How would you analyze the causal impact of a marketing campaign at Ahf when randomization is not possible?
