Net Revenue Retention

When software was sold once and shipped on a disc, counting revenue was simple: add up how many copies went out this quarter. Subscriptions changed the arithmetic. Whether the customer who signed this month is still paying a year from now decides most of the revenue.

For a while, though, the only number on the meeting agenda was new business. "Twenty new accounts this quarter" got reported every time; how many of last year's eighty accounts were still around got reported by nobody. As long as new accounts kept arriving the company looked like it was growing. Then sales had one slow quarter and revenue dropped off a cliff. Everyone had been measuring how fast water went into the bucket and never measured the hole in the bottom.

So companies started looking at existing customers on their own. Take the accounts that were already there a year ago, leave every new signup out of the calculation, and ask what that fixed group pays now. One more piece gets added: in subscription products, the customers who stay often pay more than they did last year by adding seats or moving to a higher plan, and net revenue retention (NRR) counts that increase too. Because it counts money rather than people, it splits off from retention, which tracks whether users keep coming back.

Put what a fixed group of customers paid a year ago in the denominator, and what those same customers pay now in the numerator. Revenue from accounts signed during the year goes in neither place.

The numerator is built from four items.

1. Starting revenue — what that group paid per month a year ago. It doubles as the denominator.
2. Expansion — money added by buying more seats or moving up a plan. This portion is called expansion revenue.
3. Contraction — money lost when an account cuts seats or drops to a cheaper plan. The customer stays; the amount shrinks.
4. Churn — the full amount that cancelled accounts used to pay.

The formula is (starting + expansion − contraction − churn) ÷ starting. If the group paid $80,000 a month a year ago, then lost $4,000 to cancellations and $2,400 to downgrades while the remaining accounts added $12,000 in seats, the numerator comes to $85,600 and the result is 107 percent.

Because there's something to add, this figure can pass 100 percent. Clearing it means revenue grows even in a year with no new customers at all, which is why investors ask for this number before almost anything else.

Two things get settled before the first calculation: which date defines the customer group, and what counts as revenue. Mixing in one-time money such as implementation or training fees lets non-recurring amounts swing the ratio. Once both are set, they stay set — changing either one turns this quarter's figure into something that can't be compared with last quarter's.

The dividing line is 100 percent. Below it, new business is patching a leak in the existing base, and revenue shrinks the moment sales pauses. Above it, the existing base grows on its own and new business stacks on top.

Expectations differ by customer size. Products sold to small teams do fine between 90 and 100, because whole teams cancel often and there isn't much room to add seats. Products sold to large enterprises are expected in the 110 to 120 range, where a deployment in one department keeps spreading to the next. Pagedeck carries both kinds of customers, so it reports them separately: 96 percent for teams that upgraded from the free plan, 118 percent for enterprise accounts on annual contracts.

This number is always read next to gross revenue retention (GRR). A company at 80 percent gross and 105 percent net is growing on the surface while losing one customer in five every year, with the gap covered by a handful of large accounts expanding. Lose one of those accounts and both numbers collapse together.

It's also worth checking whether one account is carrying the result. If a single customer out of 200 tripled its spend, the other 199 can be flat and the figure still clears 110. Recalculating without the top few accounts makes that obvious.

And the number confirms late. It settles only after renewals close, while the cause sits two or three quarters back in usage data. Time to first value, the stretch between signing and seeing a first result, and usage depth, how many features an account actually touches, move earlier.

"Net revenue retention was 112 percent last year. We grew 12 percent without a single new account."

The existing base alone is producing growth — usually the first line in an investor deck.

"112 is good. What's gross retention?"

Asking how much churn the expansion figure is covering up.

"Can you run it again without that one account?"

Checking whether a single customer lifted the number by itself.

At Pagedeck, Grace and Sofia review this figure at the close of every quarter. Sofia brings the values by customer segment and Grace lines the two retention numbers up side by side. Last quarter the overall figure looked healthy at 108, and then dropped to 97 once the top five accounts came out. What the meeting decided wasn't a higher revenue target — it was to go read the usage logs for thirty mid-sized accounts first.

A high number gets read as proof the customer success team is doing well. Existing customers are paying more, so the inference feels natural. But the figure rises just as neatly from a price increase. Apply a 10 percent uplift across the base and it jumps almost 10 points with nothing improved. Report that as performance and next year arrives with no pricing room left and the target still standing. Recording expansion-driven gains separately from price-driven gains keeps the two apart.

Reading it as the company's growth rate is the other common slip — seeing 115 percent and writing down that revenue grew 15 percent this year. New business is excluded from this calculation entirely, so it isn't the company's growth. A company signing plenty of new accounts grows much faster than this number suggests, and a company whose new business has stopped can shrink while still posting 115. Filing both under one heading makes it impossible to tell which side the problem is on, and sales usually takes the blame by default.