A/B tests

You can experiment with web funnels by splitting traffic between different web funnels.)

Creating experiments

1. Create control and test version of a web funnel.
2. Publish the release version of funnels.
3. Create an experiment in the Experiments section
4. Specify an experiment name and URL where you want to run traffic.

5. Select control Funnel A and variant funnel B and set traffic split. You can create up to 4 funnels per A/B test, allowing you to compare multiple variations at once. And create an experiment.

Once created, an experiment is live and traffic will be distributed between the control and variant funnels.

❗️Make sure your funnels are published before creating experiments.

Stopping experiments and rolling out variants

When you decide to stop an experiment, click Finish Experiment and choose which variant to roll out.

🚧

Once the experiment is finished, the winning funnel will automatically receive the experiment URL as its funnel URL.

Running multiple experiments for one funnel

You can create multiple experiments for one funnel (with the same experiment URL). Traffic will be distributed between experiments equally.

Analyzing experiment results

FunnelFox sends ff-experiment and ff-experiment-variation user properties to all connected analytics systems for funnels included in an experiment.

To analyze experiment results you need to:

• filter out all users included in the experiment by the ff-experiment user property
• and group events by the onboarding parameter to filter out variations included in the experiment.

Learn more about analytic events [here.](https://docs.funnelfox.io/docs/tracking-analytics