Subscription Churn Rate: Divide by Day 1, Read by Charge

Subscription churn rate is the share of subscribers who stop paying in a set period. Divide subscribers lost this month by subscribers active on day 1. Then read it 3 ways: monthly, annual and by cohort.

Most churn numbers mix those up. Check what they divide by and how long the period is before you trust one.

How do you calculate subscription churn rate?

Use this formula.

Subscription churn rate = subscribers lost in the month ÷ subscribers active on day 1 of the month × 100.

Two rules make it work.

1. Leave out anyone who signed up during the month. They were not there at the start, so they are not in the base.
2. Count "lost" the way our subscription churn article sets out. A subscriber is lost when they miss a charge they were due. That covers cancels, cards that failed and never recovered, and pauses.

The second rule also settles timing. A subscriber who cancels in March with a charge due in April is lost in April. Ordergroove's guide makes the same call. In its words, "the churn doesn't go into effect until they're up for renewal the following month."

Here is a made-up example to show the math. It is not real data.

You start March with 2,000 active subscribers. 150 of them miss a charge. 400 new subscribers join during March. Your churn rate is 150 ÷ 2,000 = 7.5%. The 400 new sign-ups stay out of it.

Why does your app show a different number?

Because each source divides by something different. Here is what their own pages say.

Source What goes on the bottom New sign-ups Bold "total subscribers at the start of the period" Left out Ordergroove Subscribers from the period before Left out Loop Active before the month plus created in the month Added to the bottom Recharge "The sum of all daily total active customers" in the period Counted on each day they are active Stripe 2 formulas on one page: all subscriptions, or subscriptions at the start Depends on which you pick

Adding new sign-ups to the bottom makes the rate smaller. So a store that grows fast looks like it churns less. It can lose the same people and still show a better number.

Then there is the date range. Recharge's rate follows your date filter. Its help page says a range of October to December gives "a quarterly churn rate." Pick a quarter and you get a number close to 3 times a monthly one. Pick a week and it shrinks.

So before you compare 2 churn numbers, ask 2 things. What is on the bottom? How long is the period?

How do you turn monthly churn into annual churn?

Do not multiply by 12. Churn compounds. Each month you lose a share of the subscribers who are left, not of the ones you had in January.

Annual churn = 1 minus (1 minus monthly churn) to the 12th power.

Bold's glossary shows the gap. At 4% a month, times 12 gives 48%. The compound math gives about 39%. Here is our math on a few more rates.

Monthly churn Times 12 (wrong) Compounded (right) 3% 36% about 31% 5% 60% about 46% 7.5% 90% about 61% 10% 120% about 72%

Times 12 says a 10% store loses everyone and then some. It does not. About 28% of its subscribers are still paying after a year.

What does a cohort view of churn show you?

A cohort is a group of subscribers who started in the same month. A monthly rate blends every cohort into one number. New subscribers sit next to ones who have paid for 2 years. Recurly notes that "new subscribers tend to churn at a higher rate." So the blend hides where people leave.

None of the pages above gives a cohort formula. Flowium's guide comes closest. It says to "build cohort curves (churn by cycle number) rather than one blended rate."

Here is how we run it. Line cohorts up by charge number, not by calendar month.

1. Group subscribers by the month of their first charge.
2. For each cohort, count who paid charge 2, charge 3 and charge 4.
3. Churn at each charge = subscribers who missed it ÷ subscribers who paid the one before.
4. Read down each column. The column that loses the most is your problem.

Why charge number? A monthly subscriber reaches charge 2 in month 2. An every-other-month subscriber reaches it in month 3. Calendar months mix them. Charge numbers line them up. Recharge's help page makes a related point. It "recommends accounting for the subscription frequency when calculating your churn rate."

Here is a made-up example. It is not real data.

Cohort Started Paid charge 2 Paid charge 3 Paid charge 4 January 500 400 (20% lost) 360 (10% lost) 342 (5% lost) February 600 450 (25% lost) 405 (10% lost) 385 (5% lost)

Both cohorts lose the most at charge 2. So the fix sits between the first charge and the second: the first box, the first delivery, the first bill. A blended monthly rate would not tell you that.

Which view should you use?

All 3. Each one answers a different question.

View Use it to Watch out for Monthly See if this month got better or worse Keep the base as subscribers active on day 1 Annual Plan the year Compound it. Never times 12 Cohort by charge Find the charge where people leave You need a few months of cohorts first

Comparing your rate to other stores is a separate trap. Our average churn rate for subscription services article covers which numbers to trust. If you sell one-off orders too, our customer churn analysis article shows the grid for those buyers. For how every piece fits together, start at the ecommerce customer retention hub.

Fix charge 2 before the first charge

Look at who you put on a subscription. A buyer who has finished 1 order knows if they want more. Our guide's subscription conversion flow waits for that. It sends 3 emails at 21, 28 and 35 days after a first order, only for products with a subscription option.

Get the whole system

The subscription conversion flow is 1 of the 12 in the guide, 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.