Your question is Pitfalls in A/B Testing. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
What pitfalls can invalidate an A/B test, such as novelty effects, network interference, or sample ratio mismatch?