> ## Documentation Index
> Fetch the complete documentation index at: https://docs.funeli.top/llms.txt
> Use this file to discover all available pages before exploring further.

# Run A/B experiments on Funeli funnels

> Split live traffic between funnel variants in Funeli, read revenue per visitor and the chance a variant is better, and copy the winner into the original funnel.

A Funeli experiment splits the traffic of a live funnel between two or more variants, each a published funnel of the same project. The ad link stays the same: visitors who open the original funnel are assigned to a variant.

## Set up an experiment

**1. Make the variant**

Duplicate the live funnel and change what you want to test: a price, a headline, the order of steps. Publish the copy.

**2. Create the experiment**

Open the original funnel's menu (**⋯**) and choose **Experiments**. Name the experiment, write the hypothesis and add the variants: the original and the copy, each with a variant id. Starting an experiment splits real traffic.

**3. Check a variant**

Open the original funnel's link with `?variant=<variant id>` to see a specific variant.

## Read the result

For each variant the experiment shows **Conversion**, **Revenue / visitor** and the **Chance better** than the original. Decide on revenue per visitor rather than conversion alone, and give the experiment enough traffic before you pick a winner; an early lead often disappears.

## Finish

Stop the experiment and choose the winner. **Copy winner into the host draft** copies the winning variant into the original funnel's draft; review it and publish.

You can also ask an AI agent to set up, read and stop experiments; it asks before it splits traffic.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.