Net Revenue Retention Explained: NRR Formula, Benchmarks, and Why It Drives SaaS Valuation

Net revenue retention is arguably the single most important metric for evaluating the health of a subscription software business. It tells you whether existing customers are growing, staying flat, or shrinking, all in one number. A company with strong NRR can grow its revenue without acquiring a single new customer. A company with weak NRR is running on a treadmill, constantly backfilling revenue it loses from existing accounts.

This guide breaks down the NRR formula, explains why it matters more than many other SaaS metrics, walks through cohort analysis, and connects NRR to the revenue growth and valuation multiples you see in public market comps.

What Is Net Revenue Retention?

Net revenue retention (NRR), also called net dollar retention (NDR) or net revenue retention rate, measures how much revenue a company retains from its existing customer base over a period, including expansions, contractions, and full churn.

If a company had $10 million in annual recurring revenue (ARR) from a specific cohort of customers at the start of a year, and that same cohort generated $11.5 million by year end after accounting for upgrades, seat additions, downgrades, and cancellations, the NRR is 115%.

The critical insight is what NRR above 100% implies: the company's existing customer base is growing on its own, independent of any new customer acquisition. That is sometimes called "negative churn" because revenue from expansions outpaces revenue lost from cancellations and downgrades.

The NRR Formula

NRR is calculated from four inputs measured over a defined period, typically one month or one year:

• Beginning ARR: Revenue from the customer cohort at the start of the period
• Expansion revenue: Upsells, cross-sells, seat additions, usage growth
• Contraction revenue: Downgrades, seat reductions, pricing concessions
• Churned revenue: Revenue from customers who cancelled entirely

The formula:

NRR = (Beginning ARR + Expansion - Contraction - Churn) / Beginning ARR

Working through a concrete example:

• Beginning ARR from existing customers: $20 million
• Expansion revenue added during period: $4 million
• Contraction revenue lost: $1.2 million
• Churned revenue: $1.8 million
• Ending ARR from same cohort: $21 million

NRR = $21 million / $20 million = 105%

The company retained 105% of its starting revenue from existing customers. It grew that cohort by 5% without any new customer acquisitions contributing.

Gross Revenue Retention vs. Net Revenue Retention

GRR and NRR measure different things and should be read together, not interchangeably.

Gross revenue retention (GRR) measures what fraction of starting revenue the company held onto, counting only losses. It explicitly excludes any expansion revenue. The formula is:

GRR = (Beginning ARR - Contraction - Churn) / Beginning ARR

GRR is always equal to or lower than NRR, and it can never exceed 100%. A GRR of 90% means the company lost 10% of its starting revenue from cancellations and downgrades before any expansion is counted.

GRR tells you about the quality of the core product and customer satisfaction. If customers are consistently churning or downgrading, no amount of expansion revenue can mask the underlying problem permanently.

NRR tells you the net result after the full customer lifecycle plays out. High NRR with low GRR means the company is papering over significant churn with aggressive upsells - a less durable growth profile than high NRR with high GRR.

The best SaaS businesses have both high GRR (above 85% for SMB, above 90% for enterprise) and high NRR (above 110% for most categories, above 120% for best-in-class enterprise software).

Why NRR Above 120% Is Exceptional

A 120% NRR means the company grows its existing customer cohort by 20% per year without any new customer acquisition. That creates a structurally advantaged growth model.

Consider two companies with the same starting ARR of $50 million and the same new customer acquisition rate of $10 million per year:

• Company A: 90% NRR (loses 10% of existing ARR annually)
• Company B: 120% NRR (grows existing ARR by 20% annually)

After three years:

Company A: Year 3 ARR = roughly $63 million (expansions don't offset churn; growth is primarily from new logos)

Company B: Year 3 ARR = roughly $103 million (existing base self-compounds; new logos add on top)

The math compounds powerfully. High NRR means a company's installed base is itself an engine of organic revenue growth. This is why enterprise software businesses with deeply embedded products and extensive land-and-expand models, such as Snowflake, Datadog, or Twilio in their early growth phases, can sustain 30 to 50% revenue growth even when their sales teams are not adding proportionally more new customers.

NRR above 120% is rare. In a broad sample of publicly traded SaaS companies, fewer than 15 to 20% consistently sustain 120%+ NRR over multiple years. The threshold signals strong product-market fit, effective customer success, and genuine value expansion within accounts over time.

NRR Industry Benchmarks by Sector

NRR varies significantly by customer segment and product category. These benchmarks reflect typical ranges observed across public and private SaaS companies:

SMB SaaS benchmarks are lower because small businesses have higher inherent churn rates. Business closures, budget cuts, and switching to cheaper alternatives are all structurally more common in the SMB segment. Expecting 120% NRR from an SMB-focused product is generally unrealistic; 95 to 100% is a healthy range.

Enterprise SaaS NRR is higher because contracts are larger, switching costs are greater, and products tend to become more deeply integrated over time. Land-and-expand models that start with one business unit and grow across an organization naturally produce high NRR.

Usage-based pricing (where customers pay per API call, per query, per GB processed) often produces the highest NRR because there is no artificial constraint on expansion. As customers use the product more, revenue grows automatically. Snowflake and Datadog are canonical examples.

How NRR Interacts With Revenue Growth

NRR and new customer acquisition together determine total ARR growth. The relationship is:

Total ARR Growth Rate = NRR Growth from Existing Base + New Customer ARR / Beginning ARR

If a company has 115% NRR and adds new ARR equal to 25% of beginning ARR each year:

Total growth = 15% (expansion from existing) + 25% (new customers) = 40%

This decomposition matters because growth rates that look identical can have very different quality profiles. Two companies both growing at 40% could look like this:

• Company A: 120% NRR + 20% from new customers - stable, self-reinforcing
• Company B: 85% NRR + 55% from new customers - fragile, dependent on constant acquisition

Company B must continuously acquire new customers just to replace lost revenue, and must dramatically outspend Company A on sales and marketing to reach the same growth rate. As markets mature and customer acquisition costs rise, Company B will decelerate much faster.

Investors pay a significant premium for NRR-driven growth because it signals unit economics improve over time. The contribution margin from an existing customer in year three is typically far higher than in year one; the acquisition cost has already been paid, and support and success costs often scale sub-linearly.

Cohort Analysis: The Underlying Foundation of NRR

NRR is a blended average across all existing customers. Cohort analysis disaggregates that average to show how different generations of customers behave over time.

A cohort is a group of customers who started in the same period, typically the same quarter or the same year. Tracking cohort revenue over time produces a cohort retention curve.

A healthy cohort retention chart shows a pattern sometimes called a "smile" or "expanding fan": revenue from each cohort initially drops slightly (some early churn) but then stabilizes and grows as expanded contracts more than offset attrition.

An unhealthy cohort chart shows consistent decay: every cohort is worth less revenue each successive quarter than the one before, signaling structural churn that no amount of new customer acquisition will solve.

What to look for in cohort analysis:

• Are older cohorts still growing? If the 2021 customer cohort is generating more revenue in 2025 than it was in 2023, that is a sign of deep product embedding and successful expansion.
• Does early-stage churn stabilize? Almost every SaaS product loses some customers in the first six to twelve months. What matters is whether the curve flattens. A cohort that starts at $1 million and falls to $800,000 but holds there for four years has a different risk profile than one that continues falling to $400,000.
• Are newer cohorts behaving like older ones? Degrading cohort performance is a leading indicator that NRR will fall in future periods, often before the blended NRR metric shows obvious deterioration.

NRR as a Leading Indicator vs. a Lagging Metric

NRR as reported in quarterly earnings is a trailing measure, typically calculated on a trailing twelve-month basis. By the time you see deteriorating NRR in a press release, the underlying customer dynamics that caused it happened 6 to 12 months earlier.

This is why analysts monitor secondary signals that precede NRR movement:

• Net promoter scores and customer satisfaction trends: These often move before churn does.
• Usage data: Declining active user counts or feature adoption within existing accounts is a warning signal even when renewal rates look stable.
• Expansion quota attainment: If the customer success team is consistently missing upsell targets, NRR compression is likely 1 to 2 quarters out.
• Cohort performance of recent vintages: If 2024 cohorts are churning faster in their first two quarters than 2022 cohorts did, the blended NRR will show the impact once those cohorts age into the trailing twelve-month window.

For public market investors reading quarterly earnings, the combination of management commentary on expansion activity, gross retention data, and any disclosed cohort metrics provides the best forward read on NRR trajectory.

How NRR Maps to Valuation Multiples

NRR is one of the strongest predictors of EV/NTM revenue multiples in the SaaS universe. The intuition is straightforward: a company with 130% NRR is, all else equal, worth more per dollar of current revenue than one with 95% NRR because the existing revenue base is itself a growth engine.

Empirical data from public SaaS comparables consistently shows:

These ranges are wide because NRR is one variable among many. Growth rate, free cash flow margin, total addressable market, and competitive moat all interact with NRR in determining where a specific company lands in a valuation range.

The cleanest use of NRR in relative valuation is as a screener within a peer group. When comparing two companies with similar revenue growth rates, the one with higher NRR and higher GRR deserves a premium because its growth is more durable and less capital-intensive.

In DCF modeling, high NRR supports lower near-term customer acquisition cost assumptions in the revenue model and higher long-run margin assumptions, both of which raise intrinsic value. A company that grows 20% from existing customers needs to spend less on sales and marketing to hit the same total growth target than one that grows 20% entirely from new customer acquisition.

Common NRR Presentation Issues to Watch For

Trailing vs. spot NRR: Some companies report trailing twelve-month NRR (a smoother but lagging measure) while others report point-in-time quarterly NRR. Make sure you are comparing like to like when benchmarking.

Cohort definition differences: Some companies include all contracted customers; others exclude customers below a revenue threshold, which can flatter the metric by cutting out small, high-churn accounts.

Lack of GRR disclosure: A company that reports NRR but not GRR is hiding information. A high NRR built on low GRR (lots of churn papered over by aggressive upsells) is structurally riskier than the blended number implies. Push for both figures when evaluating any SaaS business.

Geographic or segment mix shifts: If enterprise customers (naturally higher NRR) are growing as a share of the customer base while SMB shrinks, reported NRR will improve mechanically even if the underlying quality of each segment is unchanged. Segment-level NRR disclosure is more informative.

Key Takeaways

• NRR measures how much revenue a company retains from its existing customer base after accounting for expansions, contractions, and full churn. It is the cleanest single metric for evaluating SaaS business quality.

• The NRR formula: (Beginning ARR + Expansion - Contraction - Churn) / Beginning ARR. NRR above 100% means the existing base is growing; above 120% is best-in-class.

• GRR measures only losses and can never exceed 100%. Reading GRR alongside NRR reveals whether growth is driven by genuine retention or by upsells masking underlying churn problems.

• Benchmarks vary by segment: enterprise SaaS should aim above 110 to 120%; SMB SaaS above 90 to 100%; usage-based infrastructure above 120 to 130%.

• High NRR dramatically reduces the sales and marketing spend required to hit growth targets, improving long-run unit economics and free cash flow margins.

• Cohort analysis is the foundation beneath reported NRR. Healthy businesses show cohort revenue that stabilizes and grows over time. Deteriorating cohort performance is a leading indicator that blended NRR will compress in future periods.

• NRR is one of the most reliable predictors of EV/NTM revenue multiples within a SaaS peer group. Companies with higher, more durable NRR consistently command premium valuations.