Net Revenue Retention (NRR / NDR)

How net revenue retention is calculated

The formula is NRR equals starting recurring revenue plus expansion minus contraction minus churned revenue, all divided by starting recurring revenue, expressed as a percentage. You fix a cohort, the customers who existed at the start of the period, and follow only those accounts forward. Revenue from customers acquired during the period never enters the calculation.

Because it nets three opposing forces, NRR can move above or below 100%. Expansion pushes it up; contraction and churn pull it down. A base that started at one million dollars, added two hundred thousand in expansion, and lost fifty thousand to downgrades and eighty thousand to churn ends at 1.07 million, or 107% NRR. The mechanics look simple, but the inputs are not standardized across companies. Some report an adjusted figure, some blend logo effects into the number, and some normalize multi-currency revenue to a single currency. Two companies quoting the same percentage may have computed it differently, which is why the stated methodology matters as much as the result.

NRR vs. gross revenue retention

Gross revenue retention (GRR) answers a narrower question: of the revenue you started with, how much did you keep before any upsell. It subtracts contraction and churn but ignores expansion entirely, which caps it at 100%. NRR layers expansion on top, so it can exceed 100% and it can mask underlying leakage that GRR would reveal.

Read alone, a high NRR can flatter a company. A base losing meaningful revenue to downgrades and cancellations can still print NRR above 100% if a handful of large accounts expand aggressively. GRR strips that expansion away and shows the floor, meaning how sticky the base is on its own merits. The two are best read together rather than treated as competing headline metrics: GRR describes durability, NRR describes durability plus growth. A wide gap between them signals a company whose reported retention depends heavily on expanding its biggest customers.

Reading NRR benchmarks by customer segment

There is no single good number, because NRR scales with the type of customer being retained. Benchmarking sources broadly treat below 100% as concerning, 100 to 120% as good, and above roughly 120 to 130% as exceptional for enterprise SaaS, but those bands shift with segment and company stage.

Segment mix is the biggest confounder. One benchmarking source cites median NRR of roughly 118% for enterprise accounts with annual contract value above 100,000 dollars, around 108% for mid-market between 25,000 and 100,000 dollars, and about 97% for SMB below 25,000 dollars. Larger customers expand more and churn less, so an enterprise-heavy company will structurally outscore an SMB-focused one selling a similar product. Comparing a rival's NRR to your own without accounting for who each of you sells to produces a misleading read. Stage matters too: a young company with a small base can swing on a few accounts.

What NRR tells a competitive-intelligence team

For public SaaS companies, NRR is disclosed in S-1 filings, quarterly earnings, and investor decks, which makes it one of the rare unit-economics figures a competitor hands over directly rather than one you have to estimate. That makes it worth tracking as a reported series, not a modeled one.

The trendline is where the signal lives. A rising NRR usually means growth is coming from expansion, whether an active upsell, cross-sell, or land-and-expand motion, while a falling NRR can indicate the reported figure is masking churn beneath new-logo growth. Because NRR feeds valuation multiples, a sustained shift can foreshadow funding events, M&A activity, or changes in go-to-market spend. Movements often precede pricing changes, new bundling, or repackaging aimed at lifting expansion. Monitoring a competitor's filings, pricing pages, and hiring alongside its disclosed NRR turns a single reported number into a leading indicator of strategy, provided you first confirm how that competitor defines the metric.

Common mistakes and limitations

The most common error is treating NRR as a clean cross-company comparison. Because calculation methods vary, whether adjusted figures, blended logo effects, or currency normalization, two disclosed numbers are only comparable after you check each company's stated definition and customer-segment mix.

NRR is dollar-weighted, so it says nothing about how many customers you kept. A company can lose a long tail of small accounts and still post NRR above 100% by retaining and expanding its largest ones, which is why logo retention should be read beside it. It also describes only the existing base; it ignores acquisition cost and gross margin, so it is an input to broader unit-economics measures like LTV, not a substitute for them. And it is a trailing snapshot: a strong figure reflects last year's cohort behavior and can lag a deterioration already underway in the current book of business.

Frequently Asked Questions

Related terms