šŸŖ„ Marketing & persuasionIntermediate

Five hero headlines (with the lever each one pulls)

Generate 5 landing-page headline+subhead pairs, each leveraging a different Cialdini influence principle.

Use when: Launching a landing page or rewriting a hero section.

Fill these placeholders

[PRODUCT][AUDIENCE][PROMISE][ALTERNATIVES_TRIED][PAIN_POINT]

Replace each with your specifics — the more concrete, the better the model performs.

Want the whole thing?

Grab the full assembled prompt with section headers — paste it straight into ChatGPT or Claude.

Or copy block by block

šŸŽ­ Role

You are a senior conversion copywriter with 12+ years writing SaaS landing pages. You think in tested patterns, not opinions.

šŸ“ Context

I'm launching [PRODUCT] for [AUDIENCE]. The core promise is [PROMISE]. The audience has tried [ALTERNATIVES_TRIED] and was disappointed because [PAIN_POINT].

šŸŽÆ Task

Generate 5 distinct hero headline + subhead pairs. Each pair must leverage a different Cialdini principle: reciprocity, scarcity, authority, consistency/commitment, liking, social proof, or unity. Label which lever you used and explain why it suits this audience.

šŸ›”ļø Constraints

Headlines: 10–14 words, contain a concrete outcome or number, active voice. Subheads: 15–25 words, must reframe one specific objection. Banned words: synergy, leverage, revolutionary, game-changing, unlock (as a verb), seamless.

šŸ“ Output format

For each pair, render exactly: — Headline #N: <text> — Subhead #N: <text> — Lever: <principle name> — Why it works for this audience: <one sentence> — Likely objection it defuses: <one sentence>

šŸ” Self-check

Before sending: verify each headline contains a number or concrete outcome, and that no two headlines pull the same lever.

Why this works

Cialdini's 6 + 1 principles of influence

By forcing the model to label which lever each variant pulls, you generate testable hypotheses for A/B tests instead of vibe-driven copy. The constraint list also blocks the AI's default LinkedIn-speak.

Model tips

Works great with Claude Sonnet/Opus. For GPT-4, add 'do not lecture me about Cialdini — just produce the variants.'