What is the proper way to measure significance in an A/B test, when a visitor can potentially "convert" more than once?
I understand a chi-square test or z-test are most often used for A/B testing, but from what I understand that is only if it is categorical, like if a visitor purchased or not.
In my case, customers can purchase multiple times in a test period, and I want to make sure I'm using the right significance test for this.
For example:
test A: 800,000 visitors, 50,000 purchases
test B: 45,000 visitors, 2,700 purchases
I saw a similar question here, but it was asking about revenue: AB Testing other factors besides conversion rate. In my case, should I be using a Wilcoxon test also?
