Skip to main content

Launching an A/A test in the Manager

Joshua Barnard
  • Updated

This article describes what an A/A test is and how to launch one in the Manager

What are A/A tests?

An “A/A test” is an empty optimisation that collects visitor and conversion data without rendering any changes.

This guide assumes you have access to the Evolv Manager.

You can use an A/A test to:

  1. Validate the Evolv snippet implementation.
    • Confirm data is appearing in the Evolv dashboards.
  2. Check data being captured by third-party analytics tools, like Google Analytics and Adobe Analytics.
  3. Gauge traffic numbers and conversion rates.
    • These figures help you understand how many variants you can test.
  4. Test context and environment-level code.

Pre-launch checklist

Before you can launch your A/A test, ensure you have completed the following steps.

  1. Install Evolv’s Experience Accelerator snippet on the target website.
  2. Log in to the Manager.
    • Create an environment.
      • Create a “Production” environment at a minimum.
      • You might also like to create a QA or UAT environment.
      • Environments are linked to the ID used when implementing the Evolv snippet.
    • Create an audience.
      • Customize the audience based on the scope of your A/A test.
    • Create an event.
      • Define the optimization metric for your test.

How to build an A/A test in the Web Editor

Using Evolv’s proprietary Web Editor, follow these steps to create your project.

  1. Install the Web Editor if you haven’t already.
  2. Create a new project.
  3. Create a new context by selecting the CREATE CONTEXT button.
    • Paste in a URL into the Reference URL field, where you want the optimization to run.
    • We recommend using Full Domain as the Domain Match type.
    • Configure the path matching criteria based on the scope of your optimization.
    • Add some additional reference URLs where you do and do not want the optimization to run - this helps you validate the page matching criteria.
  4. Create a variable and call it “noop”.
    • The variable name doesn’t impact the test, but it’s helpful to know it’s an A/A test when reviewing the project in the Manager.
  5. Create a variant and call it “noop”.
    • When you create a new variable, the Web Editor automatically creates a control variant.
    • After creating the new variant, you should see “control” + “noop” in the variant list.
    • Add a console log to the JavaScript window: 
      console.log('My variant is working!')
      • Remove the code once you have confirmed the variant is working correctly.
      • There should be no CSS or JavaScript code for the “noop” variant when you start the test.
  6. Add event tracking for your optimization metric.
    • Your optimization metric is an event defined in the Manager that the optimization uses to record a conversion.
    • Create a new context based on the page where you want to trigger the event.
    • If your optimization metric is based on a page visit:
      • Add event code to the context-level JavaScript window: 
        evolv.client.emit('your-event-key-goes-here');
    • If your optimization metric is based on a button click or an interaction with another element:
      • This implementation requires a JavaScript function to add a click listener to the desired elements.
      • The click-listener includes a function that triggers the same code that you would fire inside a context.
  7. Enter QA Mode to test your project.
    • Verify the context is matching correctly.
    • Check the variant activates.
      • This is where a console log in the variant’s JavaScript window is helpful. Remember to remove the code before you finish.
  8. Save the project. This will save a .yml file that can be shared or imported to the Evolv Manager.
  9. Go to the Manager and import the project.
  10. Move the project to live when you’re ready for the A/A test to start.

Related articles