---

name: conversion-optimization description: > Optimize conversion rates — funnel analysis, hypothesis generation, test design, landing page review, CTA optimization, and results analysis. TRIGGER when: user says /conversion-optimization, needs to improve conversion rates, or asks about CRO and A/B testing. argument-hint: "[relevant context]" user-invocable: true

Conversion Optimization

Process

Step 1: Map the Funnel

Stage	Page/Step	Current Rate	Benchmark
Visit → Engage	Landing page	X%	X%
Engage → Lead	Form/CTA	X%	X%
Lead → Qualify	Nurture sequence	X%	X%
Qualify → Convert	Purchase/sign-up	X%	X%

Step 2: Identify Drop-Off Points

Signal	Analysis Method
High bounce rate	Landing page relevance, load speed
Form abandonment	Form analytics, field analysis
Cart abandonment	Checkout flow analysis
Pricing page exit	Pricing clarity, comparison

Step 3: Generate Hypotheses

For each drop-off:

IF we [change], THEN [metric] will [improve] BECAUSE [reason].
Priority: Impact (H/M/L) × Confidence (H/M/L) × Ease (H/M/L)

Step 4: Design Tests

Element	Test Type	Variants
Headlines	A/B	2-3 variants
CTA text/color	A/B	2 variants
Page layout	A/B	2 variants
Form length	A/B	Short vs long
Social proof	A/B	With vs without
Pricing display	A/B	Monthly vs annual first

Step 5: Run and Analyze

Metric	Control	Variant	Lift	Significance
Conversion rate	X%	X%	+X%	p = X
Revenue/visitor	$X	$X	+X%	p = X

Step 6: Implement Winners

• Ship winning variants to 100%
• Document learnings for future tests
• Move to next priority hypothesis
• Track cumulative impact over time

Output Format

## CRO Report: [Page/Flow]
### Funnel: [Current rates by stage]
### Opportunities: [Prioritized hypotheses]
### Test Results: [Results table]
### Impact: [Cumulative conversion lift]

Quality Checklist

• Funnel mapped with current conversion rates
• Hypotheses prioritized by impact × confidence × ease
• Tests have adequate sample size for statistical significance
• Results analyzed correctly (avoid peeking)
• Winning changes documented and shipped

Edge Cases

• For low-traffic sites, use qualitative methods (user testing, heatmaps) instead of A/B tests
• For B2B with long sales cycles, track micro-conversions
• If running multiple tests, ensure no interaction effects