How to Calculate Churn Rate: Formula and Retention Rate

To calculate churn rate, divide the customers you lost during a period by the customers you had at the start of that period. The formula is simple and gets misapplied all the time: new customers in the denominator, monthly churn multiplied by 12, one churn rate for every source. This guide is for dating advertisers who want to keep subscribers and paying users, and for publishers whose RevShare depends on how long their users stay. You get the churn rate formula and the retention rate formula, a worked dating example, the conversion from monthly to annual churn, and how churn connects to LTV and traffic quality.

What is churn rate? Meaning and formula

Churn rate is the share of customers you lose in a given period. In dating, these are usually paying users: subscribers who do not renew, or credit pack buyers who stop topping up.

Churn rate = customers lost during the period ÷ customers at the start of the period × 100

Settle three things before you calculate:

• Period: usually a calendar month. Only compare periods of the same length.
• Customer definition: paying subscribers, all payers or registered users. Each group gives a different churn rate.
• Loss event: Does a subscriber count as lost on cancellation or only when the paid term ends? For revenue planning, the end of the paid term usually works better; what matters is applying one rule consistently.

Besides customer churn (user churn), there is revenue churn, the share of recurring revenue you lose:

Revenue churn = recurring revenue lost during the period ÷ recurring revenue at the start × 100

Lost revenue includes cancellations and downgrades. The two figures can diverge: if mostly users on the cheapest plan cancel while premium subscribers stay, revenue churn is lower than user churn. This guide covers user churn; the same rules apply to revenue churn.

Monthly churn rate: a worked dating example

A dating platform with monthly subscriptions starts March with 2,000 paying subscribers (hypothetical numbers). During the month, 500 new subscribers join and 300 subscribers from the starting base do not renew.

Churn rate = 300 ÷ 2,000 × 100 = 15%

The 500 new subscribers stay out of the denominator. They were not at risk at the start of the month, because their first paid month usually ends in April. Count them anyway and you get 300 ÷ 2,500 = 12%, understating churn by a fifth. The error grows with acquisition volume: the more traffic you buy, the better a miscalculated churn rate looks.

Growth also hides churn: the base grew by 10% even though 15% of existing subscribers left. New subscribers who leave within the same month, for example after a refund, belong in their own cohort, not in the churn of the starting base.

Monthly vs annual churn rate

A 15% monthly churn rate does not mean 180% a year. Churn compounds in the wrong direction: each month only hits the customers who survived the month before.

Annual retention = (1 − monthly churn rate)¹²

Annual churn rate = 1 − (1 − monthly churn rate)¹²

At 15%, annual retention is 0.85¹² ≈ 14.2%, so annual churn is about 85.8%. More values:

This conversion assumes constant monthly churn. Dating rarely delivers that, because new payers leave faster than long-standing ones. For forecasts, cohorts work better.

The retention rate formula: period and cohort retention

Retention rate is the flip side of churn rate. For a single period:

Retention rate = 1 − churn rate

In the example, that is 1 − 0.15 = 85%: 1,700 of the 2,000 subscribers from the start of March renewed. Period retention describes your whole base but mixes old and new customers. Cohort retention tells you more. A cohort groups all users who arrived in the same period from the same source:

Cohort retention (checkpoint X) = active users of the cohort at checkpoint X ÷ users in the cohort at the start × 100

What counts as active depends on the cohort:

• Registrations: logged in on day 1, day 7 and day 30 after sign-up. These early checkpoints show whether users engage with the product at all.
• Payers: still paying in month 1, 2 and 3 after the first payment, through a renewed subscription or a repeat purchase.

A hypothetical example: 1,000 users pay for the first time in March.

The cohort loses 40% in the first month; after that, monthly churn drops sharply. The curve also explains why base churn can rise after a strong acquisition month without any change to the product: fresh payers leave more often. Look only at period churn and you can mistake that mix effect for a product problem.

Customer churn in dating: subscriptions, credits and success

Dating has quirks that affect every churn calculation:

• Subscriptions and credit packs. Subscriptions have a clear loss event: the missed renewal. Credit buyers never cancel; they simply stop buying. Define their churn through an inactivity window, such as “no purchase in 60 days”, keep that window fixed, and report both groups separately.
• High early churn. Many payers try one month or a starter pack and leave, as the cohort table shows.
• Successful users leave, too. Someone who found a partner cancels because the product worked. You cannot and should not prevent this churn. Ask for a cancellation reason so you can separate it from dissatisfaction.
• Voluntary and involuntary churn. Voluntary churn is a user cancelling. Involuntary churn is losing a user to a failed payment: an expired card, insufficient funds or a declined charge. Report them separately, because the fixes have nothing in common.
• Refunds and chargebacks. A refunded payment is a loss even if the subscription is still formally active. Decide whether refunds count as churn and stick to it.

Churn rate, LTV and ARPPU: how they connect

Churn determines how long a payer stays, and therefore what that payer is worth. With constant churn, average lifespan in months equals 1 ÷ monthly churn rate. That gives the simplified LTV formula per payer:

LTV ≈ ARPPU × gross margin ÷ monthly churn rate

ARPPU is average monthly revenue per paying user. ARPU, by contrast, spreads revenue across all users, including those who never pay; our ARPU guide explains the difference. With $30 ARPPU, a 70% gross margin and 25% monthly churn (hypothetical numbers), LTV is $30 × 0.70 ÷ 0.25 = $84 over a lifespan of 4 months. Cut churn to 20% and LTV rises to $105, a quarter more from the same monthly revenue.

This is a simplified steady-state formula: it assumes constant churn and constant ARPPU. With a curve like the one in the cohort table, it misleads. Plug in the 15% from month 3 and you get $30 × 0.70 ÷ 0.15 = $140, even though the cohort lost 40% in its first month. For bids and payouts, cohort LTV is more reliable; our customer lifetime value formula guide walks through the calculation. To value a registration, multiply LTV per payer by the payer rate.

Traffic sources and churn: judge sources by cohort retention

How quickly users leave also depends on how they arrived. Users who were rewarded for signing up rarely pay and rarely stay. The same goes for users whose prelander promised something the offer does not deliver, or who were sent to the wrong vertical. Such sources can look fine on registrations and even on first payments and still lose money.

A hypothetical example: two sources each deliver 200 first-time payers at the same price. After three months, 90 users from source A are still paying (45%), but only 50 from source B (25%). At the first payment, both looked identical.

Judge sources by cohort retention and cohort ROAS, not by registrations or first payments alone; our ROAS formula guide covers choosing the time window. When retention is suspiciously low, check for fake or incentivized sign-ups (see click fraud in dating traffic) and for prelanders that set the wrong expectations (see dating prelanders).

For publishers on RevShare, churn is income: their share flows only as long as the users they referred keep paying. The shorter those users stay, the less RevShare earns compared with a fixed CPA. See the break-even math in RevShare vs CPA.

How to reduce churn

Four general levers target the usual causes:

Onboarding. Early churn happens when users see no value in their first days. A short path to a complete profile and first relevant contacts addresses that.

Matching quality. Relevant suggestions within reach and consistent moderation against fake profiles give users a reason to stay.

Payment retries and dunning. Against involuntary churn, use automatic retries after failed payments, reminders before a stored card expires, and alternative payment methods. This lever targets users who wanted to stay.

Plan design. Longer terms lower the monthly churn rate but move churn to the end of the term. Compare plans by cohort retention, not by the monthly rate. Transparent renewal terms belong here, too: users who feel caught off guard by a charge may not just cancel but also file a chargeback.

Common churn rate calculation mistakes

1. New customers in the denominator. In the example, this pushes churn from 15% down to 12%.
2. Multiplying monthly churn by 12. 15% a month is not 180% a year but about 85.8%.
3. An unclear loss event. Cancellation date, end of term and inactivity windows give different results. Pick one definition and document it.
4. One churn rate for everything. Split by product type, plan, source and GEO; an average hides weak sources.
5. Mixing voluntary and involuntary churn. You end up fixing the product when the problem is payments.
6. Applying the steady-state formula to new payers. Using the churn rate of long-standing subscribers overstates the LTV of new cohorts.

FAQ

How do you calculate churn rate?

Divide the customers lost during a period by the customers at the start of that period and multiply by 100. Lose 300 of 2,000 subscribers and your churn rate is 15%. New customers from that period stay out of the denominator.

What is the retention rate formula?

For a single period, retention rate = 1 − churn rate, so 15% churn means 85% retention. Cohort retention tracks a fixed group over several checkpoints, such as day 1, 7 and 30 or month 1, 2 and 3.

How do you convert monthly churn to annual churn?

Work through retention: annual churn rate = 1 − (1 − monthly churn rate)¹². A 5% monthly churn rate becomes about 46% a year, not 60%.

Conclusion

Churn rate is quick to calculate but only meaningful with clear rules: the starting base in the denominator, one fixed loss event, and separate figures for subscriptions, credit packs and sources. In dating, cohort retention shows how long paying users really stay, which tells you what they are worth and which traffic source pays off. If you want to buy dating traffic on transparent terms and judge your sources by cohort retention, talk to Leadpartner.