What is a SaaS business worth?

Our SaaS rows come from two kinds of source, and they behave nothing alike. Small auctions publish a closing price beside a seller-stated profit figure, so those deals can carry a multiple. Acquisition journalism and first-party founder posts publish a price and, usually, no profit line at all — they raise the deal count without touching the median.

That split is not a rounding detail. Nearly every profit multiple in our SaaS sample comes from the auction end, where the typical row is a few hundred dollars. The million-dollar acquisitions in the same table sit outside the median entirely, because there is nothing to divide by. So the median above is an accurate, cited, arithmetically correct statement about the small-auction end of SaaS — and it is not a statement about what a SaaS business with audited books and two years of retention data is worth.

A median below 1× reads like a data error until you look at what is being sold. The profit figure is the seller's own claim about software the bidder has never operated; the closing price is what an anonymous buyer would risk against that claim. Two numbers, two completely different standards of evidence — and the gap between them is the finding.

The recurring reasons, all of them visible in listings rather than inferred:

• The revenue is asserted, not attested. A dashboard screenshot is not read-only access to the payment processor, and bidders price the difference.
• The history is too short to have a trend. Most of the SaaS rows we can compute a multiple for were under a year old at sale — see annualized months.
• Churn is invisible. A month of subscriptions tells you nothing about how many of those subscribers renew, and a buyer who cannot see retention assumes the worst.
• The product may be a thin wrapper. When the software's core is somebody else's API, the buyer inherits that vendor's pricing and terms along with the code.
• The buyer pool is thin. Three-figure auctions attract bidders pricing an option, not operators pricing a business.

Our SaaS sample also contains multiples above 30×, and they are not evidence of a hot market. They are what happens when you divide a real price by a nearly-zero profit figure.

This is the single most useful thing our SaaS data says. A multiple is a ratio, and a ratio with a tiny denominator is noise wearing the costume of a valuation. Before you quote one — yours or anybody's — look at the profit figure underneath it and ask whether it is large enough to divide by.

One deal in our SaaS sample sits at real size with both numbers published: an Instagram growth SaaS sold for $300,000 against $18,000 in monthly recurring revenue and $9,000 in monthly profit — 2.78× profit, 1.39× revenue — reported by They Got Acquired and sold through a broker. It is a 2018 sale, it is one deal, and we would not build a median out of it. What it demonstrates is the shape of the difference: recurring revenue a buyer can confirm, a broker-run process with more than one interested party, and a price that survives being divided by a real profit figure.

The distance between that row and a $75 auction is not the multiple. It is everything the multiple was calculated from.

None of these are answered by a multiple, and all of them move a price more than the choice of multiple does.

1. Can revenue be verified without the seller? Read-only access to Stripe or the app store beats every screenshot, export and spreadsheet.
2. What is net revenue retention over twelve months? Gross MRR with 15% monthly churn is a different asset than the same MRR with 2%.
3. How concentrated are the customers? One account carrying a third of revenue is a discount a buyer will insist on, and the one sellers argue with most.
4. What does the stack depend on? Upstream API pricing, rate limits and terms transfer with the product — including the right to keep using them commercially.
5. Who owns the code? Contractor work without a written assignment is an unresolved claim on the asset, not a detail for later.
6. How much of the product is the founder? Support, sales and the one undocumented deploy script all price in as transition risk.
7. Month by month, not trailing twelve. A single annual total hides the launch spike, the decline after it, and the month the churn started.
8. How many buyers are actually bidding? Price is set by the second-most-motivated buyer; one interested party and a live auction are different pricing mechanisms.