Uniform Optimize supports Google Analytics as a tracking destination for all personalization and A/B test-related events. This allows to collect and analyze the performance of your personalization and A/B testing activities in the Google Analytics dashboard. This doesn't require the configuration of the Google Analytics integration, as the event collection is happening from the front-end using Uniform Tracker, as described here.
Where this Google Analytics integration capability becomes instrumental is when you want to surface the insights around personalization and A/B testing from Google Analytics inside the Uniform Optimize dashboard. This makes it easier for the marketers to track how different A/B testing variances are performing and understand how the intents, configured in the same dashboard, are performing.
This integration extends the Dashboard with two more tabs:
- Testing - surfacing the insights from your A/B tests
- Personalized Experiences - surfacing the insights from your personalization activities
In order to utilize this integration you will need to confirm the following prerequisites are met:
- You have at least one project created in Uniform dashboard.
- You have access to a Google Analytics account that is already configured for the website you are associating with the Uniform project and you are seeing some data in your GA dashboards (outside of Uniform).
- In order to see anything in the Testing dashboard, you need to have at least one Test component configured in your code and deployed, and at least one session in Google Analytics collected with the test active. For reference, refer to the Testing with React section if you have not done this already.
- In order to see anything in Personalized Experiences dashboard, you need to have at least one intent configured and have you application enabled for Uniform Tracking and Personalization (see the getting started guides for React and Vue).
Step 1: Google Analytics configuration
The following configuration is required on Google Analytics front before enabling the integration with Uniform.
Configure two custom dimensions in Google Analytics (see the doc here). The names do not matter but the scope must be set to "Session".
These two new custom dimensions will be used to power data for the Personalization dashboard (Strongest Intent Match Dimension and All Intent Matches Dimension) and must not be used for any other scenarios.
Step 2: Enable the Google Analytics integration
This guide will walk you through setting up Google Analytics integration for Uniform Optimize for an existing site.
Navigate to your project in the Uniform dashboard.
Click on the Settings tab and select Integrations from the sidebar.
At the bottom of the page listed under Add Integrations click on the plus icon next to the Google Analytics card:
Click Add within the popup:
Now click Access Google Analytics to authorize read only access to your GA data:
Google OAuth consent popup will be displayed. Please review it and accept to continue.
After the popup is closed, click Change to configure which Google property you want to link to this site configuration:
Select Account name, Property name, and View options, click OK to close:
The GA id property will update, click the Save button to apply this change:
In order to collect personalization events, Uniform needs to know which Google Analytics dimensions can be used to push this data. Specify the index of the Google Analytics custom dimension created earlier to be used for the Strongest Intent Match Dimension and All Intent Matches Dimension (these must correspond to the dimensions that you are not using currently for anything else):
If the save is successful, the success message will appear, which means that you have now configured the Google Analytics integration: ![success message](./assets/success message.png) and that the new links to both Testing and Personalized Experiences dashboards should now appear in the main tab area of your project and two new cards on the Dashboard tab:
Using Testing dashboard
This dashboard allows for both the configuration of new tests in Uniform Optimize and viewing of the A/B test analytics, as well as picking a winning A/B test variance.
Adding a new test
In order to add a new test, click the red plus-sign button.
In the Add a Test box, assign test name - ensure the test name matches the name of the test component in code. If the names do not match the test will not produce results needed and the test will not be visible on your website. Once name has been assigned, click Create.
Afterwards, you should be able to see the data for the given test pulled from Google Analytics. It shows how many variants are configured in a test and experiments have been run.
Viewing A/B test results
Select the title of the test to view a breakdown of the variant test results:
On this page you can see how different Variants are performing (according to the Goal Conversion Rate and Bounce Rate), see the total number of experiments running for the current test, and see the confidence score. The Confidence calculator reports which variant is performing best, when applicable.
You can also perform such operations as selection of a default variant and closing an active test and pick a winning Variant.
Selecting a default variant
In the Select Default Variant field select which variant you are declaring as the default. The other variants are handled as treatments to the selected default variant.
Once assigned, click Submit.
Closing test / Picking a winner
To close the test to declare the winner, click the red button Close Active Test:
In the Close Test field, select the winning variant and finalize the test by clicking Close Test.
The Test winner will be displayed on the dashboard:
After the winning variance is selected, you must Publish, which will trigger site re-deployment. This will update the Uniform manifest JSON and prevent the test from running after site re-deployment is successful. The winning variant will be shown to all visitors after this publish takes place, regardless of which variant they saw before. In order to trigger Publish, click the red Publish button in top right corner:
Using the "Personalizated Experiences" dashboard
The Personalized Experiences dashboard provides insights related to your personalization activities on a given project. You can see the breakdown of personalized vs. non-personalized sessions:
as well as see the breakdown of the most matched intents:
Personalized Experience analytics
You are generally running multiple personalized experiences on your website. In order to see how a particular experience is performing, click on its name in the table below the charts:
On this page you will find such key insights as the total number of experiments running for this experience, confidence level and the list of all variances associated with your intents:
Each Personalized experience associated with a given variance (shown in blue box) is measured against the control group that can be configured in
Project Settings > Measurement Group (green box), allowing you to see the impact of your personalization against a group that is behaving like a particular intent but is not seeing the effect of the personalization. You can quickly understand the key performance metrics of your personalized variance (such as Goal Conversion Rate, Bounce Rate) as well as the Page Views per session where a given personalization variance was active: