Marketing Cro

CRO — CONVERSION OPTIMIZATION OS (OPERATIONAL)

Built as a no-fluff execution skill for systematic conversion rate optimization.

Structure: Core CRO fundamentals first. Advanced testing in dedicated sections. AI/ML optimization in clearly labeled "Optional: AI / Automation" sections.

Modern Best Practices (January 2026)

• Google Optimize sunset: Use VWO, Optimizely, or PostHog
• Statistical significance: https://www.evanmiller.org/ab-testing/
• CXL Institute: https://cxl.com/
• Baymard Institute UX: https://baymard.com/

When to Use This Skill

• Landing page optimization: Hero, CTA, proof, form optimization
• A/B testing: Hypothesis design, sample size, statistical significance
• Funnel analysis: Drop-off identification, micro-conversion mapping
• Form optimization: Field reduction, multi-step forms, friction removal
• Trust/credibility: Social proof, security signals, guarantees

Core: CRO Framework

The CRO Process

1. ANALYZE → Identify conversion problems (data + qualitative)
2. HYPOTHESIZE → Form testable hypotheses
3. PRIORITIZE → Score by impact/effort (ICE/PIE)
4. TEST → Run A/B tests with statistical rigor
5. LEARN → Document results, iterate
6. IMPLEMENT → Roll out winners, test next

Conversion Rate Benchmarks

Note: Benchmarks vary significantly by industry. Use as directional only.

Core: Landing Page Optimization

Above-the-Fold Checklist

Every landing page needs these elements visible without scrolling:

Headline Formula

[Outcome] + [Timeframe/Ease] + [Without Pain Point]

Examples:
"Get 10 qualified leads per week without cold calling"
"File your tax return in 15 minutes with expert review"
"Double your email conversions without hiring a copywriter"

CTA Button Best Practices

CTA Button Optimization:

• Size: Large enough to tap on mobile (min 44px height)
• Color: Contrasts with page background
• Position: Above fold AND after key sections
• Text: First person ("Get My...") often outperforms second person

Trust Elements Hierarchy

STRONGEST TRUST SIGNALS (use at least 3):
├─ Customer logos (recognizable brands)
├─ Review score (4.5+ stars with count)
├─ Security badges (SSL, payment, compliance)
├─ Money-back guarantee
└─ Phone number visible

SUPPORTING TRUST SIGNALS:
├─ Customer testimonials (with photo, name, company)
├─ Case study snippets (specific metrics)
├─ "As seen in" media logos
├─ Team photos (for services)
├─ Live chat widget
└─ Physical address (for services)

Core: Form Optimization

Form Field Rules

Field Priority (Ask Only What You Need)

Multi-Step Form Pattern

Step 1: Low commitment (email)
├─ "What's your email?"
├─ Progress indicator: 1 of 3
└─ CTA: "Continue"

Step 2: Qualifying info
├─ Company size / Industry
├─ Progress indicator: 2 of 3
└─ CTA: "Almost there"

Step 3: Contact info
├─ Name / Phone (optional)
├─ Progress indicator: 3 of 3
└─ CTA: "Get My [Deliverable]"

Multi-step benefits:

• Commitment and consistency principle
• Captures partial data (even if abandoned)
• Feels less overwhelming
• Can qualify leads progressively

Core: A/B Testing Methodology

Hypothesis Template

IF we [change/add/remove X]
THEN [metric] will [increase/decrease] by [estimate]
BECAUSE [reasoning based on data/research]

Example:
IF we add customer logos to the hero section
THEN form conversion will increase by 15%
BECAUSE trust signals reduce perceived risk for new visitors

Sample Size Calculator

Minimum sample size formula (simplified):

n = (16 × p × (1-p)) / MDE²

Where:
- n = sample per variant
- p = baseline conversion rate
- MDE = minimum detectable effect (e.g., 0.10 for 10% lift)

Example:
Baseline CVR: 3% (0.03)
MDE: 20% relative lift (looking for 3.6% or higher)

n = (16 × 0.03 × 0.97) / (0.006)²
n ≈ 12,933 per variant

Total traffic needed: ~26,000 visitors

Quick reference:

Per variant. Multiply by 2 for total traffic needed.

Statistical Significance

Requirements for valid test:

• 95% confidence level (minimum)
• Run for at least 1-2 full business cycles (7-14 days)
• Don't peek and stop early (increases false positives)
• Document before test: hypothesis, primary metric, sample size, duration

Test Prioritization: ICE Framework

ICE Score interpretation:

• 8-10: High priority, test immediately
• 5-7: Medium priority, add to queue
• 1-4: Low priority, revisit later or skip

Core: Funnel Analysis

Funnel Diagnostic Framework

STEP 1: Map your funnel
Page Visit → Key Action → Form Start → Form Complete → Confirmation

STEP 2: Measure drop-off at each step
├─ Page Visit to Key Action: ___% (bounce rate inverse)
├─ Key Action to Form Start: ___%
├─ Form Start to Complete: ___%
└─ Complete to Confirmation: ___%

STEP 3: Identify biggest drop-off
Biggest percentage drop = highest priority to fix

STEP 4: Diagnose root cause
├─ High bounce? → Relevance, load speed, messaging
├─ Low engagement? → Content, CTA visibility
├─ Form abandonment? → Form friction, trust
└─ Checkout drop? → Pricing, shipping, trust

Micro-Conversion Mapping

Heatmap & Recording Analysis

What to look for:

• Click heatmaps: Are users clicking CTAs? Clicking non-clickable elements?
• Scroll maps: Where do users stop scrolling? Key content below fold?
• Session recordings: Where do users hesitate? Rage clicks? Form confusion?
• Form analytics: Which fields cause abandonment? Error patterns?

Core: Page Speed Optimization

Speed Impact on Conversion

Source: Google/Akamai research, directional only

Core Web Vitals Targets

Quick Speed Wins

Quick Reference

Decision Tree (CRO Triage)

Low conversion rate?
├─ Check page speed first (if >3s, fix that)
├─ Check bounce rate
│   ├─ High bounce (>70%) → Message/audience mismatch
│   └─ Normal bounce → Continue diagnosis
├─ Check scroll depth
│   ├─ Low scroll → Above-fold problem
│   └─ Good scroll → Below-fold CTA/proof issue
├─ Check form analytics
│   ├─ Low form starts → CTA/offer problem
│   └─ High abandonment → Form friction
└─ Check mobile vs desktop
    ├─ Mobile worse → Mobile UX issues
    └─ Same → Universal issue

Form abandonment?
├─ Check which field causes drop-off
├─ Reduce fields to minimum
├─ Add progress indicator
├─ Add inline validation
└─ Test multi-step

Traffic but no clicks?
├─ CTA not visible → Move above fold
├─ CTA not compelling → Test copy
├─ Too many options → Simplify, single CTA
└─ Page too long → Add sticky CTA

Operational SOPs

Weekly CRO Review (30 minutes)

1. Check conversion metrics

   • Landing page CVR vs last week
   • Form completion rate
   • Funnel stage conversion rates

2. Review active tests

   • Current test status
   • Statistical significance check
   • Call winner if significant

3. Prioritize next tests

   • Score new ideas (ICE)
   • Plan next week's tests
   • Document learnings

Monthly CRO Audit (2 hours)

1. Full funnel analysis

   • Map all conversion points
   • Identify top 3 drop-off points
   • Quantify opportunity (traffic × CVR gap)

2. Heatmap/recording review

   • Review 10-20 session recordings
   • Analyze click and scroll heatmaps
   • Document friction points

3. Test performance review

   • Win/loss ratio
   • Cumulative lift from tests
   • Update testing roadmap

4. Competitive analysis

   • Review competitor landing pages
   • Note new patterns/trends
   • Identify test ideas

Templates

Anti-Patterns

Optional: AI / Automation

Note: Core CRO fundamentals above work without AI. This section covers optional AI capabilities.

AI-Powered CRO Tools

When to Use AI

HAVE thousands of daily conversions?
├─ YES → AI personalization can optimize in real-time
└─ NO → Stick to traditional A/B testing

NEED many variants quickly?
├─ YES → AI copy generation for variant ideation
└─ NO → Human copywriting with hypothesis

HAVE limited traffic?
├─ YES → Qualitative research over testing
└─ NO → Statistical A/B tests

Related Skills

• marketing-leads-generation — Lead capture strategies
• marketing-paid-advertising — Traffic sources
• marketing-seo-technical — Page speed, Core Web Vitals
• software-ui-ux-design — Design patterns
• software-ux-research — User research methods

Usage Notes (Claude)

• Stay operational: return checklists, audit results, test plans
• Always include statistical significance requirements for testing
• Recommend qualitative research for low-traffic sites
• Use benchmark ranges, not absolute numbers
• Do not invent conversion data; state "varies by industry" when uncertain