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

Debug ModeAdd ?avsb_debug=1 to any page and the snippet says what it did, and why each experiment did not run. Start here.

Status PanelAdd ?avsb_status=1 to any page and read the answer on the page: key, config age, every experiment decision, consent, and the events just sent.

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.

A Feature or Action Is UnavailableA feature has disappeared, your organization is suspended, you have reached a plan limit, or A vs B is down for maintenance.

Before you start troubleshooting

Whatever the symptom, the fastest first step is debug mode: add ?avsb_debug=1 to the page and the snippet prints what it did, including the exact rule that stopped each experiment that did not run.

Most A vs B issues fall into one of four 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.
Access problems: nothing is misconfigured, but a feature is switched off, your organization is on hold, or you have reached a limit on your plan. Start with A Feature or Action Is Unavailable.

Still stuck?

If the Dashboard, your results, or event tracking seem unavailable rather than misconfigured, open the A vs B status page at status.avsb.cloud before you file a ticket. It shows the live health of the Dashboard, event collection, snippet delivery, and scheduled jobs, and it refreshes on its own every five minutes. Any recent day on which one of them failed every check is listed under Incidents. If something there is marked as degraded or down, the problem is on our side and we are already working on it, so you can wait for the all-clear instead of retracing your own setup. If everything is green, the guides above are your fastest path to a fix.

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.