Calculating Subscription Churn Rate Properly

Most subscription businesses compute churn in a way that makes it move for reasons unrelated to retention.

The naive formula and its problem

Churn = subscribers lost in period ÷ subscribers at start of period

The problem is composition. New subscribers churn far faster than tenured ones. So this number is a weighted average whose weights are your recent acquisition volume.

Grow fast, and blended churn rises — because you added a lot of cycle-one subscribers, who are the most likely to leave. Slow down, and it falls. Neither movement tells you anything about whether your product got better or worse.

Churn by cycle number

The honest version. For each signup cohort, what proportion survive each cycle?

Cycle transition	Retained
Signup → cycle 2	86%
Cycle 2 → 3	92%
Cycle 3 → 4	95%
Cycle 4 → 5	97%

This shape is typical and it is the most useful diagnostic you have. It says: your problem is the first transition. Fix cycle-one retention and everything downstream improves.

It also tells you not to spend on retention campaigns aimed at cycle-six subscribers, who are already 97% likely to stay.

Voluntary versus involuntary

Split every churn event by cause:

• Voluntary — the customer cancelled
• Involuntary — a payment failed and was never recovered

They need opposite interventions, and involuntary is often 20–40% of the total. Reporting them together guarantees you will apply product fixes to a payments problem. Involuntary churn.

Customer churn versus revenue churn

Customer churn = subscribers lost ÷ subscribers
Revenue churn  = MRR lost ÷ MRR

They diverge in informative ways. If revenue churn is much lower than customer churn, you are losing small subscribers and keeping valuable ones — less alarming than the headcount suggests. If it is higher, your best subscribers are leaving, which is urgent.

What a good number looks like

For consumer ecommerce subscriptions:

• Monthly churn 8–12% — typical
• Under 5% — strong
• Over 15% — investigate urgently

Replenishment generally churns less than curation, because the value does not depend on each cycle’s selection being good. Why.

These are rules of thumb, not benchmarks to manage to. Your own cohort curve is worth more than any industry average.

Where to look when it is high

In rough order of frequency:

1. Cadence mismatch — delivering faster than consumption. The largest single cause in consumer subscriptions, and fixable with frequency options and easy skipping.
2. Involuntary churn — payment failures nobody recovered.
3. First-delivery disappointment — expectations set at checkout not met.
4. Hidden skip and pause options — people cancelling because they cannot find the alternative.
5. Price — genuinely, but less often than merchants assume, and discounting is the wrong first response.

How to reduce it.

Frequently asked questions

Should I measure churn by customer or by revenue?

Both. Customer churn tells you how many people are leaving; revenue churn tells you whether the ones leaving were valuable. A business losing many low-value subscribers and retaining high-value ones has healthy revenue churn and alarming customer churn.