Google Ads Landing Page A/B Test Plan

Test Hypothesis

Because current data shows homepage converting at 17.15% vs /start-your-case at 12.42% (27.6% gap), and this gap varies significantly by campaign intent,

We believe matching landing page to campaign intent (brand → homepage for trust-building, cold prospecting → /start-your-case for direct action, DIY → both tested)

Will cause higher conversion rates and better lead quality (measured by case completion rate, not just initial conversions)

For users from different campaign types (brand, prospecting, DIY)

We'll know this is true when we see statistically significant improvements in conversion rate AND downstream metrics (case started, payment received) by campaign type over 21-30 days.

📊 Current State Analysis

Key Findings from Historical Data (90 days)

Campaign Type	Campaign	Current Landing	Clicks	Conv Rate	Cost/Conv
Brand/High-Intent	Search Only Campaign	/start-your-case	1,491	10.00%	$7.69
Brand/High-Intent	Search Only Campaign	/family-law/	35	11.43%	$81.73
DIY/Self-Service	DIY - For PK	/start-your-case	1,258	16.53%	$18.63
	DIY - For PK	Homepage	487	17.15%	$17.10
	Self-Rep & Court Prep	/start-your-case	153	13.07%	$32.48
Cold Prospecting	Family Law - Search	/start-your-case	104	3.85%	$118.41
	AI Legal Help	/start-your-case	111	5.41%	$59.57
	AI Legal Help	Homepage (www)	29	24.14%	$13.79

⚠️ Critical Insight: Homepage Wins on Cold Prospecting

AI Legal Help campaign:

Homepage: 24.14% conversion (29 clicks)
/start-your-case: 5.41% conversion (111 clicks)
346% performance gap

This aligns with your "leaky bucket" problem — /start-your-case may be letting low-intent browsers in who never complete cases.

🎯 Test Design

Test Structure: Campaign-Type Based Testing

We'll run separate A/B tests for each campaign type because they have different user intent and traffic volumes.

Test Group 1: Brand/High-Intent → "Search Only Campaign"

Current: 100% /start-your-case (10.00% conv rate)
Test: 50% Homepage / 50% /start-your-case
Rationale: Brand searches may not need immediate CTA — users already know ezylegal. Homepage could build more trust before asking for commitment.

~50 Clicks/day

21 days Minimum test duration

525 Clicks per variant

50-53 Expected conversions

Minimum Detectable Effect (MDE): 20% relative lift (10% → 12%) with 80% power

Test Group 2: Cold Prospecting → "Family Law - Search" + "AI Legal Help"

Current: Mostly /start-your-case (3.85-5.41% conv rate)
Test: 50% Homepage / 50% /start-your-case
Rationale: Historical data shows homepage converts 4-6x better for cold traffic. This is your biggest opportunity.

~8 Clicks/day (combined)

30 days Minimum test duration

120 Clicks per variant

5-29 Expected conversions

MDE: 50% relative lift (5% → 7.5%) with 80% power — or detect homepage's 4x advantage

Note: Lower traffic = longer test duration needed. Consider combining both campaigns for faster results.

Test Group 3: DIY/Self-Service → "DIY - For PK"

Current: Mixed (16.53% on /start-your-case, 17.15% on homepage)
Test: 50% Homepage / 50% /start-your-case
Rationale: Currently performs similarly on both pages (only 3.8% gap). Test will validate which is better long-term.

~42 Clicks/day

21 days Minimum test duration

441 Clicks per variant

73-76 Expected conversions

MDE: 10% relative lift (16.5% → 18.2%) with 80% power

📏 Metrics & Measurement

Primary Metric

Conversion Rate (ad click → case started conversion in Google Ads)

This is the existing Google Ads conversion you're already tracking.

Secondary Metrics (Critical for Full Picture)

Metric	Why It Matters	Where to Track
Case Completion Rate	Addresses "leaky bucket" — do conversions become paying cases?	PostHog: conversion → case paid
Time to Case Start	Does homepage delay or accelerate decisions?	PostHog: landing → case started
Bounce Rate	Are users engaging with the page?	GA4
Pages/Session	Homepage users explore more, /start-your-case more direct	GA4
Revenue per Visitor	Ultimate business metric — which page generates more revenue?	PostHog (if revenue tracking enabled)

Guardrail Metrics

Cost per conversion — shouldn't increase significantly
Quality Score — watch for drops indicating poor ad-page match
Support tickets — increased confusion would indicate poor experience

🛠️ Implementation

Option 1: Google Ads Experiments (Recommended)

Pros:

Built into Google Ads — no dev work required
Automatic 50/50 split with consistent user assignment
Statistical significance calculated automatically
Can run multiple experiments simultaneously (one per campaign)

How to set up:

In Google Ads, go to campaign → Experiments
Click "+" and select "Campaign experiment"
Name it: "Landing Page Test - [Campaign Name] - Mar 2026"
Set 50/50 split
In the experiment, change all ad final URLs from current landing page to test variant
Launch experiment

Timeline: 15 minutes per campaign = 1 hour total setup

Option 2: Manual URL Parameter Split + PostHog

Pros:

Track full funnel behavior in PostHog
More flexible for custom analysis
Can segment by device, location, etc.

How to set up:

Create two ad variations per ad group (50% each)
Variant A: Current landing page + ?test=control
Variant B: Test landing page + ?test=variant
Use PostHog to track conversions by test parameter

Timeline: 2-3 hours dev work + PostHog setup

Recommended Approach

Use Google Ads Experiments for speed and simplicity. You can still track full funnel in PostHog by campaign name — no additional dev work needed.

📅 Test Timeline

Phase	Duration	Activities
Week 1: Setup	2-3 days	• Create Google Ads experiments • Verify tracking (conversion tags, PostHog events) • QA both landing pages • Document baseline metrics
Week 2-4: Run Test	21 days	• Monitor daily for technical issues • Check traffic distribution (should be 50/50) • No peeking at results (yet!) • Document any external factors (holidays, PR, major news)
Week 5: Analysis	2-3 days	• Pull results from Google Ads • Analyze secondary metrics in PostHog/GA4 • Segment analysis (mobile vs desktop, time of day, etc.) • Calculate statistical significance • Document learnings
Week 6: Decision & Rollout	1-2 days	• Make go/no-go decision per campaign • Implement winning variants • Update documentation • Plan follow-up optimizations

Total timeline: 4-6 weeks from start to full rollout

📊 Sample Size & Statistical Power

Detailed Calculations by Campaign Type

Test Group	Baseline CR	Target CR	Relative Lift	Sample Size	Days at Current Traffic
Brand (Search Only)	10.00%	12.00%	20%	550/variant	22 days
DIY (For PK)	16.50%	18.15%	10%	1,800/variant	43 days
Cold Prospecting	5.00%	7.50%	50%	550/variant	69 days

All calculations assume 95% confidence level and 80% statistical power

⏰ Traffic Considerations

Cold Prospecting campaigns have low traffic (8 clicks/day combined). At this rate, you'd need 69 days to reach statistical significance.

Options to speed up:

Increase budget for Family Law - Search and AI Legal Help campaigns temporarily
Accept lower power (70% instead of 80%) = 400 clicks/variant = 50 days
Accept larger MDE (100% lift = 5% → 10%) = 200 clicks/variant = 25 days
Directional decision after 21 days with note about statistical limits

🎯 Expected Outcomes & Decision Framework

Decision Matrix

Campaign Type	If Homepage Wins	If /start-your-case Wins	If No Clear Winner
Brand (Search Only)	Switch all brand ads to homepage Rationale: Build trust before asking for commitment	Keep /start-your-case Rationale: Direct path works for brand-aware users	Keep /start-your-case (lower risk) Test homepage with trust elements added
Cold Prospecting	Switch to homepage immediately Expected outcome based on historical 4x lift	Surprising! Dig into why /start-your-case worked Check if ad copy pre-qualified users differently	Run longer test or increase budget Low traffic = hard to detect signal
DIY (For PK)	Switch to homepage Add "DIY resources" section to make it clear	Keep /start-your-case DIY users want to get started immediately	Directional decision: Use winner at 90% confidence Difference is likely small (~1-2 percentage points)

Beyond Conversion Rate: Lead Quality Analysis

Critical: A landing page that converts 20% with 5% case completion is worse than one that converts 15% with 15% case completion.

Calculate Lead Quality Score:

Lead Quality Score = (Conversion Rate × Case Completion Rate) × 100

Example:

/start-your-case: 12.42% conv × 30% completion = 3.73 quality score
Homepage: 17.15% conv × 50% completion = 8.58 quality score

In this example, homepage is 2.3x better despite only 38% higher conversion rate.

🎯 Your Real Goal

You mentioned the core problem: "customers enter the system but don't fully commit — we're spending lawyer time on cases that go nowhere."

Track case completion rate by landing page. This test isn't about more conversions — it's about more quality conversions that become paid cases.

✅ Pre-Launch Checklist

Google Ads experiments created for all 3 campaign types
Traffic split set to 50/50 for each experiment
Conversion tracking verified (test conversions fire from both landing pages)
PostHog events tracking landing page parameter (for case completion analysis)
GA4 tracking verified (for bounce rate, pages/session)
Both landing pages QA'd on desktop and mobile
Baseline metrics documented (current conversion rates, traffic volumes)
Decision criteria agreed upon (what constitutes "winning")
Team notification: No changes to campaigns during test period
Calendar reminder: Check for technical issues daily, analyze results after 21 days

🚀 Next Steps

Immediate Actions (This Week)

Set up Google Ads Experiments (1 hour)
- Test Group 1: Search Only Campaign (brand)
- Test Group 2: Family Law - Search + AI Legal Help (cold prospecting)
- Test Group 3: DIY - For PK (DIY/self-service)
Verify tracking setup (30 min)
- Test conversion on both homepage and /start-your-case
- Check PostHog captures landing page in session properties
Document baseline (15 min)
- Screenshot current conversion rates
- Note any external factors (upcoming holidays, PR plans, etc.)
Launch experiments and set calendar reminder for April 8 to check results

Optional: Quick Win for Cold Prospecting

Historical data shows homepage converts 4-6x better for cold prospecting (AI Legal Help: 24.14% vs 5.41%). While the test runs, consider:

Immediate switch Family Law - Search to homepage (currently 3.85% — hard to get worse)
Keep AI Legal Help in test to validate the 4x lift is real and sustainable

This gets you immediate gains while collecting data to confirm the pattern.

📚 Resources

Tools & Calculators

Sample size calculator: Evan Miller's A/B Test Calculator
Statistical significance: Chi-Squared Test
Google Ads Experiments: Official Documentation

Analysis Templates

After the test completes, use this structure to document results:

Executive Summary: Winner, magnitude, recommendation
Primary Metric Results: Conversion rate by variant, statistical significance
Secondary Metrics: Case completion, bounce rate, pages/session, revenue
Segment Analysis: Mobile vs desktop, new vs returning, time of day
Guardrail Check: Cost per conversion, quality score, support tickets
Decision & Next Steps: Rollout plan, follow-up tests

🧪 Google Ads Landing Page A/B Test Plan