Expansion MRR

For a while the report card for a subscription business was a single number: monthly recurring revenue (MRR), the total that is contracted to arrive again every month. Last month's $36,500 became this month's $38,000, somebody wrote down "up $1,500," and the meeting ended.

One question that number can't answer kept coming back. Grace told an investor the company was growing and got asked where the $1,500 came from. Was it new customers, existing customers buying more, or both, minus whoever left along the way? Pagedeck had nothing on hand that could answer it.

Split apart, the same number describes completely different companies. Take two companies that both netted $4,500 this month. One took in $8,300 from new customers and lost $3,800 to departures. The other took in $5,300 and lost $800. The totals match, but the first is pouring water into a leaking bucket and the second has an intact bucket. Cut the marketing budget and only the first one collapses.

So the month-over-month change gets broken into four pieces: new MRR from customers who just arrived, expansion MRR from customers who were already here, churned MRR from cancellations and downgrades, and net new MRR, which is the three combined. Expansion MRR is the piece that counts growth arriving without a new customer attached to it.

Take only the customers who were present last month and are still present this month, and add up the amounts that went up. Three steps.

1. Pull the customers who paid in both months. Anyone who paid for the first time this month belongs to new MRR and drops out here.
2. For each remaining customer, subtract last month's recurring amount from this month's.
3. Add up only the positive differences. Negative ones move over to churned MRR.

Three customers make it concrete. A went from $400 a month to $600, up $200. B stayed at $250. C went from $300 down to $200, off by $100. Expansion MRR for the month is A's $200 and nothing else. C's $100 is recorded as churned MRR rather than subtracted from expansion. Net the two against each other into a single $100 and both the customer who grew and the customer who shrank disappear from view.

Three things go wrong most often in the counting. First, one-time payments don't belong here. The $2,300 charged for setting up an account doesn't arrive again next month, so it isn't recurring. Second, annual contracts get divided by twelve and only one month goes in. Third, a customer who cancelled and came back months later isn't expansion. They weren't there last month, so step 1 already excludes them, and they're usually tracked separately as a reactivation or counted as new.

Compare it against churned MRR before looking at its size. When expansion MRR exceeds churned MRR, revenue grows from the existing customer base alone, with no new customers at all. Written as a ratio that relationship is net revenue retention (NRR). Start the month at $38,000 with $2,300 of expansion and $3,800 of churn and the ratio is 96 percent, meaning the base shrinks a little every month. Flip it to $4,600 of expansion against $2,300 of churn and it's 106 percent.

The second thing to look at is expansion's share of total growth. Early on, when there are few customers, almost all of the increase coming from new business is normal. As customers accumulate and the product settles into daily work, that share climbs, and in businesses selling to companies it commonly passes 30 percent. A share stuck in single digits for several quarters says the product is attached to one task on one team, which is worth fixing before selling any more expansion.

A large number isn't automatically a good one. Sell seats nobody intends to use and this month's expansion MRR comes back as churned MRR at the next contract date. That's why the two numbers belong on the same screen over the same period.

"Net new looks fine but there's almost no expansion. That number is all new business."

Growth that stops the moment the advertising stops.

"Expansion MRR passed churn. We could take new business to zero and revenue still wouldn't shrink."

The existing base now sustains revenue on its own.

"That $2,300 is a one-time setup fee, not expansion. Take it out."

Money that won't arrive again got counted as recurring.

Pagedeck started tracking this piece separately after Grace fielded the same question three times in the monthly metrics meeting. Until then the screen held one line of monthly recurring revenue and the change from the prior month, and every time someone asked why it moved, Sofia opened accounts one by one to find out. Once the change was split four ways, the meeting stopped discussing the total and started discussing the shape of the pieces. In any month where expansion catches up to new business, that month gets examined first.

A common reading is that expansion MRR larger than new MRR marks a healthy company. Existing customers buying more on their own certainly sounds like one. But the same relationship appears when new business dries up. In a company where almost no new customers arrive, a handful of existing accounts adding seats flips the ratio, and a year in that state leaves no new customers to expand later, so expansion stalls too. That's why the share matters less than the size of both numbers together.

Using it directly as a performance metric for sales or customer success is nearly as common. Expansion MRR doesn't distinguish money that someone persuaded a customer into from money the customer added on their own. To see only the portion produced by an upsell — nudging a customer onto a higher plan — you have to log which accounts received a proposal and count those separately. Attach the undifferentiated number to evaluations and reps put their names on accounts that were already growing, while the accounts that actually need attention go unclaimed.