Subscription Churn: Count Who Didn't Pay, Not Who Cancelled

Subscription churn is the share of subscribers who stop paying you. Most subscription apps count only cancels. Count every subscriber who missed a charge they were due. Then split them by reason: cancelled, card failed or paused.

That one change shows you churn your app report leaves out.

What is subscription churn?

Churn is when a customer stops buying from you. In a subscription, the customer has a charge coming on a set date. Subscription churn is when that charge stops happening.

There are 2 kinds. Voluntary churn is a choice. The subscriber clicks cancel. Involuntary churn is a payment problem. A card expires or the bank declines it, and the charge fails.

Most guides stop there. They skip the third case: the pause.

How do subscription apps calculate churn?

Each app counts a different thing. Here is what their own help centers say.

App What it counts Pauses Failed cards Recharge Customers. Churned only when "they have no active subscriptions remaining" Not addressed Counted, once the app marks the subscription inactive Loop Cancellations ÷ active at the start of the month plus new sign-ups "A paused subscriber is an active subscriber" in cohort reports Not stated in the formula Appstle Subscriptions. Cancels ÷ active at the start Not mentioned Not stated Skio Gives 4 states: active, failed, paused, cancelled Paused means "no orders processing until resumed" "Failed" sits in retries. It becomes cancelled after all retries fail

So the same store can show 3 churn rates. Recharge counts people. Appstle counts subscriptions. Loop adds new sign-ups to the bottom of the division, which makes the rate smaller.

None of them says a paused subscriber counts as churn. Loop even files pauses under "Deactivated subscribers" next to churned ones. But its churn formula counts cancels only.

Does a paused subscription count as churn?

Here is our rule. Count a pause as churn until the next charge clears.

A paused subscriber pays you nothing. Neither does a subscriber whose card is stuck in retries. A churn rate built on cancels counts neither one. Your churn rate looks fine while your charges drop.

Pause is still a good tool. Darkroom's guide to subscription churn makes the case for it. In its words, a pause keeps "the payment method, the account, and the habit." It is a save when the subscriber comes back and pays. Until then, you do not know.

The same guide says to judge a save at 90 days. In its words, "a save that cancels next month was a deferral, not a save."

How do you calculate subscription churn without fooling yourself?

Start from charges, not from statuses.

1. List every subscriber who had a charge due this month.
2. Mark who paid it.
3. Everyone else is churn for this month.
4. Split them by reason: cancelled, card failed and not recovered, or paused.

Churn = subscribers who did not pay a charge due ÷ subscribers with a charge due.

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

1,000 subscribers had a charge due in March. 40 cancelled. 30 cards failed and never recovered. 60 paused. Your app shows 40 ÷ 1,000 = 4%. By our rule, 130 did not pay. That is 13%.

Next month, check the 60 pauses. The ones who paid move back to active. The rest stay counted.

Skips need a call too. A skip moves 1 charge. It does not end the subscription. Leave skipped charges out of "charges due" for that month. Watch a subscriber who skips month after month. That is a pause in all but name.

Reading this by cohort is a bigger topic. A cohort is a group of subscribers who started in the same month. For now, get the count right.

What does each reason tell you to fix?

The split matters more than the total. Each reason has a different fix.

Reason What it means What to fix Cancelled The subscriber decided to leave Your cancel flow. Offer skip, swap or a slower schedule first Card failed They did not decide anything Your dunning. Dunning means the retries and emails that follow a failed charge Paused They are not sure The resume. Remind them before the pause ends, and show what is coming

A store with high card failures does not need a better cancel offer. A store with high cancels does not need more retries. One number for both hides which problem you have.

How do you reduce subscriber churn before it starts?

Look at who you put on a subscription in the first place.

A buyer who subscribes on day 1 has never used the product. A buyer who has finished one 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. It runs only on products with a subscription option. Then it stops asking.

If you sell both one-off orders and subscriptions, keep the numbers apart. Our article on repeat customer rate vs retention rate shows how. For one-off buyers, churn works on a different clock. You set a churn line for each customer, as our customer churn analysis shows. Our churn prevention article puts the moves before that line in order.

For how every piece fits together, start at the ecommerce customer retention hub.

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.