What sample size would you need to detect a meaningful change in conversion rate for a pricing experiment at Seatgeek?
