System Design · Amazon · Hard
A medication-subscription reminder feature was rolled out to different groups at different times — the schedule varied by market and by delivery channel (push notification, email, or SMS). Assignment to the reminder was never randomized, so the analysis has to rest on observational data. The outcomes of interest are perceived user experience, measured as a CSAT score from 1 to 5, and whether the user is still subscribed four weeks later. It is plausible that one person's…
Checking your access…