Monthly Recurring Revenue (MRR): How to Calculate It

Monthly recurring revenue (MRR) is the normalized value of your subscription revenue for a single month: the number that tells you exactly how much predictable income your business collects every 30 days. You calculate it by taking every active subscription's contract value, converting anything billed quarterly, annually, or over multiple years down to its monthly equivalent, and summing the result across your whole customer base.

MRR is the operating pulse of a subscription business. Where annual recurring revenue is the number investors and boards want, MRR is the number operators watch weekly to catch problems before they show up in a quarterly report. This article covers the ground ARR doesn't: how to normalize different contract lengths into one monthly figure, how to build the MRR movement waterfall, what belongs in the number and what quietly inflates it, the difference between committed and booked MRR, and the mistakes that cause finance and sales to argue over whose MRR is correct.

Key Facts: monthly recurring revenue

• Median net revenue retention is 103% and median gross revenue retention is 91% for bootstrapped B2B SaaS companies between $3M and $20M ARR, meaning a healthy chunk of a company's MRR base already lives inside its existing accounts before a single new deal closes (SaaS Capital, 2026).
• Median growth across private B2B SaaS companies slowed to 22% in the 2026 reading, down from 25% in 2024, per SaaS Capital's 15th annual survey of more than 1,000 private B2B SaaS companies (SaaS Capital, 2026).
• Acquiring a new customer costs five to 25 times more than retaining an existing one, and a 5% improvement in retention can lift profits 25% to 95%, which is why protecting existing MRR is almost always cheaper than replacing it with new MRR (Amy Gallo, Harvard Business Review, October 2014).

What is monthly recurring revenue (MRR)?

MRR is the total recurring revenue your business is contracted to collect in a given month, normalized so that every subscription (regardless of its actual billing schedule) is expressed as a monthly value. It excludes anything that won't repeat: setup fees, one-time services, and usage spikes that aren't part of a committed minimum.

The core formula:

MRR = Sum of (normalized monthly value) across all active subscriptions

If you already track ARR, the shortcut version is:

MRR = ARR / 12

That shortcut only works cleanly if your ARR figure is itself accurate. If your ARR number already has one-time fees baked in, dividing by 12 just spreads the same error across every month. Get the underlying contract math right first.

Normalizing quarterly, annual, and multi-year contracts to MRR

This is where most MRR numbers go wrong. A subscription business rarely bills everyone the same way: some customers pay monthly, some quarterly, some annually upfront, and some sign multi-year deals. Each of those needs to be converted down to its monthly equivalent before it can be added to the total.

The rule is simple: divide the contract's total value by the number of months it covers.

Notice that all four rows above land on $3,000 or less in MRR per month, even though the cash that arrives at signature looks completely different: $500 shows up monthly, $108,000 might land as a single upfront wire transfer. MRR is a recurring-value metric, not a cash-flow metric. The cash timing matters for your bank account; it doesn't change what belongs in MRR.

This is also where the most common inflation mistake happens: booking the full annual prepayment as one month's MRR. A customer who prepays $36,000 for the year did not just add $36,000 to this month's MRR. They added $3,000, the same as if they'd paid monthly. If you book the full amount in month one, your MRR chart shows a spike that vanishes for the next 11 months, and nobody downstream can trust the trend line.

The MRR movement waterfall

MRR isn't a static number you recalculate from scratch every month. It's a running balance that moves through five components: new, expansion, contraction, churned, and reactivation. Understanding the waterfall tells you where growth (or decline) is actually coming from, instead of just watching a total go up or down without knowing why.

• New MRR: revenue from first-time customers signing their first contract.
• Expansion MRR: additional revenue from existing customers who upgrade, add seats, or move to a higher tier.
• Contraction MRR: revenue lost when an existing customer downgrades rather than cancels.
• Churned MRR: revenue lost when a customer cancels entirely.
• Reactivation MRR: revenue from a previously churned customer who comes back.

Net New MRR = New MRR + Expansion MRR + Reactivation MRR, minus Contraction MRR and Churned MRR.

Worked example

Say a company starts the month with $500,000 in MRR. Over the course of the month:

• New customers sign contracts worth $45,000 in new MRR
• Existing customers expand by $28,000
• A previously churned customer comes back, adding $6,000 in reactivation MRR
• Existing customers downgrade, losing $12,000 in contraction MRR
• Customers cancel outright, losing $31,000 in churned MRR

Net New MRR of $36,000 (45,000 + 28,000 + 6,000 minus 12,000 minus 31,000) confirms the ending balance: $500,000 plus $36,000 equals $536,000. If your waterfall and your ending balance don't reconcile to the dollar, one of your five components has a data problem, usually contraction or reactivation, since those two are the ones teams forget to track separately.

The same math connects directly to churn rate: churned MRR divided by starting MRR is your gross revenue churn rate for the month. And expansion plus reactivation minus contraction and churn, relative to starting MRR, is the input to net revenue retention. MRR waterfall discipline is what makes both of those metrics trustworthy.

What counts (and what doesn't) in MRR

The waterfall only works if the inputs are clean. Getting MRR wrong is usually a scoping problem: revenue that isn't truly recurring gets counted anyway, or a discount gets ignored, and the total no longer reflects what the business can actually count on next month.

The test that cuts through most edge cases: will this exact revenue renew next period under the current contract terms, without a new statement of work or a one-off event? If yes, it's MRR. If it depends on a separate agreement or a variable trigger, it isn't.

Usage overages are the trickiest line. A customer with a committed minimum of 10,000 API calls per month who occasionally uses 12,000 shouldn't have that overage folded into MRR, because it isn't guaranteed to repeat. If the contract has a true minimum commitment (say, they're billed for at least 10,000 calls regardless of actual usage), that minimum belongs in MRR. Everything above it doesn't, until it becomes part of a renegotiated committed tier.

Committed MRR vs booked MRR

Two teams can report two different MRR numbers for the same customer base, and both can be technically correct, if they're measuring committed MRR and booked MRR without labeling which is which.

Booked MRR is the value sales records at the moment a deal closes, often the full contracted value across the term, even if the contract ramps up over its first few months. Committed MRR is the value that's actually contracted and billing right now, at the current point in the term.

A worked example: a customer signs a contract that starts at $6,000 per month for the first three months, then steps up to $10,000 per month for the remainder of the term. Sales books the deal at $10,000 in MRR, because that's the contract's steady-state value. Finance, reporting committed MRR for month one, correctly shows $6,000, because that's what's actually billing. Neither number is wrong. The mistake is reporting one of them as "MRR" without saying which one, then being surprised when the two teams' dashboards don't match.

MRR per customer: ARPA and cohort views

Once you have a clean, reconciled MRR total, dividing it by your customer count gives you average revenue per account (ARPA), sometimes written ARPU when measured per user instead of per account.

ARPA = Total MRR / Number of paying customers

Using the ending MRR from the waterfall example above: $536,000 in MRR across 335 paying customers gives an ARPA of $1,600 per month. Track ARPA over time and it tells you whether your growth is coming from adding more customers, or from getting more value out of the customers you already have. A rising ARPA alongside flat customer counts usually means expansion motions are working; a falling ARPA alongside rising customer counts usually means you're moving downmarket.

Cohort views take ARPA a step further by grouping customers by signup month and tracking how each cohort's MRR changes over time, independent of new customer acquisition. Illustrative example: a cohort of 25 customers that signed in January contributed $40,000 in starting MRR ($1,600 ARPA). By month six, two of those customers had churned, but the remaining 23 expanded enough to push cohort MRR to $52,000. That cohort's net revenue retention is $52,000 / $40,000, or 130%, even though two logos left. A blended, company-wide MRR number would never surface that detail; only a cohort view does.

Cohort MRR tracking pairs naturally with the growth-lever math in deal size optimization: if your best cohorts expand at 130% and your newest cohorts expand at 95%, the difference tells you whether a recent change to onboarding, pricing, or targeting is helping or hurting.

MRR vs ARR: which one to lead with

MRR and ARR measure the same underlying thing (recurring subscription value) at two different time scales, and the annual recurring revenue article covers the full formula, worked examples, and the four-component breakdown of ARR in detail. The short version for deciding which to lead with:

Most B2B companies with any annual-contract volume track both: MRR internally for operating cadence, ARR externally for board and investor reporting. Neither one substitutes for the other, and a company that only tracks ARR often misses early warning signs that a monthly waterfall would have caught weeks sooner.

Common MRR reporting mistakes (and how to audit your own number)

Most MRR errors aren't malicious, they're the result of an inconsistent definition applied by different people on different teams. Here's what to check for before you trust a monthly report.

To audit your own number, pull the waterfall for the last three months and check two things. First, does starting MRR plus net new MRR equal ending MRR, to the dollar, for each month? If not, one of your five components is missing data. Second, pick five customers at random and trace their contract terms through your MRR report by hand: does a quarterly customer show up as one-third of their quarterly fee, or as the full amount in the month they paid? That single spot-check catches most normalization errors before they reach a dashboard.

The same reconciliation discipline supports revenue predictability: a forecast built on a shaky MRR number inherits every one of its errors, and those errors tend to surface at the worst possible time, right when finance is trying to close the books.

Where MRR movement comes from

The waterfall components don't appear out of nowhere. New MRR is a direct output of closed-won deals, and it's worth pairing your new MRR trend with customer acquisition cost to see whether the deals you're closing pay back fast enough. Expansion MRR usually traces back to specific upsell and cross-sell motions; see upsell vs cross-sell for how the two differ and which one tends to be cheaper to run. And because new MRR is what most sales teams are compensated on, the shape of your sales compensation plan has a direct, measurable effect on which waterfall line grows fastest, new logos or expansion.

Tying MRR back to LTV:CAC ratio closes the loop: ARPA feeds directly into lifetime value math, and a clean MRR number is the foundation that both metrics depend on.

MRR is only as useful as the discipline behind it. Build the waterfall every month, keep one-time revenue out of the total, normalize every contract to its true monthly value, and agree on a single definition for committed versus booked MRR that sales and finance both sign off on. Get that discipline in place and MRR stops being a number you argue about, and starts being the fastest, most honest signal you have for where the business is actually heading.

Related reading

• Annual Recurring Revenue
• Churn Rate
• Net Revenue Retention
• Customer Acquisition Cost
• LTV:CAC Ratio
• Revenue Predictability
• Deal Size Optimization
• Upsell vs Cross-Sell
• Sales Compensation