Net Revenue Retention Formula: Net Dollar Retention, GRR

Net revenue retention (NRR) measures how much recurring revenue you keep from an existing group of customers over a period, after expansion, downgrades, and cancellations, and excluding any revenue from new customers. The formula is starting recurring revenue plus expansion, minus contraction, minus churn, divided by starting recurring revenue. An NRR above 100 percent means your existing customers grew enough to more than replace everyone who left or downgraded.

Net revenue retention is the single number investors ask for first, and the one operators most often calculate wrong. The arithmetic is not hard. What trips teams up is scope: which customers belong in the cohort, whether new logos sneak into the numerator, and whether a discount counts as contraction or as churn. Get the scope wrong and you can publish a 118 percent NRR that quietly includes new business, which is not a retention metric at all.

This page covers the formula, a full worked calculation with real dollar movements, how NRR relates to gross revenue retention and to customer retention rate, whether net dollar retention is a different metric (it is not), and what the published benchmarks actually say once you stop comparing yourself to a blended median.

What is net revenue retention?

Net revenue retention is the percentage of recurring revenue retained from customers you already had at the start of a period, once you account for the money they added and the money they took away. Take a fixed cohort, the customers on the books on day one, and follow only their revenue to the end of the period. Upgrades and additional seats push the number up. Downgrades and cancellations pull it down. Revenue from customers who signed after day one is excluded entirely.

That exclusion is the whole point. NRR isolates the health of the base. A company can grow ARR 60 percent a year on the strength of new sales while its existing customers shrink, and NRR is the metric that exposes it.

What is the net revenue retention formula?

There are four retention metrics in this family, and they differ only in what they let into the numerator.

Two rules make the formula behave. First, the starting cohort is frozen: only customers present on day one are in it, for the entire calculation. Second, expansion means additional recurring revenue from those same customers, seats, tiers, usage, or added modules. A one-time implementation fee is not expansion because it is not recurring.

How do you calculate net revenue retention?

Work it in dollar movements rather than trying to read an ending balance off a report, because the ending balance almost always has new customers baked into it. Here is a quarter for a B2B software company with a fixed starting cohort.

NRR is $406,000 divided by $400,000, which is 1.015, or 101.5 percent.

GRR strips the expansion out of the numerator: $400,000 minus $12,000 minus $28,000 is $360,000. Divided by $400,000 that gives 90.0 percent.

The $95,000 of new business never appears in either calculation. If you had let it in, the numerator would have been $501,000 and you would have reported an NRR of 125 percent, a number that says nothing at all about retention.

Note the gap: 101.5 percent net against 90.0 percent gross, an 11.5 point spread. That spread is the amount of leakage the upsell motion is papering over. A company with NRR 101.5 and GRR 90 has a very different problem from one with NRR 101.5 and GRR 99.

Converting a quarterly rate to an annual rate

Retention compounds, so you cannot multiply. A quarterly NRR of 101.5 percent annualizes to 1.015 to the fourth power, which is 1.0614, or 106.1 percent. Simply multiplying 1.5 points by four would have given 106.0 percent, close enough here because the rate is near 100. At a quarterly NRR of 90 percent, though, compounding gives 0.90 to the fourth, which is 65.6 percent, while naive multiplication suggests 60 percent. The same compounding trap applies in reverse when you annualize a monthly churn rate.

Where the numbers come from

Most teams pull cohort revenue out of the billing system, and this is where a two-hour analysis becomes a two-day one. You need MRR by customer at two points in time, plus a movement log that classifies every change as expansion, contraction, or churn. If your billing platform or payment processor only hands you a PDF statement rather than a queryable export, budget the first hour of every reporting cycle for turning those statements into a workable spreadsheet before anyone calculates anything. Teams that automate that step tend to move from an annual NRR number to a monthly one, and a monthly one is what actually changes behavior.

Net revenue retention vs gross revenue retention

NRR includes expansion revenue. GRR does not. That single difference is why the two metrics answer different questions and why serious operators track both.

GRR measures leakage. It is a defensive number, the floor of your business, and it can never exceed 100 percent because there is no term in the formula that adds revenue. The best possible GRR is a quarter in which nobody downgraded and nobody left.

NRR measures leakage net of growth inside the base. It is an offensive number and it can, in principle, run well past 130 percent if a usage-based pricing model expands with the customer.

Read them together. High NRR with low GRR means you are losing customers steadily and hiding it by growing the survivors, which works until the pool of survivors runs dry. High GRR with NRR near 100 means you keep everyone but sell them nothing more, which is a pricing and packaging problem, not a retention problem. The fix in each case is completely different, which is why a single blended number is dangerous.

Why can gross revenue retention never exceed 100 percent?

Because every term in the GRR numerator either subtracts revenue or leaves it alone. Starting MRR minus contraction minus churn cannot be larger than starting MRR, since contraction and churn are never negative. If someone shows you a GRR of 104 percent, they have let expansion into the numerator and computed NRR by another name. This is the fastest way to check whether a retention dashboard is trustworthy.

Is net dollar retention the same as net revenue retention?

Yes. Net dollar retention (NDR) and net revenue retention (NRR) are the same calculation with different labels. The terms are used interchangeably, with NDR more common in investor decks and NRR more common in finance and customer success teams. Neither term implies a different scope, cohort, or time window. If a company reports both and the values differ, the difference comes from their definitions of the cohort, not from the metric.

The one place the vocabulary matters is diligence. Ask any company quoting NDR two questions: does the cohort exclude customers acquired during the period, and does the numerator exclude one-time fees? Roughly stated NDR figures fall apart on both.

How is net revenue retention different from customer retention rate?

Customer retention rate counts logos. NRR counts dollars. They routinely disagree, and the disagreement is informative.

Take the same quarter. The cohort started with 250 customers and 12 cancelled, so CRR is 238 divided by 250, or 95.2 percent. Meanwhile NRR came in at 101.5 percent. Losing 4.8 percent of accounts while growing revenue 1.5 percent tells you the accounts you lost were small and the accounts you kept expanded. That is the classic profile of a company moving upmarket, and it is a good outcome, right up until the small accounts you are shedding turn out to be the top of your funnel.

Invert the picture and it is an alarm: CRR of 99 percent with NRR of 92 percent means almost everybody stayed and almost everybody bought less. That is a value delivery problem that no customer success outreach will fix.

What is a good net revenue retention rate?

It depends almost entirely on what you charge, and comparing yourself to a blended cross-industry median is the most common benchmarking mistake in SaaS. Retention rises with contract value, because larger contracts come with more seats to add, more expansion surface, and buyers who cannot easily switch.

The figures below are reported ranges from published benchmark studies, not universal facts. Treat them as orientation and rebuild the comparison against companies with your pricing model.

Two specific anchors are worth knowing. SaaS Capital, which surveys private B2B SaaS companies annually, reported a median NRR of 102 percent for the $25,000 to $50,000 ACV tier, with the top quartile at 111 percent and the bottom quartile at 97 percent, and found that higher net retention correlates with higher contract values. Separately, the widely quoted Bessemer framing calls 100 percent good, 110 percent better, and 120 percent best, though that scale was built with growth-stage enterprise software in mind and flatters nobody selling to small businesses.

The practical read: at SMB pricing, holding 100 percent is a genuine achievement. At enterprise pricing, 100 percent is a warning.

Common mistakes when calculating NRR

• Letting new customers into the numerator. The most common error, and it inflates the number the most. If ending MRR came off a dashboard rather than a cohort query, assume it is contaminated.
• Counting one-time revenue as expansion. Implementation fees, professional services, and overage settlements are not recurring, so they do not belong in either side of the ratio.
• Reporting a single company-wide NRR. A blended figure across SMB and enterprise hides both. Segment it, then act on the segments.
• Annualizing by multiplication. Retention compounds. Raise the periodic rate to the power of the number of periods.
• Ignoring involuntary churn. Failed cards and expired payment methods land in the churn term of the formula and drag GRR down even though those customers never chose to leave. A functioning dunning and failed-payment recovery process is one of the few retention levers that pays back in weeks rather than quarters.
• Measuring NRR without measuring the leading indicators. NRR tells you what already happened. By the time a renewal is lost, the decision was made months earlier.

How do you improve net revenue retention?

NRR moves through exactly three levers, and it helps to know which one you are pulling.

Reduce churn (lifts both GRR and NRR). This is the highest-quality improvement because it raises the floor. Most churn is decided in the first ninety days, which is why a structured onboarding process outperforms any save play attempted at renewal. Fix involuntary churn first: it is the cheapest revenue on the table.

Reduce contraction (lifts both). Downgrades usually mean a team bought more capacity than it activated. Track seats licensed against seats actually used, and intervene when the gap opens, not at renewal when procurement finds it.

Grow expansion (lifts NRR only). Expansion is a product and packaging question before it is a sales question. Usage-based components, tiered feature access, and additional modules give the base somewhere to grow. Expansion built on a shaky GRR is a treadmill.

Underneath all three sits the same requirement: you need to know which accounts are drifting before the renewal date. That is the job of a customer health score, and the forum for acting on it is the quarterly business review. NRR is the scoreboard. Those two are the game.

Where NRR sits in the wider metric set

NRR is a lagging financial metric. It should be the last thing you look at, not the first. The operational metrics that predict it, activation rate, time to first value, support burden, sponsor engagement, all move weeks or months ahead of the revenue. Teams that run a small set of operational metrics that predict churn alongside NRR get to intervene. Teams that only report NRR get to explain.

If you are building the reporting layer from scratch, start with the movement log, because every metric on this page is derived from it, and none of them can be reconstructed later. Then decide what the numbers are for. NRR belongs in the board pack. GRR belongs in the retention review. Health scores belong on the desk of whoever owns the account. Getting the right number in front of the right person is the whole of customer experience operations.