Skip to main content

​
A/B Testing Images

This guide walks you through creating an A/B test experiment, embedding it in your site or emails, tracking events, and completing the test.

​
Prerequisites

  • A Pictify API key
  • At least two templates to test

​
Step 1: Create an Experiment

Create an A/B test with two variants: 
curl -X POST https://api.pictify.io/experiments/api \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Email Banner CTR Test",
    "type": "ab_test",
    "slug": "email-banner-ctr",
    "variants": [
      {
        "id": "control",
        "name": "Current Design",
        "weight": 5000,
        "templateUid": "tmpl_current123"
      },
      {
        "id": "new-design",
        "name": "New Design",
        "weight": 5000,
        "templateUid": "tmpl_new456"
      }
    ],
    "goalConfig": {
      "type": "click_through",
      "destinationUrl": "https://example.com/landing"
    }
  }'
The experiment is created in draft status.

​
Step 2: Start the Experiment

curl -X POST https://api.pictify.io/experiments/api/exp_abc123/start \
  -H "Authorization: Bearer $API_KEY"

​
Step 3: Embed in Your Content

​
Email

<!-- Variant image with click tracking -->
<a href="https://api.pictify.io/s/email-banner-ctr/click">
  <img src="https://api.pictify.io/s/email-banner-ctr.png" alt="Special offer" />
</a>

<!-- Tracking pixel for open tracking -->
<img src="https://api.pictify.io/s/email-banner-ctr/pixel.gif" width="1" height="1" />

​
Website

<img src="https://api.pictify.io/s/email-banner-ctr.png" alt="Hero banner" />
Each viewer automatically receives a variant based on the weight distribution.

​
Step 4: Track Events (Optional)

For server-side or custom tracking: 
curl -X POST https://api.pictify.io/s/events \
  -H "Content-Type: application/json" \
  -H "X-Write-Key: $WRITE_KEY" \
  -d '{
    "event": "conversion",
    "experiment": "email-banner-ctr",
    "variantId": "new-design"
  }'
Always include a write key to prevent third parties from injecting fake events into your experiment.

​
Step 5: Complete the Experiment

Once you have enough data, declare a winner: 
curl -X POST https://api.pictify.io/experiments/api/exp_abc123/complete \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "winnerVariantId": "new-design" }'
After completion, the experiment URL (/s/email-banner-ctr.png) serves only the winning variant.

​
Tips

  1. Check your quota first: GET /experiments/api/quota shows your plan limits
  2. Run for at least 7 days to account for day-of-week variations
  3. Use at least 1,000 impressions per variant before drawing conclusions
  4. Enable auto-optimization with banditConfig.enabled: true for low-traffic tests
  5. Set a fallback image so the experiment URL works even after the experiment expires