Google Optimize services are meant first and foremost for increasing conversions using the best method available: running constant experiments, i.e. changing different elements on you landing pages. To see if the changes increase desired user behavior such as add to cart or add a review. Before Optimize and similar tools, you needed a developer to create the second variant of your page, in an expensive and cumbersome manner. With Optimize, you get an elegant WYSIWYG editor layered on top of your site, allowing instant changes.
As part of the Google Marketing Platform, Optimize is tightly integrated with both Tag Manager and Analytics. Implementing is done via GTM, while A/B (or multivariate) test reports are viewed within Analytics.
Implementing Optimize usually requires a few separate steps:
- Deploy GTM code (if not already installed) or Optimize code snippet if not using GTM
- Deploy page-hiding snippet (though not mandatory, it is advised)
- Set-up the experiment using Optimize
- Connect to GA
- Choose test type, audience etc
- Create the B version of your page
- View reports in Google Analytics
Running Successful AB Tests
Not all websites are made equal. If your site has low traffic volumes, tests may take significant amounts of time before the system decides on a conclusive winner. So as a general rule of thumb, smaller sites need to focus AB testing on relatively big changes from A-B versions, and not subtle changes like many larger websites usually make.
Determining Which Elements to Change
Most obvious candidates to start with are your CTA, call to action, elements. It may be a “buy now” button, “subscribe here” box, or banner placement on the page. Whatever you consider as a conversion and the funnel leading to it.
Tests can be as simple as changing a button color and text for all users who visit the B version or more complex, targeting multiple audience types and showing each one a different version of your landing pages.
The most basic experiments target all your traffic, regardless of the device they use, their traffic source and other factors, and simply redirecting 50% of them to the alternate version. This is fine in many cases, but often it is a good idea to define an audience. Since mobile users often convert less effectively than desktop users, it is highly recommended running A/B tests targeting users with the appropriate attributes.
In another example, you may want to create different tests for new vs returning visitors. Returning visitors may already contemplate a conversion, while new visitors are often only checking and comparing different brands and require a different call to action.