Verified against Claude · 2026-07-08
Draft A/B test headline variants for a landing page
Get a set of genuinely different headline angles for a landing page test, each with the hypothesis it is testing — not ten synonyms of the same sentence.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Landing page purpose: Landing page for a free trial signup for an invoicing tool aimed at freelancers. Current headline (if any): "Send invoices in seconds". Primary benefit to communicate: Get paid faster because clients can pay directly from the invoice. Audience: Freelance designers and consultants who currently use Word or Excel invoices. Write 6 headline variants for an A/B test, each testing a genuinely different angle — not the same sentence reworded. Use these 6 angles specifically: 1. Outcome-led (what changes for them) 2. Pain-led (the problem they're currently living with) 3. Specific-number-led (a concrete, honest number — not a made-up stat) 4. Social-proof-led (implies validation without a fabricated claim) 5. Speed/simplicity-led 6. Direct/no-frills, stating exactly what the product does For each headline, add one line: what hypothesis this variant tests, so I know what I'll actually learn depending on which one wins. Keep every headline under 60 characters so it doesn't wrap awkwardly on mobile. Do not invent a specific statistic, review count, or customer number I haven't given you — if the social-proof angle needs a real number, mark it as a placeholder I need to fill in, don't guess one.
Customize the highlighted detailsoptional — the prompt above already works
Why this works
Headline tests fail to produce a learning when every variant tests the same underlying idea with different words — you learn 'people liked headline B' without knowing why. Assigning each variant a distinct, named angle (outcome vs. pain vs. proof vs. speed) turns the test into something that answers a real question about what resonates, and the explicit ban on inventing statistics closes off the most common way AI-generated landing page copy quietly becomes false advertising.
What you get back
1. Outcome-led: "Get paid the same day you send it" — tests whether speed-of-payment beats speed-of-sending as the hook. 2. Pain-led: "Stop chasing clients for payment" — tests whether the pain of chasing (not the invoice itself) is the real trigger. 3. Number-led: "Paid on average 9 days faster" [placeholder — insert your real figure] — tests whether a concrete number outperforms a feeling-based claim.
Verified against
Claude Claude Opus 4.5 · 2026-07-08
ChatGPT GPT-5.1 · 2026-07-09
Changelog
- 2026-07-09 — Initial version.
Need this built into your business?
If a prompt isn't enough — Google Ads management, built and maintained for you — that's Scult's day job.
EXPLORE GOOGLE ADS MANAGEMENT
