The experiment attribute that is added to all events contains the experiment id and experiment variant concatinated by a dot – for the very first experiment on the page.

Why is it useful?

Multiple experiments may be present on one single page and unfortunately this attribute will not cater for those cases. But many websites do not run ( and should not run – due to insufficient traffic amount ) more than 1 experiment per page. Since experiment data is addded to all events it becomes really easy to segment reports by users who have seen one or the other version. You can usee this parameter to see differences in events performed.

Best bit is that regardless the A/B tresting framework that you use this attribute is called experiment, so even if you make changes to your tech-stack, with a couple simple moves you are able to view long-time usage behavior trends within one page, and how they were affected by the experiments that you were running.