Net revenue retention (NRR / NDR)
Best single proxy for compounding/expansion.
- Formula
- (Start + Expansion - Contraction - Churn) / Start, starting cohort only, TTM; can exceed 100%
- Unit
- %
- Models
- SaaS, Usage-based, Subscription
| private company median | 100%–106% | KBCM; OpenView/High Alpha; ChartMogul; SaaS Capital; S-1 filings |
| private company top quartile | 110%–125% | KBCM; OpenView/High Alpha; ChartMogul; SaaS Capital; S-1 filings |
| Usage-based | 120%–130% | KBCM; OpenView/High Alpha; ChartMogul; SaaS Capital; S-1 filings |
What it is
Net revenue retention (NRR), also called net dollar retention (NDR), measures how much revenue a fixed starting cohort of customers generates over time, including expansion, contraction, and churn. The formula is (Start + Expansion − Contraction − Churn) / Start, measured on a TTM basis from a starting cohort. It can exceed 100% when expansion outpaces losses.
How to calculate it
Take a cohort of customers as of the start of the trailing twelve months. Sum their starting ARR, add any expansion (upsells, seat additions, usage growth), subtract contraction (downgrades), and subtract churn (full cancellations). Divide the result by the starting ARR. Only the original cohort is included — no new customers added during the period.
Why it matters
NRR above 100% means existing customers are growing their spend faster than others are leaving, producing compounding revenue even without acquiring a single new customer. It is the most important signal of product-market fit depth in B2B SaaS and usage-based businesses. High NRR dramatically improves capital efficiency and long-run LTV.
Benchmarks & pitfalls
Based on KBCM, OpenView/High Alpha, ChartMogul, SaaS Capital, and S-1 filings (2024–26): private B2B SaaS median NRR is roughly 100–106%, with top-quartile companies reaching 110–125%+. Usage-based models typically run higher at 120–130%+ (e.g. Snowflake declined from 158% to ~125%, Datadog from >130% to ~120% as they scaled). NRR is weak for low-ARPA consumer products where individual churn is noisy and expansion is limited. It is not meaningful for marketplace businesses, which lack a recurring revenue cohort structure to measure. A common pitfall is including new customers added during the measurement period, which inflates the metric — the denominator must be the starting cohort only.