Cold direct response ads for the $7 two-day live event at webinar.studiohawk.com. Everything here is lifted from the 140-slide deck Harry actually presents, which means the ad promises what the room delivers, and every statistic already has the source on the slide it came from. Eighteen angles exist on the film day page; these twenty are new and none of them repeat those. Built to Jeremy's original SOP-31, one reason to buy per ad filmed whole, and SOP-27's timing. Sixteen are scripted in full with a 30 and a 60 second cut.
webinar.studiohawk.com, $7, and the date. Every 60-second cut names the date and "live only, no replay" in the last beat.Each one is a different reason, taken from a specific slide, so the ad and the room agree. Slide numbers refer to the deck's own order. The eighteen existing angles are listed at the bottom so nobody films a duplicate.
| # | Reason to buy | From | SOP-31 category | Hook style | Status |
|---|---|---|---|---|---|
| W21 | ChatGPT reached 100 million users in two months. Nothing has ever moved this fast. | Slide 2 | Urgency | Shock stat | scripted |
| W22 | Six in ten US adults read AI summaries, and most would never tell you they use AI | Slide 4 | Problem agitation | Shock stat | scripted |
| W23 | You did not lose a ranking. You lost the click. | Slide 6 | Problem agitation | Myth-busting | scripted |
| W24 | Three prompts, six minutes, and you will know where you stand | Slide 7 | Tactical | Call-out | scripted |
| W25 | Recognition is not recall: knowing you exist and volunteering your name are two assets | Slide 8 | Curiosity | Myth-busting | scripted |
| W26 | There is no page two of an AI answer | Slide 9 | Problem agitation | Bold claim | scripted |
| W27 | The machine just wrote your real competitor list, and it is not the one on your whiteboard | Slide 9 | Envy | Curiosity | scripted |
| W28 | Being named for the thing you do not want to sell is a positioning problem | Slide 9 | Curiosity | Specific question | scripted |
| W29 | The rent went up, the street got quieter, same landlord | Slide 20 | Problem agitation | Pain | scripted |
| W30 | Rented, borrowed, owned: sort every channel you pay for | Slide 22 | Tactical | You already know | scripted |
| W31 | Being number one and being named are two different competitions | Slide 23 | Myth-busting | Bold claim | scripted |
| W32 | You were sold exercises, not a programme | Slide 25 | Frustration | Analogy | scripted |
| W33 | Sourced, considered, invisible: most well-run businesses are the middle one | Slide 26 | Curiosity | Specific question | scripted |
| W34 | The sale is decided before you are in the room | Slide 27 | Urgency | Shock stat | scripted |
| W35 | The three companies with the most to gain from you giving up are telling you not to | Slide 30 | Social proof | Objection-first | scripted |
| W36 | Four things being sold to you right now that you do not need to buy | Slide 40 | Tactical | Objection-first | scripted |
| W37 | Three questions to ask whoever you are currently paying | Slide 41 | Frustration | Tactical | brief below |
| W38 | Stop publishing more. There is no page eleven. | Slide 41 | Myth-busting | Bold claim | brief below |
| W39 | Why I am doing this at all, answered before you ask | Slide 16 | Objection-first | Objection-first | brief below |
| W40 | One question now fans out into dozens you never see | Slide 34 | Curiosity | Curiosity | brief below |
Already filmed on the film day page, do not duplicate: Samsung found us through ChatGPT · Stop paying twice for the same buyer · What optimizing for ChatGPT means · Learn SOURC-E not more AI tips · Ten Claude Skills yours to keep · Does AI name you or your rival · Get found like QuickBooks and Weber · Get in before the answer sets · Hold your agency accountable on AI · Your impressions rose, clicks fell · Two days live, 90 minutes a day, $7 · $7 to test it before you fund it · 621% more mentions in AI answers · Own the channel, stop renting clicks · Bring a case not another AI article · You have never once checked this · No developer, no rebuild, get cited · Bring your market, ask us live.
Reason to buy: the speed of the shift is the argument for showing up now rather than next year. Nothing in consumer technology has moved this fast, and the numbers make the case without a single adjective.
Instagram took two and a half years to reach a hundred million users. TikTok took nine months. ChatGPT took two.
Three figures on screen, one at a time. Small source line: UBS, reported by Reuters.
Two months. That's how fast the place your customers ask questions changed, and most businesses have never once checked what it says about them. On the twenty-third and twenty-fourth I'll show you how to check, and what to do about the answer.
Two days live, ninety minutes each, seven dollars. Link below.
Reason to buy: the deck's sharpest line about the hidden nature of this shift. Your customers are using AI to decide and you will never see it in a survey or a support call.
Six in ten American adults read AI search summaries. Most of them would never tell you they use AI.
Flat. Source line on screen: Pew Research Center, Americans and AI, June 2026.
That's why this feels invisible. Nobody rings to say they picked your competitor because a machine named them. There's no line in your analytics for the customer who never arrived. It just quietly stops happening.
Two days live, ninety minutes each, seven dollars. Twenty-third and twenty-fourth. Link below.
Six in ten American adults read AI search summaries. Three in ten say they don't. And most of that six would never describe themselves as an AI user.
That's Pew, this year. It's also why this change is so easy to miss from inside a business. Your customers aren't announcing it. They're just asking a machine first, getting three names, and calling one of them.
There's no report for the customer who never arrived. No line in analytics, no lost-deal note, no bad review. The first time most owners find out is when a quarter comes in soft and nobody can say why.
Underplayed. Do not dramatise; the fact is enough.
So over two mornings we make it visible. We run your category live, in the room, so you see what the machine currently says when someone asks who to use. Then the framework for changing it, the same one we run for QuickBooks and Weber.
Seven dollars, ninety minutes a day, and the whole toolkit is yours to keep.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay, one cohort a month. Below.
Reason to buy: explains a drop the owner has already seen and could not account for. The best kind of ad: it names something they have already noticed.
If your impressions held and your clicks fell, here's what happened. You didn't lose a ranking. You lost the click.
When Google shows an AI Overview, people click a normal result about eight percent of the time. Without one, fifteen. That's Pew's behavioural study. Same position, half the traffic, and nothing in your rankings report changed to warn you.
On screen: 15% then 8%.
Two days live on what to do about it. Seven dollars, twenty-third and twenty-fourth. Below.
The Mirror is the deck's strongest live moment: three prompts, six minutes, and everyone in the room finds out where they stand. It yields five distinct reasons to buy, because what people find is different every time. SOP-26's real-time contextual hook applies: a viewer who runs the prompt has just become the person this ad is for.
Reason to buy: the fastest audit anyone can run, done before the ad ends. Whatever they find is the reason to take the seat.
Do this before the ad ends. Open ChatGPT and ask it three things about your own business.
Phone in hand, thumb typing, screen hidden.
One: best whatever you sell, in your city, with citations. Two: who should I hire for the thing you do, and why them. Three: what do you know about your brand. Write one word for each. Named, missing, or wrong. Most businesses come back missing, and that's the whole reason for these two days.
Bring your three answers. Seven dollars, twenty-third and twenty-fourth. Below.
Reason to buy: the distinction almost nobody has drawn for them. "It knows I exist" is the test everyone passes and the one that means nothing.
Recognition is not recall, and the difference is worth more than any ranking you hold.
Here's the test everyone runs: ask an AI what it knows about your company. It comes back with something reasonable and you feel fine. Almost everybody passes that. You've been online for years, of course there's something in there.
Here's the test your buyer runs: best whatever you sell, for whoever they are, in their city. Nobody mentioned your name. The machine had to pick you out of your entire category and put you in a sentence. That's a completely different asset, and most well-run businesses do not own it.
Over two mornings I'll show you how the engines decide which two or three businesses go in that sentence, we'll run your category live in the room, and you'll leave with the framework and the whole toolkit.
Seven dollars, ninety minutes a day, and it's yours to keep.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay. Below.
Reason to buy: the structural fact that makes this urgent rather than interesting. Classic search had ten winners and a consolation tail. This has one to three and no consolation prize.
There is no page two of an AI answer.
Hold. Let it sit before the next line.
Old search had ten winners on page one and a long tail of consolation traffic underneath. An AI answer names one to three businesses and stops. Nobody scrolls to find the one the machine left out. If you're not in the sentence, you weren't in the running, and you'll never see the search that decided it.
Two days on getting into the sentence. Seven dollars, twenty-third and twenty-fourth. Below.
Ten winners just became one to three. That's the whole change, and it happened without an announcement.
Page one of Google had ten results, and even position eight got something. There was a consolation prize. An AI answer names one to three businesses in a sentence and then it's finished. There is no page two. Nobody scrolls to find the business the machine left out.
And you'll never see it happen. There's no impression, no click, no line in a report. Somebody asked, three names came back, one of them got the call. If it wasn't you, nothing in your analytics will ever tell you.
Flat. This is the fear beat, so underplay it.
Over two mornings we go through exactly how those one to three get chosen, run your own category live in the room so you can see where you sit, and hand you the framework we run for QuickBooks, Weber and HelloFresh.
Seven dollars, ninety minutes a day, toolkit included.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, one cohort a month. Below.
Reason to buy: the machine writes your competitive set from the same source material your buyers see, and it is rarely the list on the whiteboard. Envy, and it is factual.
Ask an AI who to hire in your category and read the names it gives back. That's your real competitor list.
Not the one on your whiteboard. The machine just wrote it, from the same source material your buyers are reading, and it's rarely who you think. Sometimes it's a business you've never heard of. Sometimes it's the one you stopped worrying about two years ago.
Two days on taking that position back. Seven dollars, twenty-third and twenty-fourth. Below.
Reason to buy: the case nobody plans for. Being cited enthusiastically for the service you are trying to move away from is a positioning problem, and no amount of SEO fixes it.
What if the AI does name you, and names you for completely the wrong thing?
It happens more than the missing case, and it's worse, because it looks like a win. You're in the answer. You're just in it as the budget option, or for the service line you've been trying to shut down for two years, or with a description of your business that stopped being true in 2022.
That is a positioning problem wearing a visibility problem's clothes. No amount of technical SEO fixes it, because the machine isn't reading your intentions. It's reading what everybody else has published about you, and that's a different job with a different method.
Over two mornings we go through how an answer about your brand actually gets assembled, why it gets things wrong, and the specific work that changes it. Then you run it on your own business in the room.
Seven dollars, ninety minutes a day, and you keep the toolkit.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, one cohort a month. Below.
Both numbers here are on the slide with their sources, which settles the open question from the topic test: cost per lead is up 72 percent between 2022 and 2025, $40.74 to $70.11 (WordStream, Google Ads Benchmarks), and paid clicks are down 68 percent where an AI Overview appears (Seer Interactive, AIO Impact on Google CTR, September 2025). Use these, not the unsourced 57 percent that circulated earlier.
Reason to buy: the lease is getting worse from both ends at once, and the owner feels one end already. Naming the second end is the insight.
Your cost per lead is up seventy-two percent since 2022. And on the same searches, paid clicks are down sixty-eight percent.
Two figures on screen. Sources in small type: WordStream, Seer Interactive.
Forty dollars seventy-four to seventy dollars eleven, on average. The rent went up and the street got quieter, and it's the same landlord doing both. That's not an argument for turning ads off. It's an argument for stopping treating one rented channel as the whole portfolio.
Two days on building the one you own. Seven dollars, twenty-third and twenty-fourth. Below.
The rent went up. The street got quieter. Same landlord.
Average cost per lead on Google Ads climbed seventy-two percent between 2022 and 2025, forty dollars seventy-four to seventy dollars eleven. Over the same period, paid clicks fell sixty-eight percent on searches where an AI Overview shows up. Both ends of the lease got worse at once.
Sources on screen throughout this beat.
When organic slides, most businesses fall back on paid, because it's fast and measurable. But you don't own that traffic, you rent it. It works the day you pay and stops the day you stop, and someone else sets the price.
Over two mornings you'll do a ninety-second exercise on your own numbers, work out what your annual rent actually is, and then spend the rest of the time on the channel nobody can price you out of. Framework and toolkit included.
I'm not telling you to turn the ads off. I'm telling you not to run one rented channel as the whole portfolio.
Seven dollars. September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay. Below.
Reason to buy: a sorting frame they can apply to their whole marketing budget in sixty seconds, which is a genuinely useful thing to receive from an ad.
Here's a sixty-second exercise that will tell you more about your marketing than most quarterly reviews.
Three columns. Rented: paid search, paid social. It works the day you pay and it stops the day you stop, and the price is set by someone else. It's gone up seventy-two percent since 2022.
Borrowed: marketplaces, platforms, somebody else's audience. Real customers and real revenue, with a landlord who can change the terms on a Tuesday. Owned: your site, your list, and your brand travelling on its own. Slowest to build. The only one nobody can price you out of.
Now put your actual numbers against the three. For most businesses it's almost all rented, a bit borrowed, and a column three that hasn't been touched in years, which was survivable when search still sent traffic and is a different proposition now.
Two mornings on building column three, with the framework we run for QuickBooks and Weber and the toolkit to keep going after.
Seven dollars, September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay.
The deck's spine: AI search is built on top of SEO, not instead of it, but being number one and being named are different competitions. These four are the myth-busting cluster, and they are the ones most likely to stop a marketing lead mid-scroll because they contradict something they said in a meeting last week.
Reason to buy: resolves the question every owner has quietly: does my existing SEO still count? Yes, as the foundation, and no, not as the finish line.
If someone tells you SEO is dead, they haven't looked at the data. Here's what it actually says.
Ninety percent of Google's AI Overviews cite at least one page from the organic top ten. More than half cite the number one result itself. Ninety-seven percent cite something in the top twenty. That's a 432,000 keyword study by seoClarity. AI search is built on top of SEO, not instead of it.
Which is exactly where businesses get caught. Ranking is the entry fee. It gets you into the pool the machine draws from. It does not decide which one to three names go in the sentence, and that second decision runs on different signals that most agencies have never had to work with.
Over two mornings we go through those signals one at a time, six of them, and you'll run your own site against each. The framework we use for QuickBooks, Weber and HelloFresh, and it's the one Samsung found us through.
Seven dollars, ninety minutes a day, toolkit yours to keep.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay. Below.
Reason to buy: the best line in the deck. It reframes years of disappointing SEO spend without blaming the buyer or calling anyone a fraud, and it makes sequence the product.
Two people join the same gym. Same equipment, same fees, same twelve months. One changes shape, one doesn't.
Nobody concludes gyms are a scam. The difference was the programme. Same with your SEO. A blog post here, some links there, a technical fix in March: every one of those is a real exercise a good trainer might prescribe. Done in no particular order, they're great exercises for the wrong person.
Two days on the programme, in order. Seven dollars, twenty-third and twenty-fourth.
"Does SEO actually work?" is a question that has never had a useful answer. Here's the one that does.
Two people join the same gym on the same day. Same equipment, same fees, same twelve months. One changes shape and one doesn't. Nobody concludes gyms are a scam, and nobody concludes gyms are magic. The difference was the programme.
You've been sold exercises. A blog post here, some links there, a technical fix in March. Every one is a real exercise a good trainer might prescribe. Done in no particular order, for the wrong individual, they produce exactly what you've experienced.
So the question isn't does SEO work. It's whether anyone was ever running a programme, or whether you were being handed exercises and invoiced. You already know which one you've been buying.
This is the line. Say it plainly and stop.
Two mornings on the programme, in the order it goes in, for AI search and search together. Seven dollars, and the toolkit's yours.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, one cohort a month. Below.
Highest-potential ad in the bank. It validates a real grievance without accusing the current supplier, which is what lets a business owner forward it to their own marketing manager. Film it three times until the gym story sounds told rather than performed.
Reason to buy: gives them a category to place themselves in, and the middle one is deliberately uncomfortable because it is where the competent businesses sit.
There are three places your business can be in an AI answer, and only one of them pays.
Sourced: you're named inside the answer. One to three businesses per question, that's the whole prize. Samsung found us that way. No ad, no referral, nobody in the room who knew us.
Considered: you rank, you're crawlable, the machine knows precisely who you are, and it named somebody else. That's where most professional, well-run businesses are sitting right now, and it's the worst place to be, because from the inside everything looks healthy.
Invisible: not in the conversation at all. The answer gets written as though your business doesn't exist, and you will never see the search that decided it.
Over two mornings you'll find out which one you are, in the room, and then work through what it takes to move up. Framework and toolkit included, seven dollars.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay. Below.
Reason to buy: converts an AI-search argument into a pipeline argument, which is the version a B2B owner or a sales-led business actually acts on.
Sixty percent of a B2B buying journey is done before a seller is ever contacted. Eighty percent of the time, the first vendor contacted wins.
And ninety-five percent of winners were on the shortlist from day one. That's 6sense, about four thousand buyers. Which means the shortlist is the sale, and it now gets written by a machine, in an answer you never see, before anyone has heard of you.
Three figures on screen in sequence.
Two days on getting onto that shortlist. Seven dollars, twenty-third and twenty-fourth.
This isn't a marketing number. It's a pipeline number, and it's the reason AI search matters to your revenue this quarter.
Sixty percent of the B2B buying journey is complete before a seller is contacted. Eighty percent of the time, the first vendor contacted wins the deal. And ninety-five percent of winners came off the day-one shortlist. That's 6sense, roughly four thousand buyers.
Put those together and the shortlist is the sale. Everything after it is paperwork. So the only question that matters is how that shortlist gets built, and increasingly the answer is: a machine writes it, in a conversation you're not part of, from sources you didn't know it was reading.
Over two mornings we go through exactly how it's assembled, run your own category live in the room so you can see whether you're on it, and hand you the framework we run for QuickBooks, Weber and HelloFresh.
Of your last ten deals, how many arrived already knowing who you were? That's the number this is about.
Seven dollars. September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, no replay.
Reason to buy: answers "is SEO dead, should I stop?" with the least self-interested proof available: the platforms that would benefit most from you quitting are telling you not to.
You've probably been told SEO is dead. Look at who would gain most from you believing that.
If businesses stopped investing in organic search tomorrow, three companies would do very well out of it. And all three are on the record telling you not to. Shopify is publicly telling merchants to keep investing. Microsoft says traditional SEO fundamentals matter more than ever. Google says best practices continue to be relevant because it's all built off the search index.
That's about as disinterested as evidence gets. The traffic didn't vanish, it changed shape, and the businesses reading "SEO is dead" as permission to stop are handing their category to whoever kept going.
Over two mornings we go through the new shape: how the engines choose which one to three businesses to name, the six signals that decide it, and your own site run against each one, live in the room.
Seven dollars, ninety minutes a day, and the toolkit is yours to keep.
September twenty-third and twenty-fourth, twelve to one-thirty Eastern. Live only, one cohort a month. Below.
Reason to buy: the ad that saves them money before they have bought anything. It also positions Harry against the people currently pitching them, without naming anyone.
Four things are being sold to you right now for AI search, and you don't need to buy any of them.
One: getting you listed in ChatGPT. There is no listing. If a pitch promises you a slot, ask them to show you the slot. Two: an llms.txt file. No major assistant has committed to reading it. Three: chopping your site into FAQ blocks. Four: buying the AI visibility dashboard first. A tool that tells you weekly that you're not cited is not the same as being worth citing.
Two days on what actually works. Seven dollars, twenty-third and twenty-fourth.
SOP-26 wants a refresh every two to four weeks and these are the reserve. Each has its hook, its body argument and its slide; write the beats when the first sixteen have run.
| # | Hook | Body argument | Tag |
|---|---|---|---|
| W37 | "Three questions to ask whoever you're currently paying, this week." | Show me a page on my site the way an AI crawler receives it. Which third-party pages get cited in my category, and what's our plan for them. What did we get named in last month that we weren't named in before. Give the good answer and the bad answer for each, exactly as the deck does. Ends: a good agency will welcome these questions, and you'll know inside a week. Slide 41. | WEB_37_Three_Questions |
| W38 | "Stop publishing more. There is no page eleven." | The reflex when traffic drops is volume. But there's no page eleven now, and twice as many pages saying the same thing is twice as much of nothing. The work that moves an AI answer is not more content, it is content that could only have come from you. Slide 41 plus chapter 8 of the book. Good pairing with W36 as a "stop wasting money" cluster. Slide 41. | WEB_38_No_Page_Eleven |
| W39 | "Why am I doing a seven dollar event at all? I'll answer it before you have to ask." | Three reasons, verbatim from the deck: we want strategy calls and a handful of people in the room will be a good fit; this shift is mispriced right now and businesses that move in the next twelve months get a position that businesses moving in three years won't; an educated buyer is a better client. "That is the whole motive. There is nothing behind it." The objection-first ad for the price-suspicious. Slide 16. | WEB_39_Why_Doing_This |
| W40 | "You asked one question. The machine asked twelve." | Query fan-out: your buyer types one question and the engine silently expands it into a dozen sub-questions, retrieves for each, and assembles one answer. So you are not competing for one query any more, you are competing for the dozen behind it, and you cannot see any of them. Credit Aleyda Solis for the term on the slide. Slide 34. | WEB_40_Query_Fan_Out |
| Setting | For the webinar |
|---|---|
| Objective | Conversions on the registration event, not on a tap-fireable one. SOP-04's pixel-conditioning rule matters less here because the $7 payment is itself the qualifier: optimise to the purchase, not the page view. |
| Placements | Manual. Audience Network excluded. This is the rule that round one of the topic test broke, and it cost roughly $1,100. |
| Audience | A real one: interests, a lookalike from the client list or past registrants, retargeting. Exclude anyone already registered for this cohort. |
| Dynamic creative | Per ad set: 2 to 4 videos, the 3 headlines and 2 primary texts above, plus one static of the graphic text on a plain card. One reason per ad set. |
| Cadence | Cohort-shaped rather than evergreen. Push hardest in the ten days before the date, and let the date do the urgency. The monthly refresh is largely a date change plus swapping in two reserve reasons. |
| Benchmark | SOP-04: aim for a 30 to 50 percent opt-in rate on the registration page. Below that, the page is the problem, not the ad. Watch cost per registration as the primary metric, with CTR and frequency as the diagnostics. |
| Retargeting | Registered but not attended is the Hammer Them audience for this funnel, and the window is short. The library applies with the CTA swapped for "see you on the twenty-third". |