Valuation is more black art than science but there are generally 3 accepted methods
- replacement cost basis ... basically what you've spent (or someone else would spend) to get to the same point
- discounted cashflow
- market peers
You can immediately see the problem(s)
- in era where technology is moving so fast, new tools/frameworks make the next generation faster cheaper better so often the replacement cost is a fraction of the actual sunk costs
- negative cashflow and no revenue means crystal-ball gazing is now a first order predictive tool
- if you are a first mover or creating a completely new category, who are your peers?
So investors avoid the issue (along with the tedious yelling/name-calling) by issuing convertible notes which is basically a loan (at high interest rate) with debt-equity swap at discretion of investors. This gives them a chance to see your performance and leaves the tedious arguments about value to a more mature stage. The years of (painful) startup funding people have come up with rough rules of thumbs like $1M / engineer, or $X per unique visitor or $Y per conversion rate. However, this varies so much across domains that only detailed study might come up with a methodology which is acceptable accounting-wise. If you're fortunate enough to be at that stage to pay for such expertise, congratulate yourselves on a value-creation joy-ride.