NRR (Net Revenue Retention)

NRR (net revenue retention) is a metric that measures what percentage of the recurring revenue from your customer base is retained and expanded over a period, excluding any revenue from new customers. It answers a key question for any subscription or recurring-contract business: if I stopped selling, would my current base grow or shrink on its own?

It is calculated by taking the recurring revenue at the start of the period, adding expansions (upsells, cross-sells, plan upgrades) and subtracting contractions (plan downgrades) and churn (cancellations). The result is divided by the starting revenue. An NRR above 100% means the base grows on its own, even while losing some customers, because the ones who stay spend more. It is one of the indicators tracked live on a Metrix dashboard alongside churn and LTV.

Unlike gross revenue retention (GRR), which only measures how much you lose and can never exceed 100%, NRR can go above 100% because it builds growth into the calculation.

NRR has become one of the most closely watched indicators at recurring-revenue companies because it sums up, in a single number, the health of the business without the noise of acquisition. While new sales depend on marketing and the sales team, NRR measures something deeper: whether the value you deliver justifies customers staying and paying more over time. A business with a high NRR is, in practice, a business that compounds on itself.

How it is calculated

The formula starts from the recurring revenue at the beginning of the period (typically the cohort's monthly MRR or annual ARR) and applies the movements of that same base:

• Starting revenue: what your customer base was billing at the start of the period.
• (+) Expansion: plan upgrades, upsell, cross-sell and additional seats or usage within existing accounts.
• (-) Contraction: customers who downgraded or reduced volume but are still active.
• (-) Churn: revenue lost from accounts that canceled entirely.

NRR = (Starting revenue + Expansion - Contraction - Churn) / Starting revenue × 100. The golden rule: new customers are never added; you only measure what happened with the ones you already had.

A concrete example

An Argentine business management software company starts the quarter billing $10,000 a month to its customer portfolio. During the period, several customers buy additional modules and add $2,000 in expansion; two accounts downgrade by $500 and one cancels for $800. The calculation: (10,000 + 2,000 - 500 - 800) / 10,000 = 107%. Even though it lost an entire account, its base grew 7% on its own, without selling to anyone new. If instead it had had no expansion, the same churn and contraction would have left it at 87%, a base that is bleeding out and forces the company to sell more and more just to stay level.

Why it matters so much

NRR defines how much sales effort you need just to grow. With an NRR of 120%, each month you start 20% ahead before closing a single new deal; the sales team works on a floor that keeps rising. With an NRR of 85%, you are on a treadmill: part of the new work goes into plugging holes. That is why it is one of the metrics investors look at most closely in funding rounds, because it projects future growth better than the billing of a single isolated month.

NRR vs GRR

The most common confusion is mixing up NRR with its cousin, gross revenue retention (GRR). The difference defines what you are measuring:

Looking at both together avoids a trap: a high NRR can hide a lot of churn that the expansion of a few large accounts is covering up. If your NRR is 110% but your GRR is 80%, you are losing a fifth of your base every year and surviving on concentrated upsells, a fragile balance.

Common mistakes when measuring it

The first is including new customers in the calculation, which inflates the number and strips it of all meaning (that is the total growth metric, not retention). The second is mixing very different segments into a single NRR: enterprise accounts and small customers often behave in opposite ways, and averaging them hides the problem. The third is not defining the cohort and the period properly: comparing different bases month to month produces numbers that cannot be interpreted. The fourth, and subtler, is celebrating a high NRR without looking at concentration: if a single giant account contributes all of the expansion, the indicator is an illusion that breaks the day that account leaves.

That is why NRR really pays off when it is tracked by segment, by cohort and alongside GRR and churn, on a dashboard where the data updates on its own instead of being built by hand in a spreadsheet every quarter.