Hoping to get some additional help/insight into how Total Sales (Control Offer) is determined for a brand-new test. Do we need to run the test for a few days ahead of the full test to get an idea of what the control offer generates? Is there best practice guidance into how this is done? Any help here would be appreciated!
Hi @data_dude_dolan ,
Before launching a new test, the best practice would be to conduct an A/A test for the intended audience. This initial step will provide valuable insights into the total sales generated by the control offer. Additionally, an A/A test would help in identifying any flaws in the experiment configuration and will check for sample ratio mismatch (SRM) on the page (if any).
The total sales data obtained from the A/A test can also be used in the sample size calculations for your upcoming test.
Note: To accurately track total sales during an A/A test, ensure that the goal metrics are configured correctly as Revenue -> Total Sales in the Goals and Settings section.
Thanks and Regards,
Vaibhav Mathur
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.