> ## Documentation Index
> Fetch the complete documentation index at: https://withlooksy.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# A/B Testing Virtual Try-On

> Learn how to run A/B tests to optimize Looksy virtual try-on performance. Test button copy and product selection strategies to maximize conversions.

<Warning>
  **Built-in A/B testing tools are coming soon.** Currently, you'll need to use your existing A/B testing platform (Google Optimize, VWO, Optimizely, etc.) to test virtual try-on variations. This guide shows you what to test and how to set up experiments.
</Warning>

## Why A/B Test Virtual Try-On

A/B testing helps you optimize virtual try-on for maximum impact. Small changes can drive **10-30% improvements** in engagement and conversions.

**How to A/B test now:** Use your existing A/B testing tools (Google Optimize, Shopify's built-in tests, or third-party apps) to test the variations recommended below.

## What to A/B Test

### What You Can Test

<CardGroup cols={2}>
  <Card title="Button Copy" icon="pen">
    Test different call-to-action text via dashboard
  </Card>

  <Card title="Button Colors" icon="palette">
    Test different colors via dashboard
  </Card>

  <Card title="Product Selection" icon="check-circle">
    Test which products benefit most from try-on
  </Card>

  <Card title="Promotion Strategy" icon="megaphone">
    Test different ways to promote the feature
  </Card>
</CardGroup>

### Other Tests

* Product selection strategy (which products to disable)
* Promotional tactics (homepage banners, emails, social media)
* Multiple product image angles (max 4 per product)
* Button corner position (top left vs top right)

<Tip>
  Start with button copy tests – easy to do via dashboard and can improve engagement.
</Tip>

## Test 1: Button Copy

### Hypothesis

**Question:** What call-to-action text drives the highest engagement?

**Hypothesis:** Action-oriented, benefit-focused copy ("See it on you") will outperform generic copy ("Try On").

### Test Setup

**Variant A (Control):** "Try On"
**Variant B:** "See it on you"
**Variant C:** "Virtual try-on"
**Variant D:** "Preview on yourself"

**Traffic split:** 25% each
**Duration:** 2 weeks

### Success Metrics

**Primary:** Click-through rate on button
**Secondary:**

* Try-on completion rate
* Conversion rate

### Expected Results

**Typical outcome:**

* Variant A: 35% engagement (baseline)
* Variant B: 41% engagement (+17% – **WINNER**)
* Variant C: 32% engagement (-9%)
* Variant D: 37% engagement (+6%)

**Winner:** "See it on you" (more personal, benefit-focused)

<Note>
  Copy performance varies by audience. Test what resonates with YOUR customers.
</Note>

## Test 2: Supporting Messaging

### Hypothesis

**Question:** Does adding context around the button increase engagement?

**Hypothesis:** Brief explanatory text will increase engagement by reducing uncertainty.

### Test Setup

**Variant A (Control):** Button only, no additional text

**Variant B:** Button + text above:
"See how this looks on you in seconds"

**Variant C:** Button + text below:
"AI-powered virtual try-on – no account needed"

**Variant D:** Button + icon (camera) + text:
📷 "Try On virtually"

**Traffic split:** 25% each
**Duration:** 2 weeks

### Success Metrics

**Primary:** Engagement rate
**Secondary:**

* Completion rate (did context set proper expectations?)
* Conversion rate

### Expected Results

**Typical outcome:**

* Variant A: 35% engagement
* Variant B: 41% engagement (+17% – **WINNER**)
* Variant C: 37% engagement (+6%)
* Variant D: 39% engagement (+11%)

**Winner:** Variant B (benefit-focused, addresses speed concern)

## Test 4: Product Selection Strategy

### Hypothesis

**Question:** Which products benefit most from virtual try-on?

**Hypothesis:** Products with higher return rates will see bigger conversion lift from try-on.

### Test Setup

**Segment A:** Products with \< 15% return rate
**Segment B:** Products with 15-25% return rate
**Segment C:** Products with > 25% return rate

**Measure:** Conversion rate lift for each segment

**Duration:** 4 weeks (longer to account for returns)

### Success Metrics

**Primary:** Conversion rate lift %
**Secondary:**

* Return rate reduction % (track via Shopify analytics, not Looksy dashboard)
* ROI per segment

### Expected Results

**Typical outcome:**

* Segment A: 8% conversion lift, 10% return reduction
* Segment B: 15% conversion lift, 20% return reduction
* Segment C: 22% conversion lift, 35% return reduction (**HIGHEST ROI**)

**Insight:** Focus try-on on high-return products for maximum impact

<Check>
  Products with fit/style uncertainty benefit most from virtual try-on.
</Check>

## How to Run A/B Tests

### Using Shopify Theme Editor

**For button copy and styling:**

1. Create multiple product page templates with different button configurations
2. Assign products randomly to each template
3. Track performance via Looksy analytics
4. Compare engagement rates

### Using Third-Party Tools

**Recommended tools:**

* **Google Optimize** (free, integrates with GA)
* **Optimizely** (enterprise, advanced features)
* **VWO** (visual editor, easy setup)
* **Convert** (privacy-focused)

**Setup:**

1. Install A/B testing tool on your store
2. Create variants in the tool
3. Set up goal tracking
4. Run experiment

### Native Looksy Testing (Pro Plan)

**Built-in A/B testing features:**

* Button copy variants
* Button styling tests
* Automated winner selection
* Statistical significance calculation

**Access:** Looksy Dashboard → Experiments

<Tip>
  Use native Looksy testing if available – it's pre-configured for virtual try-on optimization.
</Tip>

## Statistical Significance

### How Long to Run Tests

**Minimum requirements:**

* **At least 1,000 visitors per variant**
* **At least 2 weeks** (account for weekday/weekend patterns)
* **95% statistical confidence** before declaring a winner

**Calculator:**

```
Required sample size =
  (Z-score² × p × (1-p)) / (margin of error²)

For 95% confidence, p=0.5, 5% margin:
  ≈ 385 visitors per variant minimum
```

**Rule of thumb:** Run tests until you have 1,000+ visitors per variant and clear winner emerges.

<Warning>
  Declaring a winner too early leads to false positives. Be patient and wait for statistical significance.
</Warning>

### Significance Calculators

Use online tools to verify significance:

* Optimizely Stats Engine
* VWO's Bayesian calculator
* AB Testguide calculator

**Check:** Is the p-value \< 0.05? (95% confidence)

## Analyzing Results

### What to Look For

**Clear winner:**

* One variant significantly outperforms others
* Results are consistent across days
* Statistical significance achieved

**No clear winner:**

* Variants perform similarly
* High variance in results
* Need more data or different test

**Unexpected results:**

* Variant performs worse than control
* Inconsistent patterns
* Check for implementation bugs

### Making Decisions

**If there's a clear winner:**

1. Implement winning variant for all traffic
2. Document learnings
3. Plan next test

**If results are inconclusive:**

1. Extend test duration
2. Increase traffic to experiment
3. Or move on to different test

**If all variants underperform:**

1. Check for technical issues
2. Review test setup
3. Consider different hypothesis

## Sequential Testing Strategy

### Recommended Test Order

**Phase 1: Foundation (Weeks 1-4)**

1. Button placement
2. Button copy
3. Button styling

**Phase 2: Optimization (Weeks 5-8)**
4\. Supporting messaging
5\. Mobile-specific optimizations
6\. Product selection strategy

**Phase 3: Advanced (Weeks 9-12)**
7\. Promotional tactics
8\. Image quality improvements
9\. Multi-variate tests (combine winners)

<Tip>
  Test one element at a time. Don't change multiple things simultaneously or you won't know what drove the improvement.
</Tip>

## Common A/B Testing Mistakes

### Pitfalls to Avoid

<AccordionGroup>
  <Accordion title="1. Stopping test too early">
    **Problem:** Declaring winner before statistical significance

    **Solution:** Wait for minimum sample size and 95% confidence
  </Accordion>

  <Accordion title="2. Testing too many variants">
    **Problem:** 6+ variants dilutes traffic, takes forever

    **Solution:** Limit to 2-4 variants maximum
  </Accordion>

  <Accordion title="3. Changing test mid-flight">
    **Problem:** Adjusting variants during test invalidates results

    **Solution:** Plan thoroughly, don't change once live
  </Accordion>

  <Accordion title="4. Ignoring external factors">
    **Problem:** Traffic spike from campaign skews results

    **Solution:** Note external events, extend test if needed
  </Accordion>

  <Accordion title="5. Not documenting learnings">
    **Problem:** Forgetting what was tested and why

    **Solution:** Maintain test log with hypotheses and results
  </Accordion>
</AccordionGroup>

## Test Results Template

### Document Every Test

**Test Name:** Button Placement - Above vs. Below Fold

**Date:** Jan 15 - Jan 29, 2026

**Hypothesis:** Button above fold will increase engagement by 30%

**Variants:**

* Control: Button below description (28% engagement)
* Variant B: Button above Add to Cart (38% engagement) ✓ WINNER

**Results:**

* Sample size: 2,400 visitors per variant
* Improvement: +35.7% engagement
* Statistical significance: p \< 0.01 (99% confidence)
* Winner: Variant B

**Action:** Implement Variant B sitewide

**Learnings:** Visibility matters more than placement within product details. Test mobile sticky bar next.

**Next test:** Button copy optimization

## Advanced: Multi-Variate Testing

### Testing Multiple Elements Simultaneously

**Example:** Test button placement AND copy together

**Variants:**

* A: Above fold + "Try On"
* B: Above fold + "See it on you"
* C: Below fold + "Try On"
* D: Below fold + "See it on you"

**Pros:** Find optimal combination faster
**Cons:** Requires 4x more traffic

**Recommendation:** Only do multi-variate after single-variable tests establish baselines

## Next Steps

<CardGroup cols={2}>
  <Card title="Dashboard Overview" icon="chart-line" href="/analytics/dashboard-overview">
    Track your test results in real-time
  </Card>

  <Card title="Best Practices" icon="star" href="/product-setup/best-practices">
    Implement optimization strategies
  </Card>

  <Card title="Calculating ROI" icon="calculator" href="/analytics/calculating-roi">
    Measure the impact of your tests
  </Card>

  <Card title="Custom Styling" icon="palette" href="/integration/custom-styling">
    Customize button styling for your brand
  </Card>
</CardGroup>