Where it sits
Half of churn analysis is deciding which side of this line each departure belongs on.
Tell it apart
| Aspect | Churn | Network effect |
|---|---|---|
| Measures | Departures | Value of staying |
| Direction | Out | In |
| Observable | Directly | Inferred |
| Relationship | One tests the other | — |
Network effects raise the value of staying; churn measures how many leave anyway. A genuine network effect should show up as low VOLUNTARY churn — if it doesn't, the network effect is a story rather than a mechanism.
What it doesn't mean
Models distinguish VOLUNTARY churn, where the customer chose to leave for controllable reasons like billing or support, from INVOLUNTARY churn — relocation, death, circumstance. Most predictive models deliberately exclude the involuntary cases, because nothing you do would have changed them.
It converts directly into customer lifetime. A 25% annual churn rate equals 75% retention and an average customer lifetime of FOUR YEARS. That single conversion turns a percentage into a number you can plan around.
What it really means
The word is from a butter churn — a vessel in which cream is agitated continuously — and that origin is more precise than most business metaphors manage.
Churn does not mean leakage. It means turbulence: customers moving in and out of a base, continuously, in both directions. A company with 20% churn and 25% acquisition is growing and also losing a fifth of its customers every year, which are two very different facts about the same business.
The most useful thing you can do with a churn number is convert it, because a percentage is hard to think about and a duration is not.
25% annual churn = 75% retention = an average customer lifetime of four years.
That conversion changes decisions. We lose a quarter of customers annually sounds like a problem to monitor. Our average customer is with us for four years is a planning input — it tells you what you can afford to spend acquiring one, and how long you have to earn it back.
The distinction that separates useful churn analysis from panic is voluntary versus involuntary. Voluntary churn is the customer choosing to leave, for reasons that are frequently controllable: pricing, billing failures, support, a competitor. Involuntary churn is circumstance — relocation, death, a business closing. Most predictive models deliberately exclude the involuntary cases, because no intervention would have changed them, and including them makes the addressable problem look larger than it is.
For scale: industry benchmarking puts typical monthly mobile churn at 0.75–3% for postpay and 3–5% for prepay. Five percent a month compounds to roughly half the customer base per year.
And the standing asymmetry behind all of it: retaining an existing customer costs considerably less than acquiring a new one. Which is why churn is the metric that quietly determines whether growth is compounding or just replacing.
Where it comes from
the butter churn — a vessel in which cream is agitated continuously
The analogy is AGITATION rather than simple leakage. A churn describes continuous movement of customers in and out of a base — turbulence rather than simple leakage.
- 1990s onward
Churn becomes a core operating metric in contractual industries: telecoms, subscription media, software and utilities.
- 2018
STL Partners' telco benchmarking puts typical monthly mobile churn at 0.75–3% for postpay and 3–5% for prepay — the prepay end compounding to roughly a third to a half of the customer base in a year.
In real life
- A streaming service with 10,000 subscribers that loses 500 of them in a month has a monthly churn rate of 5%.
- A customer who cancels after a billing mix-up is voluntary churn; one who moves abroad, where the service doesn't operate, is involuntary.
- The staff sense: 'We've had so much churn on the support team this year that half the faces at the Christmas party were new.'
Questions
Q01Where does the word come from?
The butter churn — a vessel in which cream is agitated continuously. The analogy is turbulence and constant movement, not simple loss.
Q02What's the formula?
Customers lost divided by total customers in the period, times 100. The useful step is converting it: 25% annual churn implies an average customer lifetime of four years.
Q03What's voluntary versus involuntary churn?
Voluntary is the customer's choice, driven by controllable factors. Involuntary is circumstance — relocation, death — and is usually excluded from predictive models because it is not addressable.