Well, shoot. This guy never heard of bucket testing? Or controls? You have three changes to test. So you split your users into four groups: unchanged, A, B, and

Well, shoot. This guy never heard of bucket testing? Or controls? You have three changes to test. So you split your users into four groups: unchanged, A, B, and C, where each group sees one changed feature. Test & compare.