Revenue multiple vs profit multiple

A profit multiple is the sold price divided by trailing-twelve-month profit. A revenue multiple is the sold price divided by trailing-twelve-month revenue. Same numerator, different denominator, so for any single deal the profit multiple is the larger of the two — profit cannot exceed revenue, and a smaller denominator makes a bigger quotient. A business selling for $10,000 on $20,000 of revenue and $5,000 of profit is simultaneously a 0.5× revenue deal and a 2× profit deal. Both statements are true, and quoting one without saying which you mean is where most small-deal pricing arguments start.

That relationship is a useful integrity check, and we run it as one: across every row in our ledger that carries both multiples, not one has a profit multiple below its revenue multiple. If you ever see that inversion in a comp set, the profit figure is above the revenue figure and something in the underlying listing is wrong.

Coverage is not the same on both bases, and it is not the same across asset types. The table below is the count of comparables carrying a usable multiple on each basis, measured over the same 200-deal window a paid report uses, alongside the median figure each denominator is drawn from. Where a count falls under our five-comp floor we publish nothing further from it — the median cell stays empty rather than quoting a figure we would not price against.

The two counts in a row are not two views of one sample. A listing that publishes profit but not revenue lands in the first count only, and vice versa — which is exactly why the next section exists.

The most tempting mistake this page can prevent: taking a per-type median profit multiple and a per-type median revenue multiple, dividing one by the other, and reading the result as an implied margin. Those two medians are computed over different populations. A deal whose listing stated profit but no revenue contributes to one and not the other, so the ratio between them is partly a fact about disclosure habits, not about economics.

The effect is large enough to invert the ordering. On 29 August 2026 our SaaS rows produced a median profit multiple of 0.90× across 59 usable comps and a median revenue multiple of 1.88× across 46 — an ordering no individual deal can produce, and entirely an artefact of the two medians resting on different rows. Ledger-wide on the same day, 87 rows carried a usable profit multiple with no revenue multiple to pair it with, against 8 rows the other way round. A within-deal comparison is the only sound one: compare bases on the same deal, or compare the same basis across deals, never one median against the other.

There is a further problem specific to the size band we cover. On marketplace listings for three- and four-figure businesses, the seller frequently enters one earnings figure and it ends up in both the revenue and the profit field. The row then carries two multiples that are identical, and the "profit multiple" has had no cost of any kind subtracted from it.

This is not an edge case in our data. Of the rows that carry both multiples when this page was written, the two were the identical number on 44% of them — 98 rows out of 222. On mobile apps the share was higher still. We keep those rows: the price is real and cited, and discarding evidence because it is imperfect would bias the sample toward sellers who fill in forms carefully. What we do instead is publish both bases and their counts, so the pattern is visible rather than buried inside a single headline multiple. Pro subscribers can reproduce the whole calculation from the comps CSV export.

Practically: when a listing states the same figure twice, treat it as a revenue multiple. It tells you what buyers paid per dollar of top line. It tells you nothing about what the business earned.

Our own rule is deliberately boring and keyed to the subject rather than the sample: if the business being valued states a positive trailing-twelve-month profit, we price it on the profit basis; otherwise we price it on revenue. The estimate is then built only from comparables carrying a usable multiple on that basis, after discarding multiples at or below zero and at or above 100× as data errors and trimming 5% from each tail once the sample is large enough to stand it, over the most recent 200 comparable deals — the window a paid report uses; the free calculator reads a wider one. Below five usable comps we publish no estimate, and where the subject is more than four times the size of the median comparable we say so instead of letting the range stand unqualified.

For your own pricing, three questions settle it. Does the business state a profit figure at all, and does anything net out of the revenue to produce it? Are the comps you are pricing against disclosed on the same basis? And if you switch bases, does your answer move more than the negotiating range you had in mind — because if it does, the basis, not the multiple, is the number you are actually arguing about.