A streaming service's monthly churn has doubled in a year. How would you find out why?

First, a definition. Is churn here the share of paying subscribers who cancel in a month, and does it include people whose payment failed? Before I look for causes, I'd find out who is leaving. So I'd cut churn four ways. By signup cohort, to see whether recent joiners leave faster than older ones did at the same point. By plan, meaning ad-supported, standard and premium. By tenure, because leaving in month two and leaving in year three are different problems. And by acquisition channel, such as a free trial, a paid ad or a bundle with a phone plan. Then there are the causes I'd test. One is price. Was there an increase, and did cancellations jump after it? Two is content. Did a big show end or a sports contract lapse? Three is competition. Is there a new rival or a cheaper bundle? Four is involuntary churn. That's people who never chose to leave. Their card expired or the payment failed. Once I know which group is leaving, I'd match it to one of those causes. I'd start with monthly churn by cohort and channel. Can we get that?

First, a definition. Is churn here the share of paying subscribers who cancel in a month, and does it include people whose payment failed? I'll assume it does, and that churn went from 3% a month to 6%. Before I look for causes, I'd find out who is leaving. So I'd cut churn four ways. By signup cohort, to see whether recent joiners leave faster than older ones did at the same point. By plan, meaning ad-supported, standard and premium. By tenure, because leaving in month two and leaving in year three are different problems. And by acquisition channel, such as a free trial, a paid ad or a bundle with a phone plan. I'd also check the mix. If the service doubled its promotional signups last year, average churn can rise a lot while no group behaves any differently. Then there are the causes I'd test. One is price. Was there an increase, and did cancellations jump after it? Two is content. Did a big show end or a sports contract lapse? Three is competition. Is there a new rival or a cheaper bundle? Four is involuntary churn. That's people who never chose to leave. Their card expired or the payment failed. It gets missed, and it's the cheapest to fix, with payment retries and card reminders. Each cause leaves a different mark in those cuts. A price problem hits all tenures at once, when the new price lands. A content problem hits the viewers of one show. Failed payments show up as accounts that end with no cancel click. My hypothesis is that most of the rise sits with subscribers in their first three months who came in on promotions, with failed payments as the second piece. I'd start with monthly churn by cohort and channel, split into voluntary and involuntary. Can we get that?