A/B Testing Overview
Run controlled experiments to find the prevention settings that convert best for your store
A/B Testing is currently available through your Account Manager. Contact your account manager to set up an experiment for your store. Self-service A/B testing is rolling out to all merchants soon.
A/B testing lets you compare two different prevention configurations side-by-side to find what actually converts for your store. Instead of guessing whether a 10% or 15% discount performs better — or whether offering a discount at all makes a difference — you run a controlled experiment and let real shopper data decide.
How It Works
NavonaAI splits your incoming cart traffic 50/50 between two variants:
- Control — typically your current configuration (or AI prevention turned off)
- Treatment — the configuration you want to test
Each cart is deterministically assigned to one variant, so every shopper sees the same experience consistently across their session. The split is sequential, ensuring an exact 50/50 distribution over time. Assignment happens at the moment the prevention popup is about to appear.
You define a primary goal KPI when creating the experiment — such as conversion rate or revenue — and NavonaAI tracks results for both variants until the experiment ends.
Only one experiment can run per store at a time.
Next Steps
- What You Can Test — configurable settings per variant
- Experiment Lifecycle — from draft to winner promotion
- Metrics & Results — tracked KPIs and how winners are determined
- Best Practices — tips for running effective experiments