What Is NRR in SaaS? The Net Revenue Retention Formula Explained

What Is NRR in SaaS? The Net Revenue Retention Formula Explained - hero image

If you want to know what NRR in SaaS is in one sen­tence: NRR (Net Rev­enue Reten­tion) is the per­cent­age of recur­ring rev­enue you keep from your exist­ing cus­tomers over a year — after sub­tract­ing can­cel­la­tions and down­grades, and adding back any upsells and expan­sion. Above 100% means your exist­ing cus­tomer base spends more this year than last year, even if you nev­er sign a sin­gle new cus­tomer. Below 100% means it shrinks. That one num­ber is also among the first three or four met­rics any acquir­er or investor asks for, because it tells them whether your rev­enue com­pounds or decays on its own.

Here is why that mat­ters more than most founders real­ize. A SaaS com­pa­ny with net rev­enue reten­tion above 100% has, in effect, a growth engine that runs with­out fuel. I once worked with a founder run­ning a rough­ly $10M-a-year busi­ness he found bor­ing — no chaos, no inno­va­tion, noth­ing to fight. When he final­ly cal­cu­lat­ed his NRR, it came back at 140%. I did the math in my head: at 140% reten­tion, with zero new cus­tomers, that “bor­ing” $10M busi­ness 10x’s to $100M in rev­enue in well under a decade — rough­ly sev­en years of pure com­pound­ing off the exist­ing base. He had been ignor­ing the most valu­able asset he owned.

This arti­cle answers “what is NRR in SaaS” from the ground up — the plain-Eng­lish def­i­n­i­tion, the exact for­mu­la, a worked exam­ple using real­is­tic num­bers at $10M ARR, cur­rent bench­marks with the con­text that makes them use­ful, the mis­takes that qui­et­ly cor­rupt the num­ber, and a short FAQ. If you already know the basics and want the strate­gic play­book — how NRR dri­ves val­u­a­tion and how to push it high­er — read the deep­er guide on net rev­enue reten­tion. This piece is for get­ting the def­i­n­i­tion right first.


What NRR Actually Measures

NRR (Net Rev­enue Reten­tion) answers a decep­tive­ly sim­ple ques­tion: take the group of cus­tomers you had at the start of a peri­od — usu­al­ly 12 months ago — and ignore every­one who signed up after that. How much recur­ring rev­enue is that same group gen­er­at­ing today, com­pared to what they gen­er­at­ed back then?

The word “net” is the impor­tant part. NRR nets togeth­er every­thing that hap­pened to that cohort:

  • Some cus­tomers churned — they can­celled entire­ly, tak­ing their rev­enue with them.
  • Some cus­tomers con­tract­ed — they down­grad­ed to a cheap­er plan or dropped seats, so they still pay you, just less.
  • Some cus­tomers expand­ed — they upgrad­ed, added seats, or bought more, so they pay you more than they did a year ago.

If the expan­sion from the cus­tomers who stayed and grew out­weighs the loss­es from the ones who left or shrank, your NRR is above 100%. If it does­n’t, you’re below 100%.

The rea­son NRR is so close­ly watched is what it implies about the future. NRR above 100% means your exist­ing base is a self-fund­ing growth machine — it grows on its own, before you spend a dol­lar acquir­ing any­one new. This is the engine behind com­pa­nies like Slack and Drop­box: once a team is in, they only add more seats and more data over time. There is no real­is­tic sce­nario where they use Slack less next year. That dynam­ic is what makes the life­time val­ue of those accounts effec­tive­ly infi­nite, and it’s why those busi­ness­es com­mand the mul­ti­ples they do.


The NRR Formula

Here is the for­mu­la, writ­ten out explic­it­ly:

NRR = (Start­ing MRR + Expan­sion MRR − Con­trac­tion MRR − Churned MRR) ÷ Start­ing MRR × 100%

A quick def­i­n­i­tion of each term, because the for­mu­la is only as good as your under­stand­ing of the inputs:

  • Start­ing MRR is the Month­ly Recur­ring Rev­enue (MRR) — the pre­dictable month­ly sub­scrip­tion rev­enue — from your cohort of exist­ing cus­tomers at the start of the peri­od.
  • Expan­sion MRR is the addi­tion­al month­ly rev­enue those same cus­tomers added through upgrades, seat addi­tions, or cross-sells.
  • Con­trac­tion MRR is the month­ly rev­enue lost when those cus­tomers down­grad­ed but did not leave.
  • Churned MRR is the month­ly rev­enue lost when those cus­tomers can­celled entire­ly.

You can run the same cal­cu­la­tion on Annu­al Recur­ring Rev­enue (ARR) — your year­ly sub­scrip­tion rev­enue — instead of MRR. The result is iden­ti­cal as long as you stay con­sis­tent and use the same unit for every term. Whether you use MRR or ARR, the rule that breaks the for­mu­la is mix­ing the two. (Some sources, like Wall Street Prep’s NRR break­down, fold con­trac­tion into the churn term; the math is the same either way, but sep­a­rat­ing the two makes down­grades eas­i­er to see and fix.)

The sin­gle most impor­tant thing the for­mu­la does not include is new cus­tomer rev­enue. NRR mea­sures the exist­ing base only. The moment you let rev­enue from cus­tomers acquired dur­ing the peri­od sneak into the numer­a­tor, you are no longer mea­sur­ing reten­tion — you’re mea­sur­ing growth, and you’ll fool your­self into think­ing a leaky busi­ness is healthy. More on that mis­take below.


Net Revenue Retention worked example — a seedling growing inside a sealed glass vessel with no outside inputs

A Worked Example at $10M ARR

Num­bers make this con­crete. Con­sid­er a B2B SaaS com­pa­ny at $10M ARR — square­ly in the range where this ques­tion actu­al­ly mat­ters. To keep the arith­metic clean, that’s rough­ly $833,333 in MRR at the start of the year ($10M ÷ 12).

Over the next 12 months, here is what hap­pens to that start­ing cohort — and only that cohort, ignor­ing any new logos signed dur­ing the year:

ComponentMonthly AmountWhat Happened
Starting MRR$833,333The existing base, 12 months ago
Expansion MRR+$125,000Upgrades, added seats, cross-sells
Contraction MRR−$33,333Downgrades from customers who stayed
Churned MRR−$58,333Customers who cancelled entirely
Ending MRR (same cohort)$866,667What the original base pays now

Plug those into the for­mu­la:

NRR = ($833,333 + $125,000 − $33,333 − $58,333) ÷ $833,333 × 100% = $866,667 ÷ $833,333 × 100% = 104%

So this com­pa­ny has 104% net rev­enue reten­tion. The exist­ing base grew about 4% on its own over the year, before any new sales. That’s a healthy, self-sus­tain­ing num­ber — mod­est, but it means the busi­ness is not run­ning uphill.

It’s worth not­ing what Gross Rev­enue Reten­tion (GRR) would be for the same com­pa­ny. GRR mea­sures the same cohort but excludes expan­sion — it only counts what you kept before any upsells:

GRR = ($833,333 − $33,333 − $58,333) ÷ $833,333 × 100% = $741,667 ÷ $833,333 × 100% = 89%

The 15-point gap between 89% GRR and 104% NRR is the con­tri­bu­tion of expan­sion. That gap is one of the most diag­nos­tic things you can look at — it’s cov­ered in depth in the com­par­i­son of gross rev­enue reten­tion ver­sus net.


A worked example of Net Revenue Retention — abstract translucent stacked glass tiers of increasing height and brightness

NRR Benchmarks (and the Context That Makes Them Useful)

A raw NRR num­ber means lit­tle with­out a bench­mark — and bench­marks mean lit­tle with­out con­text. Here is the gen­er­al inter­pre­ta­tion scale:

NRRInterpretation
Below 90%Leaky bucket — the base is in net contraction
90–100%Stable, but the base isn't growing on its own
100–110%Healthy — the existing base grows without new sales
110–130%Strong — expansion-driven growth
Above 130%Elite — a powerful upsell and cross-sell engine

For 2026, the medi­an NRR for pri­vate B2B SaaS sits in the low-to-mid 100s. SaaS Cap­i­tal’s 2026 bench­mark­ing sur­vey of more than 1,000 pri­vate B2B SaaS com­pa­nies puts the medi­an NRR for boot­strapped firms in the $3M–$20M ARR range at 103%, with 90th-per­centile per­form­ers reach­ing about 118%. But the medi­an hides enor­mous vari­a­tion by cus­tomer seg­ment, and this is where most founders mis­read their own num­ber:

Segment (by deal size)Typical Median NRR
Enterprise (ACV above $100K)~118%, top performers 130–135%
Mid-market ($25K–$100K ACV)~108%
SMB (under $25K ACV)~97%

The same NRR means dif­fer­ent things at dif­fer­ent scales. Hold­ing 100%+ at SMB scale is gen­uine­ly strong — small cus­tomers churn more and expand less, so stay­ing flat is an accom­plish­ment. At enter­prise scale, 100% would be a warn­ing sign, and even 120% is mere­ly “good.” Before you cel­e­brate or pan­ic over your NRR, com­pare it to com­pa­nies with a sim­i­lar cus­tomer size, pric­ing mod­el, and con­tract struc­ture. A usage-based busi­ness will nat­u­ral­ly post high­er NRR than a flat-seat busi­ness; com­par­ing the two tells you noth­ing.

A note on the num­bers: Bench­mark fig­ures shift year to year and vary by source and method­ol­o­gy. The ranges above reflect 2025–2026 con­di­tions and are includ­ed to show rel­a­tive dif­fer­ences across seg­ments, not as fixed tar­gets. Pull the lat­est data for your spe­cif­ic seg­ment before set­ting goals against it.


The Mistakes That Corrupt Your NRR

NRR is sim­ple to state and sur­pris­ing­ly easy to cal­cu­late wrong. These are the errors I see most often, and each one makes a strug­gling busi­ness look health­i­er than it is.

  1. Count­ing new cus­tomers in the numer­a­tor. This is the car­di­nal sin. NRR tracks the exist­ing cohort only. If you fold in rev­enue from cus­tomers acquired dur­ing the peri­od, you’ve turned a reten­tion met­ric into a growth met­ric and lost the entire point. The whole val­ue of NRR is that it iso­lates whether your base grows on its own.
  2. Ignor­ing con­trac­tion. A cus­tomer who down­grades from $1,000 to $600 a month is not “retained” — they rep­re­sent $400 of month­ly con­trac­tion. Treat­ing par­tial loss­es as zero qui­et­ly inflates the num­ber.
  3. Let­ting NRR hide a churn prob­lem. A com­pa­ny can post 120% NRR while its GRR sits at 75% — mean­ing it los­es a quar­ter of its base rev­enue every year and papers over the hole with expan­sion from a hand­ful of accounts. Always read NRR and GRR togeth­er. NRR alone can mask a leaky buck­et. If churn is the real issue, the fix lives in the work on reduc­ing SaaS churn, not in more upsells.
  4. Mix­ing time peri­ods or units. Com­pare a month­ly fig­ure against an annu­al one, or MRR against ARR, and the cal­cu­la­tion is mean­ing­less. Pick one peri­od (12 months is stan­dard) and one unit, and hold them con­stant.
  5. Blend­ing all seg­ments into one num­ber. A com­pa­ny-wide NRR aver­ages your healthy enter­prise cohort with your leaky SMB cohort and hides both. Seg­ment it — by deal size, ver­ti­cal, and acqui­si­tion chan­nel. In my expe­ri­ence, 100% of the time there are sig­nif­i­cant vari­ances across seg­ments, and the blend­ed num­ber con­ceals the ones that mat­ter.

Why acquirers value Net Revenue Retention — a magnifying glass focused on a single glowing gemstone among plain stones

Why Acquirers Care So Much About NRR

NRR is not just an oper­at­ing met­ric — it’s a val­u­a­tion lever. When an investor or acquir­er eval­u­ates a SaaS busi­ness, NRR is among the first num­bers they request, because com­bined with your growth rate and gross mar­gin they can ball­park what the busi­ness is worth in about ten min­utes.

The log­ic is straight­for­ward. A busi­ness with NRR above 100% is de-risked — it will be larg­er next year even if sales stalls com­plete­ly, because the exist­ing base keeps expand­ing. A busi­ness below 100% is the oppo­site: it has to win new cus­tomers just to stand still, and any dis­rup­tion to its sales engine means decline. Acquir­ers pay a pre­mi­um for the first kind and a dis­count for the sec­ond. The spread is real — com­pa­nies in the 100–110% NRR band tend to com­mand mean­ing­ful­ly low­er rev­enue mul­ti­ples than those above 120%.

This con­nects direct­ly to how you should think about build­ing toward an exit. NRR is one of the clean­est sig­nals of the kind of recur­ring-rev­enue qual­i­ty that dri­ves a high mul­ti­ple — a theme that runs through how NRR com­pares to ARR as a mea­sure of busi­ness health. If you’re build­ing to sell, NRR isn’t a met­ric you mon­i­tor; it’s an asset you cul­ti­vate.


Frequently Asked Questions

What does NRR stand for in SaaS? NRR stands for Net Rev­enue Reten­tion. It mea­sures the per­cent­age of recur­ring rev­enue retained from your exist­ing cus­tomer base over a peri­od (typ­i­cal­ly 12 months), after account­ing for churn and down­grades and includ­ing upsells and expan­sion.

What is a good NRR in SaaS? Above 100% is the thresh­old for “healthy” — it means your exist­ing base grows on its own. The 2026 medi­an for pri­vate boot­strapped B2B SaaS at $3M–$20M ARR is about 103%. But “good” depends on seg­ment: 100% is strong for SMB-focused busi­ness­es, while enter­prise-focused com­pa­nies should tar­get 110%+ and elite per­form­ers exceed 130%.

What’s the dif­fer­ence between NRR and GRR? GRR (Gross Rev­enue Reten­tion) excludes expan­sion rev­enue — it only mea­sures how much of your exist­ing rev­enue you keep before any upsells, so it can nev­er exceed 100%. NRR includes expan­sion, so it can rise above 100%. Read­ing them togeth­er reveals whether healthy NRR is hid­ing a churn prob­lem under­neath.

Does NRR include new cus­tomers? No. This is the most com­mon mis­take. NRR mea­sures only the cohort of cus­tomers you had at the start of the peri­od. Rev­enue from cus­tomers acquired dur­ing the peri­od is exclud­ed entire­ly — includ­ing it turns NRR into a growth met­ric and defeats its pur­pose.

How is NRR cal­cu­lat­ed? NRR = (Start­ing MRR + Expan­sion MRR − Con­trac­tion MRR − Churned MRR) ÷ Start­ing MRR × 100%. You can use ARR instead of MRR as long as you use the same unit for every term and mea­sure over a con­sis­tent peri­od.

What time peri­od should NRR cov­er? Twelve months is the stan­dard, because it’s the most intu­itive to bench­mark and it smooths out sea­son­al swings. The crit­i­cal rule is con­sis­ten­cy: use the same cohort for the start­ing and end­ing mea­sure­ment, and keep the peri­od and unit fixed.


The Bottom Line

NRR in SaaS is the per­cent­age of recur­ring rev­enue you keep and grow from your exist­ing cus­tomers over a year. Above 100% and the base com­pounds on its own; below 100% and it decays, forc­ing you to acquire new cus­tomers just to stay flat. Cal­cu­late it on the exist­ing cohort only, read it along­side GRR, and seg­ment it before you trust the blend­ed num­ber. Get those three things right and you’ll have one of the most hon­est sig­nals avail­able about whether your busi­ness is the kind that com­pounds — the kind acquir­ers pay the most for.

Facebooktwitterlinkedinmail
author avatar
Vic­tor Cheng
Author of Extreme Rev­enue Growth, Exec­u­tive coach, inde­pen­dent board mem­ber, and investor in SaaS com­pa­nies.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top