Monthly Churn vs Annual Churn: Two Numbers, Two Decisions

Monthly churn judges a fix. Annual churn plans your sign-ups. Why times 12 and compounding are both right, and which number goes with which decision.

Michal Hlavacek, founder of WDNN · LinkedIn
10 years building retention systems for US ecommerce brands.
Monthly churn is the share of subscribers active on day 1 who stop paying that month. Annual churn is the share gone after 12 months. Use monthly churn to judge a fix. Use annual churn to plan sign-ups.
Monthly churn is the share of subscribers active on day 1 who stop paying that month. Annual churn is the share gone after 12 months. Use monthly churn to judge a fix. Use annual churn to plan sign-ups.
Most guides give you both numbers and one rule: never multiply by 12. That rule is half right. Times 12 and compounding answer 2 different questions, and you need both.
What is monthly churn?
Monthly churn = subscribers lost in the month ÷ subscribers active on day 1 of the month × 100.
Stripe's guide uses the same formula. A subscriber is lost when they miss a charge they were due. Our subscription churn article sets out that rule. Our subscription churn rate article covers why the base is day 1 and why new sign-ups stay out.
This article is about what you do with the number once you have it.
Should you multiply monthly churn by 12?
The sources do not agree. Here is what their own pages say.
Source What it says Churnkey Annual churn is "about 12 times the monthly rate" when monthly churn is 1% to 3% eightx "The trap is multiplying by 12. The right math is compounding." Prospeo "Don't multiply by 12." Chargebee docs "It is incorrect to compare reports with Monthly and Yearly granularity."Here is our take. Both camps are right. They answer different questions.
Compounding answers this: of the subscribers you had on 1 January, how many are gone by 31 December? Each month you lose a share of the ones left, so the losses shrink.
Times 12 answers this: how many new subscribers do you need this year to stay the same size? If you refill the base each month, each month starts at the same size. So each month loses the same number.
Which number tells you how many subscribers to add?
Times 12. Here is a made-up example to show the math. It is not real data.
You have 1,000 subscribers. Monthly churn is 5%.
Question Math Answer How many of today's 1,000 are gone in a year? 1 minus 0.95 to the 12th power About 460 (46%) How many sign-ups do you need to stay at 1,000? 50 lost a month × 12 600 How many sign-ups do you need to grow to 1,200? 600 to stay flat, plus 200, plus churn on the new ones More than 800So a store that plans with the compound number plans for 460 sign-ups. It needs 600 to stand still. It ends the year smaller and does not know why.
Use the times 12 number for the acquisition plan. Use it for the ad budget and for the sign-up target you give your team.
Use the compound number for a different job. It tells you how long one group of subscribers lasts. That is the number for a cohort view, which our subscription churn rate article walks through.
Which number tells you if a fix worked?
Monthly churn. Annual churn moves too slowly. A change you make in March is 1 month out of 12 in a yearly number, so it hardly moves it.
Here is how we read it.
- Write down your monthly churn for the last 3 months, counted the same way each time.
- Make 1 change. A new dunning email, a new skip option, a change to the first box.
- Read monthly churn for the next 3 months.
- Compare it to the 3 months before. Not to another store.
Expect a lag of about 1 month. A subscriber who cancels in March with a charge due in April counts as lost in April. So a fix you ship in March shows up in April's number first.
Keep the base the same each month, or the number moves for reasons that have nothing to do with your fix. Comparing your rate to other stores is a separate trap. Our average churn rate for subscription services article covers it.
What about subscribers on annual plans?
Leave them out of monthly churn. Track them by renewal.
A subscriber who prepays for a year has no charge due for 11 months. By our rule they cannot be lost until month 12. So their monthly churn reads close to 0 for 11 months. Then it jumps in the renewal month.
eightx puts it plainly: "you have to plan for the cliff." A blended monthly number hides it. Track annual-plan subscribers as their own group, and read one number for them: the share who renew.
The same goes for app reports. Chargebee's docs say a yearly report can show less churn than the 12 monthly ones added up. The reason is people who come back. In Chargebee's words, "Reactivations offset previous churns within the same year." So do not expect a yearly report to match 12 monthly ones.
Which number goes with which decision?
Decision Number How to read it Did this month's fix work? Monthly churn Against your own last 3 months How many sign-ups do we need this year? Monthly churn × 12 × base On the base you want to keep How long does a new subscriber last? Compounded churn, by cohort Follow one start month forward Is the annual plan working? Renewal rate Annual-plan subscribers onlyFix monthly churn first, because every other number on this table moves with it. Our churn prevention article lists the moves to try before you discount. For how every piece fits together, start at the ecommerce customer retention hub.
Get the whole system
Churn is the number. Flows are what move it. The guide has all 12, with every send time written out.
12 flows, 37 emails, one job: repeat purchases. Every trigger, every send delay, every exit condition, written out so you can build them this month.
