SKILL.md
$27
Step
Action
1. Research
Map funnel; identify high-traffic, low-conversion pages
2. Hypothesize
Form testable hypothesis (if X, then Y because Z)
3. Prioritize
Score by Potential, Importance, Ease (PIE)
4. Test
A/B or multivariate; adequate sample size
5. Analyze
Statistical significance; implement winner
PIE Prioritization Framework
Score each test idea 1–10:
Factor
Question
Potential
How much improvement is possible?
Importance
How much traffic does this page get?
Ease
How easy to implement?
Rank backlog by total score; run highest-impact tests first.
A/B Testing Best Practices
Practice
Guideline
Sample size
Calculate minimum before launch; 95% significance without adequate sample = false positives
Duration
Run full week cycles; account for day-of-week effects
One variable
Test one element per experiment (or use MVT for multiple)
Mobile separate
Mobile converts ~50% of desktop; test mobile independently—thumb reach, form complexity differ
Low traffic
Use Bayesian testing for faster, actionable results
Key Testing Areas
Page Type
Test Ideas
Homepage
Search bar prominence; personalized content; hero CTA; social proof placement
Landing page
Headline; form length; CTA copy; above-fold layout
Product/Category
Quick view; descriptions; add-to-cart placement
Checkout
Form fields; progress indicator; trust badges; guest checkout
Pricing
Plan order; anchoring; CTA per tier
Personalization: Personalized experiences generate ~41% more impact than generic ones.
Commercialization Infrastructure
Module
Purpose
Data & BI
Data warehouse; user behavior events; agile surveys
A/B testing
Experiment platform; statistical significance; backend-controlled variants
User education
Help docs (multi-language); update notifications; EDM
Attribution
Ad pixels; attribution model; impression-to-click-to-sale tracking
Avoid: Intrusive interstitials; popups that block content. Prefer non-intrusive ad formats.
Foundational Requirements
- Analytics: Map funnels; identify drop-off points (analytics-tracking, traffic-analysis)
- Qualitative: Heatmaps, session recordings, user tests—understand why drop-off occurs
- Technical: Dedicated resources for 2–4 tests/month; maintain momentum
Output Format
- Funnel map (stages, conversion rates, drop-off)
- Hypothesis (if X, then Y because Z)
- Test plan (variant, metric, sample size, duration)
- Implementation checklist
Related Skills
- landing-page-generator: Landing page structure and copy
- cta-generator: CTA design and placement
- analytics-tracking: GA4, events, conversion tracking
- traffic-analysis: Attribution, funnel analysis
- copywriting: Headline, CTA copy for tests