relevant products:
  • Optimizely Classic

This article will help you:
  • Integrate AT Internet Analytics to read data from Optimizely
  • See reporting on Optimizely experiments in AT Internet Analytics

If you use AT Internet Web Analytics for analytics on your website, you may wish to view analytics details for the visitors who have seen a given variation in your Optimizely experiment. As with Optimizely's Google Analytics Integration, you can easily integrate Optimizely with AT Internet. 

Enabling the AT Internet Integration

Optimizely uses AT Internet's "MV Testing Tags" to tag your visitors with the experiments and variations to which they've been added. Once the integration is enabled, you will be able to access the Multivariate Testing report in your Digital Workspace and click on any test name to see the visitors to the test broken out by variation. 

In order to be able to use the MV Testing tags with AT Internet, you will need to install the MV Testing plugin

To enable AT Internet integration, go to the Integrations tab on the Home page. Select the AT Internet Integration and toggle the button to On in the right panel.

Optimizely needs 2 pieces of information from you in order to integrate successfully with AT Internet:

  • Account Number: Your AT Internet tag account number
  • Log Server URL: The log server URL used by your AT Internet tag

These boxes can be found in the Settings section of the right-hand panel in the Integrations tab. To learn how to find this information, see the section below on finding your Account Number and Log Server URL.

You can also check the box to enable this integration by default for all new experiments. 

To enable or disable the integration for an individual experiment, go to that experiment in the Editor and click Options > Integrations. Then check the AT Analytics box.



Make sure that your Optimizely experiment is running for this integration to work.

Finding your account ID and log server URL

In order to Integrate your Optimizely Experiments with AT Internet Analytics, you'll need to enter your Account ID and Log Server URL in the Optimizely Editor.

You can find these by looking at the source code of your page.

Here is an example of an AT Internet tag on a webpage. The values were looking for are found in the <noscript> <img> tag at the end of the tag.

The 'src' attribute of the <img> tag will start with the log server URL (everything before /hit.xiti). This is the Log Server URL.

The 'src' attribute of the <img> tag will also have a query parameter 's', which contains the account id we need.

The values in the screenshot below show these 2 values:

 
Note:

The log server URL is different than the xtsd variable at the beginning of the script. Make sure to use the URL at the beginning of the <img> tag. 

It's possible you will see two different Log Server URLs - one for HTTP (i.e. 'logc404.xiti') and one for HTTPS (i.e. 'logs1404.xiti.com') requests. Make sure to use the corresponding protocol with the log server URL.

The assumptions upon which the integration is built are that:

  • AT Internets client-side logic (xtcore.js) is hosted by the customer and is loaded asynchronously. 
  • Since there is no event that occurs when xtcore.js is loaded on the page. To give xtcore.js enough time to load (so that we can use the xt_mvt function that is defined on the window element within xtcore.js), we wait for window.onload before verifying if xt_mvt is available and using that function to send information to AT Internet. 
  • This function is a handler that is executed on window.onload (event listener is attached in the makeThirdPartyRequests function if there is data to be sent and window.xt_mvt does not already exist).
  • The code of the integration available here: https://github.com/optimizely/optimi.../integrator.js

Viewing Experiment Variations in Digital Workspace

Once you've enabled the Optimizely AT Internet integration and started your experiment, you can view this data in your Digital Workspace of your AT Internet account.

Once logged in, click Usability > Multivariate Testing on the left hand navigation menu at the bottom. You can then select Tests to see a list of all the experiments that have sent data to AT Internet and see their metrics broken down by variation names.