NRR stands for Net Revenue Retention. It measures how much recurring revenue you keep from existing customers over a period, including expansions, upgrades, downgrades, and cancellations, but excluding anyone who became a customer during that period. The formula is (Starting MRR + Expansion − Contraction − Churn) ÷ Starting MRR × 100. Above 100% means your existing base grows on its own. The blended private B2B SaaS median is about 101% (Benchmarkit 2025), and Bessemer's fundability scale puts 100% at good, 110% better, 120%+ best.
What is Net Revenue Retention (NRR) in SaaS?
Net Revenue Retention measures how much recurring revenue you keep from existing customers over time, including expansions, upgrades, downgrades, and cancellations.
One number tells you whether your customer base is an appreciating or a depreciating asset. At 90% NRR you have to sell $10 of new revenue for every $100 of base just to stand still. At 130% the base pays you $30 more per $100 before your sales team wins a single new logo.
NRR, NDR, and net retention rate are the same metric
The acronym soup is real, and it costs founders time in investor conversations. Net Revenue Retention (NRR) and Net Dollar Retention (NDR) are the same metric with the same formula (SaaS Metrics Standard Board). So are "net retention rate" and "net revenue retention rate". In practice, private companies and VCs lean toward NDR, while public filings and underwriters lean toward NRR. If an investor asks for NDR and you have NRR, you already have the answer.
Two terms that are not the same thing:
- Net recurring revenue is a dollar amount. NRR is a percentage.
- Logo retention (or customer retention) counts accounts. NRR counts dollars. More on that gap below.
Why is Net Revenue Retention important?
It is growth you do not have to buy. Expansion revenue carries no new-customer acquisition cost, so it lands at a far better unit economics profile than net-new revenue.
It sets your valuation floor. High-NRR companies (120%+) that are also growing fast (80%+ YoY) have commanded premium revenue multiples of 18-22x; more typical 120%+ NRR companies trade around 7-9x, while those below 100% NRR sit closer to 4-6x. Investors read NRR as evidence of product-market fit that survived contact with a renewal date.
It compounds. A single point of NRR is worth more every year you hold it, which is why it separates companies that scale from companies that run on a treadmill.
How to calculate the Net Revenue Retention rate
There are two accepted methods. They answer the same question and, applied correctly to the same data, give the same answer. The SaaS Metrics Standard Board treats the cohort method as preferred for annual measurement and the formula method as the practical choice for monthly or quarterly tracking.
Method 1: the formula method
NRR = (Starting MRR + Expansion − Contraction − Churn) ÷ Starting MRR × 100Includes expansion, can exceed 100%
Where:
Starting MRR = Monthly Recurring Revenue at period start, from customers who existed at period start
Expansion = additional revenue from those same customers (upsell, cross-sell, seats, usage, price increases)
Contraction = revenue lost from downgrades by those customers
Churn = revenue lost from cancellations by those customers
Method 2: the cohort method
NRR = Ending MRR from the starting cohort ÷ Starting MRR of that cohort × 100Preferred for annual measurement
Freeze the list of customers who were paying on day one. Twelve months later, look at what that exact list pays you now, and divide. You never have to classify a dollar as expansion or contraction, which removes the most common source of error.
Worked example: the formula method
Monthly view:
- Starting MRR: $100,000
- Expansion: $30,000
- Contraction: $5,000
- Churn: $10,000
($100,000 + $30,000 − $5,000 − $10,000) ÷ $100,000 × 100 = 115%Worked example, healthy growth
Annual view, same method applied to ARR:
- Starting ARR from the base: $2,400,000
- Expansion: $480,000
- Contraction: $120,000
- Churn: $240,000
($2,400,000 + $480,000 − $120,000 − $240,000) ÷ $2,400,000 × 100 = 105%Annual NRR, formula method on ARR
Worked example: the cohort method
Say 180 customers were paying you a combined $500,000 MRR in January. Twelve months later some of those 180 have cancelled, a few have downgraded, and the rest have bought more seats. Whatever is left of that original 180 now pays $540,000 MRR.
$540,000 ÷ $500,000 × 100 = 108%Cohort method, no component classification needed
Note that the customer count almost certainly fell while revenue rose. That is normal, and it is the whole point of measuring dollars rather than logos.
What counts, and what does not
Most NRR disputes are definition disputes. The Standard Board's guidance:
| Item | In or out |
|---|---|
| New customers won during the period | Out. They are not in the starting cohort |
| Expansion from existing customers | In |
| Downgrades and cancellations from existing customers | In |
| Price increases on existing contracts | In, counts as expansion |
| Win-backs inside your defined win-back window (typically 30-90 days) | In, as a renewal rather than a new customer |
| Reactivations after the win-back window | Out, treat as new |
| One-time fees, services, setup | Out. NRR is a recurring revenue metric |
| Revenue recognition accelerated by a cancelled contract | Out |
One more definitional trap worth knowing about: several widely read sources publish NRR as (Starting + Expansion − Churn) ÷ Starting, folding contraction into churn. The arithmetic is fine if you are consistent, but it means two companies can report different NRR on identical data. When you benchmark, check which variant the source used.
Monthly, quarterly, or annual NRR?
The period changes the number, so state it. Common practice at Series A and beyond is to calculate monthly for internal trend-spotting, report quarterly to the board, and use annual cohorts for investors and benchmarking. Almost every published benchmark is annual.
To convert a shorter period to annual, compound it:
Annual NRR = (Monthly NRR)12 | Annual NRR = (Quarterly NRR)4Compound, do not multiply
This is where small monthly numbers get loud. A quiet 1% monthly net revenue loss compounds to roughly 89% annual NRR (0.9912 = 0.886). Running the other way, 115% annual NRR is only about 1.2% net expansion per month. Monthly NRR looks unremarkable in both directions, which is why the annual view is the one that gets funded.
What is a good Net Revenue Retention rate in SaaS?
The honest answer is that "good" depends almost entirely on your average contract value. A 97% NRR is under-performing for an enterprise vendor and top-quartile for a $9-per-month product. Here is the picture from sources that publish real datasets.
The fundability ladder
Bessemer's State of the Cloud lists NRR among six fundability benchmarks: 100% good, 110% better, 120%+ best. Useful shorthand, but it was framed for growth-stage cloud companies, so treat it as the enterprise-flavoured bar rather than a universal one. The same framework puts logo retention at above 85% good, above 90% better, 95%+ best.
Medians from private-company surveys
| Cut | NRR | Source |
|---|---|---|
| Blended private B2B SaaS median | ~101% | Benchmarkit 2025 |
| Enterprise (ACV above $100K) | ~118% | Benchmarkit 2025 |
| SMB (ACV under $25K) | ~97% | Benchmarkit 2025 |
| Bootstrapped, $3M-$20M ARR, median | 103% | SaaS Capital 2026 (1,000+ private B2B SaaS) |
| Bootstrapped, $3M-$20M ARR, 90th percentile | 117.9% | SaaS Capital 2026 |
Early-stage reality check
If you are pre-$15M ARR, the ladder above will make you feel terrible for no good reason. ChartMogul's platform data puts top-quartile net retention at 94% for companies between $1M and $3M ARR, 99% between $3M and $15M, and above 105% only once you pass $15M. Early-stage companies commonly sit around 79%.
ARPA matters even more than stage. In ChartMogul's dataset, B2B companies with ACV above roughly $6,000 posted top-quartile NRR of 109.3%, and 41.1% of them exceeded 100%. For products under $10 per month, top-quartile net retention was 65.1%, and just 2.7% cleared 100%. Same metric, different universe.
Reading your own number
| Your NRR | What it means |
|---|---|
| Under 90% | The base is shrinking fast. Every new sale partly refills a leak. Fix retention before scaling spend |
| 90-100% | Net revenue loss from the base. Normal for low-ARPA and very early-stage, a problem for enterprise |
| 100-110% | The base grows on its own. Solid for SMB, mid-pack for enterprise |
| 110-120% | Strong. Expansion is a real motion, not an accident |
| Above 120% | Best in class. Check GRR before celebrating, see below |
Why NRR compounds, and why it belongs in your model
NRR is not a scoreboard metric. It is an input that changes every forecast downstream of it. Take a $100K MRR base and add zero new customers:
| Annual NRR | Base after 1 year | After 3 years | New MRR needed each year to hold flat |
|---|---|---|---|
| 90% | $90,000 | $72,900 | $10,000 |
| 100% | $100,000 | $100,000 | $0 |
| 115% | $115,000 | $152,100 | None, base adds $15,000 |
| 130% | $130,000 | $219,700 | None, base adds $30,000 |
The spread between the top and bottom row on a $100K base is a $40,000 annual swing in MRR, before sales does anything. Over three years it is the difference between $72.9K and $219.7K of recurring revenue from the identical starting customer list. This is why a change in assumed NRR moves runway, hiring plans, and the size of the raise you need more than almost any other assumption in a SaaS financial model.
What’s the difference between NRR & Gross Revenue Retention (GRR)?
| NRR | GRR | |
|---|---|---|
| Expansion included? | Yes | No |
| Can exceed 100%? | Yes | Never |
| Answers | Is the base growing? | How leaky is the bucket? |
| Formula | (Start + Expansion − Contraction − Churn) ÷ Start | (Start − Contraction − Churn) ÷ Start |
| Recent median | ~101% (Benchmarkit 2025) | ~88%, down from 90% over three years (Benchmarkit 2025) |
Short version:
- GRR = how good you are at keeping what you have.
- NRR = how good you are at growing what you have.
The gap between them is the real diagnostic
Report both, always. A company at 115% NRR and 75% GRR is not the same business as one at 115% NRR and 95% GRR. The first is losing a quarter of its revenue base every year and covering it with a handful of large expansions, so the moment one big account stops growing, the headline number collapses. The second has a genuinely sticky product. NRR alone cannot tell those two apart.
Worth knowing that the floor is moving: the 2026 Benchmarkit B2B SaaS and AI-native metrics report puts median GRR at 84%, down from 88%, with the 75th percentile falling from 95% to 91% (reported by The SaaS CFO). Retention is getting harder, which raises the value of every point you hold.
NRR vs customer (logo) retention
These move independently, and confusing them is a classic board-meeting mistake. You can keep 95% of your logos and still lose 20% of your revenue if the accounts that left were your biggest. The reverse is just as common in product-led businesses: a long tail of small accounts churns constantly while a handful of expanding mid-market accounts push NRR above 100%.
Track both. Logo retention tells you whether the product delivers for the median customer; NRR tells you what it does to your revenue. When they diverge, the divergence is the story, and it is usually about revenue concentration.
How to improve Net Revenue Retention
1. Build expansion into pricing, not into a quarterly upsell push
This is the biggest structural lever. Companies on usage-based or hybrid pricing tend to post NRR in the 115-130% range, while flat per-seat models cluster nearer 95-105% (m3ter). The reason is mechanical: when revenue scales with a metric the customer's own success drives, expansion happens without anyone having to sell it. Adding one usage dimension to an existing subscription, such as records processed, API calls, or seats past a threshold, creates that path without repricing the whole book.
2. Surface the upgrade at the moment of need
An in-product prompt when a team hits 80% of its tier limit converts better than the same conversation in a quarterly review, because the value is obvious right then. This needs usage instrumentation before it needs a playbook.
3. Catch contraction before renewal, not at renewal
Declining usage shows up 60 to 90 days before a downgrade or cancellation lands. That window is the entire opportunity. Once a customer has decided, a save attempt usually just converts churn into contraction. See churn reduction techniques and churn prediction.
4. Fix onboarding, because early value predicts renewal
Accounts that never reach first value churn regardless of what customer success does in month ten. Onboarding is a retention lever with a long delay on its payoff, which is exactly why it gets under-resourced.
5. Put escalators in contracts
Modest annual uplifts on auto-renewal, commonly in the mid-single digits, add expansion revenue with no sales motion attached. Unglamorous, and it compounds.
6. Segment before you optimise
Blended NRR hides the answer. Split it by ACV band, acquisition channel, and cohort vintage. It is common to find one segment above 120% subsidising another below 85%, which turns a vague retention project into two specific decisions: double down, or stop selling into that segment.
Common NRR mistakes
- Including new customers. The most frequent error, and it inflates NRR badly. A customer who signed in month three was not in the starting cohort and does not belong in the calculation.
- Mixing cohort vintages. Using total prior-year revenue as the denominator instead of the frozen starting cohort's revenue produces a number that is not NRR.
- Switching between MRR and ARR mid-calculation. Pick one and stay in it.
- Comparing across ARPA bands. Your 98% next to a competitor's 125% means nothing if they sell $200K contracts and you sell $500 ones.
- Reporting NRR without GRR. Hides the leaky-bucket case described above.
- Using strong NRR to justify cutting acquisition. Expansion has a ceiling per account. A base that is never replenished eventually stops expanding.
- Counting one-time and services revenue. It is not recurring, and it makes the metric unusable for forecasting.
Net Revenue Retention FAQ
What does NRR stand for?
Net Revenue Retention. It is the percentage of recurring revenue you keep from your existing customer base over a period, counting expansion, downgrades, and cancellations, and excluding customers acquired during that period.
How do you calculate NRR?
(Starting MRR + Expansion − Contraction − Churn) ÷ Starting MRR × 100. $100K start + $30K expansion − $5K contraction − $10K churn = 115% NRR. Or use the cohort method: ending MRR from the starting customer list divided by that list's starting MRR.
Is NRR the same as NDR?
Yes. Net Revenue Retention and Net Dollar Retention are the same metric with the same formula. Private companies and VCs tend to say NDR, public filings tend to say NRR.
What is a good NRR?
It depends on contract size. Bessemer's scale puts 100% at good, 110% better, and 120%+ best for growth-stage cloud companies. The blended private-SaaS median is around 101%, with enterprise near 118% and SMB near 97% (Benchmarkit 2025). Below $15M ARR, ChartMogul's data shows top-quartile net retention at 94% to 99%, so early-stage founders should benchmark against their own stage and ARPA, not against public enterprise vendors.
What’s the difference between NRR and GRR?
NRR includes expansion revenue and can exceed 100%; GRR excludes expansion and caps at 100%. NRR shows growth; GRR shows pure retention strength. Report both, since a wide gap between them means a leaky base masked by a few large expansions.
Can NRR be over 100%?
Yes, and that is the point of the metric. NRR above 100% means expansion from existing customers more than covers everything lost to downgrades and cancellations, which is the same condition as negative churn. GRR cannot exceed 100% by construction.
How often should I calculate NRR?
Monthly for internal trends, quarterly for the board, annual cohorts for investors and benchmarking. To compare a monthly figure to published benchmarks, compound it: monthly NRR to the 12th power, quarterly to the 4th.
Why is my NRR different from a competitor’s reported number?
Usually definitions rather than performance. Check four things: the period (monthly compounded vs annual cohort), whether contraction is separated from churn or folded into it, whether win-backs count as renewals, and whether one-time or services revenue crept in. Also check ARPA, since NRR expectations scale with contract size.
Why do investors care about NRR?
NRR above 100% means the business grows even without new customers, a strong signal of product-market fit and efficient, compounding growth, which drives premium valuation multiples.
Related retention metrics
- Revenue Retention Rate, the overview metric, start here.
- Gross Revenue Retention (GRR), revenue kept from existing customers, excluding expansion.
- Net Revenue Retention (NRR), revenue kept including expansion/upsell. (you are here)
- SaaS Churn & Retention, the pillar guide covering both sides of the equation.
- Expansion revenue and account expansion rate, the numerator behind NRR above 100%.
