The NRR formula, broken down with your numbers
Every analyst memorises the same equation: NRR = (Starting ARR − Churn − Contraction + Expansion) ÷ Starting ARR. What trips people up is the cohort definition — NRR is a statement about a fixed group of customers measured at the start of a period, not the company's revenue at large. Take a $20M ARR baseline. Burn $800K of churn, $400K of contraction, and add $3.2M of expansion, and the cohort ends at $22M. The nrr formula collapses that to a single number: 110%. The hero number on the calculator above is that division done in real time.
The reason VCs read this number first — before MRR, before logo count, before CAC payback — is that it isolates the compounding leverage inside a SaaS business. If your nrr is above 100%, the existing book of business grows itself between renewal cycles. If it is above 130%, the book of business roughly doubles from expansion alone every 30 months. That compounding is what justifies the revenue multiples premium SaaS companies trade at, which is why the calculator exposes the Valuation Premium tile beneath the hero.
What is Net Dollar Retention, and why VCs use it as the #2 SaaS metric
Net dollar retention and net revenue retention are the same metric under two different vendor labels — both use the cohort formula above and both express the result as a percentage. Public-company filings tend to favour "net dollar retention" (Snowflake, Datadog, Cloudflare); board memos and private-company decks drift toward "net revenue retention" or just NRR. The math is identical.
The reason it ranks second only to growth rate on most VC scorecards is that it is the cheapest dollar of growth a SaaS company can buy. Expansion ARR carries near-zero CAC because the customer is already on the platform — the marginal cost of a seat upgrade or a feature add-on is a couple of hours of CS time. Compare that to the $1,200–$18,000 CAC required to land a brand-new logo at most stages, and a company that runs at 125% NRR is, in effect, getting 25% top-line growth at gross-margin economics. That asymmetry is what bends the valuation multiple.
Gross retention vs net retention — what the side-by-side actually tells you
Reading gross retention vs net retention together is where the diagnostic value sits. GRR strips out expansion and shows only the cohort's leakage: (Start − Churn − Contraction) ÷ Start. It is capped at 100% by definition. NRR adds expansion back and is uncapped. The distance between the two is your expansion engine's workload — and that workload tells you which lever your business is actually pulling.
Two companies can both report 118% NRR and have nothing in common. Company A: GRR 96%, expansion 22pp — the customer base is sticky and the expansion engine works on a healthy foundation. Company B: GRR 84%, expansion 34pp — the expansion engine is doing 34 points of work to mask 16 points of underlying churn. Company A is durable. Company B is fragile: the day the expansion engine stalls (price-lift backlash, champion turnover, macro slowdown), NRR falls off a cliff. The GRR/NRR twin tile above forces this distinction into view on every calculation.
Net retention SaaS benchmarks by stage — what "good" looks like at your size
Stage matters more than industry for net retention saas benchmarks. The cohorts below are calibrated to the three industry surveys most commonly cited in board decks — Bessemer Cloud Index, KeyBanc Private SaaS Survey, and OpenView — for 2024:
- Pre-Seed / Seed (< $1M ARR): p25 ≈ 95% · p50 ≈ 105% · p75 ≈ 115% · p90 ≈ 128%
- Series A ($1–10M): p25 ≈ 98% · p50 ≈ 108% · p75 ≈ 118% · p90 ≈ 132%
- Series B ($10–50M): p25 ≈ 102% · p50 ≈ 112% · p75 ≈ 122% · p90 ≈ 136%
- Series C+ ($50M+): p25 ≈ 105% · p50 ≈ 116% · p75 ≈ 128% · p90 ≈ 142%
Notice the upward drift. As ARR scales, the top decile climbs faster than the median — Series C top-decile NRR (142%) is fully 14 points higher than seed-stage top-decile (128%). The reason is mechanical: at scale, a 1% expansion lift generates much more absolute ARR, which creates the budget for the CS and product-led growth investments that drive further expansion. The percentile dot on the Stage Benchmark chart above shows where this cohort's NRR sits within the stage you select.
Dollar retention rate vs NRR — terminology disambiguation
Dollar retention rate is the umbrella term that covers both halves of the formula. "Gross dollar retention rate" is GRR expressed in dollar terms (as opposed to logo count); "net dollar retention rate" is NRR expressed the same way. In practice, most SaaS practitioners use "dollar retention" and "retention" interchangeably because revenue, not logo count, is what the multiple applies to. A 50-customer cohort that retains 47 logos but loses its three biggest accounts has 94% logo retention and 65% dollar retention — and only the second number affects enterprise value.
Dollar-based net retention vs logo retention — why VCs read both
Dollar based net retention tells you whether the revenue is durable; logo retention tells you whether the product is loved across the customer base. They diverge in two telling patterns. First, a healthy product with concentration risk: 95% logo retention but 70% dollar retention means a small number of big accounts churn while the long tail stays. Second, an unhealthy product with sticky whales: 75% logo retention but 105% dollar retention means small accounts churn rapidly while a handful of large customers absorb the loss with expansion. The first looks healthier on a board slide. The second is structurally fragile: if any whale leaves, the floor collapses.
How to use this calculator — the five-input cohort methodology
- Pick the period and stage. The period toggle (quarterly / annual) changes the band conversion math; the stage selector drives benchmark percentiles. If you're running a board prep memo, default to annual.
- Enter Starting ARR — the cohort baseline. This is the ARR of the customers on the platform on day one of the period. Not total ARR; the existing book only.
- Enter Churned + Contracted ARR. Churned ARR is dollars lost from full cancellations. Contracted ARR is dollars lost from downgrades within accounts that stayed. Their sum drives GRR; keep them separate because the levers to address each are different (saves and renewals for churn; pricing and packaging for contraction).
- Enter Expansion ARR (and optionally split it). Expansion is upsell, cross-sell, seat growth, or usage expansion inside the cohort. Open the Expansion Mix accordion to attribute it across the four sources — single-source concentration above 70% shows up as a yellow flag in the report card.
- Read the hero, the twin, and the Premium Tile. The hero is NRR with its zone band. The GRR/NRR twin shows where the leak is and where the expansion is. The Premium Tile applies the directional multiple curve to show valuation lift at your current band.
The hidden valuation lever — how NRR band shifts your revenue multiple
Sitting beneath the hero is the Valuation Premium tile — an internal feature of the calculator, not the focus of the page. The concept is widely cited in venture decks: companies running above 130% NRR consistently trade at meaningfully higher revenue multiples than companies in the 100–110% band. The tile applies a directional model — roughly 0.85× the base multiple at 90–100% NRR, 1.00× at the 100–110% acceptance line, 1.30× at the 120–130% premium band, and 1.50× at the > 130% best-in-class threshold. These coefficients reflect the documented multiple expansion of high-retention SaaS, not a single cited equation, and they shift with the funding cycle.
What makes the tile useful is the leverage it surfaces. On a $20M ARR base with a 6× base multiple, the difference between 105% NRR and 125% NRR is the difference between a $102M valuation and a $156M valuation — a $54M swing from a 20-point retention move. That is the math that justifies the cost of a CS hire, a price-lift project, or an onboarding redesign. The What-If Simulator above lets you model each lever and watch the Premium Tile delta update live.
NRR vs Churn — when to use which calculator
We ship two retention-side tools because they answer different questions. The companion Churn & NRR Calculator is built around the leaky-bucket framing — it projects churn and expansion rates forward over 12 months, shows MRR drift over time, and includes an Expansion Advisor for routing investment between churn reduction and expansion lift. Reach for it when the question is "what does my retention engine do to MRR over the next year?"
This calculator is the answer to a different question: "what is my NRR right now, why is it that, and what is it worth in valuation terms?" It is the board-memo speed tool — five inputs, one screen, formula transparency. Most teams use both: the churn calculator for forward planning, this one for the quarterly close and the investor update. They share zero math; they overlap on no SEO surface.
The 4 expansion sources — why composition matters for NRR durability
The Expansion Mix doughnut breaks the expansion line into four sources. Each carries a different durability profile:
- Upsell — moving customers up the pricing tier. Durable if the higher tier delivers genuinely more value; fragile if it is a temporary discount unwind.
- Cross-sell — adding a new product module to the same account. Most durable expansion type — once integrated, multi-product accounts churn at roughly half the rate of single-product accounts.
- Seat growth — adding users at the same per-seat price. Durable when tied to customer hiring; fragile during layoff cycles (visible in 2023–2024 NRR compression across horizontal SaaS).
- Usage expansion — consumption-based revenue growth. The PLG default; highly durable when usage tracks workload, but volatile quarter-to-quarter.
When any single source exceeds 70% of expansion, the report card flags it as a concentration risk. A vertical SaaS with 90% of expansion coming from seat growth is one bad customer-layoff quarter away from NRR collapse. A PLG company running 90% on usage expansion is one macro slowdown away from the same outcome. Healthy expansion is diversified — usually three of the four sources contributing meaningfully.