Most founders learn how to grant stock options the way they learn fire safety: after something has gone wrong. A grant that was never board-approved. A strike copied from an old spreadsheet. An offer letter that "granted" 1% with nothing behind it. Each is cheap to prevent, expensive to fix, and each surfaces, years later, in diligence.
The good news: granting options correctly is a repeatable checklist with five parts: a plan and pool, a board approval, a strike price, an acceptance, and a record. This lesson covers each with real numbers, then the common failure modes.
Every grant needs board approval
Options are granted under an equity incentive plan, a document your board (and usually your shareholders) approved when the option pool was created. It sets the reserve, the kinds of awards allowed, default terms, and who administers it. No plan, no grants: fix that first.
Under the plan, each individual grant must be approved by the board, in a meeting recorded in minutes, or by written consent between meetings. This is not ceremony: legally, the grant does not exist until the board approves it. An offer letter can promise the company "will recommend to the board" a grant of 20,000 options; it cannot itself grant them.
The practical cadence: batch new-hire grants for each board meeting, or circulate a written consent when a hire can't wait. Either way, the approval must name the recipient, share count, strike price (or the mechanism, "fair market value on the date of this consent"), and vesting schedule.
Why investors care: in every financing and acquisition, someone reconciles the option ledger against the board record, grant by grant. Grants with no matching approval get flagged, and the fix costs legal fees and leverage at exactly the wrong moment.
Grant date vs vesting commencement date
Two dates get conflated constantly, and the difference matters for tax and fairness:
- The grant date is the day the board approves the grant. The strike is fixed as of this date; you cannot choose it retroactively.
- The vesting commencement date is the day the vesting clock starts, usually the first day of work, often earlier than the grant date.
Backdating the vesting commencement to the start date is normal and fair: a January hire whose grant is approved in March shouldn't lose two months of vesting to board scheduling. Backdating the grant date to capture a lower strike is a different animal; see the mistakes section.
Concretely: an engineer starts January 5, the board signs a consent March 2. Grant date: March 2, strike = FMV that day. Vesting commencement: January 5, so the one-year cliff falls the following January 5. Put both dates in the paperwork explicitly. A missing commencement date is a classic defect in early-stage option files.
The strike price: fair market value, no exceptions
The strike (exercise) price must be at least the fair market value of the common stock on the grant date. In the US this is the Section 409A rule; the safe way to comply is an independent 409A valuation less than 12 months old that predates any material event (a new term sheet, a signed round). The valuation comes from an outside provider; having a current one on file before you grant is the company's job.
Two details trip people up. First, it's the FMV of common stock, typically far below what investors paid for preferred, which is why early-stage options have real upside. Second, a below-FMV strike is not a favor to the employee. It's a tax trap aimed at them.
Worked example: the discounted-strike tax bomb
A company's latest 409A puts common FMV at $0.60. It grants an engineer 10,000 options at a $0.60 strike. Two years later, after a Series A, a new 409A puts FMV at $2.40. Fully vested, the paper gain is 10,000 × ($2.40 − $0.60) = $18,000, taxed under normal option rules at exercise and sale. That's the system working.
Now suppose the strike had been set at $0.10 "to make the offer attractive" while FMV was $0.60. Under US 409A rules the option can be taxed as it vests on the spread, before the employee can sell anything, plus a 20% additional federal penalty tax on top of ordinary rates, plus interest, plus withholding failures for the company. The 50-cent "discount" becomes a recurring tax bill on illiquid paper gains.
The same logic applies outside the US with local mechanics: UK EMI schemes agree a valuation with HMRC in advance; Indian ESOP taxation leans on a merchant-banker FMV. Whatever the jurisdiction: keep a current, defensible valuation on file and set every strike from it. More in the 409A and valuations lesson.
From offer letter to accepted grant
The correct sequence has four steps, and the order is the point:
- Offer letter. Promises the company will recommend a grant of N options to the board, subject to board approval and the plan. It states the count and vesting schedule, not a strike, because the strike doesn't exist yet.
- Board approval. Fixes the grant date and the strike.
- Grant paperwork. A grant notice and option agreement (plus the plan) reflecting exactly what the board approved.
- Acceptance. The recipient signs. Now, and only now, is the grant a complete, enforceable contract.
Step 4 is where companies quietly fail. Grants go out, nobody signs, nobody notices, until diligence. An unaccepted grant is a genuine problem: without a signed agreement, an acquirer's counsel can't confirm the holder is bound by the plan's terms (exercise mechanics, termination provisions, transfer restrictions), and the terms stay disputable. A departed employee who never signed can claim different vesting was promised, with no executed document to answer them.
Treat acceptance like an open invoice. Track every grant from "approved" to "signed," and chase anything unsigned after 30 days. A 10-minute recurring check keeps the ledger diligence-clean; the alternative is hunting signatures from former employees the week your acquisition is supposed to close.
This is a workflow problem more than a legal one, and worth automating. In Vquity, recipients accept grants through a public link that produces a signed PDF, each grant carries its own vesting ledger, and a pool-utilization chart shows what's committed, so "which grants are unaccepted?" has a one-screen answer instead of an email archaeology project.
Telling candidates what their options are worth
An option count is meaningless without a denominator. "25,000 options" could be 2.5% of a company or 0.025% of one. Honest communication gives candidates enough to do the math themselves: the share count and fully diluted total (or the resulting percentage), the strike and current common FMV, and scenario math: what the grant could be worth at different exits, caveats attached.
Worked example: an honest offer conversation
A candidate is offered 25,000 options at a $0.60 strike. The company has 10,000,000 fully diluted shares, so the grant is 0.25%.
Exit at $50M: naive per-share value is $50M ÷ 10M = $5.00, so the grant grosses 25,000 × ($5.00 − $0.60) = $110,000 before tax.
Exit at $200M: $20.00 per share → 25,000 × $19.40 = $485,000 before tax.
No exit: $0, statistically the most common outcome for any individual startup.
Two honest caveats to say out loud. Dilution: two more rounds might take 0.25% down to roughly 0.17%, scaling every payout above down by about a third (see the dilution math lesson). Preferences: in lower outcomes, liquidation preferences are paid before common, so common's per-share value is less than exit ÷ shares (see the waterfall lesson).
Never state a future value as fact ("this will be worth $500K"), quote the preferred price as if it were the value of the candidate's common, or promise a percentage without saying fully diluted and as of when. The equity for employees lesson covers this conversation from the recipient's side.
Record-keeping that survives diligence
For every grant, be able to produce a complete file on demand:
- The board consent or minutes approving the grant (recipient, count, strike, schedule)
- The grant notice and the executed option agreement, with acceptance date
- The plan document and any amendments in force at the grant date
- The valuation report that supported the strike, current as of the grant date
- The vesting schedule with an explicit commencement date
- Any later amendments: repricings, exercise-window extensions, accelerations
Alongside the per-grant files, keep one ledger that reconciles the pool: authorized, minus granted, plus forfeitures returned, equals available. Reconcile quarterly. Pool math drifts when forfeitures go unrecorded, and granting more than the plan reserves means going back to the board and shareholders. US companies relying on Rule 701 should also track rolling 12-month sales against its USD 1 million, 15%-of-assets, and 15%-of-class tests, with enhanced disclosure planning if aggregate sales may exceed USD 10 million. Where this paper lives, and how it gets checked, is the subject of the due diligence lesson.
Common screw-ups, and what each one costs
Almost everything that goes wrong with option granting falls into six patterns:
- Backdating the grant. Picking an earlier grant date to capture a lower strike is not a gray area. It misstates compensation expense, creates the below-FMV tax problem above, and fueled a wave of corporate scandals in the 2000s. Backdating vesting commencement is fine; backdating the grant is fraud.
- Granting without board approval. Offer letters that "hereby grant" options, or founders promising grants the board never sees. Fixable by ratification, but if FMV has risen, the strike must reflect the ratification date, and the employee eats the difference.
- Oral promises. "We'll get you 1%" said in a hallway is not a grant, but it is a claim, and disputed equity claims resurface during acquisitions, when settling them costs real money. If you mean it, put it in an offer letter with the subject-to-board-approval language; if you don't, don't say it.
- Granting on a stale valuation. A 409A over 12 months old, or one that predates your new term sheet, doesn't support a strike. Pause grants, refresh, then batch-approve.
- Percent-versus-shares ambiguity. Promising "1%" without fixing the share count at a point in time invites a dispute over pre- vs post-round. Grants are made in shares; percentages are context.
- Letting acceptances lapse. Covered above: the quietest defect here, and one of the most common.
None of this is hard. How to grant stock options correctly comes down to sequence and records: plan, board approval, FMV strike, signed acceptance, complete file. Run the checklist every time and grants become routine. The next lesson covers a moment that is anything but: closing a round.