For years the first page of every software story was the growth rate. The first page now is what stays. The numbers behind the shift are not subtle. Bain's April 2026 read on the sector has revenue growth that ran near 20 percent a year now trending at half that, and net revenue retention down about eight points since 2021. Growth bought with discounting and easy expansion left as quickly as it arrived, and buyers paid for it. They now pay for revenue that behaves as if it will still be there years from now.
Public markets set the reference price. Meritech's April 2026 software pulse put the median implied ARR multiple for public software at 3.2 times, against the pre-ZIRP median of 8.7 times in Meritech's series. The top ten companies still trade at 11.7 times. The spread between the median and the top is where durability is priced.
Durable has a specific shape in diligence. Retention that holds across cohorts, not just in the blended number. Expansion that comes from usage and product rather than from repricing a captive base. Workflows deep enough in the customer's operations that removal is a project, not a decision. A customer list without concentration cliffs. Pricing power demonstrated, not asserted.
The bar moved for reasons that are not going away. The ten-year Treasury yield has spent most of the past three years above 4 percent, against under 2 percent through 2020 and 2021, and that resets what a dollar of future revenue is worth today. AI has made parts of the software landscape contestable: buyers now ask which workflows a customer could plausibly rebuild or replace, and which are protected by data, integration depth and switching cost. Carta's first-quarter 2026 data shows early-stage SaaS valuations softening while 83 percent of SaaS venture dollars went to AI-native companies. Durability is questioned harder because more of it is genuinely in question.
The buyers are still there for businesses that can prove it. Kroll counts a record 2,897 software deals announced in 2025, with $291 billion of value, and strategic acquirers paid 5.6 times trailing revenue against 4.4 times for sponsors, the widest premium in a decade. That premium went to companies that could show what stays.
Positioning has to move with the bar. Lead the narrative with quality of revenue, and make the model prove it: cohort behavior, expansion by driver, margins with costs fully loaded. Claims the data room can verify in an afternoon are worth more than claims that need a meeting.
Headline growth still matters. It sets the ceiling on enthusiasm. But retention sets the floor on price, and processes are won and lost at the floor. Companies that show both, and can explain the machinery behind them, are the ones clearing today's bar at full value.
