Creating Tests with Google Optimize

4 min read

Table of Contents

Step 1. Select Your Container
Step 2. Create Your Experience
Step 3. Add a Variant and Build Your Test
Step 4. Set Targeting Rules
Step 5. Set Experiment Goals
Step 6. Quality Assurance

Step 1. Select Your Container #

When you log into Optimize, you will see a list of “containers” for each website that we test on. Find your desired site and click into the container.

Step 2. Create Your Experience #

Look for the blue “Create experience” button in the top right corner of the container’s main page. When you click this, you will be prompted to name your test and editor page. Choose a short, descriptive name for your experience (i.e. “Change Form Headline on Sell Timeshare Page”) and enter the full URL of the page you’re targeting in the editor page field.

Next, select an experience type. (for A/B split tests, select “A/B test.” We will rarely use the other options here).

Hit “Create.”

Step 3. Add a Variant and Build Your Test #

Once you create an experiment, you will be taken to the experiment page. The status of your experiment will be automatically set to Draft mode. Select the “Add variant” button to get started (in most cases, you can keep the default name of “Variant 1” for your variant name).

You will then see two options under “Variants”: Original and Variant. Select “Edit” next to the Variant row to be taken to the visual editor.

Use the visual editor to add your changes. Visit Google’s resource on the visual editor for a more detailed walkthrough on features and tools: https://support.google.com/optimize/answer/6211957?hl=en

Step 4. Set Targeting Rules #

After you have finished designing your experiment, it is time to set the targeting rules. By default, the experiment will target only the editor page you provided in Step 1 (seen in the section titled “Page targeting”). If your test targets multiple pages, you can change this rule by hitting the pencil icon and using the predefined logic parameters to target the pages you want.

For example, if you wanted to run a test on the Sell Timeshare page and the What’s My Timeshare Worth page, you could apply the following logic:

Path > contains > /sell-timeshare/ OR /what-is-my-timeshare-worth/

In “Audience targeting,” you can set rules to expose your experiment only to users who meet specific circumstances. Most often, we will use the “Device Category” feature to separate out mobile, desktop, and tablet. However, in more advanced experiments we can target by custom GA audiences, UTM parameter, country/state/city, and even specific paid search campaigns.

If your test targets all users, you can leave this section blank.

Step 5. Set Experiment Goals #

Every experiment needs a target goal. You can identify this goal under the “Measurement and Objectives” section. The Google Analytics information should already be set, so skip to “Objectives.”

Select Add objective > choose from list.

A list of all goal types currently being tracked on the website will appear from the right. Choose your desired goal type based on the type of test and editor page. For example, if you are testing on the What’s My Timeshare Worth page, your primary objective should be “Market Value Survey.”

You can also set additional objectives. Sometimes, it is good to track other goal types to see how your test affects conversions in other areas. This won’t always apply, so use your best judgement on whether to select additional objectives.

Choose from a number of predefined goals.

Step 6. Quality Assurance #

Your test is now ready for the QA phase. Take a moment to preview your test to make sure everything appears as it should. You can do this by entering “Preview” mode (found next to the Variant you created in Step 3).

Be sure to do the following when QAing your experiment:

Check your experiment in desktop, mobile, and tablet views to ensure cross-device compatibility (this does not apply if you are only testing on one device).
Use the “share preview” option and copy the link provided to check your experiment in Incognito mode. Sometimes, cookies may be stored on your browser that affect experiment changes. Using Incognito mode will give you the experience that a first-time user will have.
Ensure that no other elements on the page are being altered accidentally by your changes. Occasionally there can be errors with CSS that end up impacting other things on the page.