Contraction

For a long time the only question asked about a subscription business was how many accounts left. If 95 of last month's 100 customers were still there, you wrote down five percent and moved on. A cancellation is easy to see, easy to count, and everything still on the list got filed under fine.

The trouble is that a customer who stays doesn't necessarily keep using what they used to. A company that signed for 80 seats can renew for 50 and the customer count is still 100. Nobody left, so the loss rate reads zero, while that account's payment drops by 37 percent. Let the same thing happen at ten accounts and the customer list looks untouched while the bank balance does not.

So the monthly review kept producing the same standoff. Customer success reported a quarter with zero losses. Finance reported that recurring revenue was down. Both reports were accurate. Contraction is the name given to the gap between them. Once the in-between state — not gone, not unchanged — gets counted on its own, revenue decline shows up months before the customer count would have shown anything.

You measure it by comparing what the same customer pays in one cycle against what they pay in the next. There are three ways that number comes down, and they differ in how much money moves and in what tends to follow.

1. Seat count drops. An account that had 80 people renews for 50. This is the most common form and the one that moves the largest amounts.
2. Scope narrows. A tool the whole company used gets trimmed back to a single team. It shows up as a seat reduction, but a whole department walking away makes a full departure the following year far more likely.
3. Add-ons come off. Extras that were billed separately, such as long-term audit log retention, get dropped at renewal. The amounts are small, but they usually mean the person who insisted on that feature is no longer in the role.

The arithmetic is the same in all three cases. Take the monthly payment before the change and subtract the monthly payment after it. A customer who paid $3,000 a month and now pays $1,900 contracted by $1,100. Roll a month of those figures together and you have one component of churned monthly recurring revenue (MRR), the total recurring revenue lost that month to cancellations and reductions combined. The two parts are recorded separately, because the response to each is different: a cancelled customer has to be won back, while a contracted one needs the remaining people to use the product more.

Decide in advance when you measure. Contraction usually surfaces on the renewal date, but a contract that lets the customer adjust seats monthly bleeds a little all year. So values that move every month, seat counts above all, get recorded monthly, and any account below last month's number gets pulled into its own list.

This number gets used when revenue retention is calculated, when the risk on an upcoming renewal is judged, and when something has to be handed back to the product team.

First, it's what you need to read net revenue retention (NRR) properly. NRR divides what last year's customers pay today by what they paid then, which means growth and decline cancel each other out inside a single figure. One account adding $1,500 of seats while another sheds $1,500 lands at exactly 100 percent. Counting contraction separately tells you whether that 100 percent is a quiet quarter or two large movements that happened to offset.

Second, it's an early signal going into a renewal conversation. Renewal forecasting sorts expiring accounts into safe, watch and at-risk, and any account whose seats dropped at all in the past six months usually lands in the last bucket. An account that went from 80 seats to 70 barely moved the invoice, but it did already hold an internal discussion about whether to keep paying for the tool.

Third, it's evidence for the product team. Group the contracted accounts and look at which departments the lost seats belonged to and patterns appear. If most of the removed seats were design, that team's work has probably moved to a competing tool. A customer who left entirely is often unreachable; a customer who contracted is still under contract and can be asked.

"We only lost two accounts this quarter, but net revenue retention is 94 percent. Contraction did that."

Few customers left, but the ones who stayed are paying for less than they were.

"This account cut 30 seats at the last renewal. We can't let this one ride."

An account that has contracted once is a real risk at the next renewal, so get in front of it now.

"Don't book it as churn. The contract is live, they just dropped seats."

A correction about which column the number belongs in.

At Pagedeck, Sofia opens the seat-change list in the first week of every quarter. It's a table of accounts whose seat count fell against the previous quarter, and this time six of them showed up. Four had cut headcount themselves, which has nothing to do with the product. The other two kept their people and reclaimed the accounts anyway. Those two are the ones Sofia cares about: if the people are still there and the licences came back, those people weren't using the product.

From there the conversation moves to the number Grace watches. Two consecutive quarters where the customer count rises and recurring revenue stays flat means existing accounts are draining faster than new ones fill them. What Grace asks is not how many customers left but whether the total from the ones who stayed went up.

The most common mistake is treating contraction as a softer version of cancellation. The customer is still there and money still arrives every month, so it feels like half a win. In practice it's closer to advance notice that the account will be gone next year. Cutting seats means the company already reviewed whether to keep the tool and settled on keeping half of it. Skip past that moment and the contract simply ends at the next renewal with no warning, by which point there's no time left to change the outcome.

The other misreading is hearing contraction as a pricing complaint. Customers drop seats while saying they need to cut budget, which sounds like the product costs too much. But the first things cut from a budget aren't the expensive tools, they're the unused ones. If 80 seats became 50, chances are those 30 people barely logged in last quarter. Discount the deal and the same 30 still don't use it, the invoice is just smaller — and the next renewal repeats the conversation from a lower number. The question to chase is why those people stopped, not what the price should be.