A/B Testing

GoPayLocal’s A/B testing lets you test different discount levels for the same country to find the optimal balance between conversion rate and revenue.

Why A/B Test Discounts?

A 30% discount might convert better than 20%, but is the extra conversion worth the 10% revenue per sale? A/B testing helps you answer this.

Example question: Does a 40% discount for India convert enough more than 30% to justify the lower per-sale revenue?

How It Works

1. Create two campaigns for the same product with different discount settings
2. GoPayLocal randomly assigns visitors to one of the campaigns (50/50 split)
3. Track conversions for each variant
4. After sufficient data, pick the winner

Setup

1. Create Variant Campaigns

Create two campaigns for the same product:

Campaign A — Control (30% discount for India):

Name: Pro Plan PPP - Control
Pricing Mode: Group
Tier 4 (India): 30%

Campaign B — Variant (45% discount for India):

Name: Pro Plan PPP - Variant
Pricing Mode: Group
Tier 4 (India): 45%

2. Enable A/B Testing

In the dashboard, link the two campaigns as an A/B test:

1. Go to Campaign A > Settings > A/B Testing
2. Click Create A/B Test
3. Select Campaign B as the variant
4. Set the traffic split (default: 50/50)
5. Click Start Test

3. Monitor Results

View A/B test results in the analytics dashboard:

Metric	Campaign A (30%)	Campaign B (45%)
Impressions	5,100	5,080
Copies	1,428	1,524
Redemptions	204	279
Conversion Rate	4.0%	5.5%
Avg Revenue/Sale	$69.30	$54.45
Total Revenue	$14,137	$15,192

4. Pick a Winner

In this example, Campaign B (45% discount) generates more total revenue despite lower per-sale revenue, making it the winner.

Statistical Significance

GoPayLocal does not automatically calculate statistical significance, but you can export the data and use standard tools:

• Minimum sample size: Aim for at least 100 conversions per variant
• Test duration: Run tests for at least 2 weeks to account for weekly patterns
• Confidence level: 95% confidence is standard

Tip

Use a calculator like AB Test Calculator to determine if your results are statistically significant before making changes.

Best Practices

Test One Variable at a Time

Only change the discount percentage between variants. Keep everything else the same (message, colors, position) to isolate the impact.

Test the Right Countries

Focus A/B testing on countries with enough traffic to reach statistical significance. India, Brazil, and Nigeria are common high-traffic PPP markets.

Consider Revenue, Not Just Conversions

A higher conversion rate doesn’t always mean more revenue. Calculate total revenue for each variant:

Total Revenue = Conversions × (Base Price × (1 - Discount))

Don’t Test Too Many Variants

Stick to 2 variants (A/B). More variants require proportionally more traffic to reach significance.

Ending a Test

When you’re ready to end a test:

1. Go to the A/B test settings
2. Click End Test
3. Select the winning variant
4. The losing variant is paused

The winning variant becomes the active campaign, serving all traffic.

Note

A/B testing discounts is a unique GoPayLocal feature not available in ParityDeals or Evendeals. Use it to find the discount levels that maximize your global revenue.