Hey, Mark here.
I spent two days once on the best sales page I'd written that quarter. Tight lead, clean proof section, a close I was proud of.
It converted at almost nothing.
The copy was fine. The offer underneath it was the problem. Same price as three competitors, no real reason to pick this one, a guarantee so weak it read as a warning. No sentence I could write was going to fix that, and I spent two days finding out the expensive way.
The offer is the lever. Copy is the force you apply to it.
A strong offer with average copy beats a weak offer with brilliant copy, every time. And most of us were trained to start writing the moment the brief arrives, without checking which one we've been handed.
This is the system I built so I stop writing first and testing second. It runs in about twenty minutes, before a word of the page exists.
Here's the build.
You Can't Write Your Way Out of a Bad Offer
The Two Days You Lose Writing the Wrong Page
You get the brief. The offer's in there somewhere: the product, the price, the promise. You read it, it seems fine, and you start writing, because writing is the job and the deadline is real.
Three drafts in, something feels off. The proof section won't hold weight. The close keeps going soft. You assume it's your copy and you push harder, rewriting the lead for the fourth time.
The copy usually isn't the culprit. The offer was weak in a specific spot, and your draft has been straining to cover for it since the first paragraph.
Here's the part that costs you. A weak offer doesn't announce itself. It feels like a writing problem, so you spend your skill, your hours, and your best lines trying to write around a hole that words can't fill.
The fix is to test the offer before you write the page, not after the third draft tells you something's wrong. Twenty minutes up front against a system, instead of two days of drafting into a headwind.
The System, in One Paragraph
You capture the offer in a structured format, then run it through Claude as a hostile buyer across six pressure points: promise, mechanism, proof, value-to-price, risk, and speed. The model attacks each one the way a skeptical prospect would, tells you which point is weakest, and shows you the proof or the change the offer needs before any copy gets written. You finish with a one-sentence offer you can actually build a page on, or a clear note back to the client about what the offer is missing.
The system has four steps:
Capture the offer.
Run the skeptic pass.
Find and fix the weakest link.
Rebuild the offer in one sentence.
I'll walk through each.
One distinction before the build. This tests the offer, not the brief. I wrote a separate system for auditing whether a brief gives you enough to write from. This one assumes the brief is fine and asks a different question: is the thing you've been asked to sell actually strong enough to sell? A complete brief can describe a weak offer in perfect detail.
Step 1: Capture the Offer in a Format You Can Attack
You can't pressure-test an offer that lives in scattered sentences across a brief. The first move is to pull it into one structured sheet, because the gaps show up the moment you try to fill in the fields.
Run this prompt with the brief, the product page, or whatever the client gave you pasted underneath.
I'm going to paste everything I have about a client's offer. Pull it into a structured offer sheet using exactly these fields. If a field isn't covered in what I give you, write "NOT SPECIFIED" rather than filling it in with a guess.
- Product: what it literally is.
- Audience: who it's for, in one specific sentence.
- Core promise: the main outcome the buyer is paying for.
- Mechanism: the specific reason this product delivers that outcome, the thing that makes it work when other options don't.
- Proof: every piece of evidence currently offered that the promise is real (testimonials, numbers, credentials, demonstrations, guarantees).
- Price and terms: the cost, the billing, anything about contract length or commitment.
- Risk reversal: what protects the buyer if it doesn't work (guarantee, trial, refund).
- Speed: how fast the buyer reaches the outcome, and how much effort it takes them.
After the sheet, list every field you had to mark NOT SPECIFIED. Do not write any marketing copy.
[paste everything you have about the offer]
The "NOT SPECIFIED" rule is the point of this step. Most copywriters fill those gaps unconsciously while drafting, inventing a mechanism or assuming a guarantee that was never there. Forcing the model to flag the blanks shows you exactly how much of the offer you were about to make up.
The list of blanks at the end is your first finding. If "mechanism" and "risk reversal" both come back NOT SPECIFIED, you already know where two of the six pressure points are going to fail.
Step 2: Run the Offer Through a Hostile Buyer
Now you hand the offer sheet to a skeptic. Not a polite reviewer. A buyer who has been burned before, has the money, and is looking for a reason to keep it.
The six pressure points are the same six fields from the sheet, and each one is a question a real prospect asks before they pay.
Pressure point | The buyer's question |
|---|---|
Promise | Do I actually want this outcome enough to pay? |
Mechanism | Why would this work when the last thing I tried didn't? |
Proof | Why should I believe any of this? |
Value-to-price | Is what I get clearly worth more than what I pay? |
Risk | What happens to me if it doesn't work? |
Speed | How long until I see something, and how much work is it? |
Here's the prompt. Paste the offer sheet from Step 1 underneath.
Below is a structured offer sheet. I want you to act as a specific, skeptical buyer from the target audience. You have been disappointed by a similar product before. You have the money to buy this, and you are actively looking for reasons not to.
Go through these six pressure points in order. For each one, do three things: ask the blunt question you'd actually ask, give your honest reaction to how well the offer answers it right now, and rate the offer's current strength on that point from 1 to 5.
1. Promise: Do I want this outcome enough to pay for it?
2. Mechanism: Why would this work when other things I've tried haven't?
3. Proof: Why should I believe the promise is real?
4. Value-to-price: Is the value clearly greater than the price?
5. Risk: What protects me if it doesn't work?
6. Speed: How fast do I get a result, and how much effort does it take me?
Stay in character as the skeptic. Don't be polite, and don't give the offer credit for things it hasn't actually said. If a point is weak, say so plainly and explain what would change your mind.
End with the single pressure point that would most stop you from buying.
[paste the offer sheet]
Why a character instead of a checklist. Asked to "evaluate the offer," Claude defaults to a generous review and finds things to praise.
Told to be a specific burned buyer who wants to keep their money, it reads the offer the way the market actually reads it.
The skeptic catches the soft spots a neutral reviewer talks itself past.
The 1-to-5 ratings give you a quick map of where the offer is strong and where it's exposed. The final answer, the one point that would most stop the purchase, is what Step 3 attacks.
Here's what that looks like on a real offer. I ran the system on a done-for-you bookkeeping service for freelancers, $299 a month, promising "stress-free books and on-time taxes."
The skeptic scored it unevenly:
Promise — 4. The outcome was wanted, and it came fast enough.
Speed — 4. Fast enough to matter.
Value-to-price — 3. Defensible, not commanding.
Mechanism — 2. Every bookkeeping service promises clean books. Nothing said why this one delivers them when the buyer's last bookkeeper missed a filing.
Risk — 2. A vague "satisfaction guarantee" with no terms, which the skeptic read as no guarantee at all.
Then the closing line, the one I actually needed:
"I've been burned by a bookkeeper before, and nothing here tells me why this time is different, or what happens to me if it isn't."
Two days of polishing the copy would never have surfaced that. Twenty minutes did.
Step 3: Find the Weakest Link and Fix It Before You Write
An offer doesn't convert on its average. It fails on its weakest point. A buyer with a 5 on every pressure point and a 2 on risk walks because of the 2, and a stronger promise can't buy that back.
So you fix in order of weakness, starting with the point the skeptic flagged. Run this prompt in the same chat, right after the skeptic pass.
Take the weakest pressure point you just identified, the one that would most stop the purchase.
First, tell me whether this is a copy problem or an offer problem:
- A copy problem means the offer already has what it needs, but it wasn't stated, framed, or surfaced clearly. I can fix this on the page.
- An offer problem means the thing itself is missing something a buyer needs, and no wording fixes it. This goes back to the client.
Then:
- If it's a copy problem, tell me exactly what to surface and where, and what proof or detail I should pull forward that's already true.
- If it's an offer problem, tell me specifically what the offer would need to add or change to clear this pressure point, phrased as a recommendation I can take to the client.
Do this for the weakest point first, then the second weakest. Be concrete. "Add a guarantee" is not enough. Tell me what kind of guarantee, framed how, and why it answers the buyer's specific fear.
[continue from the skeptic pass above]
The copy-problem-versus-offer-problem split is the most useful line in the whole system. It tells you which weaknesses you can fix at your desk tonight and which ones you have to raise with the client before you write. Pretending an offer problem is a copy problem is exactly how you end up rewriting a lead four times.
The payoff cuts two ways:
Copy problem: free conversion lift, sitting in proof the offer already had and never surfaced.
Offer problem: a specific, defensible recommendation to send the client, the kind of strategic input that keeps you out of the commodity tier.
One note on proof, because it's the point that fails most often. The model can tell you what proof the offer needs. It must never invent the proof itself. If the fix is "add a specific result," that result has to come from the client's real records. A fabricated number in a stress-test is worse than a weak offer, because it ships.
Step 4: Rebuild the Offer in One Sentence
If the offer survived the first three steps, or you've patched the copy-level gaps, you close by forcing it into a single sentence. A strong offer can be said in one clean line. If it can't, a pressure point is still soft.
Based on the strengthened offer, write the offer in a single sentence using this structure as a starting point, not a template to follow rigidly:
[Specific audience] gets [specific, desirable outcome] through [the mechanism that makes it work], proven by [the strongest piece of real proof], for [price], protected by [the risk reversal].
Write five versions. Vary which element leads, because the strongest element should usually open the sentence. After the five, tell me which one is strongest and why, and which single pressure point is still the softest even in the best version.
Do not pad with adjectives. If an element is weak, a better sentence won't hide it, so flag it instead of dressing it up.
[continue from above]
The sentence is the spine the page hangs on, not the finished headline. Once you can say the offer in one honest line, the lead, the proof order, and the close mostly arrange themselves, because you know which element is carrying the most weight and which one needs the most support.
If the model keeps telling you the same pressure point is soft even in the best sentence, believe it. That's the offer telling you where the page will strain, and it's better to know now than on draft three.
Setting It Up So You'll Actually Use It
Four prompts, one chat, start to finish. The whole pass runs in about twenty minutes once you've done it twice.
Keep the four prompts as saved snippets, the way you'd keep any tool you reach for often. I store mine in the same prompt library as the rest, in an offer/ folder, so the stress-test sits next to the brief audit and the voice tools.
If you run a Claude Project per client, the stress-test has a natural home there. Drop the offer sheet from Step 1 into the project as a file once it's strengthened. Every piece you write for that client after that can reference the offer the page is actually built on, instead of you reconstructing it from memory each time.
Run it at one specific moment: after you've read the brief, before you've written a line. Running it after the first draft is too late, when you're already invested in the page and tempted to defend it. The whole value is in testing before you write, so the twenty minutes only pays off if it comes first.
When the offer is the client's and you can't change it, the stress-test still earns its time. The copy-versus-offer split in Step 3 hands you a clean list of what to surface on the page and a separate list of offer gaps to raise with the client. That second list, sent as a short note before you write, is often the most valuable thing you deliver that week.
What the Stress-Test Can't Do
Three honest limits before you build it.
It can't validate the offer against the real market. The skeptic is a model playing a buyer, informed by patterns in its training, not your actual audience. A 5 across the board is a good sign, not a sales forecast. The market is still the only real test, and this just stops you from shipping an offer with an obvious hole in it.
It can't change the client's pricing or product. Plenty of offer problems trace back to decisions above your pay grade. The system will name them clearly, and naming them is your job; fixing them is a conversation you start, not a line you write.
It can't supply proof that doesn't exist. The model identifies the proof gap. You and the client fill it with real evidence. There is no prompt that turns a missing case study into an honest one.
What you get for the twenty minutes is the difference between writing with the offer and writing against it. You stop spending your best hours dressing up a weak point, and you start every page knowing exactly where its strength is and where it needs you most.
The Master’s Memo
The offer was deciding your results before you wrote a word. You just weren't testing it until the draft pushed back.
Move the test to the front, find the weak point in twenty minutes, and either fix it or flag it before you write.
Run it on your next project, before the first draft, on the offer you're least sure about.
More clicks, cash, and clients,
Mark Masters
