Your question is Designing Causal Marketing Experiments. 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 design, execute, and read experiments like A/B tests and holdouts to prove causal impact?
Answer as if you are proposing the full experiment plan. Define the hypothesis, metrics, randomization unit, sample size, and analysis approach. Also call out the main pitfalls you would watch for, including peeking, sample ratio mismatch, novelty effects, and interference.