You're designing an experiment for a product used by different SMB segments, such as sole proprietors, very small teams, and larger SMBs. The team expects the feature may help some segments more than others, and you need to decide how to structure the test.
How would you think about experiment design when the treatment effect may vary by SMB segment?