good — in the same glance as the first. On a low-traffic, high-value page every visitor matters, so give the page one exit rather than six. And answer the objection at the point of contact: a face, an availability signal, the hours the phone is actually answered. The objection arrives at a specific moment, and answering it elsewhere is answering it too late. 1:16:52

Show one real customer's before and after

Pick a specific customer, show the state before and the state after, and put the customer in it. One case is more persuasive than a summary of many, because it can be pictured. Film the reaction rather than the claim — the reaction is evidence the viewer watches happen, and the claim is something they are asked to accept. Showing the client is better evidence than showing yourself, and considerably more interesting to watch. 54:21

Mark every vague phrase and name a specific instead

A concrete revision pass: read a finished draft line by line and mark every point where something could be named instead of gestured at, then name it. It is a mechanical instruction rather than an aesthetic one, which is what makes it runnable before you have taste. Anything you could have said on your first day in business is not worth saying — the test cuts exactly the claims that feel safest, the ones about care, quality and service. 1:22:11

Questions

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

Should my hook give the answer away or hold it back?

It depends entirely on how good the thing you are holding is. A promise spends the material at the top, so it only pays when the material can carry it: if you genuinely did the extreme thing, say so. A loop withholds, and it earns its keep when the plain statement reads flat — the same reservoir story that is dull as a fact is intriguing as a question. The device is chosen after the material, never before. 03:46

Why does my content feel like a list instead of a story?

Because every join in it is an "and". Put imaginary "buts" and "sos" between your points and see whether they connect; if nothing does, you have a shopping list. "And" accumulates and "but" changes direction, and direction is what makes something a story rather than a sequence. The same sentence can be rewritten with tension without a single fact in it changing — the connective does all of the work. 13:24

How many hooks does a long video need?

More than one. A single hook and a single payoff is granted as definitely true for short video, but a ten-minute piece has to re-earn attention at intervals, so the working shape is a main hook, a smaller one about a minute in, and payoffs spread through rather than saved for the end. Where the video runs long, the original hook gets restated, because an audience that has forgotten the question is no longer in a loop. Long-form structure is hook maintenance. 17:03

My video is well produced and nobody watches it — what is wrong?

Production is not a hook. A beautifully shot opening with no question in it loses the viewer anyway, because the decision to keep watching is made on interest and interest is not a function of budget. The harder verdict underneath is that craft multiplies an idea, so a zero stays a zero: nothing said well is still nothing. If the maker cannot say what the piece was meant to communicate, the piece cannot be fixed by shooting it again. 21:31

How do I work out what my product is competing against?

Start from what the buyer would otherwise do, not from the category you file yourself under. A card-based way to learn guitar is not competing with other cards; it is competing with free tutorials, books, a private teacher and passive learning. That alternative is where the objection lives, and the objection is the hook — trying to learn but unable to afford the lesson, trying to learn but unwilling to sit in a room for two hours. Map the field before you write the line. 25:29

Is "saves you time and money" a good thing to promise?

It is an outcome that gets banded around, and so many companies claim it that on its own it does little work. It is gated rather than banned: you have to be very good to use it as an outcome. The repair offered is arithmetic the buyer can do without help — take the per-session price, multiply it out into what the year costs, then show the minutes being wasted inside each session. The number is the same; what changes is that the cost is now felt. 33:00

Why do founder-to-camera videos beat produced ones?

Because loading up a camera and putting yourself in front of it is exposing, and almost nobody wants to do it. The reluctance is the barrier to entry, and the cost is what makes it a signal. Fast edits, cutaways and newspaper clippings are named as a hiding place — motion supplied where an idea should be. If you have a real idea, it survives being said plainly to a lens. 47:06

Where should a video live, and does that change what it should be?

It changes everything, to the point where the critique stalls without the answer. Homepage, search result, clickable panel, retargeting ad: each puts the viewer in a different state of mind, and state of mind decides what the opening has to do. The wit-led piece that fails as a short-form post works as an advert to someone who has already visited your site. Choosing the surface is part of making the thing, not a step after it. 50:07

How do I prove results when I am just starting out?

State where the work has got to now, however small. Point A without point Z is half a story, and the first client is a legitimate thing to celebrate. The ideal is the visible transformation — a growth graph with the purchase point marked on it — but the boundary is stated plainly on air: clients at that scale are unlikely, and not many people have them. Until then, showing the customer, the before and the after, and the artefact only you could show does the same job. 1:06:04

How do I judge my own video honestly before I post it?

Two tests, both free. Ask at what second you would have stopped watching — stopping is the one measurement that does not depend on taste, and on the session's own examples the weak one was answered with five seconds, while the strong one was never given a stopping point at all. Then ask what would happen if you posted it on a professional network: zero to one likes is an answer. Both remove your own investment from the judgement, and both can be run before anyone else ever sees the length you are asking for. 1:13:05

Sources

Live "Office Hours" with Harry Dry and Sabba Keynejad ✨

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.