A/B Testing Your Listings

Updated August 6, 2026

Test titles, tags, and prices on your real listing and let buyer behavior pick the winner.

A/B Testing Your Listings

Run controlled experiments on your real Etsy listing (PRO plan and above).

How It Works

True A/B split-serving is not possible via the Etsy API — Etsy shows every buyer the same listing. Instead, Selletto rotates each variant onto your live listing for a full week at a time (so weekday/weekend traffic stays comparable) and attributes the views, favorites, and sales earned during each variant's window. The Rotation Timeline on the experiment page shows exactly which variant was live when, and what it collected.

Creating an Experiment

Go to Dashboard > Experiments and click New Experiment.

Select the listing you want to test. The Control is snapshotted from your current listing automatically — you don't need to enter it.

Enter the Test variant's title, tags, price, or description — or click Suggest a test (1 AI credit) to have the AI pre-fill one.

Optionally set a maximum duration (4–12 weeks; the test auto-stops after this) and turn on auto-promote to have the winner applied automatically when the result is decisive.

Reading Results

The headline stat is the chance to beat your control — e.g. "the test variant has an 87% chance of beating your original." Unlike a p-value, it's meaningful at any traffic level and safe to check daily. Each variant also shows a 95% credible range for its conversion rate. The classical chi-squared p-value is still available under Advanced stats.

You'll get a notification (and an email) the moment a variant reaches a 95% chance of beating your control.

Ending an Experiment

When you end a test you choose what your listing shows next:

Apply winner — pushes the winning variant live on Etsy

Restore original — puts your listing back exactly as it was

Cancelling always restores the original. Your listing is never left mid-test on an arbitrary variant. Completed experiments with a winner also show an Apply Winner button any time after ending.

Was this article helpful?