What ARR, MRR and net revenue retention actually measure
Beyond the investor pitch deck, here is what these three numbers really capture, where they get miscalculated, and why one of them matters more than the other two.
Three numbers, three different jobs
ARR (annual recurring revenue) and MRR (monthly recurring revenue) get treated as interchangeable in a lot of founder decks, and net revenue retention (NRR) often gets bolted on as an afterthought slide. They are not interchangeable. Each one answers a different question, and mixing them up is one of the easiest ways to accidentally mislead an investor, a board, or yourself.
MRR measures the recurring revenue a business can expect in a given month, based on active subscriptions at a point in time. ARR is usually just MRR multiplied by twelve, used because it is easier to compare against annual budgets, valuations and multiples. Neither one measures cash actually collected, and neither measures total revenue if a company also sells non-recurring things like implementation, consulting or one-off hardware.
NRR measures something different again: what happens to the revenue from a fixed cohort of existing customers over a set period, with new customer revenue deliberately excluded. It is the closest thing SaaS has to a quality-of-business score.
What actually counts as recurring
The first place ARR and MRR get miscalculated is in what gets included. Recurring revenue should only capture revenue a company is contractually entitled to keep receiving without the customer doing anything further, such as subscription fees. It should not include one-off setup fees, professional services, hardware sales, or usage spikes that are not guaranteed to repeat. A common but misleading habit is to annualise a single strong month, or to include revenue from a multi-year contract as if it renews automatically at the same rate, when in practice usage-based components can swing the real figure a lot.
Another common distortion is counting bookings rather than recognised revenue. A signed contract for a future start date is not yet recurring revenue; it is a pipeline commitment. Investors and lenders will usually ask how a company defines its ARR precisely, and a founder who cannot answer that clearly, consistently, and the same way every time raises a flag.
What net revenue retention actually measures
NRR takes a fixed group of customers as they stood at the start of a period, typically twelve months, and tracks what happened to their revenue by the end of that period, with new customers added during the period excluded entirely. The calculation nets together three things: expansion (existing customers spending more, through upsells, seat growth or upgrades), contraction (existing customers spending less, through downgrades), and churn (existing customers leaving entirely).
A figure above 100% means existing customers are, on average, spending more over time even before any new sales effort. A figure below 100% means the existing base is shrinking even if new customer sales are strong, which matters because it tells you how dependent growth is on constantly refilling a leaky bucket. This is why NRR is treated differently to gross retention, which only measures churn and contraction without netting off expansion, and why the two numbers should usually be reported alongside each other rather than as one blended figure.
Where NRR calculations go wrong
The most common error is measuring NRR at the logo level (how many customers stayed) rather than the revenue level (how much of their spend stayed), which produces a very different and usually more flattering number. Another is choosing an inconsistent cohort window, comparing a rolling twelve months to a fixed calendar year depending on which makes the trend look better. A third is excluding contraction from customers who downgraded but did not fully churn, which understates the negative side of the calculation. Because there is no single mandated methodology, the only reliable way to compare NRR across companies is to ask exactly how it was calculated, over what period, and whether services or usage-based revenue were included.
Why this matters beyond fundraising
These metrics are not just fundraising theatre. They drive real operational decisions: whether to invest more in customer success versus new sales, how to price renewals, and how much revenue can realistically be forecast against fixed costs. A business with strong new sales but weak NRR is often masking a product or onboarding problem that will eventually slow growth regardless of how much is spent acquiring new customers. A business with modest new sales but NRR comfortably above 100% can often grow revenue substantially from its existing base alone.
For founders preparing to report these numbers to a board or investor, the safest approach is consistency and transparency: define recurring revenue narrowly, keep the same NRR cohort methodology every reporting period, and be explicit about what is included and excluded, since the definition matters as much as the number itself.
For guidance on how financial metrics should be presented and audited as a company scales, the Institute of Chartered Accountants in England and Wales and the British Business Bank both publish resources aimed at growing businesses.