Back to Documentation

A/B Testing & Link Variants Guide

Public Document • Documentation

Split traffic across multiple destinations to optimize conversion rates

Related Document: Conversion Tracking Guide

Track conversions for A/B tests

A/B Testing & Link Variants Guide

Split traffic across multiple destinations to optimize conversion rates

Overview

PVTLNK’s A/B testing feature allows you to:

  • Test different destinations - Compare performance of different URLs
  • Optimize conversions - Identify which destination drives more actions
  • Make data-driven decisions - Use actual click data to guide choices

Requirements

Plan Availability
Starter ❌ Not available
Pro ✅ Available
Enterprise ✅ Available

How It Works

Traffic Distribution

When you create variants, PVTLNK distributes traffic:

  • Equal split (50/50) by default
  • Weighted splits available (e.g., 70/30 for gradual rollouts)
  • Random assignment ensures unbiased results

Tracking & Analytics

Each variant tracks:

  • Total clicks per variant
  • Conversion rate (when goals are configured)
  • Device/browser/country breakdowns

Creating a Test

  1. Go to Create New Link
  2. Scroll to A/B Testing Variants section
  3. Click + Add Variant
  4. Enter destination URLs (one per variant)
  5. Optionally add a name for each variant
  6. Create the link
  1. Go to All Links
  2. Click Edit on the link
  3. Scroll to A/B Testing Variants
  4. Click + Add Variant
  5. Save changes

Variant Types

Simple Variants

Test different destination URLs:

Variant Destination
A https://landing-page-a.com
B https://landing-page-b.com

Redirect Rules

Use time-based redirects for gradual rollouts:

Rule Destination Traffic %
Rule 1 https://new-site.com 10%
Rule 2 https://old-site.com 90%

Analyzing Results

Analytics Dashboard

  1. Click Analytics on the link
  2. Scroll to A/B Testing Variant Performance
  3. View comparison metrics

Key Metrics

Metric Description
Clicks Total clicks per variant
Share Percentage of total traffic
Conversion Rate Conversions / Clicks × 100
Winner Highest performing variant (auto-identified)

Export Data

Export A/B test data for external analysis:

bash # API endpoint GET /api/v1/links/{id}/variants

Best Practices

1. Test One Variable at a Time

Good: Test different landing pages with same ad copy Bad: Test different URLs AND different ad copy simultaneously

2. Run Tests Long Enough

  • Minimum: 1,000 clicks per variant
  • Recommended: 7 days minimum
  • Consider day-of-week variations

3. Use Statistical Significance

Clicks per Variant Confidence
< 100 Low - not actionable
100-500 Medium -初步 insights
500-1000 High - reliable data
1000+ Very high - conclusive

4. Document Your Hypothesis

Before starting:

```markdown # A/B Test Document

Test Name: Homepage CTA Button Color Hypothesis: Orange button will convert better than blue Control (A): Blue button Variant (B): Orange button Success Metric: Button clicks Duration: 7 days ```

Common Use Cases

1. Landing Page Testing

Test different landing pages for the same campaign:

Primary Link → [A/B Test] → Variant A: Sales Page → Variant B: Webinar Registration

2. URL Shortener Migration

Gradually migrate from old short links:

Old Link → [A/B Test] → 10% New shortener → 90% Old shortener

3. Campaign Attribution

Test which source performs better:

Campaign Link → [A/B Test] → Variant A: facebook.com/utm_source → Variant B: google/utm_source

API Integration

Create and manage A/B tests via API:

```bash # Create link with variants POST /api/v1/links { “link”: { “destination_url”: “https://example.com”, “title”: “A/B Test Campaign” }, “variants”: [ {“destination_url”: “https://variant-a.com”, “weight”: 50}, {“destination_url”: “https://variant-b.com”, “weight”: 50} ] }

Get variant performance

GET /api/v1/links/{id}/variants

Update variant weights

PATCH /api/v1/links/{id}/variants { “variants”: [ {“id”: 1, “weight”: 70}, {“id”: 2, “weight”: 30} ] } ```

Troubleshooting

Issue Solution
Traffic not distributing Check variant weights sum to 100%
No winner identified Need more clicks for statistical significance
Variant not tracking Verify destination URL is accessible
Wrong variant showing Clear browser cache and test again

FAQ

Q: Can I have more than 2 variants? A: Yes, up to 5 variants per link.

Q: Can I change variants after creating? A: Yes, but changes only affect new clicks.

Q: What happens if a variant destination is down? A: That variant will show errors, other variants continue working.

Q: Can I use A/B testing with password protection? A: Yes, password protection works with A/B testing.