Generate 3-5 messaging variants for a value proposition, design structured A/B tests, and analyze results to determine which framing resonates most with ICP. Tests can run via LinkedIn organic posts, cold email subject line splits, or both. Pure reasoning for variant generation and analysis — the user deploys the tests through their own tools.
SkillScore82Maintained
Automatic
◆Marketing & Content pack
Last verified 8/16/2026
What it changes
Passed 18 of 22 test cases with the skill, and 17 of 22 without it.
Without skill77%
With skill82%
Checks it passed that the model alone missed
Recommends testing organic LinkedIn posts rather than paid ad campaigns.
Rejects 100-character email subject lines to prevent truncation on mobile devices.
2 fixed16 passed either way3 still failing1 broke
Pass-rate lift
+5 pp
Tokens
+75%
Agent turns
0%
Response time
-23%
Model gemini-3.6-flashMeasured 8/16/2026v118 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 2 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 2
Prompt
We can't decide how to position our AI-powered code review tool to CTOs at seed-stage startups. We currently lead with 'Catch bugs before production', but we want to test other angles against our current messaging. Can you build a set of contrasting messaging options for us to A/B test? Include the hypothesis for each, actual copy for LinkedIn and email campaigns, plus step-by-step instructions on how to run and measure the test across both channels.
What was checked, without and with the skill
Checked
Without skill
With skill
Recommends testing organic LinkedIn posts rather than paid ad campaigns.
Missed
Passed
Provides distinct positioning variants contrasting against the baseline message.
Passed
Passed
case-03 · gemini-3.6-flash · 8/16/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md253 linessingle file
Messaging A/B Tester
Stop debating which message is better — test it. Generate messaging variants, deploy them through real channels, and measure which framing actually resonates with your ICP.
Core principle: At seed/Series A, you don't have enough traffic for website A/B tests. But you do have enough LinkedIn impressions and cold email sends to test messaging angles fast.
When to Use
"Which of these value props should we lead with?"
"Test our messaging angles and tell me which works"
"I can't decide between [message A] and [message B]"
"What messaging resonates most with [ICP]?"
"Run a messaging test for [product/feature]"
Phase 0: Intake
What to Test
Core value prop — The claim or positioning you want to test (e.g., "We help growth teams run outbound 10x faster")
Test goal — What are you deciding? (Headline for website, cold email angle, LinkedIn content strategy, ad copy direction)
ICP — Who should this resonate with? (Title, company type, stage)
Current messaging — What are you using today? (Baseline to beat)
Test Channel
Where to test:
LinkedIn organic — Post variants across consecutive days, compare engagement
Cold email — A/B test subject lines or opening hooks via Smartlead
Both — Run in parallel for fastest signal
Sample size available:
LinkedIn: followers/typical impressions per post
Email: list size available for testing
Constraints
Number of variants — 3-5 recommended (more = slower signal)
Test duration — How long to run? (Default: 1 week for LinkedIn, 3-5 days for email)
Phase 1: Generate Messaging Variants
Create 3-5 variants that test different angles, not just different words. Each variant should represent a distinct strategic bet: