Seat counts explain less of software revenue

For an owner preparing a software company for a sale or a growth raise, this autumn's rebound in listed software shares does not settle the valuation question. Reuters reported on October 6 that the S&P 500 software and services index had reached its highest level since November 2025, after its largest quarterly gain since the second quarter of 2020.^[1] The same index had lost more than 26% between late January and April, on fears that companies could use AI to build applications in-house more cheaply.^[1] Our transaction conclusion is narrower than the index move: present revenue so that a buyer can trace it to the way customers now pay.

Public valuations also remain below their earlier range. Meritech's September 25 data put the median public software company at 4.0 times implied annual recurring revenue, 54% below the pre-ZIRP median of 8.7 times in its series, and found 66% of companies trading below five times next-twelve-months revenue.^[2]

Pricing has moved faster than the metrics

A survey of more than 230 software companies, fielded in April and May and published by the Growth Unhinged newsletter, found that three in four had changed pricing or packaging within the last year, and that 37% used hybrid pricing, up from 25% a year earlier.^[3] Gartner estimated on July 1 that up to $234 billion of enterprise application spending is exposed between now and 2030 to what it calls agentic arbitrage, in which AI agents complete tasks across several systems and reduce the need for users.^[4] A Gartner analyst said the shift breaks the link between user growth and revenue growth for many enterprise software vendors.^[4] The survey describes its respondents, not the whole market, and Gartner's figure measures exposure, not lost revenue. Neither shows what a particular company's customers will do.

The practical problem is measurement. When customers replace seats with usage fees or a charge per completed task, net revenue retention can move for reasons that say little about loyalty. PwC's February guidance for dealmakers warns that metrics such as net revenue retention can mask AI-driven seat contraction, and that gross revenue retention and cohort-level analysis are now essential.^[5] Bain's September 29 report says net revenue retention has dropped about nine points since 2021, and its first question on AI is whether AI is driving incremental revenue.^[6]

What buyers are paying for

Deal activity has held up better than price. Software Equity Group counted 698 SaaS transactions announced in the second quarter, up 9.6% from a year earlier, at a median of 4.0 times trailing revenue, down modestly from 4.2 times.^[7] It said buyers are placing a higher bar on strategic fit, AI relevance, workflow embedment and long-term defensibility.^[7] Lincoln International reported in May that software company values in its private market index fell 8.8% in the first quarter, entirely through lower multiples, reflecting market participants' views of longer-term AI-driven disruption.^[8] These are medians and index readings across very different businesses, not a quote for any one company.

Build a revenue bridge before launch

Start with a bridge of the last eight quarters of revenue by pricing model: seat licenses, usage, fees per outcome and services. For each customer cohort, split the change after a pricing move into price, seats removed, usage growth and churn. Report gross retention beside net retention, and show AI-priced products separately from the rest.

Where usage or outcome fees are new, show how much of next year's revenue is committed by contract and how much is estimated from usage, and how far billings have differed from those estimates month by month. If a pricing change is recent, say so plainly, and let the buyer see the cohorts that have already renewed under it.

None of this requires a view on whether seat pricing survives. A company that can show how each dollar is now earned, and how that revenue behaved through renewals, leaves a buyer less to estimate. That is a firmer basis for a process than an index level.