Is there a difference between NRR and NDR?
No. They are two names for the same calculation. Net revenue retention and net dollar retention both measure what happened to a starting cohort of customers over a period, counting expansion, contraction, and churn, excluding anyone acquired during the period.The two labels persist because different parts of the market adopted different vocabulary. Investor decks and public filings lean toward net dollar retention. Operating teams and RevOps documentation lean toward net revenue retention. Some companies use both in the same deck without noticing.
The useful question is not which term is correct. It is how the number was built, because the construction choices move NRR far more than the naming convention ever will.
What does the calculation actually measure?
It measures the revenue produced by a fixed group of customers at the end of a period, divided by what that same group produced at the start. Take the ARR from customers who were active on January 1. Look at what those same customers are worth on December 31. Divide.A cohort worth $10 million on January 1 that is worth $11.2 million on December 31 has 112 percent NRR. Some customers churned. Some downgraded. Enough of the rest expanded to more than cover both. Customers who signed in March are excluded entirely, because including them would measure sales performance rather than retention.
The threshold that matters is 100 percent. Above it, your existing base grows without a single new deal. Below it, new sales have to cover a hole before they produce any growth.
Where can the two numbers actually diverge?
On construction choices, not on terminology. These are the decisions that move the figure, and none of them are standardized.| Choice | Option A | Option B | Typical impact |
|---|---|---|---|
| Base metric | ARR or MRR at cohort start | Recognized revenue over the period | Recognized revenue understates fast growers |
| Cohort window | Trailing twelve months, rolling | Fixed fiscal year | Rolling smooths seasonality, fixed matches the plan |
| Downgrade treatment | Counted as contraction inside the cohort | Partial churn removed from the base | Removing from the base inflates the ratio |
| Reactivations | Returning customers count as expansion | Excluded, treated as new | Counting them inflates the ratio |
| Currency | Constant currency | Reported currency | FX can move it several points either way |
| Cohort scope | All customers | Customers above a revenue threshold | Excluding small accounts usually raises it |
| New customers | Excluded | Included by mistake | Inclusion inflates it badly |
Why do published NRR benchmarks disagree with each other?
Because each survey defines the cohort and the base differently, and the sample is self-selected. A benchmark set built from companies willing to disclose retention will skew toward companies with retention worth disclosing.Segment mix compounds it. An enterprise product with seat-based pricing and multi-product expansion paths has structurally higher NRR than an SMB tool with a single tier. Averaging both into one benchmark produces a figure that describes neither.
Contract length matters too. A company on three year terms shows very little churn inside a twelve month window simply because most customers had no renewal decision to make. Its NRR looks strong until the renewal cohort finally comes up.
The practical response is to treat external benchmarks as directional and your own trend as the real measurement. NRR moving from 108 to 103 over four quarters is a fact about your business. NRR sitting three points below a published median may only be a fact about methodology.
What should you check before comparing your NRR?
Five things, in this order. Each takes minutes and each can explain a multi-point gap.Confirm no new customers are in the numerator. This is the most common error and it inflates the figure most.
Confirm the base is ARR or MRR rather than recognized revenue. Recognized revenue understates a fast-growing cohort because contracts that started mid-year only recognize part of their value.
Confirm downgrades stay inside the cohort as contraction. Removing a downgraded customer from the base and treating it as a smaller new customer manufactures retention that did not happen.
Confirm the currency treatment. A base with significant international revenue can swing several points on FX alone.
Confirm the cohort scope. Many companies quote NRR on customers above a revenue threshold, which is a defensible choice and a very different number from all-customer NRR.
How do you build NRR you can actually trust?
From a monthly ARR waterfall that reconciles beginning ARR to ending ARR. The categories that matter are beginning ARR, churned customer ARR, churned product ARR, product decrease ARR, new customer ARR, new product ARR, increased product ARR, and ending ARR, where beginning ARR always equals the prior month's ending ARR.That reconciling constraint is what makes the resulting retention figure defensible. If the waterfall ties every month, the components are classified consistently, and both gross and net retention fall out of the same chart rather than being calculated separately from a different pull of the data.
The alternative, rebuilding retention quarterly from a fresh export, produces numbers nobody can reproduce and that quietly change when someone reclassifies an account. A monthly reconciliation catches misclassification in the month it happens.
How does NRR change the forecast?
It sets the starting line. An NRR of 110 percent means the existing base grows 10 percent next year before any new deal closes. An NRR of 92 percent means new bookings replace an 8 percent shortfall before producing growth. Those are two entirely different sales plans against the same revenue target.That is why net revenue retention belongs at the front of the annual plan rather than in the retention section at the back. Size the new business requirement as the gap between the target and the retained base, then work backward to pipeline and quota.
Forecasting the retention layer separately from new business also makes misses diagnosable. Accuracy on the renewal base and accuracy on new and expansion behave differently, and blending them produces a single number that cannot be improved because you cannot tell which half failed. Splitting them is one of the more reliable ways to lift forecast accuracy without touching the pipeline model at all. Our guide on how to forecast revenue covers how the retained base and new bookings stack into one model.
Frequently Asked Questions
Is NRR the same as NDR?
In practice, yes. Net revenue retention and net dollar retention describe the same calculation: revenue from a starting cohort at the end of a period divided by that cohort's revenue at the start, including expansion, contraction, and churn, excluding new customers. The two labels come from different corners of the market rather than from different math.
Why do two companies report different NRR on similar businesses?
Because the construction choices are not standardized. Cohort window, whether the base is ARR or recognized revenue, treatment of downgrades versus partial churn, currency handling, and whether reactivated customers count all move the number. Two teams can run the same customer file and land materially apart without either making an error.
Does NRR include new customers?
No. NRR measures what happened to a fixed starting cohort. Adding customers acquired during the period inflates the ratio and destroys the metric's purpose, which is isolating the behavior of the existing base. If a reported NRR figure looks unusually high, checking for new customer contamination is the first thing to do.
What is a good NRR for B2B SaaS?
Above 100 percent means the existing base grows on its own before any new sales, which is the threshold most investors care about. The specific target varies by segment, because enterprise businesses with seat-based expansion behave very differently from SMB products with limited upsell paths. Comparing your figure to a published benchmark is only meaningful if both were constructed the same way.
How does NRR affect a revenue forecast?
It sets the base your forecast starts from. An NRR of 110 percent means the existing customer base grows 10 percent next year with no new bookings at all. An NRR of 90 percent means new sales have to replace a 10 percent shortfall before producing any growth. That single coefficient determines how much of the plan the sales team has to carry.
See how ORM turns these insights into action
ORM builds custom revenue forecast models for B2B SaaS companies. Not dashboards. Prescriptive analytics that tell you what to do next.
Schedule a Demo