SaaS Retention Metrics: MRR, Churn, and NRR Explained

Retention is now the metric investors weigh first and the one founders measure last. This guide shows you how to calculate MRR, churn, gross and net revenue retention, and how to read a cohort curve, so your numbers actually tell you where the money is leaking.

6 min read7 sectionsJune 22, 2026

Start with MRR, and normalize it before anything else

Every retention metric is built on Monthly Recurring Revenue, so a sloppy MRR number poisons everything downstream. MRR is the recurring revenue you are contractually entitled to in a given month, and the word that matters is recurring. Three mistakes wreck it. First, billing cycles: a customer paying $1,200 per year is $100 of MRR, not $1,200 in the month they pay. Divide annual by 12 and quarterly by 3, every time. Second, one-time fees: setup, implementation, training, and professional services are not MRR. If your reported $50,000 includes $8,000 of implementation, your real MRR is $42,000. Third, decomposition. Total MRR alone hides the story. Split it into New, Expansion, Contraction, and Churned MRR. A company with flat MRR could be perfectly healthy or quietly bleeding, and only the components tell you which. A practical rule: source MRR from your billing and contract system, not your CRM, and reconcile it to the recognized subscription revenue on your P&L. When CRM, billing, and accounting disagree, billing wins.

Churn: count customers and revenue separately

Churn comes in two flavors and they routinely disagree. Customer (logo) churn counts accounts lost: customers lost in the period divided by customers at the start. Revenue churn measures the dollars those losses cost you. The gap matters because customers are not interchangeable. Lose ten accounts at $100 per month and that stings less than losing one $5,000 account, yet logo churn screams ten while revenue churn whispers. Watch for the inverse too: low logo churn with high revenue churn means one big account walked while many small ones stayed. Two more rules. Do not put new customers in the denominator, because it dilutes the starting base and flatters your churn during growth. And never multiply monthly churn by 12 to annualize it. Churn compounds. The correct math is 1 minus (1 minus monthly rate) to the 12th power, so 8% monthly is about 63% annually, not 96%. As a rough 2025 reference, B2B SaaS averages near 3.5% monthly churn, enterprise often sits at 0.5-1%, and SMB self-serve runs 3-7%, which is only survivable if expansion outruns it.

GRR vs NRR: the floor and the engine

Gross Revenue Retention and Net Revenue Retention answer different questions, and you need both. GRR = (Starting MRR - Churned MRR - Contraction MRR) / Starting MRR. It excludes all expansion, so it can never exceed 100%, and it tells you the floor: how much revenue you keep if nobody upgrades. NRR adds expansion back in: NRR = (Starting MRR + Expansion MRR - Contraction MRR - Churned MRR) / Starting MRR. Because expansion can outweigh losses, NRR can climb past 100%, which is the holy grail of negative net churn, where your existing base grows on its own. The trap is reading NRR alone. A company can post 115% NRR while GRR sits at 80%, meaning a handful of big upsells are papering over heavy churn underneath. That churn eventually catches up. Watch the gap between the two: a wide gap is a warning, not a victory.

What the 2025-2026 benchmarks actually say

Numbers without context mislead, so segment before you compare. Across private B2B SaaS, the 2025 median NRR is roughly 106% and median GRR lands around 88-91%, with top quartile NRR above 120-130% and GRR above 95%. But ACV is a sharper lens than headline ARR. Enterprise software (ACV above $100K) medians near 118% NRR, mid-market ($25K-$100K) near 108%, and SMB (under $25K) near 97%. So a 97% NRR is dead average for an SMB-focused product and a five-alarm fire for an enterprise one. Stage matters too: companies at $1M-$10M ARR tend to hover near 98% NRR, climbing toward 115% past $100M ARR, partly from product maturity and partly from a customer mix that shifts toward stickier accounts. Pick the benchmark cohort that matches your ACV and stage. Comparing your seed-stage SMB tool to an enterprise incumbent is how founders talk themselves into bad decisions.

Cohort analysis: read the shape, not the headline

A blended retention rate can sit comfortably flat while your newest cohorts quietly fall off a cliff, propped up by older loyal accounts. Cohort analysis fixes this by grouping customers by signup month and tracking each group over time. Read a row across to watch one cohort decay; read a column down to compare cohorts at the same age, which tells you whether onboarding is improving. The single signal that matters most is whether the curve flattens. A healthy curve drops early as poor-fit customers self-select out, then levels off (for example 95% at month 1, 84% at month 6, 79% at month 24). That flattening is the closest thing to a product-market fit signal. A curve that declines at a constant rate and never flattens is a crisis: customers keep leaving forever. The timing of the drop is diagnostic too. A sharp day-one drop points to a first-run experience problem; a steady decline that never settles points to a value problem. Two cautions: do not read immature cohorts (a March cohort has no month-6 number in April), and aggregate cohorts under ~30 customers to quarters so one cancellation does not swing the percentage.

From retention to unit economics: LTV, CAC, and the quick ratio

Retention is not an end in itself; it is the input that determines whether your unit economics work. Lifetime value is roughly average revenue per account times gross margin, divided by churn rate, which means lower churn directly lengthens lifetime and lifts LTV. Compare it to fully loaded customer acquisition cost (all sales and marketing spend divided by new customers). The common rule of thumb is LTV:CAC of at least 3:1, with healthy CAC payback inside 12 months. This matters more every year: 2025 data shows SaaS companies spending around $2.00 of sales and marketing to win $1.00 of new ARR, up meaningfully from prior years. Expanding an existing account is far cheaper than acquiring a new one, which is why high performers now get 40-50% of new ARR from expansion. One more efficiency check is the SaaS quick ratio: (New MRR + Expansion MRR) / (Churned MRR + Contraction MRR). A ratio of 4 or higher means you add four dollars of growth for every dollar lost. A caution on LTV: most LTV figures are projections, not observed data. If you have six months of history, do not project a five-year lifetime. Track cohort LTV by signup month for at least a year before you trust the formula.

Where retention numbers go wrong, and how to keep them honest

The most dangerous retention problems are not bad numbers, they are wrong numbers that look fine. Reported MRR drifts from real MRR through manual exports and spreadsheet logic: a one-off fee counted as recurring just this once, a churned customer still tagged active, an annual deal booked in full in one month. None of these look catastrophic alone, but multiplied across hundreds of accounts they distort growth, runway, and your valuation multiple during diligence. Keep it honest with explicit, written policies: when does churn count (the conservative answer is the date cancellation is submitted), what is your win-back period (most companies use 30-90 days before a return counts as new rather than renewed), how do you handle usage overages and refunds, and what seasoning period removes early trial-quitter noise. Prefer the cohort method over the formula method for anything longer than a month, because formula errors compound over the measurement window. And write your definitions down so they do not drift across teams as your pricing and product catalog grow. If your billing system cannot normalize revenue and understand contract terms, no spreadsheet formula on top of it will be correct. This is exactly the kind of plumbing we untangle in a focused sprint: getting billing, CRM, and reporting to agree on one trustworthy MRR before anyone makes a decision on top of it.

Written by

Kostadin Stamboliev

Founder and Imaginator

Founder and Imaginator at Pineido. Practitioner across IT consulting, app and web development, technical SEO, and AI / automation. Ships production code in 30-day sprints with fixed scope and clear ownership.

LinkedIn

Ready to ship?

Tell us where you are stuck. We will review it and send a focused recommendation within one business day.

Get operations insights weekly

Join service-business operators who get one actionable insight on ops, automation, and growth every week.

Checking that you're human. You can send in a moment.

No spam. Unsubscribe anytime.

You’re subscribed! Look for our next issue in your inbox.

Frequently asked questions

What is the difference between NRR and GRR?

Gross revenue retention (GRR) measures the percentage of existing recurring revenue you keep after churn and downgrades, excluding any expansion, so it can never exceed 100%. Net revenue retention (NRR) adds expansion revenue (upsells, cross-sells, seat growth) back in, so it can exceed 100%. GRR is your floor; NRR is the floor plus your expansion engine. Track both, because a high NRR can hide a weak GRR when a few large upsells mask heavy underlying churn.

What is a good NRR for a SaaS company in 2026?

It depends on your segment. The 2025 median for private B2B SaaS is around 106%, but enterprise products (ACV above $100K) median near 118%, mid-market near 108%, and SMB-focused products near 97%. Top-quartile companies across segments exceed 120-130%. The right benchmark is the one that matches your average contract value and stage, not the overall median. Anything below 100% means your existing base is shrinking and new sales are just covering the leak.

Can I annualize monthly churn by multiplying by 12?

No. Churn compounds, so multiplying overstates it badly. Use 1 minus (1 minus the monthly rate) raised to the 12th power. For example, 8% monthly churn is about 63% annually, not 96%. Even a seemingly modest 5% monthly churn loses roughly 46% of a starting base within a year, which is why getting the math right changes how you think about lifetime value.

Should I track customer churn or revenue churn?

Both, because they tell different stories. Customer (logo) churn counts accounts lost regardless of size. Revenue churn measures the dollars those losses cost. If your accounts vary widely in size, the two can diverge sharply: losing one large enterprise account barely moves logo churn but devastates revenue churn, while losing many small accounts does the opposite. Tracking only one hides which customers you are actually losing.

Why does a flattening cohort retention curve matter?

A retention curve that drops early and then flattens means you have found a stable base of customers getting real value, which is one of the clearest product-market fit signals. A curve that declines at a steady rate and never flattens means customers keep leaving indefinitely, pointing to a fundamental value problem. The shape matters more than the headline retention number, because blended averages can stay flat while newer cohorts quietly collapse.

Why don't my retention numbers match across tools?

Usually because CRM, billing, and accounting each compute MRR differently, and reported MRR drifts from real MRR through manual exports and inconsistent definitions. Common causes include one-time fees counted as recurring, annual deals not divided into monthly equivalents, churned accounts still tagged active, and undefined policies for win-back periods and refunds. The fix is to source MRR from billing and contracts, reconcile it to recognized revenue on the P&L, and write your definitions down so they stop drifting across teams.