Fifteen Headlines and One Honest Promise

A squeeze page might be a hundred words. That is precisely what makes it hard — there is no room to explain yourself, and every word is carrying weight.

It is also the piece of writing AI is best suited to, for a reason that has nothing to do with how clever the model is.

You do not want one headline

You want fifteen, so you can throw away fourteen.

That has always been the advice, and it has always been the advice people ignore, because writing fifteen headlines is tedious and by the ninth you have stopped thinking. The bottleneck in split testing was never the testing. It was producing enough options worth testing.

Generating variations is a job with no research to skip and nothing to invent. It is the safest possible use of a language model — right up until it is not.

Where it goes wrong

Short persuasive copy rewards bold claims. Ask for "punchier" headlines and a model will drift towards promises you never made.

Double your traffic in 30 days. The only guide you will ever need. Join 10,000 subscribers.

Every one reads well. Not one of them came from you. The model is not being dishonest — it has learned what high-converting copy sounds like, and high-converting copy makes specific promises. It has no way to know which specific promises are true of your business, so it supplies plausible ones.

An overstated squeeze page is worse than a weak one. It sets an expectation your freebie then fails to meet, which is an efficient way of teaching a brand-new subscriber to ignore everything you send them afterwards.

The instruction that fixes it

The usual advice is to write a better prompt. Here the useful move is narrower — tell it what it is not allowed to say:

Write fifteen headline options from the plan below. Do not state any benefit, number, timescale or social proof that is not in the plan — no subscriber counts, no results, no guarantees. If you think a claim would make the page stronger, tell me what evidence you would need instead of writing it.

That last sentence is the one that earns its place. It converts the model's instinct to embellish into something genuinely useful: a list of the proof you could go and gather, so the page gets stronger honestly rather than fictionally.

Generate a test, not a draft

Once you have the plan, ask for variations as a set rather than a single best attempt:

Give me five headline variations, each leading on a different angle: the problem, the outcome, the speed, the specific person it is for, and curiosity. Keep the same promise in all five.

Then test them one variable at a time. What you end up with is a page you chose, rather than the first one you happened to write.

What stays yours

• The promise. What your freebie actually delivers is a decision about your business.
• Numbers. Subscriber counts, download figures, results. These are the most likely to be invented, precisely because they make copy so much stronger.
• The freebie itself. A page that converts brilliantly onto a disappointing download is worse than one that converts less.
• Your privacy wording. What you do with an email address is a statement of fact, and in most places a legal one.

Start with the plan

Our course High-Converting Squeeze Pages covers what belongs on the page and what has to come off, the copy and design decisions that move conversions, the SEO and mobile basics, the mistakes that quietly cost sign-ups, and how to test your way to a better version. The first two chapters are free and need no card.

The last chapter is the one this post is about: turning that plan into a brief, so the fifteen headlines you get back are fifteen versions of something true.