landing-page-optimization

持续测试并优化托管于 Vercel 的 Hedge Edge 着陆页（hedge-edge.com），以最大化访客转试用、试用转付费的转化率。涵盖 A/B 测试、CTA 优化、定价页面实验、页面加载速度、移动端用户体验，以及借助 GA4 数据进行的漏斗分析——所有这些都针对正在评估对冲工具的对冲经纪公司交易员进行了精心调整。

SKILL.md

--- frontmatter

name: landing-page-optimization
description: |
  Continuously test and improve the Hedge Edge landing page (hedge-edge.com) hosted on Vercel
  to maximise visitor-to-trial and trial-to-paid conversion rates. Covers A/B testing, CTA
  optimization, pricing page experiments, page speed, mobile UX, and funnel analysis using
  GA4 data  all tailored to converting prop-firm traders who are evaluating hedging tools.

Landing Page Optimization

Objective

Achieve and sustain a landing-page-to-trial conversion rate of 8%+ and a pricing-page-to-checkout conversion rate of 12%+. Reduce bounce rate below 40% for paid traffic and below 55% for organic traffic. Every experiment must be grounded in prop-firm trader psychology: capital preservation anxiety, challenge-failure fear, desire for automation, and skepticism toward "too good to be true" tools.

When to Use This Skill

•Conversion rate drops below 6% for 7 consecutive days diagnose and fix.
•A new paid ad campaign launches create a campaign-specific landing page variant.
•Pricing changes are being tested (//\ tiers) run pricing page experiment.
•A major feature ships (MT4 connector, cTrader support) update hero section and feature blocks.
•GA4 funnel analysis reveals a drop-off point (e.g., 60% exit at pricing page) targeted fix.
•Quarterly CRO (Conversion Rate Optimization) sprint systematic multi-variant testing.
•Page speed score drops below 90 on Lighthouse performance remediation.

Input Specification

Field	Type	Required	Description
experiment_type	enum	Yes	hero_test, cta_test, pricing_page, social_proof, page_speed,
ew_variant, mobile_ux
hypothesis	string	Yes	Clear statement: "Changing X to Y will improve Z by N% because [reason]"
variant_specs	object[]	Yes	[{ variant_name, changes_description, mockup_url? }] min 2 (control + 1 challenger)
traffic_split	number[]	No	Percentage split per variant default: equal split
success_metric	enum	Yes	rial_signup_rate, checkout_rate, ounce_rate, ime_on_page, cta_click_rate
min_sample_size	integer	No	Minimum visitors per variant before declaring winner default: 500
duration_days	integer	No	Maximum experiment duration default: 14

Step-by-Step Process

1. Funnel Diagnosis

•
Pull current funnel data from GA4 via GA4_MEASUREMENT_ID:
- •Landing page pricing page (target: 40%+ progression).
- •Pricing page checkout initiation (target: 20%+ progression).
- •Checkout initiation checkout completion (target: 60%+ Creem.io-side).
- •Overall visitor trial signup (target: 8%+).
•Identify the largest drop-off point as the priority experiment area.
•Segment by traffic source (organic vs. paid vs. direct vs. Discord referral) different sources have different intent levels.
•Check device split: if mobile traffic > 40% but mobile conversion < half of desktop, prioritise mobile UX.

2. Hypothesis Formation

•
Ground every hypothesis in prop-firm trader behaviour:
- •"Traders who have failed a challenge are skeptical of new tools adding a money-back guarantee badge near the CTA will increase trial signups by 15%."
- •"Showing real-time user count ('487 traders hedging right now') creates FOMO and social proof will reduce bounce rate by 10%."
- •"Traders care about specific prop-firm compatibility adding FTMO/TopStep/Apex logos to the hero reduces the 'will this work for me?' objection."
- •"The current 'Start Free Trial' CTA is generic changing to 'Protect Your Funded Accounts' aligns with trader pain and will improve CTR by 20%."

3. Variant Development

•Design variant changes using the current Hedge Edge design system (dark theme, trading-UI aesthetic, green/blue accent palette).
•
Hero section tests:
- •Control: Current headline and sub-headline.
- •Variant A: Pain-led "Stop Losing Funded Accounts to Drawdown Breaches".
- •Variant B: Solution-led "Automated Multi-Account Hedging for Prop Traders".
- •Variant C: Proof-led "500+ Traders Protecting + in Funded Capital".
•
CTA tests:
- •Control: "Start Free Trial"
- •Variant A: "Protect My Accounts Free for 14 Days"
- •Variant B: "Get Hedge Edge No Card Required"
- •Variant C: "See It In Action" (leads to demo video, then trial CTA)
•
Pricing page tests:
- •Control: Three-tier horizontal cards.
- •Variant A: Comparison table with feature checkmarks.
- •Variant B: Calculator "How many accounts do you manage?" recommends tier.
- •Variant C: Annual pricing toggle with savings highlight (30% savings badge).
•
Social proof tests:
- •Control: No testimonials above fold.
- •Variant A: 3 trader testimonials with Discord avatars.
- •Variant B: Real-time counter + recent signup notifications ("John from London just signed up").
- •Variant C: Case study summary "How a 5-account FTMO trader saved 3 challenges in one month".
•Deploy variants via Vercel VERCEL_TOKEN use edge config or feature flags for traffic splitting.

4. Experiment Execution

•Implement traffic split using Vercel Edge Middleware or client-side feature flag.
•
Ensure GA4 event tracking is configured for each variant:
- •page_variant custom dimension.
- •cta_click event with ariant_name parameter.
- •
  code
  rial_signup conversion event.
- •checkout_initiated and checkout_completed events (via Creem.io webhook GA4 Measurement Protocol).
•Run experiment for minimum duration_days or until min_sample_size is reached per variant.
•Do not peek at results before minimum sample is collected (avoid peeking bias).

5. Statistical Analysis

•Calculate conversion rate per variant with 95% confidence interval.
•Use a two-proportion z-test or Bayesian A/B methodology.
•
Declare a winner only if:
- •p-value < 0.05 (frequentist) or probability-to-be-best > 95% (Bayesian).
- •Minimum sample size met for all variants.
- •Result is consistent across traffic sources (paid and organic both show improvement, or at least no degradation in either).
•If no clear winner after duration_days, extend by 7 days or declare inconclusive and document learnings.

6. Implementation & Rollout

•Roll winning variant to 100% traffic via Vercel deployment.
•Update the control baseline for future experiments.
•
Document the experiment result in Notion campaign retrospective:
- •Hypothesis, variants tested, sample sizes, conversion rates, confidence level, uplift %.
•Notify Business Strategist Agent if pricing-page changes affect revenue projections.
•Notify Content Engine Agent if new messaging/positioning won (update brand guidelines).

7. Page Performance Monitoring

•
Monthly Lighthouse audit (target scores):
- •Performance: 90
- •Accessibility: 95
- •Best Practices: 90
- •SEO: 95
•
Core Web Vitals (from Google Search Console via SEARCH_CONSOLE_KEY):
- •LCP (Largest Contentful Paint): < 2.5s
- •FID (First Input Delay): < 100ms
- •CLS (Cumulative Layout Shift): < 0.1
•If any metric degrades after a variant deployment, roll back immediately.
•Optimize images (WebP/AVIF), lazy-load below-fold content, minimize JavaScript bundles.

Output Specification

Output	Format	Destination
Experiment report	JSON { experiment_id, hypothesis, variants[], sample_sizes[], conversion_rates[], winner, confidence, uplift_pct }	Notion + Google Sheets CRM
Funnel analysis	JSON { stage, visitors, progression_rate, drop_off_rate }[]	GA4 dashboard + Notion
Page speed report	JSON { lcp, fid, cls, performance_score, accessibility_score }	Notion + alert if degraded
Winning variant deployment	Vercel production deployment	hedge-edge.com
Messaging insights	string (winning headline/CTA)	Content Engine Agent + agent memory

API & Platform Requirements

Platform	Variable	Operations Used
Vercel	VERCEL_TOKEN	Deploy page variants, manage edge config, trigger builds
GA4	GA4_MEASUREMENT_ID	Funnel analysis, variant tracking, conversion events
Google Search Console	SEARCH_CONSOLE_KEY	Core Web Vitals, mobile usability, indexed page status
Supabase	SUPABASE_URL, SUPABASE_KEY	Trial signup events for conversion attribution
Creem.io	CREEM_API_KEY	Checkout events for pricing-page experiment attribution
n8n	N8N_WEBHOOK_URL	Relay conversion events to GA4, automate experiment reporting
Notion	NOTION_API_KEY	Experiment logs, CRO sprint planning, retrospective documentation