What Is NRR and How to Calculate It (The Straight Answer)
Net revenue retention (NRR) measures how much recurring revenue you keep from an existing customer base over a defined period, including upsells and cross-sells but excluding any dollars from new customers acquired after the period start. The practical formula is (Starting MRR + Expansion MRR − Contraction MRR − Churned MRR) ÷ Starting MRR.
If you came here asking what is NRR and how to calculate it, here is the immediate worked answer: start with $10,000 MRR from current accounts. Add $2,000 from seat upgrades or add-ons. Subtract $800 lost to downgrades and cancellations. You end at $11,200. Divide by $10,000 and you get 1.12, or 112% NRR. That is the entire mechanic in one breath.
What does NRR over 100% mean? It means your existing customers alone generated more revenue at period end than at period start, before any new sales. That single property turns NRR into the most watched efficiency metric in B2B SaaS because it shows product-led gravity.
The metric is distinct from gross revenue retention (GRR), which strips out expansion. Many top results define the ratio; few show you the cohort filtering that makes it honest. We will do that next.
Why I Built a 10-Minute NRR Workflow (And the Mistake That Cost Me a Board Meeting)
When I first owned retention reporting at a Series A SaaS company, I pulled a great 121% NRR number the night before a board meeting. I had included a $15,000 annual upgrade from a brand-new logo we landed via a referral program, thinking it counted as expansion. The CFO caught it at 7 a.m. The real NRR was 104%. That humiliation drove me to build a repeatable, template-driven process.
The thing nobody tells you about NRR is that it is only as trustworthy as your cohort definition. If new-logo revenue leaks into the numerator, you are measuring growth, not retention. I now use a free Google Sheets template (linked on our Net Revenue Retention Calculator page) that forces a clean separation of cohorts by account ID and start date.
If you prefer not to build from scratch, our calculator applies the same logic and includes a copyable sheet. But understanding the mechanics prevents the errors automated tools cannot catch, especially around proration and currency.
Experience also taught me that NRR is a lagging signal of product adoption yet a leading indicator of valuation only when computed with discipline. A polished number built on sloppy inputs will collapse in diligence.
Step-by-Step: Calculate NRR in 10 Minutes With Real Numbers
Follow this exact sequence. I have used it for monthly and annual closes across three companies and it mirrors the template columns.
1. Isolate Your Starting Cohort MRR
Pick a date and a defined set of customers active on that date. For a monthly view, use the MRR from existing accounts on day one. In our running example: $10,000 from 50 customers who were subscribed on the first.
2. Capture Expansion Revenue
Expansion includes seat adds, tier upgrades, and cross-sold products from those same customers. In the example, they bought $2,000 more per month. Record only changes from existing accounts; a new division of an existing client may count if billed under same account ID, but a separate new entity should not.
3. Subtract Contraction and Churn
Contraction is a downgrade (e.g., fewer seats). Churn is full cancellation. Combine them as lost MRR. Our example loses $800 total—$500 downgrade, $300 cancel. Do not net expansion against churn before the formula; keep lines separate for auditability.
4. Compute the Ratio
Ending MRR = 10,000 + 2,000 − 800 = 11,200. NRR = 11,200 / 10,000 = 1.12 or 112%. That is your answer. The math takes seconds; the data hygiene takes the other nine minutes.
Most people don’t realize that if you use annual contracts, you should run this on ARR with the same exclusion rules. ARR vs MRR nuances matter: a $120k ARR account that downgrades to $100k mid-year shows $20k contraction; if you only snapshot year-end, you’ll miss the dip and overstate NRR.
Worked Example: Annual NRR With Mixed Contract Terms
To show the edge cases, imagine a starting ARR cohort of $1,200,000 from 120 accounts. During the year, 10 accounts expand by $300,000 total, 5 downgrade by $80,000, and 3 churn completely representing $150,000. New logos worth $500,000 are excluded entirely.
Ending ARR for the cohort = 1,200,000 + 300,000 − 80,000 − 150,000 = 1,270,000. NRR = 1,270,000 / 1,200,000 = 105.8%. If you had mistakenly added the $500k new logos, you’d report 147%—a dangerous lie to investors.
This example reveals why I insist on a cohort lock column in the sheet: filter by original contract start date <= period start. The practice eliminates the most common inflation error.
Common Calculation Errors That Quietly Break Your NRR
I have audited dozens of investor decks. These are the repeated mistakes that distort the ratio.
- Mixing new-logo revenue into expansion: The fastest way to fake a >100% NRR. Always filter by customer ID active at start.
- Using gross instead of net: Forgetting contraction. GRR excludes expansion; NRR includes it. Confusing them flips your story.
- Currency mismatches: Expanding in EUR but billing in USD? Use a fixed FX rate for the period to avoid artificial swings of several points.
- Prorating errors: A customer who churns on the 15th should count as half-month lost MRR in a daily-based model, or full-month in a snapshot model—but be consistent.
- One-time fees as MRR: Onboarding or services revenue is not recurring. Including it violates the definition.
Another trap is treating retrospective credits as negative expansion. I treat service credits as contraction in the month issued, not as a separate line, to avoid double-counting. The Investopedia definition stresses recurring revenue for this reason, yet many operators miss it.
What NRR Over 100% and 120% Reveal About Compounding
Let’s decode the specific questions: what does NRR over 100% mean, and what does 120% net revenue retention mean? An NRR of 112% (our first example) means the existing base grew 12% net. At 120%, for every $100 of starting MRR you end with $120 from the same customers, without a single new sale.
Compound that: $1M starting ARR at 120% NRR grows to $1.2M next year, $1.44M the year after, $1.73M in year three—purely from the installed base. That is why investors prize >115% NRR; it is a compounding loop. But the trade-off: high NRR can mask acquisition problems if you stop winning new logos altogether.
120% NRR is not free growth. It signals your product delivers expanding value, but you still need new customers for scale beyond saturation.
Research from the Harvard Business Review shows that a 5% increase in retention can lift profits 25%–95%, yet NRR alone doesn’t measure acquisition efficiency. Use it alongside CAC payback.
Gross vs. Net Revenue Retention: What’s a Good Benchmark by Stage?
People also ask: what is a good grr and NRR? The answer depends on your company stage and sales motion. Below is the stage-based benchmark table I use when advising founders. Ranges reflect practitioner norms, not a single definitive study, and should be adjusted for segment.
| Stage | GRR Range | NRR Range | Notes |
|---|---|---|---|
| Seed / Early (Pre-Series A) | 70%–85% | 85%–105% | Small cohort, volatile; one churn skews the average. |
| Growth (Series A–B) | 88%–94% | 105%–115% | Sales-led expansion should push NRR above 100%. |
| Enterprise (Series C+) | 95%–100% | 110%–130% | Land-and-expand; above 120% is elite. |
GRR never includes upsells, so it is a ceiling. If GRR is below 90% at growth stage, your leakage is too high regardless of NRR. A healthy SaaS pair is GRR >90% and NRR >105%. Enterprise SaaS with strong expansion can sustain 120%+ NRR for years.
Cohort-Specific Tracking and Advanced Nuances
Calculating a single blended NRR hides stories. I recommend cohort-specific tracking: group customers by acquisition quarter and measure their NRR at months 6, 12, 24.
Why Cohorts Reveal Product-Market Fit
A 2022 cohort at 130% NRR but a 2024 cohort at 90% signals onboarding or packaging problems. Blended might show 110% and mask the warning. Cohort NRR is the diagnostic that saves roadmaps.
ARR vs. MRR: Which Should You Use?
If you sell annual contracts, ARR is cleaner but lagging. MRR normalization (divide ARR by 12) works if you have mid-term changes. The key is consistency across periods; do not switch mid-flight.
Handling Credits and Refunds
Service credits reduce effective MRR. I treat them as contraction in the month issued. This prevents the ghost expansion where a credit offsets a real upsell and hides churn.
The 4-Point NRR Purity Checklist (Unique Framework)
Before you report, run this. It is the tool-first approach missing from static formula pages and the same one embedded in my template.
- 1. Cohort Lock: Verify every revenue line ties to a customer active at period start.
- 2. Recurring Only: Exclude one-time, professional services, and new logos.
- 3. Currency & Proration Consistency: Same FX and timing rules as last period.
- 4. GRR Cross-Check: Compute GRR; if NRR < GRR, your formula is broken (impossible).
This checklist would have saved me that board meeting. It takes two minutes and blocks the errors that erode trust in diligence.
Why NRR Alone Won’t Save a Leaky Funnel
A high NRR can seduce teams into ignoring acquisition. I’ve consulted for a company with 125% NRR that was growing only 8% YoY because new logos fell off a cliff. The existing base compounded, but the total pie barely moved.
The corrective lens is to pair NRR with logo retention and CAC payback. If NRR is >100% but logo retention is <80%, expansion is concentrated in few accounts—a fragile position. Diversify expansion before celebrating.
How to Extract Clean NRR Data From Your CRM
In practice, I filter Salesforce or HubSpot by Account Created Date <= Period Start and sum the MRR field from opportunity records tagged Renewal, Upsell, or Churn. A simple SQL query on your billing table works too: SELECT sum(mrr) FROM subscriptions WHERE start_date <= '2024-01-01' AND (change_type IN ('expand','contract','churn')).
The thing most operators miss is that billing systems often record mid-period changes as separate line items. You must aggregate by account to avoid double counting. That step is where the ten minutes goes; the arithmetic is trivial.
Reporting NRR to Investors: How to Present Without Misleading
When you put NRR in a deck, show the cohort definition, the period, and the GRR alongside it. I once saw a founder present 130% NRR calculated on a tiny 10-account cohort; labeling it blended hid the risk. Always footnote exclusions.
If you use our Net Revenue Retention Calculator, export the input tab so the numbers are traceable. Investors trust transparent sheets over polished charts.
Remember that NRR is a diagnostic, not a vanity metric. Calculate it cleanly, and it will tell you whether your product sells itself after the first contract.
Your 10-Minute Action Plan
Open the Google Sheet, input starting MRR, expansion, contraction, churn. Use the formula. Compare to the stage table. If you’re at 112% like our example, you’re solid for growth stage. If you’re at 120%, you’re compounding. If below 100%, existing base is shrinking and needs intervention.
Then run the 4-point checklist. In ten minutes you’ll have a defensible NRR that survives scrutiny—something the top search results rarely give you beyond the bare equation.