You changed how quickly support requests are handled, and you want to know whether that shift changed customer retention. You have pre and post periods and customer-level outcomes, but response times may also move with customer mix and seasonality.
How would you investigate whether a change in support response times affected retention?