Two founders, a term sheet, and a cap table that still lives in a shared Google Sheet. I've watched that exact scene turn into a three-month negotiation over a document nobody had even drafted yet. The investors were ready to wire. The founders were ready to spend. And everything stalled because no one could agree on what happens if one of them quits.
That's the real reason you're here, I suspect. You don't need another definition of a shareholder agreement. You need to know how to actually write a shareholder agreement for startups without lighting your own runway on fire.
I've drafted these documents with founders, sat through the redlines, and watched a couple of them blow up six months later because the drafting was lazy. Here's what I've learned about doing it properly.
Key Takeaways
- A shareholder agreement is a private contract between owners. The articles of incorporation are public. The agreement is where the real rules live.
- Draft the founder clauses (vesting, exit, deadlock) before the investor clauses. If you can't settle founder rules, investors won't fix them for you.
- Free templates are a starting point, not a finished document. Adapt the clauses to your jurisdiction or they may not be enforceable at all.
- Expect 3 to 6 weeks for a genuine first draft and 1 to 3 rounds of comments before anyone signs.
- When the agreement and the articles conflict, the articles usually win against third parties. Fix both, not just one.
How to write a shareholder agreement for startups (step by step)
Start by understanding the order of operations. The shareholder agreement comes after the term sheet and before the money moves. That window is narrow, and founders usually waste half of it arguing about valuation again.
Who drafts it first
In most early-stage deals, the lead investor's lawyer produces the first draft. That draft is not neutral. It reflects the investor's priorities, which means founder-protective clauses get softened or dropped entirely.
My advice: get your own lawyer to review it, even if it costs you. On a seed round, legal fees for a founder-side review run in the low thousands. Cheap compared to losing your vesting schedule.
If there's no investor yet and you're just formalizing a founding team, you draft it. Which brings a problem nobody warns you about.
Founders and investors want different things
Founders want operational freedom. Investors want protection against bad founder behaviour. Those two goals collide constantly, and the agreement is where the collision gets written down.
- Founders push for short vesting and broad decision rights
- Investors want veto rights on spending above a threshold, hiring senior staff, and any new share issuance
- Both sides usually agree on the big picture and fight for weeks over a single threshold number, often board size or budget approval limits
- Nobody thinks about the exit clause until the day someone actually wants out
Here's the thing: a founder who leaves at month eight with 25% of the company and no vesting is a startup-killer. Investors know this. Founders sometimes don't, until it happens.
The clauses every early-stage agreement actually needs
You'll find templates listing forty clauses. Most of them are boilerplate. These are the ones that decide whether your company survives its first dispute.
Founder vesting and the cliff
Standard structure: shares vest over four years with a one-year cliff. Leave before the cliff, you keep nothing. Leave after, you keep what's vested.
I've seen founders negotiate this down to a six-month cliff because they "trust each other." Three of them are no longer on speaking terms. Trust is not a vesting schedule.
Drag-along and tag-along
Drag-along lets a majority force minority shareholders to sell in an acquisition. Without it, one founder holding 5% can block a $40 million exit out of spite.
Tag-along protects the minority: if a majority sells, the minority gets to sell too on the same terms. Both clauses belong in the agreement. Skipping either one creates a hostage situation.
Preemption rights and anti-dilution
Preemption rights give existing shareholders first refusal on new shares. Anti-dilution protects them if a later round prices lower than theirs. These get complex fast, and the specific formula matters more than the clause name.
This is where I made a mistake early on: I accepted a full-ratchet anti-dilution clause without reading the formula. It nearly wiped out a co-founder's stake in a down round. We renegotiated, but it cost us two weeks and a lot of goodwill.
Deadlock resolution and exit terms
What happens when two 50/50 founders disagree and neither backs down? Without a written mechanism, the answer is: the company freezes.
Common fixes include a casting vote (give one person the tiebreaker), mediation before litigation, or a buy-sell provision where one party names a price and the other chooses to buy or sell at that price. Pick one before you need it.
How it relates to your articles and local law
Your shareholder agreement is a private contract. Your articles of incorporation (or equivalent) are a public filing. They overlap, and that overlap is where disputes get expensive.
In most jurisdictions, the articles govern the company's relationship with the outside world, while the agreement governs relations between shareholders. If the two conflict, third parties rely on the articles. Internally, courts often enforce the agreement, but not always, and not everywhere.
Practical rule: update both documents to say the same thing. If the agreement requires board approval for new hires and the articles give the CEO unilateral power, you've created an ambiguity that a lawyer will bill you to resolve later.
| Document | Visibility | Governs | Enforceable against outsiders? |
|---|---|---|---|
| Articles of incorporation | Public filing | Company structure and powers | Yes |
| Shareholder agreement | Private contract | Relations between owners | No, only signatories |
| Bylaws | Internal | Procedural rules, board meetings | Limited |
| Term sheet | Private, non-binding mostly | Deal terms outline | Rarely |
Free templates, and what they quietly leave out
Templates are genuinely useful. They save you from staring at a blank page. They also give founders a false sense of completion.
What a generic template usually won't handle:
- Your specific jurisdiction's mandatory provisions, which override anything you write that contradicts them
- The exact cap table you have (template share counts are placeholders)
- Investor-specific rights negotiated in your term sheet, like information rights or board observer seats
- Founder-specific vesting tied to real milestones rather than generic time periods
- Tax treatment of share transfers, which varies enormously by country and can turn a clean exit into a bill
Use a template as a checklist. Then have someone qualified adapt it. The difference in cost between a template and a reviewed document is a rounding error next to the cost of an unenforceable clause.
Can a shareholder agreement override the articles?
Generally no, at least not against third parties. Courts treat the public filing as authoritative for external matters. Between signatories, the agreement is usually enforceable, but you should align both documents rather than rely on that.
Do I need a lawyer for a startup shareholder agreement?
You can draft the first version yourself. Get a lawyer to review before signing, especially if investors are involved. A review on a seed-stage agreement is far cheaper than litigation, and it's the one document that determines who controls your company.
When should the agreement be signed?
After the term sheet, before the investment money is transferred. For pure founder teams with no outside investment, sign it as soon as equity is issued. Before the first vesting date, ideally.
How long does the whole process take?
Realistically: one week to agree on the main clauses internally, two to three weeks for a first draft and comment rounds, and another week or two for signature and any filings. If investors are involved and their counsel is slow, add three weeks. I've seen a founder-side draft sit in an inbox for eleven days waiting for a single comment.
Budget six weeks. If it closes in four, you got lucky.
The part nobody drafts for
Every clause in that agreement describes a future conflict. Vesting is a clause about someone leaving. Drag-along is a clause about an exit you can't control. Deadlock resolution is a clause about the argument you haven't had yet.
Founders sign these documents in a good mood, usually the best mood they'll ever be in together. That's exactly why you write the rules now, while everyone still likes each other enough to argue constructively.
The companies I've seen survive their founder fights are the ones where the argument happened on paper, in the agreement, months before it happened in the room. The ones that didn't write it down spent their legal budget discovering their own structure in real time.
So before you send that template back with your name filled in, ask yourself one question: if your co-founder walked out tomorrow, would this document tell you what to do, or would you be drafting it from scratch while the money burns?
