Net Revenue Retention (NRR): The Complete Guide

Net revenue retention (NRR), also called net dollar retention (NDR), is the percentage of recurring revenue a company retains from a defined cohort of existing customers over a set period, after accounting for upgrades, downgrades and cancellations, but excluding any revenue from new customers acquired during that period. It answers a narrower and more diagnostic question than overall revenue growth: if you froze new sales entirely, would your existing customer base still grow, stay flat, or shrink in revenue terms?

This distinction matters because a business can post impressive top-line growth purely by adding new logos while quietly leaking revenue from its existing base through churn and downgrades. NRR strips out that noise and isolates the health of the installed customer base, which is why it has become one of the most closely watched metrics in SaaS reporting, alongside MRR and churn rate.

A company can have NRR above, at, or below 100 percent. Above 100 percent means expansion revenue from existing customers (upgrades, seat additions, usage growth) more than offsets the revenue lost to downgrades and cancellations. Below 100 percent means the existing base is shrinking in revenue terms even before any new customers are added.

The standard formula for net revenue retention is written in plain terms as: NRR equals (Starting MRR plus Expansion MRR minus Contraction MRR minus Churned MRR) divided by Starting MRR, multiplied by 100 to express it as a percentage. Crucially, new customer MRR added during the period is left out of the numerator entirely.

Worked example: suppose a cohort of customers started the quarter contributing 100,000 in MRR. Over the quarter, existing customers in that cohort added 12,000 in expansion MRR by upgrading plans, lost 4,000 in contraction MRR from partial downgrades, and 6,000 in churned MRR from customers who cancelled outright. NRR would be calculated as (100,000 + 12,000 - 4,000 - 6,000) divided by 100,000, which equals 102,000 divided by 100,000, or 102 percent.

That 102 percent means the existing cohort alone grew revenue by 2 percent over the quarter, even though some customers left and some downgraded. If the same cohort had instead only generated 4,000 in expansion against the same 4,000 contraction and 6,000 churn, NRR would fall to (100,000 + 4,000 - 4,000 - 6,000) divided by 100,000, or 94 percent, meaning the base shrank by 6 percent despite some expansion activity.

Gross revenue retention (GRR) measures the same cohort but deliberately excludes expansion revenue, so it can never exceed 100 percent. The formula is (Starting MRR minus Contraction MRR minus Churned MRR) divided by Starting MRR. GRR shows the pure retention floor: how much revenue would remain if no customer ever expanded, only stayed flat or shrank.

Using the earlier example, GRR for the first scenario would be (100,000 - 4,000 - 6,000) divided by 100,000, which equals 90 percent. NRR for the same cohort was 102 percent. The 12-point gap between GRR and NRR is entirely explained by expansion revenue, and that gap is informative in its own right: it shows how reliant the company's retention story is on upsell rather than simply avoiding losses.

A useful way to read the two side by side: high GRR with low NRR suggests a company retains customers well but fails to grow them, often a pricing or product-adoption issue. Low GRR with unusually high NRR can be a warning sign, since it may mean a small number of large accounts are expanding fast enough to mask significant churn elsewhere in the base, which is not a sustainable pattern once those large accounts mature.

NRR is a single number, but its drivers are three distinct revenue movements, and separating them is essential for diagnosing what is actually happening inside the customer base rather than treating NRR as a black box.

Expansion MRR

Expansion is additional recurring revenue from customers already in the cohort: seat additions, plan upgrades, add-on modules, or usage-based charges that grow as the customer's own usage grows. Expansion is the only lever that can push NRR above 100 percent, so understanding what drives it (self-serve upgrades versus sales-led upsell versus natural usage growth) tells you how repeatable it is.

Contraction MRR

Contraction is a reduction in recurring revenue from a customer who remains subscribed, such as downgrading to a cheaper plan, removing seats, or reducing usage tier, without cancelling entirely. Contraction is sometimes overlooked because the customer has not churned, but it erodes NRR just as directly as churn does and can be an early warning sign that a customer is heading toward cancellation.

Churned MRR

Churned MRR is revenue lost entirely when a customer cancels their subscription. This links directly to the concepts covered in SaaS churn rate, since the dollar value of churned MRR in a period is the numerator used in revenue churn calculations. A company can have a low customer count churn rate but a high revenue churn rate if the customers who leave tend to be larger accounts.

Tracking these three components separately each month or quarter, rather than only the combined NRR figure, makes it possible to see, for example, whether a dip in NRR was caused by a slowdown in expansion (a growth problem) or a spike in churn (a retention problem), which require very different responses.

NRR is always measured against a specific starting cohort, typically defined as every customer who was active and paying at the beginning of the period, and it tracks only that fixed group forward in time, regardless of how the overall customer base changes. New customers who sign up during the period are tracked separately and only enter the NRR calculation once they become part of a future cohort's starting point.

In practice this means building a cohort table where rows represent the month or quarter a customer's subscription started, and columns represent subsequent periods, with each cell showing that cohort's MRR relative to its own starting MRR. This is the same structural approach used for cohort-based churn and lifetime analysis, closely related to the methodology used to calculate customer lifetime value.

Reporting NRR by cohort rather than as one blended company-wide number reveals whether retention is improving or deteriorating over time. If more recent cohorts consistently show lower NRR than older ones, that can indicate a shift in the type of customer being acquired, a pricing change, or a decline in onboarding quality, well before it shows up in blended company-wide figures.

Because Stripe records subscription start dates, upgrade and downgrade events, and cancellation timestamps, cohort-based NRR can be built directly from Stripe subscription and invoice data, which is the same underlying data used for Stripe revenue analytics more broadly.

Monthly NRR and annual NRR measure the same underlying concept but over different time windows, and they behave quite differently in practice. Monthly NRR is calculated using the formula above but with a one-month starting and ending point, while annual NRR compares a cohort's MRR at the start of a twelve-month period to its MRR twelve months later.

Monthly NRR tends to be noisier. A single large customer downgrading or cancelling in one month can swing the monthly figure significantly, particularly for early-stage companies with a small number of large accounts, since the starting MRR base is smaller and each event carries more relative weight. Annual NRR smooths this out because it captures a full year of expansion and contraction activity within each customer relationship, including seasonal patterns that would distort any single month.

A practical approach many finance teams use is to track monthly NRR internally to catch problems early, while reporting a trailing twelve-month NRR externally to investors and boards, since that figure is less prone to being misread due to short-term noise. When reviewing an unfamiliar company's NRR figure, it is always worth asking which measurement window was used, since a 118 percent monthly NRR and a 118 percent annual NRR do not carry the same weight or reliability.

NRR above 100 percent matters to investors because it demonstrates that a company's revenue can grow without relying entirely on new customer acquisition, which reduces both the cost and the risk profile of future growth. A business with NRR consistently above 100 percent effectively has a built-in growth engine sitting inside its existing customer base.

Consider two companies both currently at 1,000,000 in ARR. Company A has NRR of 90 percent, meaning its existing base alone would shrink to 900,000 within a year if new sales stopped entirely. Company B has NRR of 115 percent, meaning its existing base alone would grow to 1,150,000 over the same period with zero new customer acquisition. Company B needs to spend far less on new customer acquisition to hit the same overall growth target, and its revenue is inherently less fragile.

This is commonly discussed as a rule of thumb rather than a fixed rule: NRR above 100 percent is generally viewed as healthy, NRR in the 110 to 130 percent range is often cited as strong for venture-backed, mid-market or enterprise-focused SaaS businesses, and NRR below 90 percent tends to raise questions about product fit, pricing, or customer success execution. These ranges vary meaningfully depending on customer segment, contract size and business model, and should not be treated as universal thresholds.

Investors also look at NRR alongside MRR growth rate and ARPU trends to understand whether growth is coming from more customers, larger customers, or existing customers spending more, since each pattern implies a different set of risks and a different quality of revenue.

Improving NRR means either increasing expansion revenue, reducing contraction, or reducing churn, and the most effective approach usually combines several levers rather than relying on a single fix.

• Design pricing that scales naturally with customer value, such as usage-based or seat-based tiers, so that revenue grows automatically as a customer's usage or team size grows, rather than requiring a manual renegotiation.
• Build clear upgrade paths and in-app prompts that surface higher tiers or add-on modules at the point a customer is approaching a usage limit, converting natural growth into expansion revenue rather than friction.
• Invest in onboarding and early customer success, since a large share of both churn and contraction typically occurs in the first few months of a subscription when a customer has not yet reached the point of clear value.
• Monitor usage and engagement signals to flag accounts at risk of contraction or cancellation before the renewal date, giving customer success teams time to intervene. See reducing churn using Stripe data for more detail on building this kind of early warning process.
• Introduce annual contracts with modest built-in price increases or capacity step-ups, which convert what would otherwise be flat renewals into small amounts of expansion revenue.
• Segment customers by plan, size and industry to identify which segments have low NRR, since a single blended NRR figure can hide serious problems in one segment that are masked by strong performance in another.

It is worth treating contraction as seriously as churn in this process. A customer who downgrades rather than cancelling is often signalling the same underlying dissatisfaction, and addressing the root cause, whether that is a support issue, missing feature, or mismatched plan, tends to improve both contraction and churn simultaneously.

NRR interacts with valuation multiples because it is one of the clearest signals of revenue quality available to an investor, and revenue quality directly affects how much future revenue is discounted or trusted when pricing a SaaS business. Two companies with identical current ARR and growth rate can be valued very differently if one has NRR of 85 percent and the other 115 percent, because the second company's growth is inherently less dependent on continuous new sales execution.

Subscription Metric uses a straightforward reference multiple of 5x ARR (calculated as current MRR multiplied by 60) as a simple, transparent baseline for estimating enterprise value from live Stripe data. That baseline is a starting point rather than a fixed rule: in practice, buyers and investors adjust multiples up or down from any baseline based on factors including NRR, growth rate, gross margin and customer concentration. A business with strong NRR is more likely to be valued toward the upper end of whatever range a buyer is considering, while weak NRR is one of the more common reasons a multiple gets discounted during diligence.

These ranges are widely referenced rules of thumb rather than fixed industry standards, and actual valuation outcomes depend on the specific deal, market conditions and the wider set of metrics reviewed, including SaaS valuation multiples more broadly. Still, few single metrics carry as much weight in a diligence conversation as a consistent, well-documented NRR trend, because it speaks directly to whether the revenue being valued today is likely to still be there, or larger, in twelve months.

Tracking NRR from Stripe data in practice means combining subscription lifecycle events, namely creation, upgrades, downgrades and cancellations, with invoice history to compute each cohort's MRR at the start and end of a period. Stripe stores enough detail in its subscription and invoice objects to reconstruct this, but doing it manually via export and spreadsheet reconciliation is time consuming and error prone, particularly once a company has more than a handful of pricing plans or add-ons.

A dashboard connected directly to a restricted, read-only Stripe key can automate this process: pulling subscription and invoice data, grouping customers into cohorts by start date, and calculating expansion, contraction and churned MRR for each period without manual data handling. This is the same underlying data set used to compute related metrics such as MRR calculated directly from Stripe, churn rate and customer lifetime value, so building NRR reporting alongside these metrics, rather than as a separate one-off exercise, tends to be more reliable and easier to keep current. For a broader view of how these metrics fit together, see the SaaS metrics glossary.

When combining multiple Stripe accounts, for example if a company runs separate Stripe keys for different products or regions, cohorts should generally be tracked separately per account before being combined into a blended NRR figure, since merging customer bases with very different pricing structures into a single cohort can obscure meaningful differences in retention performance between the two.