Signed a two-year non-compete? What it can stop, and what the fight actually costs

The clause in Shafron v KRG Insurance Brokers restricted a departing broker from competing within the "Metropolitan City of Vancouver," a place that has never existed as a defined municipal boundary. In 2009 the Supreme Court of Canada held the phrase ambiguous, declined to redraw it into something workable, and struck the covenant down. That is the outcome most people want to hear about. What gets less attention is that the dispute ran through a trial and two levels of appeal before anyone reached that conclusion, and the broker who eventually won had been living under the clause for years while it was argued.

The employer starts behind, and that matters more than the wording

A restrictive covenant in an employment contract is a restraint of trade, and Canadian courts treat restraints of trade as presumptively void. The employer has to justify the clause, not the other way around. Justification means showing a proprietary interest genuinely worth protecting, usually trade connections or confidential information, and then showing that the restriction is no wider than necessary in three dimensions: what activity it bars, where it bars it, and for how long. Ambiguity in any of the three tends to be fatal, because a court cannot assess the reasonableness of a term it cannot pin down. The drafting burden sits entirely with the party that wrote the document.

That presumption explains why so many confident-sounding clauses collapse. A two-year bar on working "in any capacity for any competitor in North America" reads as though it has teeth, but it sweeps in jobs that touch none of the employer's customers and territory where the employer has never done business. Courts do not sever the excess and enforce the remainder in employment cases. Shafron confirmed that notional severance, reading down an overbroad term to a reasonable one, is not available where it would rewrite the bargain. The clause stands or falls as written, which is good news for the employee who reads carefully.

What a non-compete is usually allowed to stop

Less than employees assume. The governing idea since Elsley v J.G. Collins Insurance Agencies in 1978 is that a non-solicitation clause will normally protect the employer adequately, so a full non-compete is only justified in exceptional circumstances. The Ontario Court of Appeal applied that squarely in Lyons v Multari, refusing to enforce a short, geographically modest restriction on an oral surgeon because the employer had not shown why barring solicitation alone was insufficient. The exceptional cases tend to involve someone who effectively was the business to its customers, or a key person whose departure hands a competitor the relationship wholesale.

Context shifts the analysis considerably. Where the covenant comes attached to the sale of a business, as in Payette v Guay in 2013, courts apply a more deferential standard, because the buyer paid for goodwill and the seller took the money. Ontario went further on the employment side, amending its Employment Standards Act in October 2021 to prohibit non-compete agreements for most employees outright, with carve-outs for executives and for sale-of-business situations. In the United States the Federal Trade Commission has taken up the same question from the competition side. The practical point is that which document you signed, and which province and role you signed it in, decides the analysis before anyone reads the wording.

What the decision costs when you are the one deciding

An unenforceable clause is not a free clause. The employer's move is rarely a full trial; it is a letter to you and a letter to your prospective employer, sometimes followed by a motion for an interlocutory injunction heard within weeks. That motion is where the money goes. Preparing affidavit evidence, responding to a records demand, and arguing irreparable harm on a compressed timetable is expensive work, and the costs land before any court has ruled on whether the covenant was valid at all. Retainers for that kind of motion practice are quoted in the thousands, and the sum depends far more on how fast the other side moves than on how weak its clause is.

The other cost is the start date. A prospective employer that receives a demand letter will often pause the offer rather than take on litigation risk, and the pause can run a month or longer while counsel exchange positions. Some employers will indemnify a hire against the fight; many will not, and some will simply move to the next candidate. Getting an opinion letter before you sign the new offer, so the hiring manager sees a reasoned analysis rather than a threat, is the cheapest intervention available. An hour with a non compete lawyer buys a read of the actual clause, an assessment of which province's law governs, and a view on whether the employer would realistically fund an injunction over you.

Reading your own clause before anyone else does

Find the four things that decide the case: the activity restricted, the geography, the duration, and the definition of the customers or employees you cannot approach. Check whether the customer list is limited to people you actually dealt with, and within what recent window, because a bar covering every client of the firm is a common drafting failure. Note the governing law clause and your job title on the date of signing, since both can be dispositive in Ontario. Then look for the non-solicitation and confidentiality terms sitting alongside the non-compete, because those survive independently and are far more likely to bind you.

The arithmetic is usually straightforward once the clause is on the table. A covenant that is plainly overbroad, in a province that restricts them, held by an employer with no history of enforcing, is a different proposition from a tightly drawn twelve-month customer restriction signed by a regional sales director. Both may be survivable. Knowing which one you are holding, before you resign, is what turns a two-year restriction into a scheduling question rather than a year of uncertainty.