What is the difference between churn rate and retention rate?

They are complementary over the same period and the same cohort: a churn rate of 8% corresponds to a retention rate of 92%. The first counts the customers lost, the second those who stay.

How do you convert monthly churn into annual churn?

Compound the retention: annual churn = 1 − (1 − monthly churn) to the power of 12. A monthly churn of 3% gives about 31% a year, not 36%.

What is a good churn rate?

There is no universal threshold: it depends on the sector, the contract model, the commitment length and customer tenure. The right benchmark is the trend of your own cohorts and the margin lost per point of churn.

What is the difference between customer churn and revenue churn?

Customer churn counts departures; revenue churn measures the revenue lost, including reductions in spend. Increases in spend by remaining customers can make it negative in net terms.

Churn rate: calculating customer retention and what it is really worth.

SUPPORTED DECISIONManage retention by cohort and by tenure, and put a margin value on each point of churn before funding a retention action.

Key distinctions

Three conditions before calculating.

A cohort fixed at the start

The rate is calculated on the customers present at the start of the period. Mixing customers acquired along the way into the denominator artificially lowers the churn of a company that recruits fast.

Rates compound, they do not add up

A monthly retention of 97% does not give 64% over a year but 69%, because each month applies to the remaining customers. Two rates can only be compared over the same period length.

The average hides heterogeneity

The most fragile customers leave first, so a cohort's retention rises with tenure. Projecting a constant rate underestimates the lifetime of the customers who stay.

Method

Measuring and managing churn in four steps

01Define leaving
With a subscription, cancellation is observed. Without a contract, leaving is inferred from inactivity beyond a threshold, set according to the customer's usual purchase cycle.
02Build cohorts
Group customers by month or year of acquisition, then track the share still active at each tenure, rather than a global rate mixing all generations.
03Separate customers and revenue
Calculate churn in number of customers and in revenue, because increases in spend by remaining customers can offset, or mask, departures.
04Value, then test
Translate a point of retention into margin, target at-risk customers with a model validated out of sample, then measure the action against a control group.

ORIGINAL ASSET

A cohort of 1,000 customers tracked over five years

The same cohort, year after year. The churn rate falls with tenure because the most fragile customers left first.

Tenure	Customers at start of year	Lost during the year	Churn rate	Retention rate
Year 1	1,000	400	40%	60%
Year 2	600	150	25%	75%
Year 3	450	81	18%	82%
Year 4	369	52	14%	86%
Year 5	317	38	12%	88%

Same cohort, churn divided by more than three in five years.

A single average rate applied to every customer underestimates the value of loyal customers.

ILLUSTRATIVE EXAMPLE · SIMULATED DATA

An eight-to-one return on paper, to be confirmed by a test

01 · SITUATION

Illustrative example: a subscription service has 20,000 customers who each bring €30 of margin a month. The monthly churn rate is 3%. A retention action costing €15,000 a month would bring it down to 2.5%.

02 · CALCULATION

Value of a customer = €30 ÷ 3% = €1,000, against €30 ÷ 2.5% = €1,200 after the action. Customers retained each month = 20,000 × 0.5% = 100. Value created per month = 100 × €1,200 = €120,000, for €15,000 of cost.

03 · DECISION

On paper, the action returns eight times its cost. This calculation assumes a constant rate, no discounting and a causal effect of 0.5 point: the customers retained are often the most fragile, hence a test against a control group before generalising.

Acceptance conditions

What must be true to act.

01
Calculate churn on the customers present at the start of the period, without mixing in new ones.
02
Compare rates over the same period length, compounding them rather than multiplying them.
03
Track retention by cohort and by tenure, not only on average.
04
Fund a retention action on its effect measured against a control group, not on assumed churn avoided.

Limits

What this analysis does not prove.

Without a contract, leaving is not observed: the inactivity threshold chosen changes the rate.
The formula value = margin ÷ churn assumes a constant rate and ignores discounting.
A predictive model identifies at-risk customers, not those the action will make stay.
Churn in number of customers can fall while revenue churn rises, and vice versa.

Frequently asked questions

Short answers, with their limits.

What is the difference between churn rate and retention rate?: They are complementary over the same period and the same cohort: a churn rate of 8% corresponds to a retention rate of 92%. The first counts the customers lost, the second those who stay.
How do you convert monthly churn into annual churn?: Compound the retention: annual churn = 1 − (1 − monthly churn) to the power of 12. A monthly churn of 3% gives about 31% a year, not 36%.
What is a good churn rate?: There is no universal threshold: it depends on the sector, the contract model, the commitment length and customer tenure. The right benchmark is the trend of your own cohorts and the margin lost per point of churn.
What is the difference between customer churn and revenue churn?: Customer churn counts departures; revenue churn measures the revenue lost, including reductions in spend. Increases in spend by remaining customers can make it negative in net terms.

References