Gross vs net revenue retention: why you need both curves

Gross revenue retention (GRR) measures how much recurring revenue you keep from existing customers after churn and downgrades, before you count a single dollar of expansion. That makes it your floor. It can never exceed 100 percent.

Net revenue retention (NRR) starts from the same base, then adds upsells and expansion, so it can climb past 100 percent. That makes it your growth story.

Here is the trap most founders walk into. NRR above 100 percent can look healthy while it hides brutal logo and gross churn. A handful of expanding accounts paper over many that left.

You need both curves, not one. GRR tells you whether the bucket holds water. NRR tells you whether your biggest accounts are filling it faster than the rest drain it.

GRR vs NRR in one sentence

GRR is how much revenue you keep without selling anything new to existing customers. NRR is GRR plus expansion.

The only difference between the two formulas is one term: Expansion MRR. NRR adds it. GRR does not.

That single term is why they answer different questions. GRR asks whether the product holds revenue on its own. NRR asks whether the base compounds when you upsell it.

Because GRR excludes expansion, it is capped at 100 percent by definition. There is no way to hide a churn problem behind a few whales. NRR has no ceiling.

The formulas, spelled out

Both formulas start from the same point: MRR from a fixed group of existing customers at the start of the period. You never add revenue from new logos signed during the period.

GRR = (Starting MRR − Churned MRR − Contraction MRR) / Starting MRR × 100

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

Churned MRR is revenue from customers who cancelled outright. Contraction MRR is revenue lost to downgrades and seat reductions, tracked as contraction MRR. Expansion MRR is upsells, added seats, and usage growth, tracked as expansion revenue.

Two facts fall out of these formulas.

First, NRR is always greater than or equal to GRR, because the only thing separating them is a positive expansion term. If your tool reports an NRR below your GRR, the math is wrong.

Second, your gross MRR churn rate is just the mirror of GRR:

Gross MRR churn = (Churned MRR + Contraction MRR) / Starting MRR × 100 = 100% − GRR

So a GRR of 90 percent means you are losing 10 percent of revenue to churn and downgrades before any expansion. For the full breakdown of gross revenue retention and net revenue retention, the glossary entries work each one end to end.

The trap: how a strong NRR hides weak GRR

NRR above 100 percent does not mean retention is healthy. It means expansion outweighed churn in dollars, which is a different claim entirely.

The reason is that expansion is lumpy. A few large accounts adding seats can add more MRR than dozens of small accounts subtracted by cancelling. So the headline number looks like growth while the underlying base shrinks on a logo basis.

This is why logo churn, the percentage of customers lost regardless of size, can run high while NRR sits above 100. Revenue-weighted metrics give your biggest customers the loudest vote.

A company can post 120 percent NRR while losing 20 percent of its logos in a year. The board slide says growth engine. The customer base says churn crisis. Both are reading the same month.

A worked case where they diverge

Same month, two metrics, opposite conclusions. Here is how that happens.

You start with $100,000 in MRR across 100 customers paying $1,000 each.

During the month, 18 customers cancel outright. That is $18,000 in churned MRR and 18 percent logo churn. No surviving accounts downgrade, so contraction MRR is zero here and that term drops out of both formulas.

Meanwhile, a few of your largest survivors expand. They add seats and upgrade tiers, for $22,000 in new expansion MRR.

Run GRR first. Expansion is excluded:

GRR = ($100,000 − $18,000) / $100,000 × 100 = 82 percent

Now run NRR. Add the $22,000 of expansion:

NRR = ($100,000 − $18,000 + $22,000) / $100,000 × 100 = 104 percent

Metric	Value	What it says
NRR	104%	Revenue grew from existing customers with zero new sales
GRR	82%	You kept only 82 cents of every revenue dollar
Logo churn	18%	Nearly one in five customers walked
NRR minus GRR	22 points	Expansion from a few accounts is masking the rest

The NRR reads like growth without selling. The GRR and the logo churn say the bucket is leaking badly and a few expanding accounts are bailing it out. If those whales stop expanding next quarter, the 104 collapses toward the 82, and you learn your real retention was 82 all along.

That is the whole argument for charting both. One number lied. The other told the truth.

2025-2026 retention benchmarks

The 2025 private SaaS medians cluster around NRR 106 percent and GRR 90 to 92 percent. So a typical company loses 8 to 10 percent of revenue to churn and contraction before expansion does any lifting.

The most useful way to read retention benchmarks is by ACV (average contract value) first. Bigger contracts retain better, almost regardless of how big the company is.

Segment (by ACV)	Median NRR	Top performers
Enterprise (ACV above $100K)	~118%	130%+
Mid-market ($25K-$100K)	~108%	120%+
SMB (ACV below $25K)	~97%	108%+

A 97 percent NRR for an SMB-focused product is normal, not a red flag. Holding small accounts is structurally harder. Judge yourself against your ACV peers, not against an enterprise number you will never hit.

Here is what good looks like on each curve, pulled from SaaS Capital’s 2025 set and the 2025 KeyBanc/Sapphire SaaS survey:

Grade	GRR	NRR
Best-in-class	92-95%+	120%+ (enterprise 135%+)
Good	90%+	110-120%
Acceptable	85%+	100-110%
Warning	below 85%	below 100%

SaaS Capital reports roughly 91 percent median GRR for bootstrapped firms in the $3M to $20M ARR band, with median NRR near 103 percent and the 90th percentile around 118. The 2025 KeyBanc survey lands at roughly 101 percent net dollar retention and 90 percent gross dollar retention. The same 10-point expansion gap shows up in both. Our net revenue retention benchmarks post breaks NRR down by stage.

The one-number diagnostic: read GRR first

Read GRR before NRR, every time. GRR tells you whether the foundation holds. NRR tells you how much you can build on it. A great NRR on a weak GRR is a building on sand.

Then run one subtraction:

NRR − GRR = the expansion gap, in points.

If that gap exceeds 20 to 30 points, expansion from a few accounts is masking churn in the many. NRR of 130 with GRR of 85 is a 45-point gap and a flashing light, even though 130 percent NRR looks elite. The typical healthy gap is closer to 10 points. A 10-point gap on a 90 GRR gives you a 100 NRR you can trust, because the floor underneath it is solid.

A wide gap is not always fatal. But it tells you exactly where to look: at your churn and contraction, not your expansion. The growth you are celebrating is concentrated, and concentration is fragile.

Which matters more, GRR or NRR?

Neither wins. They answer different questions, and using one without the other is how founders get blindsided.

GRR is the truth-teller. It is capped at 100 percent, so there is nowhere for a churn problem to hide. It tells you whether your product holds revenue without an upsell motion propping it up.

NRR is the narrative and the valuation driver. Investors reward it directly. Public SaaS companies above 120 percent NRR have traded around 9.3x EV/revenue, versus roughly 3.1x for those below 100 percent (Software Equity Group). A 10-point NRR gain can lift valuation 20 to 30 percent with no change to ARR.

But that premium only holds on a solid GRR floor. Expansion eventually slows, and a fragile base reverts to its real retention rate, which is your GRR. The companies that compound for years carry both a high floor and a high ceiling. Our SaaS valuation and ARR multiples breakdown shows where retention plugs into the wider picture.

Track both curves in real time

The reason this gap goes unnoticed is timing. Most teams calculate GRR and NRR once a quarter for a board deck, by which point the leaky bucket has been draining for months.

Mowt charts both curves side by side straight from your Stripe data, so the moment NRR and GRR start to diverge you see it, not at the next board meeting. A strong NRR can never quietly hide a weak GRR when both lines sit on the same screen.

If you want to run the math yourself first, the net revenue retention calculator and the churn rate calculator take your starting MRR, expansion, contraction, and churn and return both numbers.

FAQ

What is the difference between GRR and NRR?

GRR measures the share of existing recurring revenue you keep after churn and downgrades, excluding expansion, so it can never exceed 100 percent and acts as your retention floor. NRR starts from the same base but adds expansion and upsell revenue, so it can exceed 100 percent. The only formula difference is the Expansion MRR term: NRR includes it, GRR does not, which makes NRR always greater than or equal to GRR.

Can NRR be above 100 percent with bad retention?

Yes, and this is the core trap. NRR can sit above 100 percent while you lose a large share of customers, because a few expanding accounts can outweigh the revenue from many that churned. NRR of 104 percent can coexist with GRR of 82 percent and 18 percent logo churn. If NRR looks healthy but GRR is below 85 percent, expansion is masking a retention problem.

What is a good GRR?

A good GRR is above 90 percent, with above 85 percent acceptable and best-in-class enterprise companies at 92 to 95 percent or higher. The 2025 private SaaS median runs 90 to 92 percent: SaaS Capital reports roughly 91 percent for bootstrapped firms, and the 2025 KeyBanc survey shows about 90 percent gross dollar retention. GRR below 80 percent signals a fundamental retention failure that no amount of expansion fully fixes.

Which matters more, GRR or NRR?

You need both. GRR is the truth-teller (can the product hold revenue without upsells?) and NRR is the growth story (does expansion compound the base?). Investors pay premium multiples for high NRR, but a strong NRR built on a weak GRR is fragile, because expansion eventually slows and the underlying churn surfaces. Read GRR first to confirm the foundation, then NRR to size the upside.

What is a good NRR benchmark in 2025-2026?

Above 100 percent means you are growing revenue from existing customers net of all churn. The private SaaS median is about 106 percent, good is 110 to 120 percent, and best-in-class is 120 percent or higher (enterprise 135 percent or higher). Benchmark by ACV: enterprise medians near 118 percent, SMB near 97 percent, so a 97 percent SMB NRR is normal, not a warning.