KeyBanc's SaaS survey is 16 years old and 71 companies deep

The record

KeyBanc Capital Markets' technology group states on its own SaaS survey access page that 'for the past 16 years, the KeyBanc Capital Markets technology group has annually surveyed hundreds of companies,' with the report distributed through a request form rather than posted openly. The 2025 edition, produced jointly with Sapphire Ventures and titled the '16th Annual SaaS Survey,' states it was 'conducted from Q2'25 close through Q3'25 close' and drew 71 respondents. Its median-metrics table reports ARR growth of 31% in 2022, 20% in 2023, 15% in 2024, and forecasts of 20% for 2025 and 25% for 2026, alongside median net dollar retention of 106%, 102%, 101%, and forecasts of 102% and 104% for the same years respectively.

What the documents establish

The access page establishes the survey's history and cadence, sixteen consecutive years, but not a specific figure, since the results themselves sit behind a distribution form; the figures above come from the 2025 PDF itself once obtained, not from the marketing page. The PDF is explicit that 2025 and 2026 figures are estimates supplied by respondents about their own current and coming year, not audited results, while 2022 through 2024 are reported as historical medians from the same respondent base. With 71 respondents, this is a considerably smaller and more concentrated sample than KeyBanc's own 'hundreds of companies' description of the survey's history suggests for earlier years, a gap worth flagging rather than assuming away.

The operating read

Editorially, the retention decline from 106% to 101% between 2022 and 2024, alongside the growth deceleration from 31% to 15% over the same window, describes a self-selected group of software executives willing to complete a co-branded bank and venture survey, not a formal census of private SaaS. A founder who sees this survey cited in press coverage should distinguish between the primary KeyBanc-Sapphire document, which states its own sample size and estimate labels, and any secondary summary that repeats a headline figure without those labels attached.

What to check before you decide

Before citing a KeyBanc SaaS survey figure, check the following.

• Is the specific year and edition named, given the metrics table spans several years within one document?
• Is a forecast figure, marked E in the source table, being presented as an already-realized result?
• If the figure came through press coverage rather than the survey PDF itself, is that secondary sourcing disclosed?

The survey is a long-running, bank-distributed benchmark with a small and self-selected respondent base in its current form, best cited with its year, its estimate-versus-actual labels, and its sample size attached.