Every software pitch leads with recurring revenue, as though the word recurring settled the question of durability. It does not. Recurring describes what happened last year: an invoice was raised, and it was paid. Durable describes what will happen next year, and the year after, when a new buyer reviews the contract with fresh eyes and asks what it is for. A subscription that renews out of habit and a subscription that renews because the business cannot function without it produce the same line in a model and behave nothing alike under pressure. The whole of our underwriting is the work of telling them apart.

The distinction matters most precisely when it is hardest to see. In a good year, almost everything renews; demand is forgiving and even weak products are carried along. The test of durability is the bad year — the budget review conducted in a downturn, the new finance chief hunting for cuts, the acquirer rationalising an overlapping stack. Revenue that survives that scrutiny was durable; revenue that does not merely recurred until it stopped. We try to underwrite for the bad year from the first meeting, because that is the only year that reveals what a customer was really paying for.

Three sources of durability

Durability is not a single property but a small set of them, and they compound. We look for at least one in force and ideally all three.

The first is embeddedness: the software is woven into how the business actually operates, not bolted to the side of it. A tool a team opens occasionally is discretionary; a system that originates the transaction, holds the system of record, or sits on the critical path of a daily workflow is not. The clearest tell is what breaks when the software is removed. If the answer is “a report is less convenient,” the revenue is soft. If the answer is “the business stops,” it is hard.

The second is switching cost that grows with use. Every integration written against the product, every year of history accumulated inside it, every workflow shaped around it raises the cost of leaving — not as a penalty the vendor imposes, but as a consequence of the customer having built on top. The best versions of this are not lock-in in the pejorative sense; they are the natural result of a product becoming more useful the longer it is used, so that the data it holds is worth more than the data any replacement would start with.

Retention is the single most honest number a software business produces, because it is the customer voting with the renewal rather than the vendor describing the product.

The third is non-discretionary spend — expenditure the business cannot easily defer because it keeps the operation running or discharges an obligation it owes to someone else. This is where one corner of our market, the software that regulated industries depend on, earns its particular conviction: when the spend answers a supervised obligation, the buyer’s question is not “is this worth it” but “can we afford to be without it,” and that is the most durable question a vendor can be on the right side of. But the principle is broader than regulation. Payroll, billing, the system that runs the warehouse — all of it is non-discretionary for the same structural reason.

What the durable businesses have in common

Read across the software companies whose revenue has compounded for a decade and the surface differences fall away. They are not united by category, by buyer or by price point. They are united by the fact that, at renewal, the customer is not making a value judgement but a continuity decision — the alternative to renewing is not reverting to a spreadsheet but re-opening a problem the software had closed.

  • High gross retention, held through at least one downturn, not just a growth year.
  • Net expansion that comes from the product doing more of the customer’s work over time, not from repricing.
  • A reason for the renewal that the customer, not the vendor, can articulate.

That is the business we want to own, and the test we apply before we underwrite anything else about it. The multiple, the growth rate, the market size — all of it is downstream of one question: when the customer next looks hard at this line, will they conclude they cannot do without it. Where the honest answer is yes, the revenue is durable, and durable revenue is the only kind that compounds.