How Do I Calculate and Read My MRR and Churn?

MRR is the revenue you can count on next month if nobody does anything: every active contract, normalized to a monthly figure. Churn is what leaks back out. Track both in dollars, not customer counts. A 2 percent monthly revenue leak compounds to roughly 22 percent of your base in a year.

How do I actually calculate MRR?

Take every recurring contract, the ones that renew without you selling again. Normalize each to a month: annual divides by twelve, quarterly by three. Add them up.

The whole trick is what you leave out.

Setup fees aren't MRR. Neither is a side project, a one-time overage, or a promo credit you never renewed. In the retainer books I clean up, mixing one-time revenue into the recurring line is the most common error I find, and it overstates MRR by 10 to 20 percent. That's what I see, not a published figure. Still enough to make a growth chart lie for a year before anyone catches it.

That last row hides a real question. If you resell software, the test isn't whether the subscription recurs, it's whether the service is yours before it reaches the client, and buyers do test that line. Control it and the whole thing is MRR. Broker it and only your fee is.

Half your book is project work? Don't force it into MRR and don't discard it. Run two lines. An MRR figure covering 55 percent of your business is a health metric only beside the swing you resell every quarter.

One more rule catches people. MRR is not cash, and it is not your tax number either. An annual contract paid up front hands you $24,000 in the bank, $2,000 of MRR, and a deferred liability that unwinds over twelve months. On the cash basis, where most $1-5M service businesses sit, tax lands on the whole $24,000 the year you receive it, so confirm timing with whoever files your return. If you have ever been profitable on paper and short in the bank, same family of problem. Read MRR beside the core reports you look at every month, never instead of them.

The five pieces that move MRR every month

MRR is not one number. It's a bridge, and every dollar of movement lands in one of five buckets:

• New: MRR from customers who weren't there last month.
• Expansion: existing customers who upgraded, added seats, or took a price increase.
• Contraction: existing customers who downgraded but stayed.
• Churned: MRR from customers who left entirely.
• Reactivation: customers who left and came back. Most books file these as New, which flatters the sales number and buries what the churn actually cost you.

Ending MRR equals beginning, plus new, plus expansion, plus reactivation, minus contraction, minus churned. One month, rounded:

The top line grew 2.4 percent. The base underneath it shrank, and no summary P&L shows you that. Build the bridge monthly and once the list is clean it takes fifteen minutes. Fifteen minutes. The first month is not: hand-classifying every account the first time through burns most of a day, and that day is the tell.

A flat MRR line looks calm. It usually isn't. Flat almost always means new business and churn cancelling out, and you paid full price to acquire the new business. You're running hard to stand still.

How do I calculate churn, and which number matters?

Three numbers, three different questions.

Settle the definition before you hand it to anybody. Plenty of lenders and buyers drop downgrades into that numerator too, a bigger figure off the identical book. Neither is wrong. Handing over one while they quietly compute the other is.

Logo churn counts how many left. Revenue churn counts how much it hurt. Lose one client out of fifty and you shrug: 2 percent, fine. Lose the client who was 12 percent of your base and logo churn reports the same 2 percent while your year quietly ends. Count heads to understand your service. Count dollars to understand your business.

Net revenue retention is the one you have probably never calculated, and the first one a buyer asks for. Above 100 percent, your base grows on its own even after the departures. Below it, you're filling a bucket with a hole. One catch, and it costs real money: the 100 percent bar everyone quotes is annual, trailing twelve months. Monthly retention of 100.5 percent is roughly 106 percent for the year, so the monthly figure undersells you by six points.

What counts as healthy churn at $1-5M?

Ranges first, caveat second. In the retainer and property-management books I clean up, monthly logo churn runs 1 to 2 percent, gross revenue churn runs lower because the accounts that leave are the small ones, and retention above 100 percent is the mark worth aiming at. Put downgrades in that numerator and it can land above logo churn instead. Shapes I see in cleanup work, not a benchmark study. A rough map, not a scorecard.

Two things distort the picture far more than size does.

Contract length is the first. An annual agreement can only churn on its renewal date, so eleven months read quiet and the twelfth reads like a disaster. If most of your book renews once a year, the monthly rate is noise. Measure the renewal cohort: of the deals that came up this month, how many renewed, and how many of that cohort's dollars? Twelve up, ten renewed, $9,000 of $11,000 kept. That's 83 percent of the logos and 82 percent of the dollars, and you can act on it Monday.

Concentration is the second. At $2M with twenty clients, one departure is a 5 percent month no benchmark can absorb.

Don't annualize by multiplying by twelve, which overstates the damage because each month you're churning a base that already shrank. Compound it. Two percent monthly is about 22 percent a year, not 24. Small gap at low rates, ugly one high up.

One last cut. Churn by service line beats churn in total, for the same reason profit by service line beats total profit: if one offering bleeds and another sticks, the blended figure hides both.

When should I actually worry?

A number you never attach a threshold to is trivia, not management, and it's why the leak usually gets spotted a quarter late. Four trip wires. Check them the same day every month:

• Retention under 100 percent two months running. One month is a big renewal landing on the wrong side of the calendar. Two is a pattern, and it calls for a save conversation, not a sales push.
• Gross revenue churn over 2 percent in a single month. Open the list and name the accounts out loud.
• Contraction climbing while logo churn stays flat. Nobody is leaving and everybody is buying less. That one surfaces months ahead of the cancellations.
• Any single client over 10 percent of MRR. Not a churn problem yet. It's what makes your churn rate meaningless the day it finally happens.

None of that needs software. It needs the bridge, built every month, and somebody in the building who actually reads it.

The honest answer

When you ask how to calculate MRR and churn, you're usually asking something else: is my revenue base leaking, and would I notice in time?

MRR alone won't tell you. It's a summary, and summaries hide the offsetting movements underneath them. The bridge tells you. Five buckets, dollars not logos, every month, next to the reports you already read.

Calculate MRR from the contracts. Judge it on net revenue retention, and say which basis you ran every time you hand it to a lender, a buyer, or your own board. It says whether you're growing or just replacing.

This breaks for one boring reason, every time. The customer list is a mess. Duplicate accounts, contracts nobody ever logged an end date on, a cancellation the books quietly never recorded. Fix that first and the rest is arithmetic.

Clean list, five buckets, dollars not logos.