You are running a growth A/B test on a music creation product and the experiment reads out on several outcomes, including activation, paid conversion, and downstream engagement. The team is excited because at least one metric appears positive, but you are concerned that testing many metrics increases the chance of false positives.
How would you handle multiple comparisons in this experiment with several metrics? Explain how you would decide which metric drives the decision, how you would control false discoveries across the rest, and how that would affect interpretation of the results.