How to Set Up A/B Testing with Google Tag Manager

A/B testing is a cornerstone of optimizing digital experiences. Whether you’re testing landing page designs, call-to-action buttons, or content layouts, a well-structured A/B test can significantly impact conversion rates. In this blog post, we’ll walk you through how to set up an A/B testing script using Google Tag Manager (GTM) and send the results to Google Analytics 4 (GA4) for analysis.

By the end of this guide, you’ll have a ready-to-use script that:

• Randomly assigns users to one of two versions (A or B).
• Redirects users to the corresponding page based on their assignment.
• Tracks experiment participants and their group in GA4.

Why Use GTM and GA4 for A/B Testing?

Google Tag Manager simplifies deploying scripts on your website without needing direct access to the codebase. With GA4’s advanced analytics capabilities, you can measure the performance of your A/B test and gain actionable insights, such as how users interact with each version.

In our case is a very important tool to allow the QA team adjust the experiment probability and change the percentage of users going to each version.

The Script: Setting Up A/B Testing in GTM

Here’s the complete JavaScript code you’ll use in GTM:

(function() {
    // Define the percentage of users entering the experiment (0 to 100)
    var experimentPercentage = 50; // Example: 50% of users enter the experiment

    // Define the probability of assigning users to version B (0 to 1)
    var bVersionProbability = 0.5; // Example: 50% of experiment users go to version B

    // Define the URLs for the A and B versions
    var urlA = "https://example.com/version-a";
    var urlB = "https://example.com/version-b";

    // Helper function to determine if a user enters the experiment
    function isInExperiment() {
        return Math.random() < experimentPercentage / 100;
    }

    // Retrieve the stored variant from localStorage to maintain consistency
    var variant = localStorage.getItem('abTestVariant');

    // If no variant is stored and the user enters the experiment, assign one
    if (!variant) {
        if (isInExperiment()) {
            variant = Math.random() < bVersionProbability ? 'B' : 'A';
            localStorage.setItem('abTestVariant', variant);

            // Send an event to GA4 for entering the experiment
            sendGA4Event("experiment_entered", { experiment_group: variant });
        } else {
            variant = 'Control'; // Not in the experiment
            localStorage.setItem('abTestVariant', variant);
        }
    }

    // Redirect users based on the assigned variant
    if (variant === 'A' && window.location.href !== urlA) {
        window.location.href = urlA;
    } else if (variant === 'B' && window.location.href !== urlB) {
        window.location.href = urlB;
    }

    // Send an event to Google Analytics 4 for the user's variant assignment
    sendGA4Event("ab_test_assignment", { ab_test_variant: variant });

    // Function to push an event to the GA4 Data Layer
    function sendGA4Event(eventName, params) {
        window.dataLayer = window.dataLayer || [];
        window.dataLayer.push({
            event: eventName,
            ...params
        });
    }
})();

Customizable Experiment Settings:

• Control the percentage of users who enter the experiment (experimentPercentage).
• Set the probability of users being assigned to version B (bVersionProbability).

Consistent User Experience:

• Once assigned to a group, users stay in the same group throughout their session using localStorage.

Google Analytics 4 Integration:

• Tracks the experiment participation (experiment_entered) and the variant assignment (ab_test_assignment).

How to Implement the Script in Google Tag Manager

Follow these steps to deploy the script:

Step 1: Create a Custom HTML Tag

1. Open GTM and go to Tags > New > Custom HTML.
2. Paste the script into the editor.
3. Name the tag (e.g., “A/B Test Redirect and Tracking”).

Step 2: Add a Trigger

1. Assign a Page View trigger to the tag so it fires on all page views.
2. Save the tag.

Step 3: Preview and Test

1. Use GTM’s Preview Mode to test the script.
2. Check if users are correctly assigned to variants and redirected accordingly.

Setting Up Event Tracking in GA4

Step 1: Define Events

The script sends two events to GA4:

1. experiment_entered: Tracks users entering the experiment. Includes the parameter experiment_group with values A or B.
2. ab_test_assignment: Tracks all users and their assigned group (A, B, or Control).

Step 2: Create Custom Dimensions

1. Go to Admin > Custom Definitions > Custom Dimensions in GA4.
2. Create dimensions for:
   • experiment_group (from experiment_entered).
   • ab_test_variant (from ab_test_assignment).

Step 3: Validate Events

1. Use GA4’s DebugView to verify the events.
2. Confirm that the events include the correct parameters.

Analyzing Your A/B Test in GA4

Once the data starts flowing into GA4, use Exploration Reports to analyze your test:

• Users Entering the Experiment: Filter by the experiment_entered event to see how many users entered the experiment.
• Variant Distribution: Break down experiment_group to measure how users are split between version A and version B.
• Conversion Analysis: Combine this data with your conversion events to determine which variant performs better.

Setting up A/B testing with GTM and GA4 provides a powerful, flexible way to experiment and optimize your website. This script ensures you can measure participation and results accurately, giving you the insights needed to make data-driven decisions.

Need help implementing this script or analyzing your results? Contact us at Deeploy for expert assistance in A/B testing, analytics, and digital marketing optimization.