38. A/B testing emaili - optymalizacja kampanii

Poziom: Średni | Czas czytania: 12 min

A/B testing (split testing) to porównywanie dwóch wersji emaila, aby znaleźć wersję, która daje lepsze wyniki. Niewielka zmiana w subject line może zwiększyć open rate o 20-30%!

Czym jest A/B testing?

Concept:

• Wersja A wysłana do 50% listy
• Wersja B wysłana do 50% listy
• Porównaj wyniki
• Winner = lepszy open/click/conversion rate

Co można testować?

1. Subject line (biggest impact!)

TEST:
Version A: "Nowe features GoHighLevel 2026"
Version B: "{{firstName}}, zobacz co nowego w GHL"

RESULT:
A: 18% open rate
B: 27% open rate → Winner! (+50%)

Why: Personalizacja + ciekawość > opis faktów

2. From name

TEST:
Version A: From "GoHighLevel Team"
Version B: From "Kamil from GoHighLevel"

RESULT:
A: 22% open rate
B: 29% open rate → Winner!

Why: Personal connection > company brand

3. Send time

TEST:
Version A: Wtorek 10:00  
Version B: Czwartek 14:00

RESULT:
A: 25% open | 3% click
B: 21% open | 5% click → Winner (conversion focus!)

Why: Audience checks email midday Thursday

4. Email length

TEST:
Version A: Short (200 words, 1 CTA)
Version B: Long (800 words, storytelling)

RESULT:
A: 4% click rate
B: 7% click rate → Winner!

Why: Our audience wants educational content

5. CTA button

TEST:  
Version A: "Learn More"
Version B: "Get Free Tutorial"

RESULT:
A: 3.2% click
B: 5.8% click → Winner!

Why: Specific + benefit > generic

6. Images vs No images

TEST:
Version A: Hero image + 3 screenshots
Version B: Text only, no images

RESULT:
A: 3.5% click (slower load time)
B: 4.2% click → Winner

Why: Plain text = more personal, faster load

7. Personalization level

TEST:
Version A: "Hi there!"
Version B: "Hi {{firstName}},"
Version C: "Hi {{firstName}}, fellow {{custom.industry}} expert,"

RESULT:
A: 20% open
B: 26% open
C: 31% open → Winner!

Why: More relevance = more engagement

Jak prowadzić A/B test w GoHighLevel?

Krok 1: Creat

e Campaign

Marketing → Emails → Create Campaign

Campaign settings:

• Campaign name: "Newsletter Feb - A/B test"
• Type: ☑ A/B Test enabled

Krok 2: Configure test

TEST SETTINGS:

What to test:
○ Subject line
○ From name  
○ Email content
○ Send time

Sample size:
○ Test 50% / 50% (for large lists >5,000)
● Test 20% / 20% + Winner to 60% (safer)

Duration:
Wait 4 hours → auto-send winner to remainder

Winning metric:
● Open rate (for subject line tests)
○ Click rate (for content/CTA tests)
○ Conversion rate (for offer tests)

Krok 3: Create variants

Version A: Control

Subject: "GoHighLevel tutorial – complete guide"
From: "Coachflow.pl Team"
Body: [original design]

Version B: Test

Subject: "{{firstName}}, see what you're missing in GHL"
From: "Kamil @ Coachflow.pl"  
Body: [original design]

Krok 4: Send & Monitor

Launch campaign → GHL automatically:

1. Sends A to 500 people
2. Sends B to 500 people
3. Waits 4 hours
4. Sends winner to remaining 4,000

Krok 5: Analyze results

RESULTS:

Version A:
Sent: 500
Opened: 95 (19%)
Clicked: 12 (2.4%)

Version B:
Sent: 500  
Opened: 160 (32%) ← Winner!
Clicked: 22 (4.4%)

Winner sent to: 4,000 remaining
Final result: 30% open rate, 4.1% click rate

Best practices A/B testing

1. Test ONE variable

❌ DON'T:

Version A: Subject + from name + body + CTA
Version B: Different subject + from + body + CTA

You won't know what made the difference!

✅ DO:

Version A: Subject line #1
Version B: Subject line #2
(everything else identical)

2. Sample size

Minimum sample size:

• Small list (<1,000): Test 50%/50%
• Medium (1,000-10,000): Test 20%/20% + winner 60%
• Large (>10,000): Test 10%/10% + winner 80%

Statistical significance formula:

• Need ≥100 opens per variant
• Need ≥20 clicks per variant (for click tests)

3. Wait time

Open rate test: Wait 4-6 hours Click rate test: Wait 24-48 hours
Conversion test: Wait 3-7 days

Give enough time for action!

4. Repeat winners

If "personalized subject" wins 3x in a row → make it default strategy

Document learnings:

Our audience responds to:
✅ Personalization (first name)
✅ Questions over statements
✅ Tuesday 10am send time
✅ Short emails (<300 words)
✅ Specific CTAs ("Download now")

5. Segment-specific testing

Don't assume results apply to all segments!

TEST: Emoji in subject

B2C Audience:
"🎉 New feature alert" → 35% open ✅

B2B Audience:  
"🎉 New feature alert" → 18% open ❌
"New feature: Automate workflows" → 28% open ✅

Przykłady real-world A/B tests

Test #1: Subject curiosity vs clarity

INDUSTRY: SaaS coaching
LIST SIZE: 5,000

Variant A (Clarity):
"How to automate lead follow-up in GHL"
Result: 22% open, 4.1% click

Variant B (Curiosity):
"This one trick doubled our conversions..."
Result: 31% open, 2.8% click

WINNER: Depends on goal!
- Want opens? → B
- Want clicks/action? → A

Test #2: Long form vs short form

INDUSTRY: Coaching/consulting  
LIST SIZE: 3,200

Variant A (Short - 150 words):
Result: 2.9% click rate

Variant B (Long - 900 words story):
Result: 5.4% click rate → Winner!

INSIGHT: Coaching audience wants depth

Test #3: Discount messaging

INDUSTRY: Course sales
LIST SIZE: 8,000

Variant A: "30% OFF"
Result: 4.2% conversion

Variant B: "Save $297 today"  
Result: 6.1% conversion → Winner!

INSIGHT: Absolute amount > percentage

Test #4: Social proof placement

Variant A: Testimonial at top
Result: 3.8% click

Variant B: Testimonial after CTA
Result: 5.2% click → Winner!

INSIGHT: Sell first, prove after

Common mistakes

Mistake #1: Testing too many things

Fix: One variable per test

Mistake #2: Small sample size

Fix: Wait for ≥100 actions before declaring winner

Mistake #3: Not re-testing

Fix: Test winners quarterly (audience changes!)

Mistake #4: Ignoring segments

Fix: Separate tests for different segments

Mistake #5: No documentation

Fix: Keep testing log → build playbook over time

Testing schedule example

WEEK 1: Subject line test
WEEK 2: Implement winner
WEEK 3: Send time test
WEEK 4: Implement winner
WEEK 5: CTA test
WEEK 6: Implement winner
WEEK 7: Content length test
WEEK 8: Implement winner

MONTH 3: Re-test all winners

Tools & Resources

In GoHighLevel:

• Built-in A/B testing (subject, content)
• Automatic winner selection
• Statistical significance indicator

External:

• Optimizely (advanced)
• VWO (visual editor)
• Google Optimize (free, being sunset)

Calculators:

• A/B test significance calculator
• Sample size calculator

---

Następny krok: 39. Analityka email - jak czytać metrics i podejmować decyzje