What is churn rate?
– Churn rate (also called customer attrition) measures how quickly customers stop buying from or subscribing to a business over a defined period. It’s usually expressed as a percentage and is most useful for subscription or repeat‑purchase models. A related concept, retention rate, is the share of customers who remain.
Common formulas (pick one and be consistent)
– Simple period churn (most common): Churn (%) = (Customers lost during period) / (Customers at the start of period) × 100
– Average‑customer denominator: Churn (%) = (Customers lost during period) / (Average customers during period) × 100, where average = (customers at start + customers at end) / 2
– Cohort churn: Track a specific acquisition cohort and compute percent who cancel after N days/weeks/months.
Note: For time‑unit consistency, use monthly churn for monthly subscriptions and quarterly churn for quarterly reporting.
Step‑by‑step: how to calculate and interpret churn
1. Define the period (e.g., month, quarter, year).
2. Decide the denominator (start of period or average during period) and apply consistently.
3. Count customers lost in that period (cancelled or inactive according to your policy).
4. Compute churn using the chosen formula.
5. Compare churn to growth (new customers / starting customers) to see net change.
6. Segment results by cohort, product, channel, or tenure to find where churn is concentrated.
7. Translate churn into financial impact (e.g., churn × average revenue per user) and compare to customer acquisition cost (CAC).
Short checklist — what to track every period
– Period definition and denominator choice documented
– Number of customers at start and end
– Number of customers lost (and reasons if known)
– Number of new customers acquired
– Churn by cohort (new vs. long‑term customers)
– Average Revenue Per User (ARPU) or lifetime revenue per customer
– Customer Acquisition Cost (CAC) and CAC payback period
– Trend vs. previous periods and industry benchmarks
Worked numeric example
Assumptions:
– Customers at start of month: 1,000
– Customers lost during month: 50
– New customers acquired during month: 120
Calculations:
– Churn rate = 50 / 1,000 = 0.05 = 5%
– Growth rate (new / start) = 120 / 1,000 = 12%
– Net change in customers = 120 − 50 = +70 → net growth = 70 / 1,000 = 7%
Interpretation:
– Positive net growth (7%) because growth (12%) > churn (5%).
– If average monthly revenue per customer (ARPU) = $10, lost revenue from churn = 50 × $10 = $500 that month.
– If CAC = $200, check payback: monthly ARPU $10 → CAC payback = 200 / 10 = 20 months. If typical customer lifetime (1 / monthly churn) = 1 / 0.05 = 20 months, LTV roughly equals CAC (ignoring margins). That suggests acquisition economics are marginal and should be improved or CAC reduced.
Why churn matters (advantages of measuring it)
– Direct signal of customer satisfaction and product fit.
– Helps quantify revenue leakage and the durability of growth.
– Reveals where to prioritize retention investments (onboarding, product improvements, pricing).
– Easier and cheaper in most cases to retain customers than to acquire new ones.
Limitations and common pitfalls
– Denominator choice changes the number; be consistent when comparing periods.
– Aggregate churn masks who is leaving: new customers often churn faster than long‑time customers.
– High churn in a period can reflect a prior period’s aggressive acquisition (trialers leaving), not necessarily worsening product quality.
– Benchmarks vary widely by industry, company maturity, and product complexity—don’t compare startups and incumbents directly.
– Churn alone doesn’t capture revenue per customer changes; revenue churn (dollar churn) can tell a different story than customer churn.
Employment churn (
employment churn (also called employee turnover) measures how quickly staff leave an organization over a given period. Separations include voluntary resignations, involuntary terminations, and retirements. Like customer churn, it’s a rate: the number of people who leave divided by the workforce base used as the denominator.
Key formulas and simple rules
– Employee churn rate (period) = (Number of separations during period / Average number of employees during period) × 100.
– Average number of employees = (Employees at start of period + Employees at end of period) / 2 (simple average). Use a more granular average if headcount swings intra‑period.
– Customer churn (period) = (Customers lost during period / Customers at the start of period) × 100. Be explicit whether you use start-of-period, end-of-period, or average in comparisons.
– Dollar churn (revenue churn) = (Recurring revenue lost from downgrades and cancellations − Expansion revenue from upsells) / Starting recurring revenue.
– Approximate average customer lifetime (periods) = 1 / churn rate (express churn as a decimal). This assumes churn is constant over time.
Worked numeric examples
1) Employee churn
– Company A: 1,000 employees on Jan 1; 980 on Dec 31. During the year 120 people left.
– Average employees = (1,000 + 980) / 2 = 990.
– Annual employee churn = 120 / 990 = 0.1212 → 12.12%.
2) Customer churn (monthly)
– Startup B starts the month with 2,000 customers. During the month 80 cancel.
– Monthly customer churn = 80 / 2,000 = 0.04 → 4% monthly churn.
– Approximate average customer lifetime = 1 / 0.04 = 25 months (assumes constant monthly churn).
3) Net MRR (monthly recurring revenue) churn
– Starting MRR = $100,000. Lost MRR from cancellations/downgrades = $8,000. Expansion MRR from upsells = $2,000.
– Net MRR churn = (8,000 − 2,000) / 100,000 = 0.06 → 6% net MRR churn for the month.
Step‑by‑step checklist to compute churn correctly
1. Define the metric precisely: customer count vs revenue vs employee separations. Specify period (monthly, quarterly, annual).
2. Choose and document the denominator (start, end, or average). Use the same method over time and across comparisons.
3. Count only defined “loss” events in the numerator (cancellations, non‑renewals, separations).
4. Adjust for reactivations or rehires (exclude reactivations from loss if you count only net exits).
5. For revenue churn, separate contraction (downgrades)
5. For revenue churn, separate contraction (downgrades) from expansion (upsells) and report both gross and net figures. Gross revenue churn = lost revenue from cancellations + downgrades, divided by starting revenue. Net revenue churn = (gross lost revenue − expansion revenue) / starting revenue. Decide whether to floor net churn at zero (some firms report negative churn when expansion exceeds losses).
6. Use cohorts to reveal underlying patterns. A cohort groups customers by a shared start date (e.g., “January 2024 signups”). Compute churn within each cohort rather than only company‑wide averages. Cohort analysis shows whether churn is concentrated among early customers, particular plan types, or acquisition channels.
7. Annualize carefully. If monthly churn rate is c_month, monthly retention r = 1 − c_month. Annual retention = r^12. Annual churn = 1 − r^12. Example: if c_month = 3% → r = 0.97 → annual retention = 0.97^12 ≈ 0.698 → annual churn ≈ 30.2%. Do not use c_annual ≈ 12 × c_month unless churn events are rare and you explicitly want a linear estimate.
8. Track both count and revenue churn. Logo (customer-count) churn = customers_lost / customers_at_start. Revenue churn (MRR/ARR churn) weights by size; a few large customer losses can create high revenue churn but low logo churn. Report both when possible.
9. Always disclose denominator, period, and sample size. State whether the denominator is starting MRR, ending MRR, or an average; whether churn excludes free trials; and how reactivations are treated. Small samples (e.g., churn based on 20 customers) are volatile—show raw counts alongside percentages.
10. Smooth and contextualize. Use rolling averages (3‑ or 12‑month) to reduce month‑to‑month noise and annotate one‑off events (major account loss, pricing changes, seasonality). Compare like‑for‑like (same month prior year or same cohort age) before declaring trends.
Quick formulas and worked examples
– Logo churn (customer count)
Formula: Logo churn = customers_lost / customers_at_start
Example: Start = 500 customers, lost = 20 → Logo churn = 20/500 = 0.04 → 4%.
– Gross MRR churn (revenue lost from cancellations + downgrades)
Formula: Gross MRR churn = MRR_lost_gross / MRR_start
Example: MRR_start = $100,000; MRR_lost_gross = $10,000 → Gross MRR churn = 10,
000/100,000 = 0.10 → 10%.
– Net MRR churn (revenue lost after expansions/upsells)
Formula: Net MRR churn = (MRR_lost_gross − MRR_expansion) / MRR_start
Example: MRR_start = $100,000; MRR_lost_gross = $10,000; MRR_expansion = $4,000 → Net MRR churn = (10,000 − 4,000)/100,000 = 6,000/100,000 = 0.06 → 6%.
Note: If expansions exceed gross losses, net MRR churn becomes negative (often called “negative churn”), meaning existing customers more than offset the revenue lost from cancellations/downgrades.
Annualizing churn (convert monthly to annual)
Formula: Annual churn = 1 − (1 − monthly_churn)^(12)
Example: monthly churn = 4% → annual = 1 − (0.96)^12 ≈ 0.396 → 39.6%.
Assumption: churn is independent and constant each month — which may not hold in practice (cohort effects, seasonality).
Quick LTV (customer lifetime value) approximation using churn
Simple formula (no discounting): LTV ≈ ARPA / churn_rate
Definitions: ARPA = average revenue per account for the period; churn_rate must be for the same period (monthly churn if ARPA is monthly).
Example: ARPA = $200/month; monthly churn = 3% → LTV ≈ 200 / 0.03 = $6,666.67.
Caveats: This ignores gross/net expansion, CAC (customer acquisition cost), margin, and discounting. Use a cohort-based discounted cash flow for precise LTV.
Cohort retention — how to build and read a cohort table (step-by-step)
1) Define cohorts by acquisition period (e.g., customers acquired in Jan 2024).
2) For each cohort, count how many customers remain active at each age (month 0, month 1, …).
3) Compute retention % at age t = customers_remaining_at_age_t / customers_at_start_of_cohort.
Worked short example: Jan cohort start = 200 customers. After 1 month = 180 remain → month‑1 retention = 180/200 = 90%. After 6 months = 150 → month‑6 retention = 75%.
Interpretation: plot retention by cohort age to see how quickly different cohorts churn and whether retention is improving over time.
Statistical significance and small-sample caution
– For a proportion p from sample size n, standard error (SE) ≈ sqrt[p(1−p)/n].
– 95% confidence interval ≈ p ± 1.96 × SE.