Documentation | Split Test Pro

Welcome to Split Test Pro

Split Test Pro gives your Shopify store the ability to run controlled A/B experiments. Instead of guessing what changes will improve your conversions, you let real visitor data decide.

The app is organized into four main areas: the Dashboard, Experiments, Settings, and Support.

Dashboard

Your starting point. The dashboard shows a summary of your active experiments and recent results at a glance.

Experiments

Create, monitor, and review all of your A/B tests. Each experiment shows its current status, the variants being tested, and live results.

Settings

Manage your subscription, billing, and account preferences. You can update your payment method or cancel your subscription from here.

Support

Access documentation, open a support request, or use the live chat to get help from the Split Test Pro team.

Setting up your first A/B test

Follow these steps to launch your first experiment. Most merchants have a test running within 10 minutes.

1

Create a new experiment

From the Experiments section, click "Create Experiment". Give it a descriptive name — for example, "Homepage CTA button color test".

2

Set your URL targeting

Define which pages the experiment runs on. You can target a specific URL (e.g., /products/my-product) or a URL pattern (e.g., /collections/*).

3

Add your variants

The original page is your control (Variant A). Add one or more variants with CSS changes that modify the page's appearance — button colors, font sizes, layouts, and more.

4

Write your CSS

Use the CSS editor to write the changes for each variant. The editor shows a live preview of your changes against your store's existing styles.

5

Start the experiment

Click Start to begin. Split Test Pro will randomly assign visitors to variants and begin collecting data immediately.

Reading your results

The results view shows how each variant is performing relative to your control. Here's what each metric means.

Sessions

The number of unique visitor sessions assigned to this variant during the experiment window.

Conversions

The number of sessions that resulted in a conversion event (e.g., a completed purchase or add-to-cart).

Conversion Rate

Conversions divided by sessions, expressed as a percentage. This is the primary metric you're trying to improve.

Probability to be Better

The statistical likelihood that this variant outperforms the control. A value above 95% is generally considered significant.

Probability to be Best

When running multi-variant tests, this shows the probability that this variant is the best performer across all variants.

When to declare a winner

We recommend waiting until your winning variant reaches at least 95% probability to be best before ending the test. Declaring a winner too early can lead to false positives. Once you have a winner, you can apply its CSS changes permanently to your theme and end the experiment.