A valuation you cannot defend is not a valuation

Most early-stage valuations are arrived at backwards. A founder decides how much money is needed, decides how much of the company is acceptable to sell, and divides. The resulting number is then presented as a valuation, when it is in fact an ask wearing a valuation’s clothes. It carries no information about what the company is worth, which is why it collapses the moment an investor who knows the sector applies any pressure to it.

The alternative is not a single correct number. It is a defensible band, arrived at by triangulation, and a willingness to state which of four things is actually true.

Every valuation resolves to one of four outcomes

Before any method is applied it is worth naming the possible answers, because a process that can only produce agreement is not a process.

01Reasonable: The number the company arrived with sits inside the band the evidence supports. It is adopted and defended as it stands.
02Too high: The number sits above the band. It is corrected downward, and the corrected number is then defended properly. This is the most common outcome and the least welcome one.
03Too low: Less common, and more expensive than founders assume. A company priced below its band gives away equity it did not need to, and signals to the market that nobody involved knew what it was worth.
04No support: There is not yet enough evidence to establish a band at all. The honest response is to say so and to go and build the evidence, rather than to invent a number and defend it with adjectives.

A process that can only produce agreement is not a process. It is a service.

Only true comparables carry the band

Comparable rounds are the strongest anchor available at early stage, and they are also the most abused. A comparable is worth citing when it matches on stage, on sector, and on roughly comparable geography and era, and when it can be sourced. Anything else is directional context, and it should be labelled as such rather than quietly counted.

The distinction matters because it changes what the band is made of. A set of three or more true comparables establishes a range the number has to sit inside. A pile of aspirational names, later-stage rounds, and adjacent-sector transactions establishes nothing at all, however impressive the logos are. In practice, if the true comparables are too few to form a band, the band has to be carried by an explicit stage, traction and market rationale, and that rationale has to be written down where an investor can attack it.

A comparable without a citable source is not a comparable. It is a recollection.

The multiple is sector-specific, and defaults are the enemy

A great deal of bad valuation work comes from applying a software multiple to a business that is not software. Sectors trade in genuinely different regimes, for structural reasons, and the regime has to be identified before any multiple is chosen.

Business software trades on recurring revenue, with the top of the range reserved for companies combining high net revenue retention with growth and margin in balance.
Medical devices and diagnostics carry a regulatory moat and high gross margin, and are valued accordingly rather than as generic hardware.
Therapeutics are frequently pre-revenue and are valued on risk-adjusted pipeline value, where a revenue multiple is not merely wrong but meaningless.
Marketplaces are valued on net revenue against the take rate, not on the gross value transacted, which is the number most marketplace decks lead with.
Deep technology carries a premium on intellectual property and on the specific people, which is the hardest of these to defend and the easiest to overstate.

Price off the right revenue

Whether a company should be priced off trailing revenue or off the next twelve months is not a stylistic choice, it is a function of growth rate. A business growing slowly is priced on what it has. A business growing quickly is priced on what it is about to have, because that is what the buyer is actually acquiring.

What is not acceptable is leaving it ambiguous. A defensible valuation states the current revenue, the forward estimate, the growth rate used to get there, whether the multiple is trailing or forward, and the multiple applied. Five facts. Their absence is usually deliberate, and experienced investors read the absence rather than the number.

Then argue against it

The last step is the one most often skipped. Once the methods have been reconciled into a band and a recommended number, the number should be attacked before it goes out: the strongest available argument that it is too high, made in writing, and answered. If it cannot be answered, the number changes.

This is uncomfortable and it is the entire value of the exercise. Every objection found internally is an objection that does not get made for the first time in a meeting, by an investor, in front of a founder who has no answer ready. The number that survives that process is worth defending. The number that was never tested was never a valuation.