Churn Rate vs Retention Rate
Churn rate and retention rate are two sides of the same coin — Retention = 1 − Churn. But customer churn, revenue churn, gross retention, and net revenue retention each reveal something different about your business health.
TL;DR
- Churn rate = customers or revenue lost ÷ customers or revenue at start of period.
- Retention rate = 1 − Churn rate. Customer retention only counts account existence; revenue retention reflects ARPA changes too.
- NRR (Net Revenue Retention) above 100% = expansion outpaces churn. Your existing customers grow your ARR without new acquisition.
- 2% monthly churn compounds to ~22% annual loss — monthly and annual churn are not simply 12× each other.
At a Glance
| Attribute | Churn Rate | Retention Rate |
|---|---|---|
| Definition | % of customers/revenue lost in a period | % of customers/revenue kept in a period |
| Formula | Lost ÷ Starting × 100 | 1 − Churn rate (or Retained ÷ Starting × 100) |
| Ideal direction | Lower is better (0% = no loss) | Higher is better (100% = no loss) |
| Customer churn formula | (Customers lost) ÷ (Starting customers) | (Customers at end − New) ÷ Starting customers |
| Revenue churn formula | (MRR lost) ÷ (Starting MRR) | Starting MRR − churn + expansion ÷ Starting MRR |
| Can exceed 100%? | No (can't lose more than you have) | Yes (NRR) — expansion pushes above 100% |
| Monthly vs annual | Compounding: annual ≠ monthly × 12 | Same — use (1 − monthly churn)^12 |
| Healthy SMB SaaS | < 3% monthly | > 97% monthly |
| Healthy enterprise SaaS | < 0.5% monthly / < 6% annual | > 99.5% monthly / > 94% annual |
| NRR benchmark (best-in-class) | N/A (use NRR) | > 120% NRR |
| Primary signal | Revenue leakage, product/fit problems | Stickiness, product value, expansion potential |
Monthly Churn → Annual Impact
| Monthly Churn | Annual Churn | Annual Retention | Avg Customer Life | Assessment |
|---|---|---|---|---|
| 0.5% | 5.8% | 94.2% | 200 months (~16.7 yr) | World-class |
| 1.0% | 11.4% | 88.6% | 100 months (~8.3 yr) | Excellent |
| 2.0% | 21.5% | 78.5% | 50 months (~4.2 yr) | Good |
| 3.0% | 30.6% | 69.4% | 33 months (~2.8 yr) | Acceptable (SMB) |
| 5.0% | 46.0% | 54.0% | 20 months (~1.7 yr) | High — review urgently |
| 7.0% | 58.0% | 42.0% | 14 months (~1.2 yr) | Critical |
Annual churn = 1 − (1 − monthly churn)^12. Average customer life = 1 ÷ monthly churn rate.
Deep Dive
Types of Churn
Customer churn counts the number of accounts lost. Simple, but ignores revenue weighting — losing 10 $50/month SMBs and 1 $5,000/month enterprise both count as 1 or 10 depending on how you aggregate.
Revenue churn (MRR churn) measures the dollar impact. 10 × $50 = $500 lost vs 1 × $5,000 = $5,000 lost. Very different business problems requiring different interventions.
Gross revenue churn counts only losses (contraction + cancellation). Capped at 100%. Net revenue churn subtracts expansion revenue — can be negative (net negative churn), meaning the existing customer base grows revenue despite cancellations. Net negative churn is a hallmark of elite SaaS companies: Twilio, AWS, and Snowflake achieve it through usage-based expansion.
NRR — The Real Retention Metric
Net Revenue Retention (NRR) captures the full picture of revenue retention including expansion. Formula: (Starting MRR + Expansion − Contraction − Churn) / Starting MRR × 100.
Example: Start at $100K MRR. Expansion +$15K. Contraction −$5K. Churn −$8K. NRR = ($100K + $15K − $5K − $8K) / $100K = 102%. The business grows 2% from existing customers alone, before any new customer acquisition. Benchmarks: NRR below 90% = serious retention problem. 90–100% = adequate. 100–110% = good. 110–120% = great. Above 120% = best-in-class (Snowflake, Datadog territory).
Gross Revenue Retention (GRR) excludes expansion. It measures the floor — how much starting revenue survives ignoring upsell. GRR above 90% is excellent; below 80% signals product-fit or support problems that no amount of upsell can compensate for.
Real-World Patterns
The Leaky Bucket Problem
A SaaS at $500K MRR with 5% monthly churn loses $25K MRR/month. To grow, they must generate more than $25K in new MRR monthly just to stay flat. At a typical $1,000 ARPA, that's 25 new customers/month from acquisition alone — before any growth. Their sales team closes 30/month (+$30K MRR), so net growth is only +$5K/month (1%). If they fixed churn to 2%, they'd only need to replace $10K MRR — the same 30 new customers now grow the business at 4%/month. Plugging the bucket is more effective than pouring in more water.
NRR > 100% as a Growth Engine
Snowflake reported NRR of 158% in FY2022 — meaning its existing customers grew revenue by 58% in a year through increased usage. Even if Snowflake acquired zero new customers, it would still grow 58% YoY. This is the power of usage-based pricing combined with strong product adoption. Enterprise SaaS companies with seat-based expansion (adding users) or usage-based pricing (more API calls, more data) often achieve NRR above 120%. Seat-based models cap at 100% NRR unless companies grow headcount or add products.
Customer vs Revenue Churn Divergence
A mid-market SaaS reports 8% customer churn but only 2% revenue churn. Investigation reveals: they're losing many small SMB customers (average $200 ARPA) while retaining all their large enterprise customers (average $5,000 ARPA). This is actually a signal of ICP evolution — the product has stronger fit with enterprise. The right response: adjust ICP, raise minimum contract size, invest in enterprise CS, and allow SMB churn to continue if the segment is unprofitable. Conflating the two churn numbers would send the wrong signal.
Involuntary Churn — The Hidden Killer
Involuntary churn (failed payments, expired credit cards) accounts for 20–40% of all SaaS churn. Many companies attribute all churn to product dissatisfaction when a significant portion is payment failure. Solutions: Stripe's smart retries, dunning email sequences, account updater for saved cards, pre-expiry card update reminders. Recovering 50% of involuntary churn at 3% total monthly churn reduces effective churn to 2.4% — equivalent to a significant retention improvement from purely operational fixes.
Verdict: Track Both, Optimise for NRR
Churn rate tells you what you're losing. Retention rate tells you what you're keeping. NRR tells you whether your existing customer base is growing or shrinking in total revenue — the definitive signal of product-market fit at scale. Most early-stage companies over-focus on new customer acquisition (the exciting metric) while under-investing in retention (the compounding metric). A 1% improvement in monthly retention creates more long-term value than a 20% improvement in new customer acquisition.
If you can only track one retention metric, track Net Revenue Retention. It synthesises churn, contraction, and expansion into a single number that captures the total health of your customer revenue base.
Decision Checklist
| Scenario | Track |
|---|---|
| How many accounts are we losing? | Customer churn rate |
| How much revenue are we losing? | Revenue churn rate |
| Are existing customers growing our ARR? | NRR |
| What % of starting revenue do we keep (floor)? | Gross Revenue Retention |
| Is our business growing from the existing base? | NRR (> 100% = yes) |
| How long does an average customer stay? | 1 ÷ Monthly churn rate |
| What's the annual equivalent of our monthly churn? | 1 − (1 − monthly)^12 |
| Is our upsell/expansion working? | Expansion MRR component |
| Are payment failures causing churn? | Voluntary vs involuntary churn split |
| Investor asks about retention | NRR (primary), GRR (supporting) |
| Cohort analysis of early vs recent customers | Cohort retention curves |
| Is SMB or enterprise churning more? | Revenue churn by segment |
Frequently Asked Questions
What is a good monthly churn rate for SaaS?
For SMB SaaS: 3–7% monthly churn is common, though 1–2% is considered good. For mid-market SaaS: 1–2% monthly. For enterprise SaaS: under 0.5% monthly (under 6% annually). Consumer subscription apps (streaming, apps): 5–10% monthly is typical. The lower the ARPA, the higher the expected churn — SMB customers are harder to retain than enterprise. Benchmark yourself within your customer segment, not against enterprise-focused companies if you're serving SMBs.
Is it better to measure customer churn or revenue churn?
Revenue churn is more meaningful for most business decisions. Customer churn counts lost accounts equally regardless of size — losing a $100/month customer and a $10,000/month customer both count as 1 churn. Revenue churn weights the loss by economic impact. A company with 10% customer churn but 2% revenue churn (because their largest customers retain perfectly while many small customers leave) has very different health than 10% customer churn with 10% revenue churn. Track both, but prioritise revenue churn for business health assessment.
What is Net Revenue Retention (NRR) and why does it matter?
NRR (also called Net Dollar Retention or NDR) = (Starting MRR + Expansion MRR − Contraction MRR − Churned MRR) / Starting MRR × 100. NRR above 100% means your existing customer base is generating more revenue than last period despite churn — expansion from upgrades, upsells, and usage growth more than compensates for losses. Best-in-class SaaS companies (Snowflake, Datadog, Twilio) have reported NRR of 120–170%. NRR is the single metric most predictive of long-term SaaS company success, because it means the business can grow revenue even without acquiring a single new customer.
How does monthly churn compound into annual churn?
Annual churn ≠ monthly churn × 12. The correct formula is: Annual churn = 1 − (1 − monthly churn rate)^12. At 2% monthly: 1 − (0.98)^12 = 1 − 0.785 = 21.5% annual churn. At 5% monthly: 1 − (0.95)^12 = 46% annual churn. At 1% monthly: 1 − (0.99)^12 = 11.4% annual churn. This compounding is why even 'small' monthly churn rates are alarming at scale — and why quoting monthly churn to an investor who is thinking annually will make your retention look better than it is.
What is the difference between gross revenue retention and net revenue retention?
Gross Revenue Retention (GRR) = (Starting MRR − Contraction − Churn) / Starting MRR × 100. It measures how much starting revenue you kept, capped at 100% (cannot exceed starting revenue). Net Revenue Retention (NRR) adds expansion: (Starting MRR + Expansion − Contraction − Churn) / Starting MRR × 100. NRR can exceed 100% when expansion outpaces churn. Investors look at both: GRR reveals the 'floor' (how well you retain if no upsell), NRR reveals the 'ceiling' (total revenue growth from existing base). Top SaaS companies target GRR > 90% and NRR > 120%.
At what churn rate does a SaaS business become mathematically impossible to grow?
There is no single mathematical limit, but high churn creates an acquisition treadmill. At 10% monthly churn, you lose ~72% of customers in a year — meaning you must replace 72% of your customer base annually just to stay flat. At $1M ARR with 10% monthly churn, you need ~$720K of new ARR per year just to maintain revenue, before any growth. This is why companies with high churn hit growth ceilings: the sales and marketing spend required to outrun churn eventually exceeds what the business can generate. For sustainable growth, monthly churn should be under the monthly new customer growth rate with meaningful margin.
How do you calculate retention rate from churn rate?
Retention rate = 1 − Churn rate (for the same time period). Monthly: 2% churn → 98% retention. Annual: 20% churn → 80% retention. The formula works for both customer and revenue metrics. Customer retention rate = (Customers at end of period − New customers acquired) / Customers at start × 100. Revenue retention (gross) = (Revenue at end − New revenue from new customers) / Revenue at start × 100. The key distinction: revenue retention includes the effect of price changes and upgrades for existing customers, while customer retention only counts whether the customer exists.
How should early-stage startups track churn when customer counts are small?
With fewer than 50 customers, individual churn events distort monthly churn percentages dramatically. Losing 2 of 20 customers is 10% churn — but it might be two unique reasons unrepresentative of the broader base. At small scale, track churn qualitatively: interview every churned customer within 48 hours of cancellation. Categorise reasons (price, competition, product gap, company shutdown, wrong-fit ICP). Revenue churn is more stable than customer churn at small scale. Quarterly cohort analysis (rather than monthly) smooths statistical noise. Aggregate monthly churn into 6-month rolling averages to identify trends without over-reacting to individual data points.
Related Comparisons
Verdict: Choose Based On Your Situation
Churn Rate
- You want to highlight customer loss problem
- You're reporting to investors on risk metrics
- You want to motivate teams to reduce loss
- You're benchmarking against industry rates
Retention Rate
- You want to emphasize positive (customers kept)
- You're tracking progress toward retention goals
- You want metrics that improve with performance
- You're focusing on growth and success