Kroll: growth pays 3x, margin above 25% pays nothing

Kroll has published its Global Software Sector Update for Summer 2026, covering software M&A and public comparables through 30 June 2026. It runs to 20 pages.

The clearest number in it is the growth cliff. Median EV to CY26 revenue for public B2B software companies growing above 20% is 7.2x. Companies in the 10% to 20% bracket get 4.1x, 0% to 10% get 3.1x, and shrinking companies get 2.4x.

Margins do not follow the same curve. Median EBITDA margins across those four buckets are 28.9%, 31.6%, 26.0% and 25.1%. The companies earning 7.2x are slightly less profitable than the companies earning 2.4x.

On Kroll's figures, roughly 25% to 30% EBITDA margin is simply what a public B2B software company looks like now. Cutting past that buys no multiple.

The EBITDA multiples tell the same story: 7.0x for shrinking companies, 9.9x for 0 to 10% growth, 14.1x for 10 to 20%, and 18.2x above 20%.

Category matters more than either input. Kroll's Engineering bucket and its HCM bucket both score 46% on the Rule of 40. Engineering trades at 5.2x, HCM at 3.0x, a 73% gap. Marketing and Cyber Security both score 38%; Marketing gets 2.1x and Cyber 5.2x. Kroll does not say what explains the spread.

SaaStr, which pulled the figures out of the report, argues the spread is either revenue durability or the market's read on which categories AI agents make redundant. It concedes the data is consistent with that rather than proof of it.

Selling currently pays better than being listed. Median EV to last-twelve-months revenue for strategic acquisitions in the first half of 2026 was 6.0x, against a ten-year average of 4.7x and a 2021 peak of 6.2x. The median public B2B company trades at 3.3x forward revenue, up 6% in the quarter and still short of the 15-year median of 5.4x.

Private EBITDA multiples rose 25% to 20.2x, near the 2021 record of 21.4x. Private equity multiples have sat at about 4.4x since 2024. Strategic buyers now account for 74% of software transactions, up from a 69% average across 2024 and 2025.

The recovery in headline volume is thinner than it looks. Annualised 2026 deal count is about 2,672, second only to 2025's 2,939, but annualised announced value is roughly $240bn — and about $120bn of that is one deal, SpaceX's $60bn purchase of Cursor in the first half, doubled by Kroll's annualisation. Strip it out and the year runs near $120bn, against $429bn in 2021. Cursor alone was about 64% of all software deal value in the second quarter, which Kroll calls the most concentrated single deal in the ten years of data it tracks.

Scale is worth less than quality. For companies under $100m in revenue, precedent EBITDA multiples run 11.5x at the first quartile, 16.1x at the median and 22.9x at the third. Above $100m the same figures are 12.4x, 17.3x and 23.9x. Crossing $100m moves the median about 7%; moving from bottom to top quartile within a size class roughly doubles it.

Multiples also shifted down-market this quarter. Medians rose 0.5x for companies under $25m in revenue, to 4.6x, and 0.5x for the $25m to $100m cohort, to 5.0x, while medians above $100m fell 0.4x to 4.2x.

For a founder pricing an exit or a raise, the practical order is: two points of growth around the 20% line, then which comp set you are placed in, then size. Deeper cost cuts below the 25% margin line are, on this data, free work. And the bid is corporate, not financial, which is a relationship question with an 18-month lead time.