Creating an experiment

Step-by-step guide to setting up a new A/B experiment.

New experiment form

Go to Experiments → New experiment and fill in:

Field	Description
Name	Internal label — merchants see this, visitors don’t
Hypothesis	What you expect to happen and why. e.g. “Showing a urgency badge will increase ATC rate”
Type	What you’re testing. See Experiment types
Traffic allocation	Percentage of visitors included in the experiment (the rest see the default)
Segment	Optional — restrict the experiment to a specific audience

Variants

Every experiment has at least two variants:

• Control — the current experience, unchanged
• Variant B — the treatment you’re testing

You can rename these anything. Traffic is split 50/50 by default. After creating the experiment, you can adjust weights per variant from the Variants tab.

Variant weights must sum to exactly 100. If they don’t, visitors in the gap won’t be assigned to any variant and won’t be tracked.

After creating

You’ll be taken to the experiment detail page. From here:

1. Go to the Variants tab and click Configure on each variant to set up the specific changes
2. Check the Overview tab to verify settings
3. Click Start experiment when ready

Editing a running experiment

You can change the experiment’s segment at any time from the Overview tab without restarting. Changing variant weights or content on a running experiment is not recommended — it can introduce bias into results collected before the change.