We are changing structure as well; if you independently look at A/B while also testing X/Y (of a different test), the results can be statistically misleading...

We are changing structure as well; if you independently look at A/B while also testing X/Y (of a different test), the results can be statistically misleading...in these cases we look at AX v. AY v. BX v. BY