Key Takeaways

• A cohort holding $2,400,000 that loses $180,000 to cancellations and $132,000 to downgrades retains $2,088,000, an 87% GRR.

• The metric caps at 100% because each account's ending revenue gets capped at its starting revenue, which makes GRR the retention number investors treat as unfakeable.

• SaaS Capital's survey of 1,500+ private B2B SaaS companies puts median GRR at 91% and calls 90% the table-stakes floor for peer parity.

• Healthy GRR rises with contract size: 90% below $25K ACV, roughly 93% above it. A blended target misleads SMB and enterprise sellers alike.

• Benchmarkit's 2026 data puts median GRR at 84%, down from 88%, and the 75th percentile at 91%. On a $10M ARR base that four-point drop costs about $400,000 a year.

Which revenue belongs in the gross revenue retention numerator and denominator?

The denominator is the recurring revenue your existing customers held at the period's start, and the numerator is that figure minus everything those customers cancelled or downgraded. Customers acquired mid-period never touch either side. GRR asks one question: how much of what you already had did you keep?

Here's what lands where:

Run it on an annual cohort and the shape shows up fast:

1. Forty accounts start the year holding $2,400,000 in ARR, the denominator.

2. Three accounts cancel outright, taking $180,000.

3. Five surviving accounts downgrade, giving back $132,000.

4. Retained revenue lands at $2,088,000, so GRR is $2,088,000 / $2,400,000, or 87%.

5. That cohort also booked $310,000 of expansion, which GRR throws away. Feed it back in and net revenue retention reads 99.9%.

Two numbers, one cohort, and only one shows that 13% of the book walked out. Cancellations from failed payments hit GRR the same as a deliberate non-renewal, so involuntary churn belongs in the same review.

What counts as a healthy GRR for SMB, mid-market, and enterprise accounts?

Healthy GRR tracks contract size rather than company size, and both major benchmark studies segment it that way. SaaS Capital's reason: companies sharing a selling price organize, sell, and support alike, so annual contract value predicts retention better than company age, ARR, or industry.

Median GRR by contract band, from SaaS Capital's private B2B SaaS survey:

Qualifiers matter before you hold a team to those figures. Sub-90% GRR correlates with slower growth: SaaS Capital found those companies below the 34% population median. Contract structure moves the number as much as segment does: month-to-month books sit at 89.5%, annual at 90%, multi-year at 95%. And the market has slipped, with Benchmarkit tracking median GRR from 90% to 88% across 2022 to 2024 and its 2026 report putting it at 84%.

Why is gross revenue retention capped at 100%, and what does that ceiling expose?

The cap is mechanical, not conventional. Each customer's ending revenue gets capped at their starting revenue, so an account that doubled its spend still contributes exactly what it began with. No account returns more than 100% of itself, so the portfolio can't either. The best possible GRR is a period where nobody cancelled and nobody downgraded.

That ceiling is what makes the metric hard to dress up:

• No amount of upselling raises it. A retention number without a ceiling gets lifted by selling more to whoever stayed, and GRR closes that door.

• It separates two stories expansion blends together. A book losing 16% while expanding 16% posts flat NRR and an 84% GRR, and only the second number names the problem.

• Silent contraction shows up here first. A contract renewing at 85% of last year's value never registers as a lost logo, so account-count retention stays clean while GRR slides.

I treat GRR as the ceiling on how ambitious a growth plan can honestly be. Fix the leak, then model the expansion. Compounding runs the wrong way in the other order.

Related terms

Retention sits inside a small cluster of metrics that read each other:

• Net revenue retention adds expansion revenue back and can exceed 100%, which is why boards read it next to GRR rather than instead of it.

• Revenue churn is the inverse view, measuring the recurring revenue lost rather than the share kept.

• Involuntary churn covers cancellations from failed payments, a GRR loss account teams miss.

• Annual contract value is the attribute both benchmark studies segment retention by.

• Contracted ARR is the committed revenue base renewal-heavy GRR calculations start from.

• Revenue leakage covers billing errors that shrink invoices without any customer deciding to spend less.

FAQ

Is gross revenue retention measured monthly or annually?

Most teams report gross revenue retention annually, comparing a cohort's recurring revenue against the same cohort twelve months earlier. Benchmarkit measures it year over year or on a trailing twelve-month basis, and SaaS Capital compares one December against the previous. Monthly GRR spots a bad quarter early, but it's noisy at low account counts.

Does gross revenue retention include price increases?

No. Gross revenue retention excludes price increases, upsells, and cross-sells, because all three raise revenue from a customer who was already there. A renewal uplift shows up in net revenue retention and disappears from GRR. The exclusion is deliberate: it stops a pricing change from masking a book that's losing accounts.

What's the difference between GRR and net revenue retention?

GRR counts only losses, and net revenue retention counts losses plus expansion. The same cohort produces both, so they diverge by exactly the upsell, cross-sell, and price increase you booked from existing customers.

Why do investors ask for gross revenue retention alongside NRR?

Investors read GRR because its 100% ceiling makes it the harder number to flatter. NRR above 100% can come from a few accounts expanding fast while the long tail quietly leaves, and GRR prices that tail. Diligence seeing 120% NRR next to 82% GRR knows the growth rests on a few accounts.