This article will help you:
  • Track conversions across multiple domains
  • Understand common issues with cross-domain traffic and workarounds

By default, Optimizely supports cross-domain tracking for any browsers/users that allow third party cookies. To implement this type of tracking, all you will need to do is add your Optimizely snippet to both domains and set up your goals in Optimizely.

 
Note:

This article deals with the implications of tracking across domains. If you're running an experiment across different subdomains, you shouldn't encounter these issues.

Issues with cross-domain tracking

Optimizely uses third party cookies to enable cross domain tracking. Although Optimizely supports cross-domain tracking, it can only track conversions on browsers that enable third party cookies. At the time of this writing, all versions of Safari and some versions of IE are blocking third party cookies by default. Mozilla has announced that it also intends to turn off third party cookies by default sometime in the future.

While this means that part of your data will not be reported, this should not impact the relative conversion rates between variations within a test because this impacts all of them equally. The results you receive should still be indicative of the difference in conversion rates between variations. The only impact on your experiment is that a percentage of overall visitors will be excluded from your experiment results.

When looking into more granular data other than the comparison of variation performance - such as total unique visitors, total unique conversions, comparing numbers against other analytic platforms, etc. - this is where cross-domain tracking will not align due to some browsers not supporting 3rd party cookies and the inability to reconcile data. 

Workarounds for cross-domain tracking issues

To avoid underreporting of data, there is an additional method that will work on all browsers and more reliably track data. By using iframes you can create a link between the two domains and pass information from one domain to another using only first party cookies.

In this model, the page view on the iframe can be used as a proxy conversion on the second domain. What this means is that you can set up a goal in your experiment that checks to see how many visitors came to www.domain1.com/conversion.html. That number should correspond to how many times the iframe was loaded on www.domain2.com, which should only happen when a person reaches the conversion page that you want to track. The key here is to use a page that is unique and is only used for the iframe to make sure that no additional visitors are included in the results.

Additionally, if you trigger the iframe to only load on the page when a certain event happens, you can track clicks, and even revenue across domains.