Your question is Use CUPED to Boost Experiment Power. 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 an online product where a new experience is being tested in an A/B experiment. The team believes there is a small lift, but the metric is noisy and the test may be underpowered.
How would you use CUPED, or another variance-reduction technique, to improve the experiment? Explain how you would define the covariate, estimate the lift, and decide whether the variance reduction is valid.