How would you design an experiment to measure whether a new product change improved a key business metric?
