Your question is Measure Interference in Customer Change Test. 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).
You work on a customer-facing change where one customer’s experience can affect another’s behavior, so a standard A/B test may not isolate the impact cleanly. You’re concerned that the treatment could create spillovers and bias the observed lift.
How would you handle network effects or interference when measuring the impact of this change? What experiment design and analysis approach would you use so the result is still decision-ready?