Customers visiting the site are randomly assigned to a particular variant unless they clear temporary data (i.e. cookies, history, cache, etc.), or the test is finished.
The second shop can run independently when the A/B test is being conducted. It has no influence on the test's results.
The statistics, i.e. Google Analytics, will always be generated for the shop conducting the test, regardless of the variant assigned. Only accessing the webpage of the second shop will generate the statistics for the shop B.
When the A/B test is being conducted, a customer assigned to the variant B can see the layout of the shop which was chosen as the variant B. Depending on the shared settings also different composition of banners, buttons and adverts can be displayed. Therefore, the variant B can be displayed in the same way as the second shop.
A/B tests perfectly check the performance of the below listed items from the administration panel:
Under ADMINISTRATION \ Installation, deinstallation and shop configuration you can manage which of the elements are to be the same in both variants and which should be set up individually. In order to proceed please use the configuration of settings sharing.