What Is NRR (Net Revenue Retention)? Formula, SaaS Benchmarks & GRR Difference

In subscription and SaaS businesses, success isn't decided at the moment of sale — it's decided by how much revenue you retain and expand after the contract is signed. The single number that captures the health of that motion is NRR (Net Revenue Retention).

NRR is one of the most important KPIs for measuring a SaaS company's growth potential and the strength of its customer base. It's used across leadership, finance, and customer success (CS) teams, and it's one of the first metrics investors check when valuing a SaaS business.

But the moment you try to apply it, questions pile up: What exactly goes in the denominator? How does it differ from GRR (Gross Revenue Retention), and which should you watch? What does 120% concretely mean? What NRR is "good enough" for my company? A dictionary-style definition won't answer these.

This article walks through NRR from the definition and formula to a fill-in calculation worksheet, the often-confused NRR vs. GRR vs. churn relationship matrix, segment benchmarks verified against primary sources, the two investment levers for improvement, and how a Digital Sales Room (DSR) turns churn-risk and upsell timing into measurable signals — all from a practitioner's point of view.

Note: "NRR" in this article refers to Net Revenue Retention, a business metric for SaaS and subscription companies. It is distinct from other uses of the abbreviation — such as the Noise Reduction Rating for hearing protection (measured in dB) or Net Run Rate in cricket. If your intent is different, please keep that in mind.

Key Takeaways

• NRR (Net Revenue Retention) shows how much your period-start existing customers retained and expanded by period-end, after adding expansion (upsell/cross-sell) and subtracting churn and downgrades. Revenue from new customers is excluded.
• The formula is "(Starting MRR + Expansion MRR − Downgrade MRR − Churn MRR) ÷ Starting MRR × 100." Above 100% means growth from existing customers alone; 120%+ is the elite "negative churn" state.
• NRR is the "offensive" retention rate (it includes expansion); GRR (Gross Revenue Retention) is the "defensive" one (it excludes expansion and caps at 100%). Watch both — otherwise a few large expansions can mask many small churns.
• Benchmarks vary sharply by segment. Various reports put Enterprise around 115–120% and consider ~100% a solid result for SMB. Among top SaaS, Snowflake disclosed 131% for FY2024 and Datadog reported a mid-110s dollar-based net retention on a trailing-twelve-month basis (both from company filings).
• To raise NRR you have two levers: "create expansion" and "reduce churn/downgrades." Which to invest in first is determined by working backward from your NRR level and its gap to GRR. A DSR's engagement data is a practical way to surface churn risk and upsell timing.

What Is NRR (Net Revenue Retention)? Meaning and Aliases

NRR is a ratio showing how much the revenue of customers who existed at the start of a period was retained and expanded by the end of that period. It stands for Net Revenue Retention.

The crucial point is that it excludes all revenue from newly acquired customers. By removing new business, you can purely assess whether your business is growing or shrinking on the strength of its existing customer base alone. If NRR exceeds 100% even with zero new customers, it means revenue is growing from existing customers' expanded usage — a powerful position.

Aliases (NDR, Net Retention Rate)

NRR appears under slightly different names across sources. They all refer to the same metric.

NRR is a revenue-based metric. To track the retention of customer counts, use a separate metric — CRR (Customer Retention Rate). Don't confuse the two: NRR measures how much revenue (MRR/ARR) remained and grew.

Distinguishing NRR from its homonyms

The abbreviation "NRR" is also used in entirely unrelated fields. To avoid intent mismatch, here's a quick disambiguation:

• Noise Reduction Rating (NRR): A number (in dB) indicating the noise-attenuation performance of hearing protection such as earplugs and earmuffs. Unrelated to this article.
• Net Run Rate (NRR): A statistic used in cricket. Also out of scope here.

This article covers only NRR (Net Revenue Retention) as a business metric for SaaS and subscription companies.

Why NRR Is the Most Important KPI for SaaS/Subscription

NRR matters because SaaS and subscription businesses have a "stock" (compounding) revenue structure. Once a customer signs, they keep paying, and revenue accumulates — which means retaining and expanding existing customers, more than acquiring new ones, is the foundation of growth.

It tells you whether you can grow on existing customers alone

New acquisition can be boosted temporarily by spending on ads and sales headcount, but it's costly and eventually plateaus. A company with NRR above 100%, by contrast, has a structure where revenue keeps growing from existing customers' expansion even if it stops acquiring new ones. This is like "adding water to a bucket with no holes" — extremely efficient growth.

Conversely, when NRR is below 100%, existing-customer revenue erodes each period, so you can't sustain revenue without acquiring new customers continuously. That's a "leaky bucket" — the moment your acquisition pace slows, revenue contracts. It's a fragile business.

This gap widens the longer you operate. A company at 110% NRR sees existing-customer revenue swell 1.1× each year without doing anything, so every bit of new acquisition stacks directly on top of growth. A company at 90%, meanwhile, has part of its newly won revenue offset by erosion in the existing base, sharply lowering growth efficiency. For the same amount of new acquisition, where you land in a few years differs entirely depending on NRR. That's why, in any conversation about SaaS growth, NRR is a decisive metric — the "fuel efficiency of the growth engine."

It's a decision lever for fundraising, new bets, and valuation

NRR is a frontline CS metric, but it's also a leadership and finance metric.

• Valuation: Investors and VCs weigh NRR heavily when valuing a SaaS company. High NRR signals predictability — "existing customers will keep growing revenue in the future" — which translates into higher valuation multiples.
• Fundraising: High NRR is read as a high probability of return on investment, making fundraising easier.
• New bets and additional investment: When an existing business's NRR is consistently high, it's easier to redirect that cash toward new businesses and new investments.

In short, NRR is a frontline KPI that ties directly into executive decisions. Within the customer-success metrics cluster, NRR is positioned as the core indicator of "how much existing-customer revenue you retained and expanded."

It shifts the dividing line between Sales and Customer Success

Prioritizing NRR means shifting the organization's center of gravity for "how revenue is made" from new acquisition toward existing expansion. Traditionally, sales was considered to have done its job once it landed a new deal. But when NRR sits at the center of the business, the contract is not the goal — it's the start.

Under this view, customer success — which accompanies the customer toward outcomes after the sale, prevents churn, and grows revenue through upsell and cross-sell — becomes as important as sales, if not more so. NRR is also the metric that prompts you to revisit the Sales/CS dividing line and the allocation of headcount and budget. A culture that evaluates "how much you grew existing accounts," not just "how much you acquired," is common to high-NRR organizations.

The NRR Formula and the Four Types of MRR

NRR is typically calculated on a monthly MRR (Monthly Recurring Revenue) basis. When viewed annually, you use ARR (Annual Recurring Revenue), but the structure is the same.

The formula

The basic NRR formula is:

NRR (%) =
  (Starting MRR + Expansion MRR − Downgrade MRR − Churn MRR)
  ÷ Starting MRR × 100

The scope here is only existing customers who were already under contract at the start of the month. The MRR of customers newly signed that month (New MRR) is excluded from both the numerator and the denominator. This is exactly what makes NRR a "pure retention-and-expansion rate for existing customers."

Breaking down the four types of MRR

Decomposing the formula's components by the "increase/decrease" of existing-customer MRR gives four types:

In other words, NRR expresses how much existing customers' Expansion (added revenue) exceeded Downgrade and Churn (lost revenue). When gains beat losses, NRR is above 100%; when losses win, it's below 100%.

Why New MRR is excluded

NRR excludes New MRR because the question it answers is "Is the existing customer base healthy?" Mixing in new business lets strong acquisition mask heavy churn, hiding quality problems in the customer base.

Overall growth including new business is measured by other metrics — ARR growth rate, or a comprehensive retention rate that includes new customers. By deliberately excluding new business, NRR isolates one thing: "Is value delivery to existing customers working?"

Consider a concrete example. A company with $1,000K of existing-customer MRR at the start of the month loses $200K to churn but acquires $300K in new contracts. Total revenue looks like it grew: 1,000 − 200 + 300 = $1,100K. But the NRR calculation excludes the new $300K, so (1,000 − 200) ÷ 1,000 × 100 = 80% — revealing that the existing base actually shrank by 20%. The "hole in the bucket" that new acquisition had concealed becomes visible only when you look at NRR. That is the single biggest reason NRR excludes new business.

NRR Calculation Worksheet (Breakdown, Examples, and Conversion Traps)

Seeing the formula isn't enough if you don't know how to plug in your own numbers. Here's a worksheet you can fill in directly, three worked examples, and the easily missed "monthly-to-annual conversion trap."

Step-by-step worksheet

Enter your figures (existing customers only; exclude new business) in ①–④ to compute NRR.

Three worked examples

Assuming Starting MRR of $1,000K, here are three representative patterns.

Pattern B works out to (1,000 + 200 − 50 − 50) ÷ 1,000 × 100 = 110%, meaning the company grows 10% per month from existing customers alone without acquiring anyone new. Pattern C, at 90%, means revenue will shrink unless new business continually backfills the erosion.

The monthly-to-annual conversion trap

This is the most common mistake in practice. Do not simply multiply a monthly NRR by 12, or convert it to annual by naive arithmetic.

NRR is a ratio, and it compounds. Even if monthly NRR holds steady at 110%, annual NRR is neither "110% × 12 months" nor "110% added 12 times." In principle, each month's retention rate compounds multiplicatively, so the result differs from simple addition or multiplication.

In practice, what matters is to standardize internally on whether you use monthly or annual NRR and never mix them. Especially when comparing against other companies' benchmarks, always confirm whether they mean a "monthly figure" or an "annual (trailing-twelve-month) figure." Lining up numbers with different period definitions makes the comparison meaningless.

The NRR / GRR / Churn Rate Relationship Matrix

To use NRR correctly, you need to understand its relationship to two neighboring metrics: GRR (Gross Revenue Retention) and churn rate. Most explainers stop at comparing two metrics in prose; here we go further with a single matrix and a "four-quadrant diagnosis" you can use in practice.

The three-metric matrix

The key point is that GRR is capped at 100%. GRR reflects only the losses from churn and downgrades and excludes gains (Expansion), so it can never exceed 100% no matter how well you do. A GRR of 100% means "no one canceled and no one downgraded" — a perfect state. NRR, by contrast, includes Expansion, so it can exceed 100%.

The NRR × GRR four-quadrant diagnosis

Watching NRR alone can hide dangerous signs. Combining NRR and GRR into four quadrants gives a more accurate read on customer-base health.

Watch especially for the "hidden weakness" (NRR high × GRR low) quadrant. For example, one large customer's big upsell can make overall NRR look like 115% while, behind it, small and mid-sized customers churn one after another. This state carries heavy dependence on a single large account — the moment that one account defects, the numbers collapse. Even when NRR is high, if GRR is low, churn prevention should be the top priority.

That's why you always look at NRR and GRR together — and at churn rate too, to see which segments you're losing customers in. If NRR is "the speed you get from combining the accelerator (expansion) and the brake (churn)," GRR isolates just "how well the brakes work." Even if you're moving fast, if that speed merely masks weak brakes through a heavy press on the accelerator, you'll eventually crash. Only by reading both gauges at once can you judge the true health of your customer base.

NRR vs. ARR, MRR, and LTV

NRR is often discussed alongside other SaaS metrics and is easily confused with them. Here's how their roles differ.

MRR and ARR are absolute amounts ("how much"); NRR and churn rate are ratios ("how much was retained or lost"). And LTV (customer lifetime value) grows larger the higher NRR is (i.e., the longer customers stay and the more they expand). Keeping in mind the causal chain — improving NRR ultimately maximizes per-customer profitability (LTV) — helps you see how the metrics connect.

For how to track these KPIs together on a single dashboard, see our guide to sales/SaaS KPI visualization.

Watch the difference between revenue churn and customer churn

The "churn rate" discussed alongside NRR actually comes in two flavors. Confusing them leads to misreading NRR, so let's distinguish them:

• Customer churn (logo churn): The percentage of customers (by count) that canceled. If 1 of 10 accounts cancels, that's 10%.
• Revenue churn: The percentage of revenue lost to cancellation and contraction. Money-based.

Since NRR is a revenue-based metric, revenue churn is the more closely related one. For example, if only small accounts churned, customer churn may be high but revenue churn is small, so the hit to NRR is limited. Conversely, if one large account cancels, customer churn may be low but revenue churn is large, and NRR drops sharply. Reading NRR by "how much you lost," not "how many you lost," is essential.

What NRR Is "Good"? A Benchmark Interpretation Guide

"What NRR is good enough for my company?" is one of the most-asked questions. The short answer: the NRR benchmark varies sharply by customer segment (company size and price band), so there's no single right number. Here we organize the meaning of each level, segment benchmarks, and real examples from top SaaS.

What 100% / 110% / 120% mean

The 120%+ state is called "negative churn." It means existing customers' upsell and cross-sell gains fully cancel out — and then exceed — losses from churn and downgrades. Even though churn isn't zero, the existing base as a whole grows revenue — a very strong customer base.

Segment benchmarks (SMB / Mid-Market / Enterprise)

The appropriate NRR level shifts with the size and price of the customers you serve. Synthesizing various benchmark reports, the broad tendencies are:

These are tendencies repeatedly shown across B2B SaaS benchmark studies of private US companies (for example, SaaS Capital's annual survey of more than 1,000 companies), synthesized here. That said, the specific numbers vary by survey year and sample, so compare your own figure on a "same segment, same period definition" basis.

The important nuance: for SMB-focused products, small businesses fail or switch more often, so churn is higher — meaning that just holding NRR near 100% is genuinely hard. Don't judge "NRR is 100%, therefore failing" across the board; evaluate against your own segment characteristics.

Real examples from top SaaS (primary sources)

It's useful to check how top-tier SaaS companies actually disclose their NRR in official filings.

• Snowflake: Disclosed a net revenue retention rate of 131% for FY2024 (fiscal year ended January 31, 2024) (source: Snowflake's Q4 and full-year fiscal 2024 results release). Its usage-based pricing model drives high expansion.
• Datadog: Described its dollar-based net retention rate as in the mid-110s for Q3 2024 on a trailing-twelve-month basis (source: Datadog 8-K, quarter ended September 30, 2024).

So even for top SaaS, roughly 110–130% is the realistic range for "elite." Figures like 140% or 150% sometimes make headlines, but those reflect specific market conditions and growth phases and aren't necessarily sustainable. For your own targets, anchor to a realistic range.

The pricing model changes where NRR "moves"

For the same NRR, the lever for improvement changes with your pricing model. Knowing which model you run reveals where to focus.

• Usage-based pricing (charged by consumption): Expansion happens automatically as usage grows, so NRR structurally tends to run higher. This is why usage-based SaaS like Snowflake more easily posts high NRR. The flip side: when customers' usage falls in step with the economy or their own performance, downgrades occur more easily, so NRR is more exposed to market swings. The improvement focus is "usage adoption and expansion."
• Flat / seat-based pricing (per-user or per-plan): Expansion depends on plan upgrades and adding seats. Since unit price doesn't rise automatically as usage grows, proactive upsell proposals drive NRR. The improvement focus is "planned upgrades to higher tiers" and "horizontal rollout inside the account (more seats)."
• Hybrid (flat + usage): Has both properties — you defend with the flat base and pursue expansion through the usage component.

In short, your NRR targets and improvement tactics must be designed around the structure of your pricing model. When comparing NRR with other companies, note that the number's meaning changes depending on whether their pricing model matches yours.

NRR Varies by Region and Market Maturity

It's natural to want "the average NRR" for a given market, but reliable, market-wide public statistics are often scarce — especially in less mature SaaS markets. Where comprehensive figures don't exist, avoid asserting a single average.

A more useful reference is the way individual listed SaaS companies disclose NRR (net revenue retention) in their earnings and investor materials. Across these, holding around 110% tends to be regarded as solidly healthy in many markets.

Bear in mind that markets differ in how long subscription and SaaS models have been established and in how mature customers' upsell behavior is. Applying a top global SaaS figure (120–130%) directly as a target in a less mature market can set an excessively high bar. Compare your figure not against global leaders but against peers in the same region, same segment, and same calculation period — and against your own historical trend.

Local business customs also matter. Where annual budgeting and internal approval cycles dictate the timing of renewals and add-on orders, you can't always increase a contract the instant usage rises. Designing upsell proposals to land when the next fiscal year's budget is being set — in line with local purchasing customs — can decide whether expansion succeeds. Rather than comparing NRR side by side with other markets, set your target level with these customs in mind.

The Two Levers to Raise NRR: Create Expansion × Reduce Churn/Downgrades

From the structure of the formula, there are simply two ways to improve NRR: increase gains (Expansion) or decrease losses (Churn/Downgrades). The key is to determine which to invest in first.

Which should you invest in? (A reverse-lookup decision flow)

Rather than doing both blindly, work backward from your NRR and GRR figures to set priorities.

The principle is "first plug the holes (reduce churn), then add water (create expansion)." Pouring effort into upsell while churn keeps leaking is like adding water to a leaky bucket — inefficient.

Tactics to create Expansion

• Upsell and cross-sell: Move customers to higher tiers; propose add-on features, additional licenses, and related products. The key is to propose at the moment the customer feels the value.
• Value-aligned pricing: Designing pricing so unit price rises naturally with usage or outcomes turns customer success directly into expansion.
• Thorough onboarding: When customers feel the value early in their deployment, later usage expansion comes more easily. See also our guide to streamlining onboarding with a DSR.

For how to think about upsell and cross-sell — their differences and how to design proposal timing — see our article on the difference between upsell and cross-sell.

Tactics to reduce Churn/Downgrades

• Early detection of churn signals: Catch warning signs early — declining usage rates, fewer logins, a champion's departure — and respond proactively.
• Running a health score: Score customers' usage and satisfaction, and intervene first with those whose scores drop.
• Eliminating onboarding failure: Much churn stems from "they didn't feel the value at first" — early dropout. Invest heavily in hands-on support during early deployment.

Whether you can run both "offense (Expansion)" and "defense (churn reduction)" on data rather than gut feel determines success or failure at improving NRR.

A 90-day roadmap to improve NRR

NRR doesn't move overnight. It takes time — as long as your renewal cycle — for the effect of initiatives to show up in the number. Still, you can build the foundation in the first 90 days. Here's a practical order to work in.

• Days 1–30: Build the measurement foundation. First, standardize the formula and the definition of the denominator internally, and get to a state where you can correctly tally Starting MRR, Expansion, Downgrade, and Churn. At the same time, compute NRR and GRR side by side to see which quadrant you're in. Skip this and you can't measure the effect of later initiatives.
• Days 31–60: Systematize churn-signal detection. Define signals such as declining usage rates and login frequency or changes in support volume, and score them as a health score. Build a flow where CS proactively intervenes with customers whose scores drop, and first plug the Churn/Downgrade "holes."
• Days 61–90: Build an expansion playbook. Define the moments a customer feels value (onboarding completion, sustained adoption of a key feature) as "upsell opportunities," and set up a mechanism that surfaces a proposal at exactly that moment. Prioritize higher tiers and add-ons for customers showing results.

The crux of this roadmap is the order: "firm up defense (churn reduction) first, then advance to offense (expansion)." Rushing into upsell without plugging the holes only backfills the leak — the number won't accumulate. And the prerequisite shared by all three phases is "being able to see the customer's state in data."

Putting NRR to Work: Visualizing Customer Health with a DSR

Improving NRR ultimately comes down to how quickly and accurately you grasp the customer's state, and whether you can act at the right moment. Miss churn signals and Churn rises; miss upsell opportunities and Expansion stalls. Many explainers say "timing is important," but unless you can translate that timing into measurable signals, the frontline can't act.

This is where a Digital Sales Room (DSR)'s engagement data helps. A DSR is a dedicated space for sharing proposals, post-sale onboarding materials, and adoption guides with customers — and it lets you see "when, what, and how much" the customer viewed.

For instance, if a customer suddenly stops accessing the room, you can intervene early on the churn-risk signal; conversely, for a customer repeatedly viewing your higher-tier materials, you can propose an upsell at the moment interest rises. Running both "offense" and "defense" on data rather than gut feel is what drives continuous NRR improvement.

In traditional CS operations, customer state relied on the CS rep's subjective sense or fragmentary notes from periodic meetings. As a result, teams often noticed only after a cancellation decision had hardened, or regretted a missed upsell after the fact. With objective signals from engagement data, you can face customers on a consistent organizational standard rather than depending on a rep's experience and intuition. Especially in "tech-touch" and "low-touch" segments — where one CS rep covers many accounts and it's impossible to give every customer equal time — being able to prioritize and act on data is the realistic key to improving NRR. The fact that customer state carries over even when reps change is another big advantage in preventing knowledge from becoming siloed.

For the full feature set and benefits of a DSR, see our complete guide to the Digital Sales Room.

Common Pitfalls When Operating with NRR

NRR is a powerful metric, but misusing it leads to wrong executive decisions. Here are the mistakes that commonly happen in the field.

• Not standardizing the denominator: If teams differ on whether to use "starting existing-customer MRR" or some other base, the numbers become incomparable. Accidentally mixing in New MRR is a classic error. Document the formula internally and keep it consistent.
• Comparing monthly and annual figures interchangeably: As noted, monthly NRR and annual NRR are different things. Lining up numbers with different period definitions makes the comparison meaningless.
• Watching NRR while neglecting GRR: Even high NRR may be a "hidden weakness" state propped up by a few large expansions. Always look at GRR together to grasp the reality of churn.
• Large-account skew distorting the overall average: When one mega-deal expands, overall NRR jumps. Without decomposing by segment and cohort, the average masks reality.
• Ignoring cohort (signing-period) skew: A bias where only customers acquired in a certain period are upselling won't show up in overall NRR alone. Tracking by acquisition cohort reveals repeatable growth drivers.

NRR offers the convenience of "one number tells the whole story," but that very convenience carries the danger of taking it at face value without decomposition. The habit of viewing it multidimensionally — combined with segment, cohort, and GRR — is essential. When you see an NRR figure in a leadership meeting, make a habit of always asking: "Whose (which segment's), which period's, and how-defined value is this?" The same "NRR 110%" might be healthy expansion, or it might be a precarious large-account dependency. Whether you can read behind the number is what separates organizations that master NRR from those that don't.

Frequently Asked Questions (FAQ)

Conclusion

NRR (Net Revenue Retention) is the single number that shows whether you can retain and expand revenue from existing customers alone — the most important KPI for SaaS and subscription businesses. Let's recap the key points.

• The formula is "(Starting MRR + Expansion − Downgrade − Churn) ÷ Starting MRR × 100." New customers are excluded.
• Above 100% means growth from existing customers alone; 120%+ is the elite "negative churn" level.
• Always pair NRR (offense) with GRR (defense, capped at 100%). NRR high × GRR low is a sign of "hidden weakness" from large-account dependency.
• Benchmarks vary by segment: roughly 115–120% for Enterprise, while ~100% is solid for SMB. Even for top SaaS, Snowflake's 131% and Datadog's mid-110s are the realistic range.
• For improvement, the principle is "plug churn first, then stack expansion." A DSR's engagement data turns churn risk and upsell opportunities into measurable signals.

NRR is a convenient metric, but taking it at face value without decomposition leads to flawed decisions. Viewing it multidimensionally — combined with segment, cohort, and GRR — and visualizing customer state in data so you can act at the right moment, is what drives continuous improvement in NRR.