Gross vs Net Revenue Retention: What Each Number Is Telling You

Gross revenue retention (GRR) measures how much of your existing monthly recurring revenue (MRR) survived the period without counting upsells or seat adds. Net revenue retention (NRR) adds expansion back in and can push past 100% where GRR never can.

I've spent the last few years consulting on retention for SaaS, e-commerce, and subscription businesses. The most consistent pattern is companies optimizing for save rate without asking why people leave.

Track both, and you'll know which one is the problem and what to do about it.

Key takeaways

• GRR measures the revenue you keep; NRR shows whether expansion MRR covers it.
• Private SaaS typically have 85% GRR and 100%+ NRR at maturity.
• A 100%+ NRR over a falling GRR means expansion is masking a churn problem.
• GRR and NRR converge for flat-rate products with no upgrade path.
• Stage changes the targets, so read your numbers against companies your size.

What is the difference between gross and net revenue retention?

Gross revenue retention measures the recurring revenue you keep from existing customers, before any upsells or seat adds. It counts only what survived cancellations and downgrades. Net revenue retention adds expansion revenue back in, which means GRR can never exceed 100% while NRR can.

GRR tells you what churn is costing you. NRR tells you whether expansion is covering it. One number without the other gives you half the picture.

The two formulas differ by a single term. NRR's numerator includes expansion MRR, where GRR's stops before it. Expansion adds revenue the period started without, so NRR can push past 100%.

GRR is capped there by construction, because the best you can do without expansion is keep it all.

Both formulas start from the same cohort and the same starting MRR. One just stops counting before the expansion line.

The two numbers converge when a business has no expansion revenue at all.

A flat-rate product with no usage-based pricing, no upgrade path, and no seats to add has nothing to put in the expansion term. Both formulas then produce the same result, which is just GRR by another name.

The gap opens up only once existing customers have a real way to spend more.

GRR formula

GRR strips expansion out of the math, so the formula only ever subtracts. You take starting MRR, subtract churned MRR and contraction MRR, then divide the result by starting MRR. Written out, the calculation is:

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

Starting MRR is the cohort's recurring revenue at the start of the period. The cohort is fixed at that moment, so any customer you sign during the period stays out of the math.

Churned MRR is what you lose when customers cancel outright. Contraction MRR is what you lose when existing customers downgrade, drop seats, or cut usage without leaving.

Note what's missing. Expansion never enters the formula at all. The metric is built to ask one question, which is how much of what you already had you managed to keep.

Say your cohort starts at $10,000 MRR, loses $800 to cancellations and $200 in contraction MRR. That's 10% of the starting base gone, so GRR is 90%. No expansion figure touches that calculation, even if the cohort added upgrades that month.

NRR formula

NRR uses the same cohort and the same subtractions, then adds expansion MRR back into the numerator. The full derivation and a worked example live in the net revenue retention glossary entry, so this article won't re-run them.

Take the same cohort from the GRR example. It added $1,500 through upsells on top of the same $1,000 in losses. GRR stays 90%, but NRR is 105%, because the $1,500 of expansion MRR more than offsets the loss. The expansion is the only input that changed.

The practical point of running both is that the spread between them is information.

GRR alone tells you the loss, and NRR alone tells you the net result. Only the two together show how much expansion you needed to get from one to the other. That spread is what warns you when expansion is doing too much of the work.

Side-by-side comparison

The table below sets both metrics next to each other on the dimensions that decide which one to act on:

GRR is the baseline, NRR is the expansion cap

GRR shows the minimum revenue your existing base keeps once expansion is stripped out. That number is the baseline your upsells build on, and NRR shows how far expansion lifts you above it. Look at GRR first, because it's the number you can always control.

A business with no way to expand lives entirely on the baseline. Its whole retention shows up in that one number. Add an upsell motion and a gap opens between GRR and NRR. The size of that gap tells you how much of your churn expansion is covering.

A wide gap means expansion is covering a lot of churn. A narrow one means the GRR is most of your story.

Two cases make the gap concrete. Picture two businesses, each adding 20 points of expansion but starting from different GRR:

Same expansion effort, two different outcomes, because the baseline underneath it differs.

The floor-and-upside split only holds with a real expansion path. A consumer app at $5 to $15 a month on a flat rate has no expansion MRR to measure. There are no seats to add, no tiers to climb, and no usage to meter.

GRR and NRR converge there, so GRR is the only diagnostic that matters. A flat-rate product reads its retention through GRR alone, while a seat-based B2B tool watches both.

When healthy NRR hides a GRR problem

A business can post NRR above 100% and a falling GRR at the same time. When expansion outpaces your churn losses, NRR looks healthy even while GRR is dropping underneath it.

Folks should be able to spot this, because the better-looking number is the one hiding the problem.

The masking happens because of the order the formula works in. NRR adds expansion before you ever see the subtraction from churn. A big enough upsell motion can produce a strong NRR even as cancellation rates climb.

An NRR well above 100% is called negative churn, the state where expansion outweighs all losses.

The weak GRR is still there. It's just covered.

The moment upsell momentum slows, the GRR you ignored shows up directly in NRR.

Suppose your GRR slides from 88% to 80% over two years while your NRR holds steady at 105%. That's a structural problem wearing a healthy disguise. You're losing 20% of your base revenue every year to churn and downgrades, and only the upsell motion keeps NRR positive.

If one quarter of expansion underperforms, NRR can fall below 100% before you can even diagnose why.

This trap needs a real upsell surface to spring. A low-priced subscription with no upgrade path can't mask a GRR problem, because the two numbers converge. There, NRR falling below 100% is a direct read on the GRR problem. No expansion exists to paper over it.

How to tell if this is happening to you

The tell is a divergence, where NRR stays flat or rising while GRR drifts down over several quarters. To catch it, plot GRR and NRR on the same timeline and track them as a pair, the way you would any linked retention metrics.

A single healthy NRR reading tells you nothing about the GRR underneath.

Trend both numbers over at least four quarters and watch the gap between them. A gap that widens because GRR is falling is the warning sign. The diagnosis then points to the fix.

If expansion is the only thing holding NRR up, the work is on the GRR, which means cutting the cancellations and downgrades that GRR measures.

Run one more check before you act. Split your GRR loss into its two parts, churned MRR and contraction MRR, since each points to a different fix.

A GRR slipping because of cancellations is an activation or value problem. If it goes down because of downgrades is usually a pricing or packaging problem. The blended number hides which one you have, so pull the split before you decide where to spend.

What moves GRR (and what doesn't)

Only two things move the GRR baseline, and they are fewer cancellations and fewer downgrades. Expansion can't lift GRR, because expansion is exactly what the formula leaves out.

Every gain has to come from keeping revenue you already have. Both levers live in the cancel and downgrade path, well away from the sales pipeline.

The first lever is recovering cancellations before they complete. A customer heading for the cancel button is lost revenue unless you intercept the decision and keep it on the floor.

Churn.io cancellation recovery works that lever directly. It catches would-be cancels in the cancel flow with the right offer at the moment of the decision. The second lever is converting downgrades into smaller saves, where a customer who would have cut their plan stays closer to their original spend.

What doesn't move GRR is anything on the expansion side. More seats sold, usage billed, upgrades closed, all of it raises NRR and leaves GRR untouched. That's the point of the metric. GRR isolates these factors so you can see whether retention is improving on its own.

One churn problem sits outside both levers. When customers leave because their own company shut down or got acquired, no save offer reaches them, and the GRR loss is structural.

The fix there is upstream, in who you sell to. The cancel screen can't reach a company that no longer exists.

GRR and NRR benchmarks: what good looks like at your stage

For private SaaS, ChartMogul's 2023 data puts a healthy GRR around 85% and a healthy NRR at 100% or higher. Both targets shift with your stage, because the expansion surface grows as the customer base matures. Read your own numbers against companies your size.

The 85% mark is a broad target, not a hard line. At smaller scale, even the top quartile runs a little under it. The figures below come from two private-SaaS datasets:

That last row is why NRR becomes the dominant growth metric at scale. Benchmarkit's 2025 data shows more than half of new revenue at $50M ARR comes from existing accounts. A GRR-only read misses the larger half of the story.

Benchmarkit's benchmarks show a $2M ARR company with no customer success team can run 85% GRR and post no NRR above 100% and still be healthy. The same numbers at $50M ARR would be a warning.

Use the NRR calculator to model your stage, and the net revenue retention guide for the levers.

GRR also tracks with growth, though as an association rather than a lever you pull. In ChartMogul's data, companies above 85% GRR grew at 40.1% a year on average, against 20.2% for the 60 to 75% range. Growing products tend to retain better, so the two rise together.

What does 120% NRR mean?

A 120% NRR means the existing base generated 20% more revenue this period than last, after all churn and downgrades, with no new customers counted. Expansion outran every loss by 20 points of the starting cohort.

Public SaaS leaders report figures like this. Atlassian disclosed 120%+ cloud NRR in FY2024.

That number isn't a meaningful target for an early-stage startup, though.

According to Atlassian's 2024 10-K, that 120% rests on seat-based pricing and a large installed base to average over. Scale and upsell surface decide what "good" means. Treat a public leader's NRR as what good looks like at scale.

The practical read of 120% NRR is that the installed base compounds on its own. Atlassian's 2024 data shows a company at that level could stop signing new logos for a year and still grow revenue 20%. That compounding is why investors weight the figure so heavily.

Can GRR exceed 100%?

No. GRR can't exceed 100%, because the formula has no term that adds revenue. It only ever subtracts churned and contraction MRR from the starting base, so the best case is keeping the full 100%.

The moment a customer cancels or downgrades, GRR drops below parity for the period.

That structural cap is the whole reason GRR and NRR are different metrics. NRR can clear 100% because it adds expansion MRR back in. If you see a "GRR" above 100% on a dashboard, the tool is folding expansion into the number, which makes it NRR wearing the wrong label.

A perfect 100% GRR is close to theoretical, since it would mean zero cancellations and zero downgrades all period. ChartMogul's research puts even strong B2B companies in the low-to-mid 90s, so treat that range as a healthy GRR range rather than a failure to hit 100%.

FAQ

What is the relationship between GRR and NRR?

GRR is always less than or equal to NRR, since the two share the same cohort and subtractions while NRR also adds expansion MRR. NRR equals GRR plus that expansion effect, measured as a percentage of starting MRR.

What does 100% NRR mean?

A 100% NRR means expansion exactly canceled out your losses, so the existing base ended the period worth what it started. Below 100% the base shrinks and acquisition has to refill it, while above 100% it grows on its own.

Are GRR and gross dollar retention (GDR) the same thing?

Yes, gross dollar retention (GDR) and gross revenue retention (GRR) are two names for one metric, calculated identically. GDR shows up more in finance-led reporting, while GRR is common in product and growth contexts.