Startup valuation: the five methods, and which one your stage actually uses

What the number actually is

Startup valuation is the process of assigning a monetary value to a company that usually has no profits and often no revenue, using comparable transactions, structured scorecards, or the arithmetic of the round being raised — producing a negotiating position rather than a measurement.

That last clause is the part worth holding onto. Valuing a mature business is an estimation problem: there are cash flows, and the question is what they are worth today. Valuing an early-stage company is a negotiation with an estimation problem attached. Two competent parties can look at the same company and land a factor of three apart without either being wrong, because most of the value sits in a future neither can observe.

Methods matter anyway — not because they produce the truth, but because they are the language the negotiation is conducted in. A founder who cannot say which method produced their number has conceded the frame.

The identity everything else hangs off

Before any method, one piece of arithmetic:

Post-money = pre-money + amount raised. Investor ownership = amount raised ÷ post-money.

Raise $2M at an $8M pre-money and the post-money is $10M; the investor owns 20%. Raise the same $2M at a $6M pre-money and they own 25%.

This identity, not a model, sets most early-stage prices. The practical sequence in a seed round is usually backwards from how it is presented: the founder works out how much money reaches the next milestone, decides what dilution is tolerable, and the pre-money valuation falls out. A term sheet that quotes a valuation without specifying pre or post is either careless or deliberately ambiguous, and the difference is real money.

The five methods

1. Comparable transactions

What did similar companies raise at, recently, in this sector and geography, at this stage?

The dominant method in practice, and the one investors actually use, because it reflects the market as it is rather than as a model says it should be.

Its whole weight rests on comparable. Matching on ARR alone is not a comparison. Two SaaS companies at $1M ARR — one growing 15% month over month with 120% net revenue retention, the other growing 4% with 85% — are not the same company and will not clear the same multiple. A comparison set that ignores growth, gross margin and retention is a list, not evidence.

This is also where benchmark hygiene matters. INITE Studio's finance lens is constrained to early-stage comparables and explicitly rejects mega-cap benchmarks, because a $1M-ARR company benchmarked against NVIDIA or Amazon produces a number that is worse than no number — it is confidently wrong and easy to dismiss.

2. Revenue multiples

A shorthand on top of comparables: ARR × a multiple drawn from the comparison set.

Fast and legible, and it fails in one specific way. The multiple is not a property of the company; it is a property of the comparison set at a moment in time. Multiples across the whole software market moved by a large factor between 2021 and 2023 without any individual company changing. Quoting a multiple from a different market cycle is quoting a different market.

Use it as a sanity check on a number you got another way, not as the derivation.

3. Scorecard and checklist methods

Structured pre-revenue approaches. Start from the average pre-money for recent seed rounds in your region, then adjust up or down against weighted factors — team strength, market size, competitive position, product stage, channel.

Their virtue is that they force the conversation to be explicit. Their weakness is that the weights are conventional rather than derived, so two people applying the same scorecard to the same company can land 40% apart.

They work best as a structure for disagreement, not as a calculator. If you and an investor apply the same scorecard and diverge on the team weighting, you have found the actual disagreement, which is more useful than the number.

4. Discounted cash flow

Project cash flows, discount them to present value at a rate reflecting risk, add a terminal value.

Mathematically valid at every stage and close to useless before revenue. Pre-revenue, almost all of the result sits in the terminal value — the assumption about what the business looks like in year seven or ten. Change that assumption slightly and the answer moves by multiples. The model does not tell you the value; it restates what you assumed, with decimal places.

It becomes genuinely useful once there is enough revenue history that the near-term projections carry real weight — which for most software companies is somewhere after Series A.

5. The round itself

Not a method so much as an admission: at pre-seed and seed, the valuation is frequently set by what the round needs to be.

Standard instruments make this explicit. A SAFE or convertible note defers the valuation entirely, often with a cap that acts as a ceiling rather than a price. The company is not valued at signing; the cap is a bet about what the priced round will look like.

This is worth naming because founders often feel they should have a defensible model when the market norm is a negotiated band informed by comparables. Knowing the band is more useful than owning a model.

Which method your stage actually uses

The pattern is a handover: as evidence accumulates, methods based on the company's own numbers displace methods based on other people's.

What actually moves the number

Four things, roughly in order of leverage at early stage.

Growth rate, above all. For revenue-stage companies, growth explains more of the multiple than absolute size does. A smaller company growing fast will out-value a larger one growing slowly, consistently.

Retention. Net revenue retention above 100% means the existing customer base grows without new sales, which changes the shape of every forecast built on it. It is the single number most likely to be checked.

Gross margin. It decides whether revenue is software revenue. A "SaaS" company at 45% gross margin will be valued closer to a services business, because that is what the margin says it is.

Competitive alternatives. Two interested investors move a valuation more than any model. This is uncomfortable and true, and it is why fundraising process design affects price as much as company performance does.

Unit economics sit underneath all four as a floor rather than a lever. An LTV/CAC ratio near 1.5 is the point below which payback becomes arithmetically impossible; 3.0 is the common definition of healthy; and CAC payback beyond roughly 36 months means the model has stopped being a business. These do not raise a valuation — but failing them caps one, whatever the growth rate says.

The three common errors

Confusing pre and post. The most expensive clerical error in early-stage fundraising. On a $2M raise, the difference between an $8M pre and an $8M post is five percentage points of ownership.

Optimising for the headline. A high valuation raised on hard terms — participating preferred, a large liquidation preference, aggressive ratchets — can leave founders worse off than a lower valuation on clean terms. The headline number is one term among several, and it is the one that appears in the press.

Valuing the projection instead of the company. A model showing $40M revenue in year five does not make the company worth a multiple of $40M today. Investors discount projections heavily and almost universally, because they have seen the base rate.

Presenting a number you can defend

Three things make a valuation survive questioning.

Name the method. "Comparable seed rounds in vertical SaaS in Europe over the last four quarters" is a position. "We think we're worth $8M" is not.

Show the comparison set. Four or five named transactions with their ARR, growth and terms. The set is the argument; the number is the conclusion.

State the dilution first. Leading with "we're raising $2M for 20%" rather than "we're valued at $8M pre" frames the conversation around the thing both sides actually care about, and signals that you understand the arithmetic rather than the headline.