Your question is Designing Statistically Sound A/B Tests. 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).
How do you approach the design of A/B tests to ensure statistically significant results?