Skill Locker
All skills
v1.3.03 loop iterations

A/B Headline Generator

Most pages live or die on the headline, and most creators write one and ship it. The A/B Headline Generator produces five variants that each pull on a different psychological lever — so the test isn't five rewordings, it's five real hypotheses.

What this skill does

Most creators write one headline, decide it's fine, and ship. That's the single highest-impact element on the page being treated as a draft. Changing a headline can swing conversion 20-100% with zero other changes — and the only way to find the lift is a structured test, not a tweak.

This skill doesn't brainstorm. It applies five distinct psychological frameworks to your core message — outcome-driven, problem-aware, curiosity gap, social proof, and direct how-to — so each variant tests a real hypothesis instead of a synonym. If three of your five variants are just the same angle reworded, you've learned nothing from the test. The framework rule forces genuine separation, and each variant ships with the lever named: "this targets loss aversion in buyers who've been burned by competitors before."

It's context-aware. An email subject line gets optimised for 41-50 visible characters and a preview-text suggestion. A landing page hero gets 5-12 words. An ad headline gets the platform's character limit. The skill won't hand you a 14-word variant for a Google Ads asset and call it a day.

Power words and number psychology are applied where they earn their place — odd numbers over even, specific over round, "without [common objection]" clauses to remove the perceived barrier — and refused where they'd feel forced. A plain headline that reads honest beats a headline stuffed with "proven, instant, guaranteed" that reads like a 2009 affiliate page.

Every variant lands with the hypothesis written down, character count checked, and at least one variant deliberately simpler than the current — because clarity wins more tests than cleverness. The output finishes with a testing plan: which variant to run first, sample size needed for statistical significance at your traffic volume, and the success metric. A test you run without sample-size maths isn't a test, it's a guess in a spreadsheet.

When this triggers

  • ·You have a current headline and traffic to test against it
  • ·Your hero is converting badly and you don't know if the angle is wrong or the wording is
  • ·You need variants for an ad set, email subject line, or landing page hero
  • ·You want each variant tied to a clear hypothesis, not vibes
  • ·You need a defensible test plan with sample size — not 'let's run it for a week'

Example

Trigger

User: 'Current headline: Marketing Automation for Modern Teams. Page sells a workflow tool to ops leads at 50-500 person SaaS. Convert at 1.8%.'

Output

CURRENT: "Marketing Automation for Modern Teams" Analysis: vague benefit, no audience signal, no differentiator. "Modern Teams" is filler. V1 — Outcome-Driven: "Cut Campaign Setup From 4 Hours to 20 Minutes" Lever: specific, measurable outcome. Targets ops leads under quota pressure. V2 — Problem-Aware: "Still Stitching Together Zaps to Run Your Funnel?" Lever: pain recognition. Targets the Zapier-fatigued. V3 — Social Proof: "The Workflow Tool 400+ SaaS Ops Teams Switched To" Lever: peer validation. [+ V4 Curiosity, V5 Direct How-To, sub-headlines, sample-size calc: ~4,200 visitors per variant for 95% confidence at the expected lift band]

Get this skill + 10 more

Get the full Websites & Landing Pages pillar (11 skills) or the complete library.

Get the full stack — $299

What you get

  • 168-line SKILL.md, ready to drop into ~/.claude/skills/
  • Tested through 3 Karpathy-loop iterations (versions v1.0.0 → v1.3.0)
  • Triggers automatically when relevant — no command to remember
  • Lifetime updates as the skill is refined further

More from Websites & Landing Pages

Browse the full library

297 skills across 31 categories. One purchase, lifetime updates.

See all bundles