Headline and Hook Generator
Generate twenty headlines across five persuasion angles, then ship the one that scores.
Generates twenty headline and hook options spread across five distinct persuasion angles, scores each for specificity and believability, and returns the three worth testing plus the reason the weaker angles failed for your audience.
Ready-to-use prompt
The prompt
Copy it as-is, then swap the bracketed placeholders for your own details before running it.
Role: You are a direct response copywriter generating headline options.
Context:
- Product or offer: {{offer}}
- Audience and what they already believe: {{audience}}
- The single outcome the offer delivers: {{outcome}}
- Proof available (numbers, names, results): {{proof}}
- Where the headline appears: {{placement}}
- Brand voice and any banned words: {{voice}}
Task: Produce twenty headlines, four each across five angles:
1. Outcome, the result stated plainly
2. Problem, the pain named in the reader own words
3. Curiosity gap, without withholding the subject
4. Proof led, opening on a number or named result
5. Contrarian, challenging a belief the audience holds
For each headline give: character count, the angle, and the one word doing the persuasive work.
Then score every headline 1-5 on specificity and 1-5 on believability, and return:
- The top three by combined score, with a matching opening line for each
- Any headline whose believability scored below 3, and what claim caused it
- The angle that performed worst for this audience and why
Rules:
- No headline may contain a superlative that the proof does not support.
- Never use a curiosity gap that hides what is being sold.
- Stay within the placement character limit and show the count.
- Do not reuse the same lead noun in more than three headlines.
- Tie each factual claim to supplied proof; label proof-led slots blocked when evidence is insufficient.
- Break combined-score ties by believability, then specificity; identify any remaining tie.
- Calculate the weakest angle using mean combined score, then explain its audience mismatch.
Finish with: the one headline you would run if you could only run one, and the risk it carries.Estimated results
Editor's note
Why this prompt matters
Most headline sessions produce twenty versions of the same idea. The wording changes, the angle never does, and the resulting test teaches nothing because every variant rests on one unexamined assumption about what the reader cares about. This prompt forces the split up front: four headlines per angle across five genuinely different persuasion routes, then a scoring pass that separates a specific claim from a vague one and a believable claim from an aspirational one. The output you act on is small — three headlines and a reason — but you get there having actually considered the alternatives.
Anatomy
Prompt engineering breakdown
Role
Context
Goal
Constraints
Output format
What you'll get
Expected output
Brief and evidence boundary
This fictional offer is a downloadable pack of six overdue-invoice email templates for freelancers who believe reminders must sound forceful. The desired outcome is drafting a clear payment request without starting from scratch. Supplied product facts establish the template count, not faster payment. Placement: landing-page headline, maximum 45 characters; voice: calm and direct.
Twenty candidates
S = specificity; B = believability. Both scores run from 1–5. Counts include spaces and punctuation, but exclude table formatting.
| # | Headline | Characters | Angle | Persuasive word | S/B | |---|---|---:|---|---|---| | 1 | Write your next invoice reminder | 32 | Outcome | Write | 4/4 | | 2 | Ask for payment without starting over | 37 | Outcome | starting | 4/5 | | 3 | Send a clear request for payment | 32 | Outcome | clear | 3/5 | | 4 | Draft a follow-up from a template | 33 | Outcome | template | 3/5 | | 5 | Still rewriting invoice reminders? | 34 | Problem | rewriting | 4/4 | | 6 | Unsure how to phrase that request? | 34 | Problem | phrase | 4/5 | | 7 | Another overdue invoice. What now? | 34 | Problem | overdue | 4/4 | | 8 | Stop chasing clients forever | 28 | Problem | forever | 2/1 | | 9 | What belongs in a payment reminder? | 35 | Curiosity | belongs | 3/5 | | 10 | A follow-up email, already drafted | 34 | Curiosity | drafted | 3/5 | | 11 | The next line in your payment request | 37 | Curiosity | next | 3/4 | | 12 | See how these payment reminders begin | 37 | Curiosity | begin | 3/5 | | 13 | 6 templates for overdue invoices | 32 | Proof | 6 | 5/5 | | 14 | 6 scripts for asking about payment | 34 | Proof | 6 | 3/5 | | 15 | 6 examples of invoice follow-up | 31 | Proof | 6 | 3/5 | | 16 | 6 reminders you can adapt | 25 | Proof | adapt | 3/5 | | 17 | Your reminder need not start from zero | 38 | Contrarian | zero | 3/4 | | 18 | A firmer email cannot promise payment | 37 | Contrarian | cannot | 3/5 | | 19 | Templates can still sound like you | 34 | Contrarian | you | 3/4 | | 20 | More pressure is not a payment plan | 35 | Contrarian | pressure | 2/4 |
Three worth testing
#13 — 10/10. Opening: “Use these six email templates as starting points for your overdue-invoice reminders.” Concrete inventory makes the offer immediately legible; this is feature proof, not evidence of payment results.
#2 — 9/10. Opening: “Adapt a drafted reminder instead of composing your payment request from a blank page.” It promises help with writing, which the product can deliver.
#6 — 9/10. Opening: “Start with an overdue-invoice email template, then adjust the wording to fit your client.” Its specificity depends on the surrounding invoice context; isolated in an ad, it would score lower.
Rejections and final choice
Below the believability threshold: #8 implies the templates permanently eliminate chasing clients. No supplied evidence supports that claim.
Weakest angle: Contrarian has the lowest average combined score. Only #18 and #20 directly challenge the stated belief, and both foreground what cannot be controlled rather than the useful deliverable. #17 and #19 drift into ordinary template benefits.
Run #13 first. Its risk is sounding like a commodity bundle. The opening should demonstrate usefulness rather than escalate to unsupported collection promises. These scores are editorial screening judgments, not predicted conversion rates.
Under the hood
Why this prompt works
The believability score is what makes this more than a brainstorm. A model asked for exciting headlines will happily write a claim your product cannot support, and that claim reads fine until a customer tests it. Scoring it separately from specificity catches the common failure where a headline is admirably concrete and quietly untrue. The angle split does the other half of the work: because variants differ structurally rather than lexically, any test you run afterwards returns information about your audience rather than about synonyms.
Model fit
Best AI models for this prompt
ChatGPT
Best default. Keeps the five-angle split intact and counts characters accurately.
Claude
Strongest on believability scoring and the most willing to reject a superlative your proof does not support.
Gemini
Useful when you paste competitor headlines and want deliberate differentiation.
When to use
- Before drafting a landing-page hero, once the offer and proof are settled.
- When an email or ad opening needs alternatives built around different motivations.
- After weak click-through results, to generate hypotheses—not diagnose the cause.
- Before a test cycle, to screen variants for duplicated promises and unsupported claims.
When not to use
- When the audience is undefined; customer interviews beat invented pain language.
- When positioning is unsettled; headline scoring cannot clarify an incoherent offer.
- For SEO title tags requiring keyword, search-intent and display-length rules.
- For regulated claims without qualified human review, or performance predictions without actual experiment data.
Get more from it
Pro tips
- 1
Supply verifiable proof and its limits. Product counts support inventory claims; customer outcomes require separate evidence.
- 2
Treat believability below 3 as a revision gate, not permission to publish with a warning.
- 3
Test different angles, not synonyms. Replace finalists that make essentially the same promise.
- 4
Keep each matching opening line attached; check that it delivers the headline’s promised subject immediately.
- 5
Put the exact character limit and surrounding context in {{placement}}; verify final counts independently.
- 6
Include audience phrases and objections in {{audience}}, distinguishing observed language from your assumptions.
- 7
For a second pass, name the failed claim or angle. Request replacements only for those slots, preserving supported finalists.
Don't ship this
Common mistakes
✗ Testing four variants that all make the same claim.
Fix — Enforce the angle split; variants must differ structurally, not lexically.
✗ Shipping the headline that sounds best in the meeting.
Fix — Use the combined score, then test — meetings are not the audience.
✗ Running it without an audience definition.
Fix — Name what the reader already believes; the contrarian and problem angles depend on it.
People also ask
Frequently asked questions
Q.Why generate twenty headlines instead of five?
Because the value is in covering five distinct angles, not in volume. Four options per angle gives you enough range to see which route suits the audience, and the scoring pass narrows it back down to three worth testing.
Q.What does the believability score actually catch?
Claims your product cannot support. A headline can be admirably specific and still overstate the result, and that is the version customers test hardest. Anything scoring below 3 should be verified or cut before it ships.
Q.Can I use this for SEO title tags?
Not directly. Title tags need keyword placement and length rules this prompt ignores in favour of persuasion. Use it for ad, email and landing page headlines, then adapt the winner for search separately.