If you sell a subscription, MRR is the one number that tells you whether the business is alive. It is the predictable revenue you can expect to arrive again next month, and its direction over time is the most honest summary of your situation that exists. You do not need an accounting background to use it. You need one formula and an awareness of about five ways people get it wrong.
The formula
MRR = number of active paying customers × average revenue per customer, per month. Thirty customers paying $20 a month is $600 MRR. That is genuinely the whole calculation.
Annual plans get normalised. A customer paying $240 once a year counts as $20 of MRR every month, not $240 in January and nothing after. The point of MRR is to describe a rate, not to track cash arriving, so you divide annual contracts by twelve regardless of when the money lands in the bank.
ARR, annual recurring revenue, is just MRR multiplied by twelve. It is the same number wearing a suit. Use it when talking to investors, use MRR when running the business, and never quote both in the same sentence as if they are separate achievements.
The five components that move it
A single MRR figure hides the mechanics. The useful version breaks the change from one month to the next into parts, because $10,000 that grew smoothly is a completely different business from $10,000 that churned half its base and replaced it.
- New MRR: revenue from customers who were not paying last month.
- Expansion MRR: existing customers who upgraded or added seats.
- Reactivation MRR: former customers who came back.
- Contraction MRR: existing customers who downgraded to a cheaper plan.
- Churned MRR: customers who cancelled entirely.
Net new MRR is the first three minus the last two. If net new is positive you grew, and the composition tells you why. A month where new MRR is strong but churned MRR nearly matches it is a warning that no top-line figure would show you. This is the difference between a leaky bucket and a growing one, and it is invisible until you split the number.
Expansion MRR deserves particular attention because it is the cheapest revenue in the business. Nobody had to be acquired. A pricing model with genuine upgrade paths, like a tiered product where growing customers move up on their own, can produce growth without any additional marketing spend at all.
The mistakes that inflate MRR
Almost every wrong MRR number comes from one of these.
Counting one-time payments. A lifetime deal is not recurring revenue. If you sell a $99.99 lifetime plan, putting it into MRR is simply false, and it will make a flat month look like a great one. Either exclude one-time payments from MRR and report them separately, or amortise them over a defensible expected lifespan and be transparent that you did. The honest approach is to keep two lines: recurring revenue and everything else.
That distinction is not academic for anyone selling a mix. seenpaid, for instance, sells monthly plans at $19.99, $39.99 and $69.99 alongside a $99.99 lifetime option, and only the first three belong in MRR. Blending them would make a month with a few lifetime sales look like a step change in recurring revenue that never repeats.
Counting trials. A user on a seven-day trial has not paid you. Including trials in MRR means your number is a forecast of your own optimism. Count them when the first charge succeeds, not when the trial starts.
Counting failed payments as revenue. A subscription can be active in your database while the card is declining. If your MRR comes from subscription records rather than successful charges, you are counting money you did not receive. Take the number from actual paid invoices.
Forgetting refunds and discounts. A customer on a 50% coupon contributes half. A refunded month contributes nothing. Small distortions individually, and they compound into a number you stop trusting.
Mixing currencies carelessly. Fix an exchange rate per month rather than converting at the moment of each charge, or your MRR will move because of the euro rather than because of your business.
Growth rate matters more than the absolute number
MRR on its own is a size, and size is not health. A business at $8,000 MRR growing 8% a month is in a completely different position from one at $8,000 MRR that has been flat for a year, and the flat one is usually in trouble even though the number looks identical.
Month-over-month growth rate is the follow-up figure: this month’s MRR minus last month’s, divided by last month’s. Track it as a line, not as a single reading, because one good month proves nothing and a three-month trend proves a lot. Small businesses are volatile enough that a single lumpy month will mislead you in both directions.
A related check is the ratio of new plus expansion MRR to churned plus contraction MRR. Above 1 you are growing, and the higher the number the more of your growth is real rather than replacement. Below 1 you are shrinking regardless of how many signups you had, which is a distinction the signup count will never show you.
What MRR tells you that revenue does not
Total revenue can be rescued by a single large one-off sale, which is exactly why it flatters bad months. MRR cannot be rescued that way, because it only counts what will recur. That constraint is the feature.
MRR also lets you do arithmetic you cannot do with a revenue total. Divide MRR by customers and you get average revenue per account. Combine it with churn and you get a rough customer lifetime value. Compare it to your monthly costs and you get runway in the most direct form available. All of that requires a rate, not a total.
From scoreboard to map
The limitation of MRR is that it tells you what happened, not why. You know you added $400 last month. You do not know whether it came from the thread that went around, the Reddit comment you left in a thread six weeks ago, the newsletter, or the customer who was always going to find you.
Without that, marketing decisions are guesswork dressed as strategy. You keep doing everything, because you cannot tell which part is working, and the effort spreads thinner every month. Most solo founders and small teams live here permanently.
Breaking MRR down by acquisition source turns it from a scoreboard into a map. Which channel produced new MRR. Which channel produces customers who expand rather than churn. Which posts, specifically, preceded payments. That is a data problem rather than a strategy problem, and it is solvable: seenpaid connects read-only to Stripe and attributes revenue back to the individual post that earned it, so your MRR growth comes with an explanation attached rather than a shrug.