Built from StudioHawk's own sales calls rather than from a whiteboard. The messaging framework is Jeremy's Messaging Mastery SOP: Circumstances, Problems, Outcomes, plus Conditions and Amplifiers. The evidence is all 222 transcripts: 175 companies, 448,355 words of prospect speech, parsed line by line so the counts below are shares of calls rather than impressions. Every quote is a real prospect, in their own words, on a real call.
Share of calls in which the buyer, not the rep, brought the theme up. DISC is the 110 conversations recorded at Discovery stage, PROP the 77 recorded at Proposal or Pitch. Lift is the difference. A positive lift means the theme is still being talked about when money is on the table; a negative lift means it is an early-conversation topic that fades.
| What the prospect raises | All | Disc | Prop | Lift | Read |
|---|---|---|---|---|---|
| Has or had an SEO agency | 52% | 57% | 48% | -9 | You are a replacement, not a first purchase. |
| AI search or LLMs, raised by them | 40% | 50% | 26% | -24 | The most important row. See below. |
| Wants ongoing, not a project | 40% | 46% | 32% | -14 | Asked early, settled early. |
| Site migrated, rebuilt or re-themed | 32% | 31% | 40% | +9 | The strongest positive lift. Still live at proposal. |
| More than one person on the call | 34% | 32% | 39% | +7 | Gets more crowded as it gets real. |
| Hard commercial deadline | 18% | 18% | 25% | +6 | Real urgency, never manufactured. |
| Explicitly comparing agencies | 15% | 28% | 17% | -11 | The shortlist is an early-stage behaviour. |
| A competitor overtaking them | 12% | 13% | 12% | -1 | Minority trigger. Do not build the bank on it. |
| Explicit agency dissatisfaction | 8% | 8% | 8% | 0 | Do not attack the incumbent. Confirmed at full scale. |
| Says outright they lack SEO expertise | 7% | 6% | 9% | +3 | Most will never admit it. Do not require them to. |
| Wants implementation done for them | 7% | 6% | 5% | -1 | Smaller than assumed. |
| Non-branded is the gap | 6% | 5% | 9% | +4 | Rises when the numbers get examined. |
| Content is the bottleneck | 5% | 6% | 5% | -1 | I previously said 80%. It is 5%. |
| Found StudioHawk via content or search | 5% | 7% | 3% | -5 | Under-counted: most never say how they found you. |
| Organic dropped, cannot explain why | 1% | 2% | 4% | +2 | Rare as stated. Vivid when it happens. |
Half of Discovery conversations mention AI search. Barely a quarter of Proposal conversations do. It falls by 24 points, the largest move on the table. Meanwhile migration goes the other way, from 31% to 40%, and it is the only substantial theme that rises.
The honest caveat first: this compares conversations at two different moments, not winners against losers. A topic can fade simply because it was asked and answered. I do not have closed-won data, so nobody should call this causation.
But the messaging implication holds either way. AI search is what gets them on the call. It is not what they are still talking about when they are deciding. So put it in the cold ads, where it is doing real work, and stop building the VSL and the proposal around it, where the buyer has already moved on to whether their own site is broken and who is going to fix it. That is a cleaner split than the one I gave you before, which was to stop selling the category everywhere. Wrong: sell it at the top, drop it lower down.
52% have or had an agency. 8% criticise them, and it does not move between stages. This is now measured across 222 calls rather than inferred from ten, and it is the single most useful constraint on the creative: an ad built on "your agency is failing you" is talking to eight percent of the market and insulting the person who hired them, who is on the call 34% of the time.
The ads that fit this reality are the ones that treat the incumbent as a fact rather than an enemy: reason 41 frames the gap structurally, and reason 42 invites the agency to the audit.
Worth measuring rather than assuming, because it decides who every script is written for. Two independent tests across 220 transcripts, neither depending on the other.
| Test | Result | Read |
|---|---|---|
| Self-stated titles sentences where someone says what they themselves do | Marketing-side 22, owner-side 8 | Roughly 2.75 to 1 toward marketing. |
| Is the owner the person talking? needs no title: does the speaker call the owner "our owner", "the directors", "my boss" | Owner is someone else in 65% of the 80 calls where it is determinable | Roughly 2 to 1 toward marketing. |
Both tests point the same way: marketing-side people are who turn up. But turning up is not the same as deciding, and I originally wrote as if it were. The 65% figure is equally consistent with the owner being the approver who did not attend the call. The ICP resolves it, because its target is a business where someone other than the founder runs marketing, which describes who does the work rather than who signs the invoice.
So the VSL is written to the owner and names the marketing person out loud in the qualifier: you own or run a business doing a million a year or more, and somebody, an agency or someone on your team, is supposed to be handling search for you. It then closes by turning to the secondary audience directly, because a third of calls have more than one person on them and this video gets forwarded: if you're the one who actually runs marketing here and this got forwarded to you, it's for you too, probably more than it's for them. One hook, one film day, both audiences addressed without either being softened.
Honest limits. Most calls state neither signal: titles appear in 13% and the owner test resolves on 80 of 220. These are directional, not a census. What makes them usable is that two unrelated methods and the ICP all agree.
The best prospect in this dataset is not the founder, and not the SEO-savvy marketer. It is the senior marketing or ecommerce lead who inherited a website they did not build, from people who did not know what they were doing, and now has a commercial deadline they cannot move. They are not confused about whether search matters. They are confused about who to trust with it, and they are running a shortlist.
The two purest examples in the set are Mukhair (haircare, 20 years old, mid-migration, hard Black Friday deadline, asked unprompted for BAU pricing) and Bstore (25-year-old Birkenstock retailer, head of SEO performance who is candid about not being an SEO). Dimple Contacts is the same shape with a founder in the chair. If the new funnel converts more of these three, it is working.
Jeremy's rule: articulate their Circumstances and Outcomes so they feel read, then present the solution against their true Problems. The gap between what they blame and what is actually wrong is the whole opportunity.
This is the narrative they arrive with. Note that in every case it is something that happened to them.
The four recurring circumstances, in order of frequency: a migration or rebuild broke something; we have no internal capability; the last agency spent our money without moving the number; everyone's traffic is down because of AI. The fourth one matters most for messaging, because it is the story they use to make the first three feel less like their fault.
Jeremy: the real problem is almost always different from the one they name. Here it is, and it is the same in nine of the ten calls.
There is no single person accountable for search inside the business, so the decisions get made by whoever is nearest: a brand agency, a theme update, a developer clearing out old products. Nothing compounds because nothing is anyone's job.
The evidence is not subtle. Bstore's homepage H1 had been broken for months and nobody noticed. Brand URLs were never redirected because "we don't carry that anymore, it doesn't really matter". Dermapenworld had an hreflang fix sitting on staging, recommended by an agency they no longer employ. This is not a knowledge problem, it is an ownership problem, and it is why the audit lands so hard.
The second real problem, and the one that kills deals: they cannot prove it upward.
This is the single most under-served problem in the current funnel. The buyer is not only buying rankings. They are buying a defensible answer to the person who controls the budget. Nothing in the VSL currently speaks to that.
| Level | What they want | In their words |
|---|---|---|
| Tiny this month | Stop the bleeding on the migration. A to-do list they can hand to a developer without having to think. | "They seem to be quite a lot of quick wins that we could implement once we've got like a full brief." (Macro Mike) |
| Milestone this year | Non-branded revenue. Every one of them is over-indexed on brand terms and knows it. | "Part of what I'm trying to do is actually combat with unbranded terms. But I lack the expertise." (Bstore) · "We're 93% branded. It's like 7% of the base, nothing but upside." (Mukhair) |
| Grand the real one | A channel they own, so they stop being at the mercy of the brands they resell, the platform, or the ad auction. | "Even if we build authority with one brand, it may or may not serve us in the long term." (Bstore) · "I want to turn things around. I want to get more money out of this." (Ultra Nature) |
Per the SOP's weight-loss-clinic example, go one step deeper than the surface outcome. The surface outcome is "more organic traffic". One step deeper is "non-branded revenue I can point at". Two steps deeper, and this is the one that actually moves them, is "a number I can take to the board that is obviously ours, that keeps growing whether or not I keep paying for it."
The SOP is blunt that demographics are the wrong lens and conditions are the right one. Two heads of ecommerce at $10m businesses are not the same buyer if one is mid-migration and the other is not. These are the conditions that actually predict a deal in this dataset.
The SOP's rule is that you never state a problem flat, you pair it with the emotional hot button. These are the five amplifiers that are actually present in the recordings, so none of them has to be invented.
Note what this is not: it is not "agencies are bad". It is churn and padding, two specific, nameable things. An ad that says "your agency is ripping you off" attracts the wrong person. An ad that says "you have had four account managers in eighteen months and none of them read the last one's notes" is precise enough to be believed.
The strongest emotional territory in the entire dataset, and the current funnel does not touch it. This person is not afraid of AI search. They are afraid of a specific Tuesday.
This is the amplifier with the most upside, because it is the one nobody else in the market is speaking to. The buyer's problem is not their own belief. It is that they are surrounded by confident, wrong people, and they do not have the ammunition to shut the conversation down.
This is the closest thing to a winning mechanism in the whole dataset, and it is already what the audit does. Bluntness beat politeness head to head against a named competitor. The VSL should promise the uncomfortable version explicitly, because the person who wants it self-selects and the person who does not is not the buyer.
The Stadium Pitch says three to four percent are in-market with a burning problem, thirty percent need convincing, and the rest never buy. Jeremy's instruction is to talk to the three to four percent first and only move to the thirty percent once that seam is exhausted. So which one are these calls?
This is not inference. They say so, unprompted, in the first ten minutes of the call.
Thirty-six percent of all 246 calls reached Proposal or Pitch stage. That is not a cold-audience number. The seam is real and it is nowhere near exhausted, which per the SOP means going after the needs-convincing thirty percent now would be the wrong move.
In this dataset, the person who needs convincing is almost never the person on the call. It is the person they report to. The buyer is already sold. Their MD, their board or their founder is not, and the buyer has no ammunition.
Dermapenworld put it in one sentence: "I'm the biggest advocate of SEO, always been in the company. Scott understands the importance but I want him to understand even better. The way that you're going to show, it's gonna be eye opening. I'm hoping at least." They had literally brought their MD onto the call to be convinced by someone else.
Ultra Nature is the same story with a worse ending: the board cut the spend because the marketer could not justify it. Tax Assure is the same story pre-emptively: they are worried their directors will hear "SEO is dead" at a business lunch.
So the second audience is not a colder version of the first. It is the same buyer, carrying an internal sceptic. That is a completely different messaging job. You are not persuading the viewer. You are arming them. And it fits exactly what you said about most consumption happening on the back end: the education content does not need to win the click, it needs to survive being forwarded to a managing director.
| Audience | Where it belongs | The job |
|---|---|---|
| In-market 3 to 4 percent | Cold ads, the VSL, the audit offer | Win the shortlist. They already believe search matters; they are choosing who. Every second spent explaining that AI search is real is a second not spent differentiating. |
| The internal sceptic the real "needs convincing" | Confirmation page, the Hammer Them layer, nurture, the two-day event | Arm the advocate. Make a piece that survives being forwarded with the line "watch this". Named sources, no hype, no red, a number a CFO respects. |
| Genuinely cold the other 30 percent | Not yet | Do not build for these until the in-market seam stops producing. The SOP is explicit and the proposal-stage rate says the seam is healthy. |
On the risk question you raised: speak to the in-market person. Going after the higher-risk needs-convincing audience first would be the wrong bet on this data, for three reasons. The in-market volume exists and converts. The needs-convincing content is back-end consumption anyway, by your own read. And the SOP's whole argument is that mismatched messaging to the wrong group is the usual cause of a bad show rate, which is a problem worth avoiding rather than creating.