Founders' agreement in Canada: the divorce paper before the marriage

What it is — and why it comes before the money

A founders' agreement is the co-founders' constitution. It sits above every later document — incorporation, shareholder agreements, term sheets — because it records the deal as it was understood when everyone was a founder and nobody was a shareholder. The classic mistake is sequencing it after the money: once a valuation exists, every percentage point has a dollar value attached, and negotiations turn adversarial. Written before revenue, the same conversation is a planning exercise among partners who trust each other.

It is also the document that prevents the most common startup dispute in Canada: the departed co-founder holding a large, inactive stake. Industry data (Carta) shows roughly a quarter to a third of two-founder teams separate within five years. Without vesting and buyback clauses, the founder who quit in month six keeps their full equity forever — un-investable dead equity that scares off future investors and poisons the remaining founder's motivation. The agreement turns that tragedy into a procedure.

Equity splits: equal, contribution-based, or something else

There is no formula — the only correct split is one every founder still feels is fair when the company is worth something. The two schools:

• Equal splits (50/50, 33/33/33): the most common choice, and Y Combinator's recommendation. They avoid resentment, keep everyone equally committed, and remove the temptation to keep score. Best when founders are genuinely comparable in role, experience, and commitment.
• Contribution-based splits: appropriate when contributions are clearly unequal — one founder full-time with no salary, the other nights-and-weekends with a day job. Score factors like role centrality, relevant experience, full-time commitment, cash invested, and risk taken (what each person is giving up).

Three rules apply to either approach. First, weight the future, not the past: the idea is worth very little; five years of execution is worth everything, so split on who will do what going forward. Second, keep cash separate from equity — a founder who puts in money should get a note or SAFE on the same terms as an outside investor, not extra percentage points. Third, reserve an option pool of roughly 10–15% for future hires before you finalize the numbers — seed investors will require one carved out of the founders' slice (pre-money dilution), so a 60/40 split with a 15% pool is really 51/34 fully diluted. And note the distinction: if you never incorporate and operate together for profit, Ontario's Partnerships Act section 24 imposes an equal split of profits and losses by default, whatever anyone remembers saying.

Vesting: the clause that saves the company

Vesting means founders earn their equity over time instead of owning it outright on day one. The market standard — recommended by Y Combinator and normal across Canadian startups — is four years with a one-year cliff:

• Years 1: nothing vests. If a founder leaves before month 12, they forfeit everything — the company repurchases all their shares at the original purchase price (typically a nominal amount like $0.0001 per share).
• Month 12: about 25% of the grant vests at once.
• Months 13–48: the remainder vests monthly (some plans use quarterly).

Every founder vests — including the CEO. Vesting protects all sides: the remaining co-founders from dead equity, future employees from a squeezed option pool, investors from an inactive co-founder on the cap table, and the departing founder, whose clean break is possible precisely because the economics were predetermined. If the company is acquired before vesting completes, decide in advance whether any unvested equity accelerates — 'double-trigger' acceleration (acquisition plus termination) is the investor-friendly norm. Document the schedule in the agreement and implement it accurately in your cap table records.

Roles and decision-making: who is the CEO?

Ambiguity about authority is the second most common co-founder conflict. Name the roles — CEO, CTO, and so on — and describe what each person is actually responsible for day to day. Then separate decision tiers: ordinary operational decisions to the relevant founder, important decisions (hiring, spending above a threshold, taking on debt) to a founder majority, and fundamental decisions (issuing new shares, selling the company, changing the business, admitting a new co-founder) to unanimity.

For 50/50 partnerships, add a deadlock breaker now, because the majority-vote rule cannot save a tied partnership. The classic is the shotgun clause: one founder names a price for the whole company, and the other must buy at that price or sell at it — it forces fair pricing because the partner naming the price doesn't know which side of the trade they'll end up on. Also name a single tie-breaker for the mundane but constant question: who has the final word when the founders disagree about a product decision? 'The CEO decides' is a complete answer, and writing it down is the point.

IP assignment: everything created belongs to the company

This clause exists because of a Canadian legal trap. Under section 13(1) of the Copyright Act, the author of a work is its first owner — copyright belongs to the creator, not the company. Section 13(3) creates an employment exception (works created in the course of employment under a contract of service belong to the employer), but pre-incorporation founders are not employees, so that exception does not protect you: the code, designs, and trademarks your co-founder created before incorporation belong to them personally. The Ontario Court of Appeal has further narrowed the exception to work within the employee's actual assigned responsibilities, so even post-hire assumptions are riskier than they look.

The agreement must contain an express assignment: each founder assigns to the company all intellectual property they create for the business — code, designs, inventions, trademarks, business plans — both past (pre-signing contributions) and future, with a promise to sign any further documents needed to perfect the assignment. Include a moral rights waiver (moral rights can't be assigned under Canadian law, only waived in writing) and a confidentiality clause covering the business's trade secrets. If contractors or freelancers contributed anything, chase their written assignments too — an independent contractor's work never falls under the employment exception.

When a founder leaves: good leaver, bad leaver, buyback

Plan the exit mechanics while the exit is hypothetical. The standard structure:

• Vesting stops on the departure date; unvested shares are repurchased at the original nominal price — automatic, no negotiation.
• Vested shares: the company or remaining founders get a buyback right. For a 'good leaver' (resignation, disability, death) the price is fair market value — set a valuation method in advance (independent appraisal, an earnings multiple, or a formula). For a 'bad leaver' (fraud, breach of the agreement, joining a direct competitor) the price can be the lower of cost or a discounted value.
• Payment terms: lump sum or instalments with interest, and how the buyback is funded if the company is cash-poor.

Add three adjacent clauses. First, a non-solicitation clause: the departing founder can't recruit your team or poach your customers for a defined period — non-solicitation survives Ontario's non-compete ban and remains enforceable when reasonable (see below). Second, a cooperation clause: the leaver signs documents needed for IP transfers and investor paperwork. Third, death and disability planning: life-insurance-funded buyouts are the classic tool, so the survivors aren't negotiating with a grieving family over an illiquid stake.

Non-compete and non-solicitation: Ontario's ban changes the calculus

Ontario bans most post-employment non-competes. Since October 25, 2021, the Working for Workers Act, 2021 (Bill 27) prohibits non-compete clauses in employment contracts and other agreements with employees — the courts confirm the ban does not void agreements entered before that date, and there are only two exceptions: 'executives' (CEO, president, COO, CFO, CIO and other chief executive positions) and the sale of a business where the seller becomes the buyer's employee. A founders' agreement with a sweeping Ontario non-compete is likely to be struck down, and including a dead clause weakens the agreement's credibility.

What still works: non-solicitation (no recruiting the team, no poaching customers) and confidentiality — both unaffected by the ban and enforceable when reasonable in scope and duration. In other provinces, non-competes remain enforceable under common-law reasonableness, but keep any restraint narrow: a short duration, a limited geography, and a defined competitive activity. The practical advice is the same everywhere: protect the company with a tight non-solicitation and a strong confidentiality clause rather than a non-compete a judge will refuse to enforce.

Dispute resolution: a ladder, not a cliff

Write the disagreement procedure while the founders still agree. The standard ladder: (1) mandatory good-faith negotiation within a fixed period; (2) mediation with a neutral mediator, costs shared; (3) binding arbitration or court as the final forum — pick one, name the seat, and state who pays. Arbitration is private and faster; courts create precedent and enforce more simply. Either beats discovering your dispute process at 2 a.m. during a founder fight.

For the 50/50 team, the shotgun clause from the decision-making section is also the dispute endgame: it converts an unresolvable disagreement into a priced buy-sell. One founder names a price for the whole business; the other must buy or sell at it. It is blunt, but it forces both sides to name a fair number — the alternative is a deadlock that kills the company while the lawyers get rich.

The contrast that should scare you: what applies without an agreement

If you never sign anything and never incorporate, you may not have a startup at all in the eyes of the law — you may have a partnership. Two or more people carrying on business in common with a view to profit are partners by default, and Ontario's Partnerships Act section 24 then supplies the rules: profits and losses shared equally regardless of contribution, no partner entitled to a salary or interest on capital, every partner able to take part in management, and — the sharpest one — any partner able to dissolve the partnership at will by giving notice. A handshake deal where one founder contributed $40,000 and the other contributed weekends becomes a 50/50 split with a free exit, enforced by statute.

If you did incorporate, the corporate defaults apply instead: whatever share classes were issued (usually equal, since nobody documented otherwise), no vesting, no buyback rights, and no way to remove a co-founder who stops contributing short of buying them out at whatever price they demand. Either way, the absence of a founders' agreement doesn't mean the absence of rules — it means the rules were written by legislators for the generic case, not by you for your company.

Mistakes that kill startups before revenue

Almost every co-founder blowup traces back to one of these avoidable failures:

• No written agreement — the Partnerships Act equal-split defaults or undocumented equal shares govern by surprise
• No vesting — the founder who quit in month six keeps their full stake as dead equity
• Splitting on past contribution instead of future work — the idea person takes 50% and then coasts
• Cash contributions counted as equity instead of a loan or SAFE — mixing investment and ownership
• No option pool reserved — founders get diluted harder than expected at the first raise
• IP never assigned — pre-incorporation code legally belongs to the author, not the company
• No buyback mechanics — the departing founder's shares become a negotiation with no anchor
• An Ontario non-compete that Bill 27 renders void — while a perfectly enforceable non-solicitation was never drafted
• No roles defined — two CEOs, or nobody willing to be the one who decides
• No dispute ladder — the first serious disagreement goes straight to lawyers