How to create a checkout experiment

This guide shows you exactly how to set up and run a custom code test using Elevate. We'll cover everything from creating your test to analyzing the results, so you can figure out what code modifications work best for your store.

Step 1: Create an Experiment

Begin by accessing your Elevate A/B Testing dashboard. From there, look for the 'Experiments' section in the main menu. Once you've found it, click on 'New Experiment' and select 'Custom Code Experiment' from the available options to start your journey.

Step 2: Define Your Test

Every successful experiment starts with clear documentation. Give your experiment a descriptive name that helps you identify it at a glance, such as "Product Page Layout Test" or "Custom CTA Button Styling." While optional, we strongly recommend adding a detailed description explaining what you're testing and writing out your hypothesis. Your hypothesis should clearly state what outcome you expect and why you expect it. This documentation will prove invaluable when analyzing your results later.

Step 3: Add Your Custom Code

For each variation, you can add custom code to modify your store:

• CSS Code: Add custom styling to change the appearance and design of elements

• JavaScript Code: Add custom functionality to change behavior and interactions

Write your code in the respective fields for each variation. Your control variation will be your current store experience without any code modifications.

Step 4: Configure Page Targeting

Choose where your custom code will run. By default, your code runs on all pages across your store. If you want more control, toggle off "Run on all pages" and specify:

• Include Pathnames: Enter specific page paths where you want the code to run (e.g., /products/summer-collection)

• Exclude Pathnames: Enter specific page paths where you want to prevent the code from running

Visitors will be marked for your experiment when they visit any of the included pages, and your code will execute only on those pages.

Step 5: Configure Traffic Distribution

By default, your traffic will be split evenly between your control and your test variation. For example, if you have two variations, each will get 50% of your traffic. You can adjust these percentages however you want - just make sure they add up to 100%.

Step 6: Set Your Experiment Goal

What do you want to measure? Pick a main goal that will determine your winning variant. You can choose from our recommended metrics: revenue per visitor, conversion rate, average order value, add-to-cart rate, or checkout start rate. Don't worry - we'll still track all other metrics for you, but your main goal helps us identify when your test has a clear winner.

Step 7: Add Audience Targeting (Optional)

You can optionally add audience targeting to refine who sees your experiment. Target based on:

• Devices: Choose between mobile, desktop, or both.

• Visitor Types: Distinguish between new and returning visitors.

• Traffic Sources: Filter traffic based on specific marketing channels or any UTM parameters (utm_source, utm_medium, utm_campaign, utm_term, utm_content, and others).

• Locations: Target specific countries or regions.

These filters allow you to run more precise tests and get targeted insights from different customer segments.

Step 8: Review and Launch

With your setup complete, click 'Create Experiment' to move into the review phase. The system will perform initial checks to ensure everything is configured correctly. Review your custom code and page targeting one final time, then click 'Launch Experiment' to begin testing. Your experiment will start collecting data immediately, and you can monitor the results in real-time through your dashboard.

Next Steps

Once your experiment is live, keep an eye on its performance through your dashboard. Let the experiment run until you have enough data to make confident decisions about your custom modifications.