What a startup option grant is really worth

The offer letter quotes one number. This is the other one: how large the exit has to be before the equity covers the salary it cost.

The offer is 40,000 options, four-year vest, one-year cliff, struck at $1.15. The last round priced the stock at $4.20. The recruiter does the subtraction out loud: $122,000, sitting there already, and a multiple of that if things go well.

The other job pays about $60,000 a year more.

The $122,000 is not money. It is the gap between two prices on a cap table, and four things stand between it and a bank account. Investors are repaid first. Later rounds shrink the stake. The shares have to be bought. The gain is taxed.

It becomes money only at an exit — a tender offer, an initial public offering, or an acquisition. The calculator treats all three the same way: as a price for the whole company, on some future date.

Set the numbers below to the offer in front of you — the grant, the company behind it, this job’s salary and bonus, and what the job you would turn down pays all in.

This page runs on javascript, and yours is off. The short version: an option grant pays nothing until the exit clears what investors put in, and it does not beat a higher salary until it also clears the pay given up in the meantime. For a typical mid-stage grant that means an exit several times the company's current valuation.

The four bites

Investors are repaid first. Preferred stock carries a liquidation preference. In an acquisition, the money that went in comes back out before common stock — employees — sees a dollar. A company that raised $48 million and sells for $40 million pays its staff nothing on their options. That is not a small win. It is a zero, and it is why the curve above runs flat on the left: down there, the offer is worth its salary and bonus and nothing more.

The stake shrinks. Options divided by fully diluted shares is the stake on the day of signing. Every round after that issues new shares. At 12 percent a year over six years, a third of a percent becomes a sixth of one. Move the dilution slider and the break-even number moves with it.

The shares cost money. Exercising means buying them, at the strike price, in cash. Forty thousand options at $1.15 is a $46,000 cheque, due before an exit turns it into anything.

The gain is taxed. What is left after the cheque is income, at a rate that depends on the type of option and how long the shares were held.

What the model leaves out

Preference terms past 1× non-participating. Participation, multiples and seniority all make the downside worse than what is drawn here, never better.
The difference between the three exits. In an IPO the preferred converts to common and the preference stack goes away, so the model runs pessimistic there and about right for an acquisition. A tender offer is usually partial, priced off the last round, and often limited by tenure.
The tax code. One blended rate stands in for ISOs against NSOs, the alternative minimum tax on the spread at exercise, holding periods, and the qualified small business exclusion. Real numbers need an accountant.
Raises, refreshers and promotions, on both sides of the comparison.
The 90-day exercise window. Leave, and in most grants the options expire three months later. That turns the grant into a bill, payable at once, on stock that still cannot be sold.

None of this argues for the safe job. Grants clear the bar all the time, and the ones that clear it tend to clear it by a wide margin. The argument is only that the bar is a number, that it can be worked out before signing, and that it is almost always larger than the one read out over the phone.