name: cro description: "Use when optimizing landing pages, signup flows, checkout processes, or any user-facing page for higher conversion rates - before A/B testing." user-invocable: true argument-hint: "[page path or component]" allowed-tools: Read, Bash, Glob, Grep

/cro - Conversion Rate Optimization

Audit pages and flows for conversion friction - identifying drop-off points, persuasion gaps, and technical blockers - then produce prioritized recommendations with testable A/B hypotheses.

Target: $ARGUMENTS

When to Use

• Before launching a landing page or signup flow
• When conversion rates are below industry benchmarks
• After redesigning a checkout or onboarding flow
• When bounce rate is high on key pages
• Before running paid traffic to a page
• When @marketing Growth mode needs structured CRO analysis
• NOT for: SEO issues (use /seo-audit)
• NOT for: general design quality (use /design-review)
• NOT for: copy creation (use /copywriting)

Workflow

Step 1: Page & Flow Analysis

1. Read the target page/component:

# Find the page/component code
find $TARGET -type f \( -name "*.tsx" -o -name "*.vue" -o -name "*.svelte" -o -name "*.astro" \) | head -20

1. Identify the conversion goal:

   • Signup (form submission)
   • Purchase (checkout completion)
   • Download (file/app download)
   • Contact (form/chat initiation)
   • Engagement (feature activation)

2. Map the user flow:

## User Flow
[Entry point] → [Step 1] → [Step 2] → ... → [Conversion]

Example:
Landing page → Pricing section → Signup form → Email verification → Dashboard

1. Identify drop-off points - Where in the flow might users leave?

Step 2: Friction Audit

Check for conversion killers:

# Count form fields (friction indicator)
grep -rn "<input\|<select\|<textarea" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | wc -l

# Check for form validation UX
grep -rn "onBlur\|onChange.*validation\|zod\|yup\|formik\|react-hook-form\|vee-validate" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -10

# Check for loading/performance issues
grep -rn "useEffect\|onMounted\|onMount" $TARGET --include="*.tsx" --include="*.vue" --include="*.svelte" 2>/dev/null | head -10

# Check image optimization
grep -rn "<img\|<Image" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | grep -v "next/image\|nuxt-img\|loading=" | head -10

Step 3: Persuasion Checklist (Cialdini's Principles)

Evaluate each persuasion principle:

# Check for social proof elements
grep -rn "testimonial\|review\|rating\|customer\|users\|trusted" $TARGET --include="*.tsx" --include="*.vue" --include="*.astro" 2>/dev/null | head -10

# Check for trust signals
grep -rn "secure\|ssl\|encrypted\|guarantee\|money-back\|badge" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -10

Step 4: Copy & CTA Audit

# Check CTA button text
grep -rn "type=\"submit\"\|<button\|<Button" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -10

# Check for "Submit" or generic CTA text (anti-pattern)
grep -rn ">Submit<\|>Click here<\|>Send<\|>Go<" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -5

Step 5: Technical CRO

Performance directly impacts conversion (every 100ms delay = -1% conversion):

# Check for skeleton/loading states
grep -rn "skeleton\|Skeleton\|loading\|isLoading\|isPending" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -10

# Check for error boundary/handling in forms
grep -rn "error\|Error\|catch\|onError\|isError" $TARGET --include="*.tsx" --include="*.vue" 2>/dev/null | head -10

# Bundle size indicators
find $TARGET -name "*.tsx" -o -name "*.ts" | xargs wc -l 2>/dev/null | sort -rn | head -10

Step 6: Prioritized Recommendations

For each finding, produce:

### Recommendation [N]: [Title]

**Finding:** [What's wrong]
**Impact:** HIGH | MEDIUM | LOW
**Effort:** S (< 2h) | M (2-4h) | L (4-8h)
**Priority:** [Impact/Effort ratio]

**A/B Hypothesis:**
> If we [change], then [metric] will [improve] because [reason].

**Implementation:**
- File: [file:line]
- Change: [specific change]

Prioritize by: HIGH impact + LOW effort first (quick wins).

Verification Protocol

Before claiming the CRO audit is complete:

1. Conversion goal was explicitly identified
2. User flow was mapped from entry to conversion
3. All 7 friction categories were checked
4. All 7 Cialdini principles were evaluated
5. At least 3 recommendations have A/B hypotheses
6. Recommendations are prioritized by impact/effort ratio
7. Technical CRO checks were performed (performance, form UX)
8. Each recommendation has a specific file:line reference

Anti-Rationalization

Rules

1. Read-only - Never modify files during a CRO audit
2. Data-informed - Base recommendations on research and benchmarks, not opinions
3. Hypothesis-driven - Every recommendation needs a testable A/B hypothesis
4. Quick wins first - Prioritize high-impact, low-effort changes
5. One CTA per view - Never recommend multiple competing CTAs
6. Genuine urgency only - Never recommend fake scarcity or countdown timers
7. Code-grounded - Every finding must reference specific files and components
8. Conversion goal explicit - Never audit without defining what "conversion" means

Output

──── /cro ────
Target: $ARGUMENTS
Conversion Goal: [signup | purchase | download | contact]
User Flow: [entry] → ... → [conversion]

Friction Score: [N]/100 (lower = more friction)
Persuasion Score: [N]/7 principles present

Findings:
  CRITICAL: N
  HIGH: N
  MEDIUM: N
  LOW: N

Top Recommendations:
1. [HIGH/S] [recommendation] → Hypothesis: [if/then]
2. [HIGH/M] [recommendation] → Hypothesis: [if/then]
3. [MEDIUM/S] [recommendation] → Hypothesis: [if/then]

Quick Wins: N (HIGH impact, LOW effort)
Total Recommendations: N

──── CRO audit complete ────