What is GRR?
Gross MRR Retention (GRR) is the percentage of recurring revenue retained from existing customers over a period, counting only losses — churn and contraction — and ignoring expansion. GRR is always capped at 100%. It shows your retention floor: how much revenue you’d keep if no customer ever upgraded.
GRR answers a different question than NRR. NRR asks “did expansion outpace losses?” GRR asks “how much of the base did we hold?” A company with 115% NRR and 85% GRR has strong expansion masking significant base erosion. If expansion slowed, revenue would fall to the 85% floor.
Investors care about both metrics, but GRR tells them how durable the business is without relying on upsell.
The GRR formula
Worked example
July 2026: Your existing customers started the month at $50,000 MRR.
| Movement | Amount |
|---|---|
| 2 customers downgraded plans | −$600 |
| 4 customers canceled fully | −$3,200 |
| 1 same-period signup canceled (joined-and-churned) | −$200 (excluded) |
| The $200 joined-and-churned MRR is subtracted from churn because it was never in the starting base. | |
At 92.8%, you retained 92.8% of base revenue before any expansion. The $600 contraction and $3,000 effective churn (after netting out the joined-and-churned MRR) reduced the base by 7.2%. Expansion MRR and reactivation MRR from the same period are not included — those appear in NRR only.
How it’s computed
GRR uses the same state-comparison engine as NRR but strips out anything positive. It compares your subscription base at the start and end of each period:
Why contraction is included
GRR counts both churn (full cancellations) and contraction (downgrades). A customer going from $500/mo to $100/mo is an 80% revenue loss even though they didn’t cancel. Both North Metric and ChartMogul include contraction in GRR — some older tools excluded it, but the industry standard is clear: GRR measures all base-revenue losses, not just cancellations.
The joined-and-churned adjustment
Subscriptions created and canceled within the same period are excluded from GRR’s churn component. GRR is a cohort metric — it measures retention of the starting cohort. A subscription that was never in the starting MRR base can’t churn from it.
Cross-validation
GRR has been cross-validated against ChartMogul across multiple months. Results: exact match or rounding-level agreement (within 0.04 percentage points) on all comparable periods. Both tools use the same formula and the same contraction-inclusion convention.
Which customers move GRR
GRR shows how much base revenue you held, not which customers are about to cancel or downgrade. North Metric works at the customer level: every paying customer gets a 0–100 churn risk score and is ranked by money at risk, so the accounts most likely to lower GRR show up while there’s still time to act.
GRR vs NRR — when to use which
GRR and NRR are complementary, not competing. GRR shows how well you hold revenue; NRR shows whether expansion outpaces losses. Use both to separate retention quality from expansion dependency.
| GRR | NRR | |
|---|---|---|
| Includes expansion | No | Yes |
| Includes reactivation | No | Yes |
| Can exceed 100% | No (capped at 100%) | Yes |
| Best for | Measuring stickiness without upsell masking losses | Overall business health, investor readiness |
Common GRR mistakes
- Excluding contraction. GRR must include downgrades, not just cancellations. A customer going from $500/mo to $100/mo is an 80% revenue loss — ignoring it paints a false picture of retention.
- Confusing GRR with NRR. NRR includes expansion; GRR does not. If your “GRR” is above 100%, you’re accidentally including expansion revenue and actually calculating NRR.
- Including same-period signups in churn. A subscription that started and ended within the same month was never in the starting base. Including it as churn deflates your GRR unfairly.
- Using end-of-period MRR as the denominator. The denominator is MRR at the start of the period. Using end-of-period MRR contaminates the denominator with the very changes you’re trying to measure.
- Comparing monthly GRR to annual GRR directly. Monthly GRR of 95% does not mean annual GRR of 95%. Over 12 months, 95% monthly GRR compounds to roughly 54% annual retention. Be explicit about which timeframe you’re reporting.
SaaS GRR benchmarks
GRR benchmarks are segmented by MRR tier. Your GRR is scored against peers at your revenue level — top 25% (75th percentile or above) and bottom 25% (25th percentile or below).
| Stage · MRR | Bottom 25% | Median | Top 25% |
|---|---|---|---|
| Seed · < $10K | 97.1% | 98.2% | 98.9% |
| Early · $10K – $50K | 98.2% | 98.9% | 99.4% |
| Growth · $50K – $100K | 98.7% | 99.3% | 99.7% |
| Scale · $100K – $500K | 98.9% | 99.4% | 99.8% |
| Enterprise · $500K+ | 99.1% | 99.5% | 99.8% |
Is my GRR good?
Is 80% GRR good?
At or better than the median at Seed stage; worse than the median at Early stage; in the bottom 25% at Growth, Scale and Enterprise stages.
Is 85% GRR good?
At or better than the median at Seed stage; worse than the median at Early and Growth stages; in the bottom 25% at Scale and Enterprise stages.
Is 90% GRR good?
In the top 25% at Seed stage; at or better than the median at Early stage; worse than the median at Growth, Scale and Enterprise stages.
Is 95% GRR good?
In the top 25% at Seed and Early stages; at or better than the median at Growth, Scale and Enterprise stages.
| Annual value | Seed < $10K MRR | Early $10K – $50K MRR | Growth $50K – $100K MRR | Scale $100K – $500K MRR | Enterprise $500K+ MRR |
|---|---|---|---|---|---|
| 80% | Median or better | Below median | Bottom 25% | Bottom 25% | Bottom 25% |
| 85% | Median or better | Below median | Below median | Bottom 25% | Bottom 25% |
| 90% | Top 25% | Median or better | Below median | Below median | Below median |
| 95% | Top 25% | Top 25% | Median or better | Median or better | Median or better |
Frequently asked questions
What is a good GRR for SaaS?
Above 95% monthly is healthy for most SaaS companies. The median GRR across all stages is roughly 98-99% monthly, meaning even well-run businesses lose 1-2% of base revenue each month to churn and downgrades. Below 90% monthly signals a significant retention problem that expansion alone cannot sustainably cover.
What’s the difference between GRR and NRR?
GRR measures only losses — churn and contraction — and is capped at 100%. NRR includes expansion and reactivation, so it can exceed 100%. GRR shows your retention floor; NRR shows whether expansion outpaces losses. A company with high NRR but low GRR is depending on upsell to mask a churn problem.
Can GRR ever be above 100%?
No. GRR is clamped between 0% and 100% by design. Since it only measures losses (churn and contraction) without counting expansion, the best possible result is 100% — meaning you lost zero revenue from your existing base. If your retention metric exceeds 100%, you’re looking at NRR, not GRR.
Does GRR include downgrades or just cancellations?
Both. GRR includes contraction (downgrades, reduced seats, plan decreases) and churn (full cancellations). This is the industry standard followed by both North Metric and ChartMogul. A customer churn rate that only counts cancellations misses the revenue impact of downgrades — GRR captures both.
How does monthly GRR compound to annual GRR?
Annual GRR = monthly GRR raised to the 12th power. A 95% monthly GRR compounds to roughly 54% annual retention — you’d retain just over half your starting revenue after a year without any expansion. A 98% monthly GRR compounds to about 78% annually. Small monthly improvements have outsized annual impact.
From metric to action
North Metric ranks your paying Stripe customers by money at risk, so you see who is likely to leave before this number moves. See how churn prediction works.