MRR: the monthly recurring revenue that anchors all your other SaaS metrics

Ask three people at your company what their MRR was last month and you have a good chance of getting three different answers. Sales adds up closed contracts, finance looks at what entered the cash register, and product pulls what the billing tool reports. The three think they are talking about the same thing. None are. And as long as the definition is not unique, any decision built on that number inherits ambiguity.

MRR, the monthly recurring revenue, is the most basic instrument of a subscription business, and precisely because it is basic, it tends to be the most mistreated. It seems simple enough to take care of, and that's where most operations start to make mistakes about their health.

What MRR is, and what it is not

MRR is the sum of all monthly recurring revenues from active customers at a given time. Recurring is the word that does the heavy lifting: what is repeated in a predictable way from month to month enters the account, while the customer maintains the subscription. A monthly software fee comes in. An annual plan comes in, but converted to the equivalent monthly installment, dividing the contract value by the twelve months it covers.

What doesn't come in is everything that happens once and isn't repeated. The implementation fee that the customer pays to start using the product is not MRR. The consulting project that your team sold on the outside is not MRR. A one-time charge for an extra service, billed once, is not MRR. These values are real and legitimate revenue, and are needed in cash, but mixing them with the recurring base corrupts the only metric that should answer the question “how much of my revenue is predictable”.

The distinction seems pedantic until the day it explodes in your hand. A company that places setup fees within the MRR sees the number rise in a month with many new sales and fall the following month, without a single customer cancelling. What changed was not the health of the business, it was the volume of non-recurring revenue that slipped into the indicator. From then on, no one can distinguish real growth from accounting noise.

Why it is the anchor metric

If there's one number that almost all others depend on, it's this one. The churn you track is a percentage of what? From MRR. How long does the CAC your procurement team reports pay for itself? In months of customer MRR. The lifetime value that underpins your unit economics is built on how much MRR each account generates and for how long. Take away the bottom floor and the entire building of advanced metrics loses its benchmark.

The reason is structural to the model. A business that sells once records revenue and closes the cycle. SaaS lives on a relationship that is renewed with each charge, so what matters is not how much you sold in an isolated peak, it is what is the recurring base that survives from one month to the next. The MRR is the snapshot of that base. It answers, in a single number, how much predictable money you have today rotating within the operation.

This predictability is the asset that differentiates the model. It allows you to plan hiring, investment and cash burn with a confidence that transactional businesses do not have. But it's only worth something if the number is clean. An MRR contaminated by one-off revenue offers a false sense of predictability, which is worse than having no forecast at all, because it leads the leader to bet big on a base that is not as solid as the panel suggests. I discuss why revenue anchors other domains in the article about what are SaaS metrics.

The four movements that transform numbers into diagnoses

An isolated MRR says little. Knowing that the base is a hundred thousand a month doesn't count whether you are gaining or losing ground. What turns the dead number into a diagnosis is decomposing it into the four movements that explain how it got to where it is. Without this decomposition, you have a thermometer that shows the temperature but doesn't tell you whether the fever is rising or falling.

The first movement is the new MRR: the recurring revenue that came in from customers that did not exist in the base in the previous month. It's what the acquisition machine produces, and it's the number that sales loves to display because it's the most visible.

The second is the Expansion MRR: the increase in revenue coming from customers who were already yours and started paying more, whether by upgrading the plan, hiring additional seats or activating paid features. This is the most underestimated movement, and it is usually the cheapest source of growth that a company has, because it grows based on a relationship that already exists and has already been paid to be conquered.

The third is the Contraction MRR: the reduction in revenue from customers who remain on the base but started paying less, by downgrading their plan or cutting seats. They didn't cancel, so they don't show up in traditional churn, but they are bleeding revenue. Contraction is the silent leak that many operations don't even track, and which usually precedes actual cancellation.

The fourth is the lost MRR: the revenue that was lost for good because the customer canceled. It's the movement that hurts the most to look at and the one that teaches the most, because every dollar lost carries the story of a promise that the product didn't fulfill.

Read together, these four movements tell a narrative that the aggregate number hides. Two companies can have exactly the same total MRR and be in opposite situations: one grows because the existing base expands faster than it loses, the other just exchanges customers who cancel for new customers, running on the treadmill in place. The difference between the two only appears in the decomposition.

The question that decomposition answers

The cross-reading of movements is where management insight lives. When expansion plus new revenue easily outweighs contraction plus cancellations, the business grows organically and healthily. When the company only grows because the acquisition of new customers plugs the hole of what is losing and shrinking, you have expensive and fragile growth, which stops the moment the marketing budget dries up.

The case that interests me most is when the expansion of the base, alone, already compensates for everything that is lost through cancellation and contraction. Then you have a machine that would grow even if it stopped acquiring new customers tomorrow, because those who stay pay progressively more. This is the most difficult engine to build and the most valuable to have, and it is the precursor to Net Revenue Retention, a metric that I delve into in the text about Net Revenue Retention.

The discipline here is to look at the four numbers every time, not just the total. The aggregate is what you report out. Decomposition is what you use to decide where to move: if the contraction is high, the problem is with delivering value in the high plans; if cancellation triggers, it is retention; If the expansion is stopped, it is pricing or product. Each movement points to a different culprit, and the total number hides them all.

Errors that distort reading

The most common mistake has already been named: stuffing non-recurring revenue into the MRR. It is worth insisting because it is silent and almost universal in young operations. Setup fees, one-off projects, one-off charges and anything that is not repeated in a predictable manner need to live outside the indicator. They are included in total revenue and cash, never on a recurring basis.

The second error is the confusion between MRR and cash. The money from an annual plan charged at once goes into the bank in full today, but the MRR only recognizes the equivalent monthly installment. Anyone who looks at the bank statement and calls it MRR overestimates recurring revenue by up to twelve times in the month of billing and disappears with it in the following eleven months. Cash is one thing, recurring revenue is another, and treating them as synonyms produces disastrous financial decisions.

The third error is the mismatch between plans with different periodicities. If you sell monthly and annually, everything needs to be normalized to the same monthly basis before adding them up, or the number becomes a meaningless mix. An annual contract worth twelve thousand reais contributes one thousand per month to the MRR, not twelve thousand at once.

The fourth mistake is treating contraction as if it didn't exist. Many teams only track entries and cancellations, and ignore customers who stayed but downsized. This blind spot masks a deterioration that, left unattended, ends up turning into cancellation a few cycles later.

Getting these four things right doesn't require a sophisticated tool. It requires a written definition, agreed between sales, product and finance, and respected by everyone. Most of the value of MRR is not in the calculation, it is in the discipline of always calculating the same way.