NRR is the single clearest signal that a product keeps the customers it already has, and that those customers spend more over time. When NRR is above 100 percent, the base alone grows the company.
NRR (net revenue retention) measures how much recurring revenue a company keeps and grows from its existing customer base over a period, excluding new logos. The formula is (starting ARR plus expansion minus contraction minus churn) divided by starting ARR, times 100. A result of 100 percent is flat, 110 percent or higher is healthy SaaS, and 120 percent or higher is world-class.
The six things to know before you quote your NRR number in a board deck, a sales deal, or a fundraise.
NRR measures the recurring revenue your company retains and grows from the customers you already had at the start of a period. It intentionally ignores new logos acquired during the period. The question it answers is simple: without any new customers, does the current customer base still grow revenue?
NRR equals starting ARR plus expansion minus contraction minus churn, divided by starting ARR, times 100. Expansion includes upsell, cross-sell, and usage growth. Contraction is downgrades and seat reductions. Churn is customers who canceled. Only the opening cohort counts, which is what makes NRR different from total growth.
Below 90 percent means the base is leaking. 100 percent means the base is flat. 110 percent or higher is healthy SaaS. 120 percent or higher is world-class territory and historically where companies like Snowflake and Datadog have printed numbers that attracted premium multiples at IPO and after.
Upsell (moving a customer to a higher tier), cross-sell (adding another product), and usage-based consumption all count toward expansion. Price increases on renewal also count. Anything that increases ARR from an existing customer during the period lifts NRR, which is why pricing design matters as much as retention design.
A customer signed on day one of the period is a new logo, and new-logo ARR is tracked separately. If you include new logos in your NRR math, the number is wrong. The whole point of NRR is to isolate the health of the existing base from the performance of the sales team acquiring new accounts.
Investors read NRR before most other metrics because it compounds. A company with 130 percent NRR grows revenue roughly thirty percent per year with zero new sales. That is the fingerprint of a product customers rely on, expand on, and renew on. NRR below 100 is the opposite signal, and no acquisition rate makes it up for long.
The NRR formula is short, but the four inputs are where most teams get it wrong. The cohort you measure, the period you measure, the currency you measure in, and how you treat partial-month moves all change the number. Agree on the definitions before you quote the result. Below is the straight version of the math, followed by the pieces that trip teams up in practice.
Freeze the list of customers who had active recurring revenue on the first day of the period, usually a quarter or a trailing twelve months. That is the opening cohort. Every expansion, contraction, and cancellation during the period only counts if it came from a customer on that opening list.
Sum the ARR of every customer in the cohort as of day one. Use the same currency and the same ARR definition you use everywhere else, usually monthly recurring revenue times twelve, excluding one-time fees, professional services, and anything non-recurring. This is the denominator.
Total every dollar of recurring revenue growth from the opening cohort during the period: tier upgrades, extra seats, additional products, higher usage tiers, and renewal price increases. Only the opening cohort counts. A customer who signed up mid-period and then upgraded is not in this number.
Contraction is seat or tier reductions that leave the customer paying less but still active. Churn is the recurring revenue of customers who canceled during the period. Both come only from the opening cohort. Add contraction and churn together to get total ARR lost from the base.
Take starting ARR, add expansion, subtract contraction, subtract churn, then divide by starting ARR, then multiply by 100. The result is the NRR percentage. A 130 percent result means the opening cohort now generates thirty percent more revenue than it did when the period began.
Opening ARR of two million. During the year the cohort adds four hundred thousand in expansion, loses one hundred thousand in contraction, and loses one hundred thousand in churn. Numerator: two million plus four hundred thousand minus one hundred thousand minus one hundred thousand equals two point two million. NRR equals two point two divided by two, times one hundred, which is 110 percent.
There is no universal NRR target. The benchmark depends on segment, pricing model, and go-to-market motion. A product-led tool selling to small businesses looks healthy at a different number than an enterprise platform selling six-figure annual contracts. The ranges below are the ones most investors and operators use when they read a SaaS board deck.
When NRR sits below 90 percent, the existing customer base shrinks faster than expansion can refill it. Every dollar of new sales has to replace lost revenue before it starts growing the company. This is the warning zone, and it usually points to a product fit problem, a pricing problem, or a retention program that does not exist.
The base is roughly flat. Expansion is covering most but not all of the losses. The company can still grow through new logos, but every new customer has to carry its own weight. This range is common in SMB SaaS where price points are low and natural attrition is high.
The base alone grows revenue. New logos stack on top of a growing floor. This is a reasonable range for mid-market SaaS and a common target for teams in year two or three of a product. Investors call this the point where retention starts doing real work on the growth rate.
The base now grows roughly ten to twenty percent on its own. This is the band that most well-run enterprise SaaS companies target. It implies a mix of good retention, deliberate expansion motion, and a pricing model that rewards usage or seat growth as the customer gets more value.
The base grows twenty percent or more per year with zero new sales. Historically this is Snowflake and Datadog territory, both of which have printed NRR north of 130 percent during peak periods. Usage-based pricing and multi-product platforms are the two structural patterns that most often produce this result.
Enterprise, mid-market, and SMB NRR benchmarks are different. SMB tends to run lower because small businesses churn more often. Enterprise runs higher because contracts are stickier and expansion lands in larger increments. Compare your NRR to companies in your segment and pricing model, not to the public SaaS leaders at the top of the chart.
NRR is deceptively simple to miscalculate. A clean formula sits on top of four inputs that each hide a modeling choice. The mistakes below are the ones that come up in diligence calls, finance team reviews, and the first time a founder tries to explain the number to a board. Fix these and the number stops moving when someone asks a follow-up question.
The most common error. If a customer signed during the period, they belong in new-logo ARR, not in NRR. Including them pushes the number up and hides whether the existing base is actually healthy. The cohort is frozen on day one of the period and nothing added later touches the calculation.
Reporting trailing twelve-month NRR one quarter and quarterly NRR the next makes the trend meaningless. Pick one period definition and hold it. Finance, sales ops, and the board deck should all pull from the same cohort rule, or the number will drift every time it is quoted.
ARR is annual, MRR is monthly, and multiplying one by twelve to get the other introduces drift during months with mid-period changes. Pick one as the system of record. If MRR is the system of record, express NRR off MRR consistently. If ARR is, express it off ARR. Switching between them is where small errors become big ones.
Professional services, implementation fees, overage charges that will not recur, and other one-time line items do not belong in NRR. The denominator and numerator should both be recurring only. Including one-time revenue inflates expansion in the quarter it lands and distorts the trend.
A customer who drops from a higher tier to a lower tier is contraction, not churn. They are still a customer. Rolling them into churn hides the downgrade motion and makes it harder to find the pricing or product issue causing the slide. Keep the two categories separate on the movement log.
Companies with international customers see NRR move with currency rates. Fix the exchange rate for the whole period, or report NRR on a constant-currency basis, so the number reflects actual customer behavior rather than how the dollar moved against the euro last quarter.
The NRR formula is easy once the underlying movement log is clean. The hard part is making sure every renewal, every upsell, every downgrade, and every cancellation is captured against the customer record as it happens. That movement log is where NRR comes from, and it is the piece a CRM is uniquely positioned to own.
Every expansion, contraction, and churn event rolls up to the company record. Multi-entity customers get a parent-child structure so a global account with ten subsidiaries is still one NRR unit when leadership reads the number. The record is the home for the full contract history.
Each renewal is a deal with a close date, an owner, and a stage. The pipeline shows what is at risk, what is likely to grow, and what is likely to shrink. Customer success and sales share the same view, so a renewal does not slip because nobody noticed it was ninety days out.
Expansion is tracked as a separate deal type, not jammed into the renewal. That keeps the renewal number clean and makes expansion visible as its own forecast. The two together are what move NRR up, and separating them is how a revenue leader diagnoses where growth is coming from.
When a customer drops a seat count or steps down a tier, the change is logged as a contraction event on the customer timeline. The ARR delta is captured, the reason code is captured, and the number flows into the NRR calculation. Contraction is a signal, not a shameful line item to omit.
When a customer cancels, the record logs the churn date, the ARR lost, and the reason. Reason codes compound into the single most useful artifact for the retention team: a honest answer to the question of why customers leave, grouped by segment, product, and tenure.
The combined log of renewals, expansions, contractions, and churn is where NRR is calculated. Finance pulls it monthly. Leadership reads it quarterly. The sales team sees it live. Because every event is tied to the same customer record, the NRR number stops being a modeled estimate and starts being a reported fact.
Strkr keeps renewals, expansion, contraction, and churn on one customer record, so the NRR number comes straight from the movement log. Pricing is published. The feature pages show exactly what ships today.
NRR stands for net revenue retention. It is sometimes called net dollar retention (NDR), especially by publicly traded SaaS companies in their investor filings. The two terms describe the same metric: how much recurring revenue a company retains and grows from its existing customer cohort over a defined period, expressed as a percentage.
GRR (gross revenue retention) measures only what is kept from the existing base, with a maximum value of 100 percent, because it excludes expansion. NRR includes expansion, so it can exceed 100 percent. GRR shows pure retention health. NRR shows retention plus the expansion motion stacked on top of it. Investors read both because they diagnose different problems.
For enterprise SaaS, 110 to 120 percent is healthy and 120 percent plus is world-class. For mid-market, 105 to 115 percent is the common target. For SMB-focused products, anything above 100 percent is strong, because SMB churn runs structurally higher. Compare your NRR to companies in your segment and pricing model, not to the top public SaaS leaders.
Most companies report NRR on a trailing twelve-month basis, updated monthly or quarterly. The trailing twelve-month window smooths out seasonal lumpiness and matches how investors evaluate the metric. Quarterly or monthly point-in-time NRR is useful for operational diagnosis, but the headline number quoted to the board and to investors is almost always TTM.
No. NRR intentionally excludes new logos acquired during the period. The cohort is frozen on day one, and only revenue movements from that opening cohort count. New-logo ARR is tracked separately so the health of the existing base can be read without being hidden by sales performance on new accounts.
Yes, and that is the goal. NRR above 100 percent means the existing customer base generates more recurring revenue at the end of the period than it did at the start, after accounting for all losses. The best public SaaS companies have historically operated at 120 to 140 percent NRR, with Snowflake and Datadog among the examples cited most often.
Customer retention rate counts customers kept, treating a tiny customer the same as a huge one. NRR counts revenue kept, so a large customer who downgrades hits NRR harder than ten small customers who leave. For revenue analysis, NRR is the stronger metric because it weights the number by economic impact.
A CRM centralizes the renewal pipeline, the expansion pipeline, and the contraction and churn log on one customer record. That shared view means renewals do not slip, expansion opportunities get worked, downgrades are seen early, and the reason codes on cancellations feed back into product and pricing decisions. The CRM is where the NRR number is calculated as a reported fact rather than a modeled estimate.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
