Every month an earn-out runs is a month the buyer, not the seller, is making the decisions that determine whether the seller gets paid in full. Length is not a neutral drafting variable — it is the single biggest lever on how much risk a seller is carrying.
Key takeaways
The risk in a long earn-out is not diffuse — it compounds through a specific mechanism. The buyer changes pricing, product lines, or service offerings, or diverts customers, referrals, or resources to another part of its business: each of those is a discrete decision the buyer can make at any point during the earn-out window, and the seller typically has no vote on any of them. A one-year earn-out gives the buyer roughly twelve such opportunities to affect the number before the measurement closes. A three-year earn-out gives it thirty-six. The dispute mechanics that follow when one of those decisions is contested are covered in dispute mechanisms for a contested earn-out payment.
Canadian law requires the buyer's discretion to be exercised honestly and in good faith, which constrains the worst version of this problem but does not eliminate it — and proving bad faith is far harder and far more expensive than drafting around it, so length itself is still doing real work even where the legal backstop holds.
Separately from the commercial risk, ITA s. 40(1)(a)(iii) caps how long a vendor can spread recognition of a deferred gain through the ordinary capital gains reserve: the reserve for proceeds not yet due is the lesser of a reasonable amount and 1/5 of the gain multiplied by the number of years remaining in a five-year window from the year of disposition. In plain terms, the ordinary reserve mechanism tops out at five years. A ten-year reserve exists, but only for specific transfers — to a child, to certain family farm or fishing property, to a qualified small business corporation share on a qualifying intergenerational transfer, or to an employee ownership trust — not for an ordinary arm's-length sale to an unrelated buyer.
An earn-out that runs longer than the reserve can shelter does not become illegal or unworkable — it simply means the vendor may need to recognize more of the gain for tax purposes before the corresponding cash has actually arrived, which is a real cash-flow consideration worth modelling before agreeing to a long earn-out window. How that interacts with the vendor's overall tax position, including the lifetime capital gains exemption, is covered in tax treatment of earn-out proceeds in Canada.
No source in this build publishes a Canadian benchmark for earn-out duration specifically. The closest comparable published figure is deavo's vendor take-back term band: typically 10–20% of the purchase price over 3–5 years. A VTB is a fixed-schedule note, not a performance-contingent earn-out, so the figure is useful only as rough market colour on how long Canadian small-business sellers are generally willing to stay financially tied to a deal they no longer control — not as a stated norm for earn-out length itself.
Absent a published benchmark, the honest guidance is structural rather than numeric: match the earn-out period to how long it genuinely takes the disputed metric to prove itself, and no longer. A revenue ramp from a specific new contract might reasonably need eighteen months to show through; a general “prove the business holds up” earn-out with no defined driver is the version most likely to run long simply because nobody defined when it should end.
The standard fix is not to avoid earn-outs but to shorten and segment them: rolling twelve-month measurement periods with their own payout inside a longer overall term, tiered thresholds that pay out partial amounts as interim targets are hit, and the operating covenants discussed in protecting a vendor during the earn-out period — requiring the buyer to operate the business in good faith, in the ordinary course, or without taking actions specifically intended to reduce the earn-out payment. Each of these narrows the window in which any one buyer decision can wipe out an entire year of earn-out, rather than the whole deferred balance.
Two sellers negotiate structurally identical $400,000 earn-outs on otherwise similar deals, tied to the same EBITDA metric. Seller A's earn-out runs one year, measured once. Seller B's runs three years, measured cumulatively at the end.
Eight months in, both buyers make a pricing change that temporarily depresses margin while a new product line ramps up. For Seller A, the measurement window closes before the ramp completes — the earn-out pays out short, and the dispute is over almost as soon as it starts. For Seller B, the same pricing decision is now one of what could be a dozen similar decisions still to come over the following twenty-eight months, any one of which could be contested, and none of which the seller has any vote over.
Seller B's total potential payout is not necessarily larger than Seller A's for taking the longer window — nothing about length alone increases the ceiling on what the metric can pay. What increased is the number of discrete opportunities for a buyer decision, made in good faith or not, to affect the outcome before the seller is paid.
No statute caps earn-out length directly. The practical ceiling comes from the tax side: the ordinary capital gains reserve under ITA s. 40 tops out at five years, which is a strong incentive not to structure an earn-out that runs meaningfully longer without modelling the tax consequence.
No. Length changes how much control-related risk the seller carries and how many opportunities exist for a disputed buyer decision — it does not, on its own, change the ceiling the underlying metric can pay out.
The ordinary reserve under ITA s. 40(1)(a)(iii) cannot shelter deferred recognition past that window in the general case, so a vendor may need to recognize gain on amounts not yet received in cash. Confirm the specific year-by-year effect with a tax professional before agreeing to a term that long.
Not automatically, but it is almost always lower-risk for the same reason a shorter loan term carries less exposure to rate movement: fewer periods means fewer discrete opportunities for a buyer decision to affect the outcome. The trade-off is that some metrics genuinely need more time to prove themselves — a shorter window can end before the thing being measured has actually happened, which produces its own kind of unfair result in the other direction.
A 30-minute call is enough to tell you whether AI pays for itself here.
Nobody publishes Canadian transaction data, so every valuation in this country quotes an American benchmark. We are building the Canadian one — multiples, asking-to-sale spreads and days on market, by sector and by city. Leave an email and you will see it first.
No pitch, no listings. One email when the first report lands.
