Docs / Run your first A/B test

Run your first A/B test

Opinions are free; data costs a variant. Here's how to run a clean test.

1. Create a variant

Open your published page's Variants panel and create Variant B - it starts as a copy of your page. Change one meaningful thing: the headline, the hero image, the offer. One change per test keeps the result readable.

2. Split the traffic

Give the variant a traffic weight - 50/50 is the fastest way to learn. Visitors are assigned sticky: the same person always sees the same version, so nobody notices the test is happening.

3. Let it run

Resist peeking-and-declaring on day one. As a rule of thumb, wait for at least a few hundred visitors per variant or a clearly consistent gap over a week - small samples lie.

4. Read the results

The analytics view shows views, leads, and conversion rate per variant, side by side with your control. The winner is usually obvious; if it isn't, the test was a tie - change something bolder next round.

5. Promote the winner

One click makes the winning variant your live page. Then start the next test - compounding small wins is the whole game.