Problem 1: What are the biggest disadvantages to using gross margin as your primary outcome metric? (Think about the variance of gross margin; statistical power; and the size of the effect you are likely to be able to detect with a fixed sample size of 10,000.)
Problem 2: Propose an alternative primary outcome metric that would still allow you to learn something directionally correct with respect to the treatment's impact on gross margin. Explain how your alternative metric would not suffer the same disadvantages as gross margin, and how that would help you as an experimenter.
Problem 3: How confident are you that your proposed metric would lead you to make directionally correct decisions with respect to gross margin? Are there any circumstances when it might not?