NRR vs GRR — what separates a $10B from a $100M SaaS company
NRR vs GRR — what separates a $10B from a $100M SaaS company
Of all SaaS metrics, NRR (Net Revenue Retention) is the one that decides whether your company becomes a unicorn or stays in the middle league. Snowflake trades at premium multiples because their NRR is ~158%. Datadog — 130%. The average public SaaS — 105-110%.
That's not a correlation. It's a cause.
Formulas
GRR (Gross Revenue Retention):
GRR = (Start MRR − Churn MRR − Contraction MRR) / Start MRR × 100%
Losses only. Doesn't count expansion. Always ≤100%.
NRR (Net Revenue Retention):
NRR = (Start MRR + Expansion MRR − Churn MRR − Contraction MRR) / Start MRR × 100%
Including expansion. Can be >100%.
Benchmarks
| Tier | NRR | GRR | ||
|---|---|---|---|---|
| Top decile public SaaS | >130% | >95% | ||
| Top quartile | 115-130% | 92-95% | ||
| Median public SaaS | 105-115% | 88-92% | ||
| Bad | <100% | <85% | ||
| Catastrophe | <90% | <75% |
That's the holy grail of SaaS. Reach it and multiples × 2-3.
Worked example
A SaaS startup, ARR $5M at start of year:
- Expansion ARR for the year: $800k (upgrades, seat growth)
- Churn ARR: $300k (10 customers left)
- Contraction ARR: $100k (5 downgrades)
NRR = (5000 + 800 − 300 − 100) / 5000 × 100% = 5400 / 5000 = 108% → healthy
GRR = (5000 − 300 − 100) / 5000 × 100% = 4600 / 5000 = 92% → ok
NRR − GRR = 108 − 92 = 16 percentage points of expansion. Better than 10pp, worse than 25pp.
Why top SaaS have NRR of 130-150%
Three mechanisms:
1. Seat-based pricing with growing customers
Snowflake bills per data warehouse compute hour. A customer starts small ($50k/year) and over 3 years grows to $500k. 10× expansion on one customer = enormous NRR.2. Tiered features with a natural upgrade path
HubSpot: Starter → Professional → Enterprise. Each tier is 2-3× the price. 30% of customers upgrade within 18 months = NRR +25-30pp.3. Cross-sell additional products
Datadog started with Infrastructure Monitoring. Then APM, Logs, Security. Each additional product on an average existing customer = +20% to ARR per customer.Why GRR is useful separately from NRR
GRR is product health. NRR is account-team health.
- GRR <90% — customers don't like the product. Fix product, not sales.
- GRR >95% but NRR <105% — product is ok, but there's no expansion mechanic. Fix pricing.
- GRR <85% — you don't have PMF. Stop scaling.
How to improve NRR — 5 levers
1. Customer Success function
Hire a CSM for every $500k-1M of ARR. Their sole KPI is the NRR of those accounts. Proactive QBRs, health monitoring, expansion identification.2. Seat-based / usage-based pricing
If you have flat per-tenant pricing — switch to a growth-friendly model. Example: Slack ($8/user/mo) grows with the customer's team for free.3. Health scoring
Track 5-10 product usage signals → red zone → CSM intervention 30 days before churn.4. Annual prepay incentives
An annual contract (vs monthly) gives you:- Mathematically cuts churn ×12
- Cash forward
- Locks the customer for a year — more time to expand
5. Multi-product strategy
A single product caps NRR at ~120% (pricing-tier expansion only). Multi-product = NRR 140%+.Bessemer's "Net Magic Number" metric
Less known, but powerful:
> Net Magic Number = (Net New ARR − Churn ARR) × 4 / S&M spend
If you spend $1 of S&M and a year later have $1.2 of new ARR but lost $0.5 to churn → the real gain is $0.7. Net Magic Number = 0.7. Worse than the gross Magic Number of 1.2.
Sequoia/Bessemer increasingly look at Net Magic Number instead of Gross.
When NRR is MISLEADING
1. Slow-burn churn models — Annual contracts. A customer signs a $100k contract, plans not to renew, but you only see it after 12 months. NRR is optimistic until the moment of non-renewal.
2. Large-account skew — one enterprise customer is 30% of ARR. If they 2× — NRR looks beautiful. If they leave — NRR craters. Concentration risk.
3. Price increases — if you raised prices +20%, NRR will be 120%+ for a year, but that's not "expansion", that's inflation.
4. Currency fluctuations — for multi-currency businesses, USD-reported NRR drifts with FX.
Bottom line
NRR is the single most important metric for SaaS valuation. Know it monthly. If it's below 100% — you have a fundamentally broken expansion strategy. If 110-115% — that's normal, keep going. If >130% — you're in the top decile, and investors will hunt for you.
Calculate your NRR below — built-in calculator + benchmark + Goal mode (what you need to hit a target NRR).
---
Further resources
- /en/grr — GRR calculator on its own
- /en/mrr — MRR, the base for NRR
- /en/quickRatio — Quick Ratio, a related metric
- /en/blog/quick-ratio-saas — Quick Ratio deep dive
- Bessemer State of the Cloud — annual NRR benchmarks