Churn Rate Benchmarks for SaaS: The Real Numbers by ARR Cohort

What "Good" Churn Actually Means (It's Not 5%)

The "5% annual churn is acceptable" rule has been circulating since around 2015, popularized by early Salesforce and SaaS Capital posts written for enterprise software companies with 3-year contracts and 500-seat minimums. If you're running a $2M ARR product-led SaaS with monthly billing and SMB customers, that benchmark has nothing to do with you.

Good churn is not a single number. It's a function of your segment, your billing cadence, your ARR, and whether you have any expansion motion at all.

The single most important distinction most founders skip: logo churn vs. revenue churn. Logo churn counts how many customers you lost. Revenue churn counts how much MRR you lost. A company with 8% logo churn but 2% net revenue churn is probably fine, because expansions from retained accounts offset the small-customer losses. A company with 3% logo churn and 6% net revenue churn is bleeding from its largest accounts and will feel it in 90 days.

At sub-$5M ARR, you almost certainly don't have enough expansion revenue to offset much. That makes gross churn the number that matters.

---

Churn Benchmarks by ARR Cohort

These ranges come from Bessemer Venture Partners' State of the Cloud report (2025) and ChurnZero's 2025 SaaS Churn Index, two of the more rigorous public data sets covering sub-$10M companies.

Under $1M ARR

Monthly logo churn of 3, 7% is common here and does not mean your product is broken. Customer fit is still being discovered. You're taking on early adopters who will naturally cycle out as you sharpen positioning. Annual churn of 30, 50% at this stage is painful but not terminal, as long as you're replacing churned customers faster than you're losing them.

Red flag: monthly churn above 10% for two consecutive quarters. That's not early-adopter noise. That's a product-market fit signal.

$1M, $5M ARR

This is the window where churn starts compounding against you in ways that are hard to see in real time. At $2M ARR, a monthly churn rate of 3% means you're losing roughly $720,000 in ARR annually before any growth. You need to acquire that revenue back just to stay flat.

Benchmark here: monthly logo churn of 1.5, 3% for SMB-focused products, 0.8, 1.5% for mid-market. Net Revenue Retention (NRR) above 100% is the goal, but most companies in this cohort sit at 90, 98%.

If your NRR is below 90% at $3M ARR, you have a retention crisis, not a churn nuisance.

$5M, $10M ARR

By $5M ARR you should be seeing some expansion revenue, either from seat growth, usage tiers, or upsells. NRR becomes the benchmark that investors and acquirers watch. SaaS Capital's 2025 data shows the median NRR at this stage is 104%, with top quartile at 115%+.

Gross logo churn at this stage: 0.5, 1.5% monthly for SMB, under 0.5% for mid-market or enterprise contracts. Annual gross churn above 15% at $7M ARR is a serious drag on valuation, typically discounting a 5x ARR multiple by 1, 2 turns.

SMB vs. Mid-Market vs. Enterprise

Segment	Acceptable Annual Logo Churn
SMB (ACV under $5K)	15, 25%
Mid-market (ACV $5K, $50K)	5, 12%
Enterprise (ACV above $50K)	2, 6%

SMB churn is structurally high. Small businesses close, get acquired, run out of budget, or switch tools every 18 months on average. If you're selling to SMBs, your job is acquisition velocity and quick time-to-value, not trying to hit enterprise-level retention numbers with an SMB product.

---

The Calculation Error That Inflates Your Churn

Most founders calculate churn like this: customers lost this month divided by customers at the start of the month. Simple enough. But three common errors make that number unreliable.

Error 1: Annualizing incorrectly. A 2% monthly churn rate is not 24% annually. It's roughly 21.5%, because you're losing customers from a shrinking base each month. The formula is 1 - (1 - monthly_churn)^12. At 2% monthly, that's 1 - 0.98^12 = 21.5%. Not a huge difference, but at scale it adds up.

Error 2: Timing of cancellations within the billing period. If a customer cancels on day 3 of a 30-day billing cycle, most billing systems still count them as "churned this month." But they paid for the full month. If you're not accounting for this, you're overstating churn and understating recognized revenue simultaneously.

Error 3: Including trials and pilots in your denominator. A 30-day free trial that converts at 40% will make your churn look 60% worse than it is if you count non-converting trials as churned customers. Separate your trial funnel from your paid customer churn calculation. Always.

---

Where Founders Lose $300K Without Noticing

Silent Churn

This is the most dangerous churn type at the $1M, $5M ARR stage. Silent churn happens when a customer stops using your product but keeps paying, usually on annual contracts. They won't tell you they've disengaged. They'll just not renew. You'll find out 11 months after they went dark.

A company at $3M ARR with 15 enterprise-ish accounts on annual contracts, where 5 have gone silent, has roughly $300K, $500K of renewal risk sitting invisible on its books. The customers aren't churned yet. But they will be.

Usage data is the only way to catch this. If you're not tracking product logins or feature adoption per account, you're flying blind on roughly 30, 40% of your retention risk.

Contraction MRR Masquerading as Retention

A customer who downgrades from a $2,000/month plan to a $600/month plan shows up as "retained" in logo churn. But you just lost $16,800 in annual revenue from that single account. If three accounts do this in a quarter, that's $50,000 gone without a single cancellation email.

Track contraction MRR separately from logo churn. If your logo churn looks stable but your MRR is flattening, contraction is likely the culprit.

The 90-Day Churn Lag

Decisions that cause churn are almost never made the month churn appears. A customer who churns in June made the mental decision to leave in March, possibly triggered by a bad onboarding experience, a missed support ticket, or a competitor demo in February. By the time it shows up in your metrics, it's too late to intervene.

This means your current churn rate is a report card on decisions made 60, 90 days ago. Fixing today's product problems will show up in churn data in Q3, not next week.

---

How to Diagnose Your Actual Churn Problem

Blended churn rates hide the signal. A cohort analysis shows you which customer segments, acquisition channels, or contract types are driving churn disproportionately.

Specifically: run churn by acquisition month (when the customer signed, not when they churned). If customers who signed up during a particular 3-month window churn at 2x the rate of other cohorts, something changed in that window: your onboarding, your targeting, your product, or your pricing.

Signals that predict churn 60 days out: - Login frequency drops below once per week for accounts that previously logged in daily - Customer has not reached their first meaningful product output (first report run, first automation triggered, first team member invited) - No response to two consecutive check-in emails from CS or automated sequences - Support ticket opened in month 1 and not resolved within 5 business days

These are not guarantees. But across a portfolio of 50, 200 accounts, they're reliable enough to prioritize outreach.

Three Levers

Most churn traces back to one of three root causes:

1. Price-to-value mismatch. The customer is not seeing enough ROI to justify the spend. Either your product is underdelivering or your pricing is too high relative to perceived value.
2. Product-market fit gaps. The customer needed a feature or workflow that doesn't exist. They found a workaround, then a better tool.
3. Success bandwidth. The customer was onboarded but never truly activated. A solo founder running a $3M ARR SaaS cannot personally QBR 150 accounts. But they can build an automated activation sequence that does it for 80% of the work.

---

What to Do This Quarter

30-day churn audit:

Pull every account that churned in the last 6 months. Categorize each churn reason into one of the three levers above. If 60%+ of reasons fall in one bucket, that's your priority. If it's spread evenly across all three, you have a systemic onboarding problem, not a product or pricing one.

Where to put effort first by ARR stage:

At under $1M ARR, focus on the first 30 days of activation. That's where most churn is decided. At $1M, $5M ARR, add silent churn monitoring through usage data. At $5M, $10M ARR, build the expansion motion so NRR becomes your buffer against logo churn.

Automating the signal.

If you don't have a customer success team (and at $3M ARR, most founders don't), you need automated triggers pulling from product usage data. MorBizAI can connect to your existing product analytics and CRM to surface at-risk accounts based on engagement signals, draft at-risk outreach sequences, and flag contraction candidates before they downgrade. The goal is not to replace a CS team. It's to make sure a 3-person founding team can see the same signals a 10-person CS org would catch manually.

Churn is rarely a surprise in hindsight. The data is almost always there. The gap is in who's watching it.