| name | co-growth-copy-variants |
| description | Copy variant generation skill. Creates messaging alternatives for experiment arms
with variable isolation, segment-appropriate tone, and A/B-ready formatting.
Use when GENERATING copy variants, writing CTA alternatives, creating messaging
for experiment arms, or adapting tone for different user segments.
|
Copy Variants
Variant messaging for experiment arms with controlled variable isolation.
Use This Skill When
- Generating copy variants for A/B test arms.
- Writing CTA alternatives for landing pages or onboarding.
- Adapting messaging tone for different user segments.
- Ensuring only one variable changes between variants.
Workflow
- Confirm the experiment hypothesis and primary metric.
- Identify the single variable to isolate (headline / CTA / value prop).
- Generate control + treatment copy (minimum 2 variants).
- Tag each variant with target segment and AARRR stage.
- Validate variable isolation: only one element differs per pair.
- Save variants to files.
Deliverables
report.md: variant rationale and isolation analysis.results/copy-variants.md: all variants with tags.
Quality Gates
If any gate fails: identify the issue, fix, and re-validate.
Gotchas
- Changing both headline and CTA simultaneously violates variable isolation — test results become unattributable.
- "Professional tone" is not specific enough. Define tone with examples (e.g., "formal, no contractions, data-driven").
- Copy variants without a linked experiment hypothesis are untestable. Require experiment-design output first.
Validation Loop
- Generate copy variants
- Check: single variable isolation, segment tags, linked hypothesis
- If any check fails → revise variants
- Proceed to experiment launch after approval
