NavonaAINavonaAI
A/B Testing

What You Can Test

See which offer settings you can compare in a NavonaAI A/B test, such as discount type and amount, minimum purchase and stacking, by plan.

Each variant in an A/B test can be configured independently across all prevention settings. This lets you isolate exactly which change drives better results.

Testable Settings

SettingWhat You Can Compare
AI Prevention toggleEnabled vs. disabled — the classic holdout test
Incentive typePercentage discount vs. fixed amount
Incentive amounte.g., 10% off vs. 15% off
Minimum purchase thresholdNo minimum vs. a specific cart value
Discount stackingWhether the NavonaAI discount combines with order, product, or shipping discounts
Product/collection exclusionsRestrict discounts to certain parts of your catalog

Testing Popup Designs

An offer test varies the settings above and shows both groups the same popup design. To compare two different popup designs, run a design test (Scale and Done-For-You plans).

Each group in a test sees one thing. The app recommends changing one thing at a time, so you know which change made the difference. The What differs panel in the test form lists every difference between Group A and Group B and tells you when you have changed just one.

Plans

Offer tests are available on the Pro, Scale and Done-For-You plans. Design tests are available on Scale and Done-For-You.

Common Experiment Scenarios

The most valuable experiment to run first:

  • Control: AI prevention OFF (no popup shown)
  • Treatment: AI prevention ON with your current settings

This tells you definitively how much lift NavonaAI is generating for your store.

Discount Amount Optimization

  • Control: 10% discount
  • Treatment: 15% discount

Find the sweet spot between conversion rate and margin impact.

Discount Type Comparison

  • Control: 10% percentage discount
  • Treatment: €5 fixed amount discount

Test which format resonates better with your customers.

Minimum Purchase Threshold

  • Control: No minimum purchase required
  • Treatment: €30 minimum cart value

Determine if a threshold protects margins without hurting conversion.

For the clearest results, change only one variable between your control and treatment. If you change the incentive type and the amount at the same time, you won't know which change drove the difference.

On this page