How to conduct your first A/B test: Automate the process with Google Optimize
Your site is up and running. New visitors are coming every day. Many are becoming your customers. You exhale, praise yourself, and enjoy the tranquility.
But one day, thoughts pop up in your head: Maybe it’s worth changing the background on the main page? What if visitors don’t like this picture? Maybe if we move this text to the right and the button to the left we’ll get more sales? What if these changes actually lower the conversion rate?
If you’re faced with this situation, it’s time to figure out A/B testing. This guide will tell you how to properly automate A/B testing, why Google Optimize is necessary for this, how to configure it, and what points you should pay attention to during your first experiment.
Table of contents
- What is Google Optimize?
- How to create your first experiment
- Reports in Google Optimize
- Wrapping up
Find out the real value of campaigns
Automatically import cost data to Google Analytics from all your advertising services. Compare campaign costs, CPC, and ROAS in a single report.
What is Google Optimize?
Google Optimize is a website interface testing service. You can use it to conduct A/B and multi-channel testing and to collect and analyze data about user behavior on your site.
For example, you can hide or change certain buttons on the site as an experiment, compare different landing pages, or test pages with different designs. As a result of these experiments, you’ll receive reports on the effectiveness of your changes and understand their effect on conversions.
How to create your first experiment
Create an account in Google Optimize
1. Go to Google Optimize and click Create Account.
3. On the next page, you can create your first optimization project.
4.In the window that opens, name your experiment, specify the address of the page you plan to test, and select the type of experiment.
There are three types of tests available on Google Optimize:
- Experiment A/B for testing multiple page variations (A/B/N testing)
- Multivariate experiment for testing options with several different sections
- Redirect experiment for testing individual web pages defined by different URLs
For an A/B test, select A/B experiment and click the Create button.
You’ll automatically be redirected to the project launch page.
Here are all the steps you need to go through to start the experiment.
Install Google Optimize (via Google Tag Manager)
Click the Connect to Google Analytics button.
Select the desired resource and view, then click Create.
Important: Using views, you can access the goals, audiences, and filters in Google Analytics. You must select at least one view for the container. If you specify several views, the container will get access to all of them, but only one can be associated with each optimization project. Learn more in the Google Optimize Help Center.
Now you need to either set up the Optimization code yourself or set it up using Google Tag Manager.
To post code through Google Tag Manager:
- Log in to your Google Tag Manager account.
If you haven’t worked with this service yet, here’s our setup guide.
- Click New tag. The tag setup screen will appear.
- Click Tag Configuration and select Google Optimize.
- Add your Google Analytics code and Google Optimize Container ID.
Important: Your Google Optimize Account ID can be found in your Google Optimize account settings, as shown in the screenshot below.
- Select an existing variable or create a new Google Analytics settings variable.
- For all your tags and changes to start working on the site, be sure to click the Publish button in Google Tag Manager.
If you’ve done everything right, information on connecting Google Analytics will automatically be pulled to the Google Optimize settings page.
If the code wasn’t installed correctly, you’ll see an error notification on the Google Optimize page.
Diagnose and correct any errors.
Important: Sending data from Google Analytics to Google Optimize can take up to 12 hours. As soon as your experiment is completed, the data will no longer be automatically sent to Optimize.
Pay attention to one more important setup step.
In order to prevent website visitors from seeing the element substitution process itself, you need to change the order of the tags so the Optimize tag works before the Google Analytics tag.
To do this, go to the advanced settings of the Analytics tag, check the Fire a tag before GA fires box, and select the Google Optimize tag from the list.
Choosing audiences to target and goals for your experiment
To select audiences and goals, first click Add Variant.
Name your option and click Done.
On the new page that loads, you’ll see two options for your site: The original option (automatically created by Google Optimize) and option 1 (which you just created).
In order to somehow determine the change for your website page for the experiment, click the Edit button.
You’ll be prompted to install an extension with which you can configure the same change. Install it.
Make sure the extension is enabled.
Now again click the Change button on the page with the options for your experiment. You’ll be taken directly to the visual editor (which will appear on your website page) where you can:
- Delete page elements
- Resize or reposition items
- Change the text on the site as well as its font, alignment, size, and inscription color
- Add a background in the form of a color fill, set a frame, or change the transparency
Select the change you want to make and click Save, then click Finish in the upper right corner of the page.
Important: It’s best to run only one test at a time. By running one test per change, it’s easier to understand how this change affects conversions.
Now we’ll set up an audience for targeting: only a certain group of users will see your experiment.
You’ll have the opportunity to choose an audience:
- The same as one you created in Google Analytics (for this you’ll need the paid version of Google Optimize 360)
- Based on Google Ads accounts, campaigns, ad groups, and keywords
- Based on devices: smartphones, tablets, or computers
- Based on the behavior of your users: new, returning, or those who clicked from specific URLs
- From a specific city, metro area, region, or country
In our example, we’ll select Behavior and set the necessary variables for this type of audience. Then click Add.
We’re done with this stage. Now let’s move to the next and final stage of configuration.
We just need to set goals yet – indicators that determine the effectiveness of our experimental changes.
Google Optimize has three types of standard goals:
- Session Duration. A session can last as long as the user’s continuous activity is monitored.
- Exit Pages. The number of visits during which the visitor viewed a particular page.
- Pageviews. The total number of pageviews by visitors, including repeated pageviews.
Choose your primary goal for your experiment.
You can also add additional goals. We recommend you first work with one primary goal to understand the principle of A/B testing.
Click Start in the upper right corner of the page and the project will be launched.
Reports in Google Optimize
You can view reports both at the end of the experiment and while it’s running. You can find them in the Reporting tab.
For example, in the report below, green values (62% to 31.022%) show that the proposed changes to the site will have a positive effect on conversions compared to the current version of the site.
Important: In order to make an informed decision, it’s best to wait until the end of the experiment or until one of the versions achieves 95% probability of superiority. Also, testing should be carried out with a sufficient amount of traffic. Therefore, it’s recommended to choose pages with a large number of visits.
If you still have questions about working with Google Optimize, we recommend you look at the recording of our webinar where we talk in detail about this service. Or you can ask questions in the comments below.
In this article, we showed you step by step how to set up your first A/B test using the free version of Google Optimize.
To get more features and run more complex experiments, we recommend to take a look at the paid version of this tool, Google Optimize 360.
If you want to learn more about the possibilities of Google Optimize 360, fill out the form below and one of our experts will contact you for a consultation.