Troubleshooting
Something not working the way you expected? This section covers the most common issues people run into with A vs B and walks you through how to diagnose and fix each one. Start with the page that matches your symptom.
Common issues
Snippet Not DetectedThe project settings page shows 'Not detected' even though you installed the snippet.
Experiment Not RunningThe experiment is published but visitors are not being bucketed into variations.
No ConversionsVisitors are showing up in the results but the conversion count stays at zero.
FlickerVisitors briefly see the original page before the variation loads (flash of original content).
SPA IssuesThe experiment works on the first page load but stops working after in-app navigation.
Sample Ratio MismatchThe results page shows a yellow or red SRM guardrail warning.
Results Look WrongThe numbers on the results page are confusing, surprising, or seem incorrect.
Before you start troubleshooting
Most A vs B issues fall into one of three categories:
- Installation problems — the snippet is not on the page, is blocked, or has the wrong key. Start with Snippet Not Detected.
- Configuration problems — the experiment targeting, audience rules, or metrics are not set up correctly. Start with Experiment Not Running and No Conversions.
- Data quality problems — the experiment is running correctly but the results are being affected by bots, bugs in variation code, or statistical noise. Start with Sample Ratio Mismatch and Results Look Wrong.
Still stuck?
If you have worked through the relevant troubleshooting guide and are still unable to resolve the issue, reach out to support. When you do, include:
- The URL of the page where the issue occurs.
- The experiment ID (found in the experiment overview panel).
- Any JavaScript errors you see in the browser DevTools console.
- Screenshots of the Network tab showing the snippet request (if relevant).
- A description of what you have already tried.