A/B Testing

Test a new version of any page against the original. Build the variation with AI, code or the visual editor, split your traffic, and let the numbers decide the winner.

Try this feature for free Read the documentation

Bigger "Buy now" button Running

Original 50 %

Trail running shoe

€ 89.00★★★★★

Buy now

Variation 50 %

Trail running shoe

€ 89.00★★★★★

Buy now

Original1.53 %

Variation1.76 %

+15 %conversion rate · 99.33 % confidence

✓ Significant

What you can do with it

Build variations with AI

Click "Ask AI" in the editor and describe your test idea. The AI makes the changes for you.

Fine-tune visually or in code

Adjust the variation in the visual editor, or add your own JavaScript and CSS where you need full control.

Test several variations at once

Add more variations to the same experiment, edit each one separately and set its share of traffic in whole percentages.

Define who sees it

Audience targeting defines which visitors take part, page targeting where the experiment runs. Mutual exclusion keeps experiments apart.

Check it before it goes live

In QA mode the experiment is not tracked and not shown to normal visitors. You switch variations on your live site and share the link with colleagues.

Decide the winner

Varify uses your existing analytics connection, such as GA4, or measures your goals itself with Varify Realtime Tracking, without delay.

How it works

Create the variation

Create a new experiment for any page. Build the variation with AI, in code or in the visual editor, and save it.

Configure the experiment

Define the page and audience targeting and the traffic allocation between original and variations.

Run QA mode

Check on your website that the variation looks right and the targeting matches, including conflicts with other active experiments.

Start and evaluate

Start the experiment and follow it in the Varify report: conversion rate, improvement and confidence for each variation.

Full guide in the documentation →

See your results right in Varify

One report per experiment, with improvement, confidence and significance for every goal.

Goals from GA4, BigQuery, Shopify or Varify Realtime Tracking
AI Results Summary
Filter by date, device and segment
Share by link or export as CSV

Varify result report with the goals Conversions and Revenue: Variation 1 improves conversions by 9.49 % at 99.04 % confidence, marked significant, and revenue by 9.73 % at 94.28 %, not yet significant

AvailabilityGrowthProEnterprise

A/B testing ✓ ✓ ✓

Multi variation testing ✓ ✓ ✓

Split URL testing ✓ ✓ ✓

Multi page testing ✓ ✓ ✓

Campaign Booster: USP bar, notification, exit intent layer, information bar, arrow up ✓ ✓ ✓

Not sure which plan fits? Compare all plans

Frequently asked questions

How are the results measured?

You choose in the tracking setup. With Varify Realtime Tracking, Varify measures your click, page view and custom goals itself and shows them in the Varify report without delay. With GA4, Varify sends its experiment events to your GA4 property and imports the data back, so you analyze the results in the Varify report or in GA4 itself. Other analytics tools such as Matomo, Piwik Pro or etracker work too; you evaluate there.

Do visitors keep their variation when they come back?

By default, yes: the assignment is kept in Local Storage for as long as the experiment is active. You can switch it so that visitors are reassigned with every new session.

How do I split traffic between several variations?

Enter a whole percentage for each variation. Whatever is left of 100 % goes to the original, a total above 100 % cannot be saved, and a variation at 0 % is not delivered to any visitor.

When is a result significant?

The Varify report highlights a variation with a green "Yes" once its confidence reaches 95 % or more. With Realtime Tracking this works out of the box; with GA4 it needs the "In Varify and GA4" option and a connected GA4 property.

← Back to all features