83. A/B Testing Landing Pages - optymalizacja konwersji

Poziom: Średni | Czas: 8 min

A/B testing = systematicznie testowanie zmian. 1% improvement tygodniowo = 67% improvement w року. Small changes compound.

Czym jest A/B Testing?

VERSION A (control):         VERSION B (variant):
Oryginalny page              Zmieniony element

        ↓                              ↓
    50% traffic                    50% traffic
        ↓                              ↓
    5% conversion                  7% conversion

WINNER: Version B (40% improvement!)

KEEP Version B → Test нову change

Co testować (w kolejności IMPACT):

1. Headline (BIGGEST IMPACT):

VERSION A:
"Welcome to Our Coaching Program"
Conversion: 3%

VERSION B:
"Transform Your Business in 90 Days (Guaranteed)"
Conversion: 7%

RESULT: 133% improvement od jednego headline!

WHY: Headline = first thing seen = decyduje CZY чита dalej

2. Call-to-Action (CTA):

BUTTON TEXT:

A: "Submit"              → CR: 4%
B: "Get My Free Guide"   → CR: 6.5%
C: "Send It To Me Now!"  → CR: 7%

WINNER: Version C (+75% vs A)

BUTTON COLOR:

A: Blue    → CR: 5%
B: Green   → CR: 5.2%
C: Orange  → CR: 6.8%

WINNER: Orange (high contrast = more visible)

3. Form Length:

A: 5 fields (Name, Email, Phone, Company, Message)
   CR: 3%

B: 2 fields (Name, Email)
   CR: 8%

RESULT: Fewer fields = 167% higher conversion

BUT: Quality может бути lower (less qualified)
BALANCE: Collect minimum, enrichuj później

4. Social Proof:

A: No testimonials
   CR: 4%

B: 3 testimonials з photos
   CR: 6%

C: "Join 10,523 customers" counter
   CR: 7.5%

WINNER: Social proof counter (+87%)

5. Images/Video:

A: Stock photo
   CR: 5%

B: Real customer photo
   CR: 6.5%

C: 2-min explainer video
   CR: 8%

WINNER: Video (+60% vs stock photo)
People resonate z real > fake

GHL A/B Testing Setup

Built-in Split Testing:

Sites → Funnels → [Your Funnel] → Split Test

STEP 1: Create Variations
- Version A: Original landing page
- Version B: Variant (clone and modify)

STEP 2: Traffic Split
○ 50/50 split (даєfast results)
○ 70/30 split (if cautious)
○ 90/10 split (minor test)

STEP 3: Primary Goal
• Form submission
○ Button click
○ Page visit
○ Purchase

STEP 4: Launch
Duration: Run until statistical significance
Minimum: 100 conversions per variant

Statistical Significance

PROBLEM: Was improvement luck або real?

EXAMPLE:

Short test:
A: 10 conversions / 100 visits = 10%
B: 15 conversions / 100 visits = 15%

Difference: 50% improvement!
BUT: Too small sample → Could be luck

Proper test:
A: 100 conversions / 2,000 visits = 5%
B: 140 conversions / 2,000 visits = 7%

Difference: 40% improvement
95% confidence = REAL difference

CALCULATION TOOL:
Google "A/B test calculator"
Enter: Visitors + conversions each variant
Output: "95% confidence" → VALID WINNER

Minimum Sample Size:

BASELINE CR: 5%
DESIRED IMPROVEMENT: +20% (до 6%)
SIGNIFICANCE LEVEL: 95%

NEEDED: ~4,000 visitors/variant (~8,000 total)

IF low traffic:
→ Test BIGGER changes (easier detect)
→ OR Run longer (gather more data)

Test Duration:

❌ TOO SHORT: 2 days
   Not enough data, day-of-week affects vary

✅ GOOD: 2-4 weeks
   Full business cycle
   Accounts for weekly patterns
   Достатньо data
   
❌ TOO LONG: 3+ months
   Market changes, offer expires

What NOT to Test (Common Mistakes):

❌ Multiple changes simultaneously
   Changed headline + CTA + image
   Win? Don't know WHICH change worked!
   
✅ Test ONE element at a time
   Headline only → Find winner → THEN test CTA

❌ Too small traffic
   10 visitors/day → Need 400+ days dla wyniku!
   
✅ Minimum 100 visitors/day dla practical testing

❌ Stopping test too early
   B leading after 1 day → Declare winner (TOO FAST)
   
✅ Wait for statistical significance

❌ Ignoring segmentation
   Mobile CR: 2% | Desktop CR: 8% → Different experiences needed!
   
✅ Separate tests for mobile vs desktop

Multivariate Testing (Advanced):

A/B Test: 1 element at time
Multivariate: Multiple elements simultaneously

EXAMPLE:

ELEMENTS TESTED:
- Headline: 2 variations
- CTA Button: 2 variations
- Image: 2 variations

TOTAL COMBINATIONS: 2 × 2 × 2 = 8 variants

TRAFFIC SPLIT: 12.5% каждая

REQUIRES: LOTS of traffic (thousands/day)
BENEFIT: Finds BEST combination faster

GHL: Use external tool like Google Optimize (free)

A/B Testing Checklist:

☑ Hypothesis defined:
  "Changing headline to benefit-driven will increase CR by 20%"
  
☑ ONE element isolated for test
  
☑ Two distinct variants created (clear difference)
  
☑ Traffic split: 50/50
  
☑ Primary goal: Form submission
  
☑ Tracking verified (test submissions yourself)
  
☑ Minimum sample size calculated (e.g., 4,000/variant)
  
☑ Test duration: 2-4 weeks planned
  
☑ Review ritual: Check daily, decide when significant
  
☑ Document результат: What worked, what didn't

Testing Priority Framework:

START HERE (Highest impact, easiest test):

1. Headline (1 week test)
   ↓ winner
2. CTA button (1 week)
   ↓ winner
3. Form length (2 fields vs 5 fields) (1 week)
   ↓ winner
4. Social proof (add vs none) (2 weeks)
   ↓ winner
5. Hero image (stock vs real vs video) (2 weeks)
   ↓ winner
6. Page length (short vs long copy) (2 weeks)
   ↓ winner
7. Colors/design (Optional, lower impact)

TIMELINE: ~10 weeks of testing
RESULT: Potentially 2-5X improvement in CR!

Analyzing Results:

TEST COMPLETE:

Version A:
2,000 visitors → 100 conversions = 5% CR
Cost per lead: $50

Version B:
2,000 visitors → 140 conversions = 7% CR (-40%)
Cost per lead: $35.71 (-28.6%)

Statistical significance: 99% confidence

DECISION:
✅ Keep Version B (clear winner)
✅ Document зміну
✅ Move to NEXT test (CTA button)

IF no clear winner:
→ Try більш dramatic change
→ OR Move to другий element

Real-World A/B Test Examples:

Example 1: Coaching Business

ORIGINAL Headline:
"Professional Business Coaching Services"
CR: 3.2%

TESTED Headline:
"Add $50K to Your Revenue in 90 Days"
CR: 7.8%

IMPROVEMENT: +144%

LESSON: Specific outcome > vague description

Example 2: SaaS Landing Page

ORIGINAL CTA:
"Start Free Trial"
CR: 5%

TESTED CTA:
"Get Full Access Free for 14 Days"
CR: 6.9%

IMPROVEMENT: +38%

LESSON: Emphasize "full access" + duration

Example 3: E-commerce Product Page

ORIGINAL:
No reviews visible
CR to cart: 8%

TESTED:
Display "3,421 5-estrela reviews" badge
CR to cart: 11.5%

IMPROVEMENT: +43.75%

LESSON: Social proof = trust = higher conversion

Advanced: Heatmap Analysis

BEFORE A/B test, USE heatmaps:

TOOLS:
- Hotjar (external)
- Microsoft Clarity (free!)

SHOWS:
- Where користувачі click (red = most clicks)
- How far scroll (where दоп off)
- Mouse movement (attention areas)

USE DATA:
"50% не scroll до форми → Move form up"
"No clicks on CTA → Not visible enough → Тест bigger button"

Informed testing > guessing

Tools for A/B Testing:

BUILT-IN GHL:
✅ Split test funnels/pages
✅ Track conversions
✅ Traffic distribution

EXTERNAL (Advanced):
- Google Optimize (FREE, integrates z GA)
- VWO (Visual Website Optimizer)
- Optimizely (Enterprise)

For GHL: Built-in достатньо dla most use cases

Common A/B Test Mistakes:

❌ Testing без hypothesis
✅ "I believe [change] will improve [metric] because [reason]"

❌ Declaring winner після 1 day
✅ Wait for statistical significance (95%+)

❌ Testing tiny changes (blue vs navy blue button)
✅ Test meaningful differences (can eyeball відмінність)

❌ Not documenting results
✅ Keep test log: Date, test, result, decision

❌ Testing without sufficient traffic
✅ Calculate minimum sample потрібен before starting

❌ Testing during atypical period (holidays)
✅ Normal business weeks only

Continuous Optimization:

NEVER STOP TESTING:

Win Rate: 50-60% of tests (many fail!)

MINDSET:
"Every page can improve"
"1% weekly improvement = 67% yearly"
"Testing = knowing > guessing"

CULTURE:
Test → Learn → Implement → Repeat

Quick A/B Test Launch Plan:

WEEK 1:
☑ Choose element to test (headline)
☑ Create hypothesis
☑ Build variant B
☑ Launch 50/50 split test

WEEK 2-3:
☑ Monitor daily
☑ Check for significance

WEEK 4:
☑ Analyze results
☑ Keep winner
☑ Document learning
☑ Plan next test (CTA)

REPEAT continuous

---

Następny krok: 84. Facebook Pixel Integration