SEO for Canadian Businesses

The bilingual technical setup, and how it usually breaks

The foundation is unglamorous: each language version has to live on its own URL that a crawler can reach and index. That means a path or subdomain structure where the English and French pages are distinct addresses, not a single URL that swaps content based on a cookie, a session variable or a JavaScript toggle. When the language lives in the session, only one version is ever crawled, and the other might as well not exist.

On top of that, hreflang annotations link the equivalent pages. Each page in the set declares all the versions including itself, the references are reciprocal, and there is an x-default for users whose language does not match either. The annotations are page-level and specific — a site-wide block pointing at the homepage does nothing for the product pages, which is a shortcut we see often. Language codes have to be right too: fr-CA and en-CA rather than a bare fr that invites the wrong regional match.

And then the defect that undoes all of it. Somebody, wanting to avoid duplicate content, sets the canonical tag on the French page to the English URL. That is not a duplicate-content fix; it is an instruction to consolidate the French page into the English one, and it takes the entire French site out of the index. It is startlingly common, it is invisible from the front end, and it is the first thing we check on any bilingual Canadian site that reports poor French performance. Each page canonicalises to itself; hreflang handles the relationship between languages.

Competing against American sites on Canadian queries

Canadian businesses face a competitive dynamic that domestic-market businesses elsewhere do not. For a great many queries the strongest pages in the world on that topic are American, with more links, more history and more content, and they surface for Canadian users because relevance beats geography for informational queries. Adding "Canada" to your title tag does not displace them, and neither does a domestic domain by itself.

What does displace them is information that is only true in Canada and that the American page cannot contain without being wrong. Prices in Canadian dollars. Shipping and delivery realities within the country. Provincial regulation, licensing and tax treatment. Canadian standards and certifications. The distinction between what applies federally and what varies by province. A page that answers a Canadian's actual question — which for most commercial queries is a question about their own jurisdiction — beats a stronger page that answers a different country's version of it.

This is also the honest test of whether a location page deserves to exist. Provincial and city pages that differ only in the place name are doorway pages, they are described as such in search engine spam policy, and they attract manual action rather than traffic. A page for a province earns its place if it says something true about operating there that the national page cannot — the regulator, the tax treatment, the local timing, the specific process. If you cannot write that, the page should not exist, and we would rather build eight substantial ones than sixty thin ones.

Getting cited by AI assistants, in both languages

A growing share of commercial questions is answered by an assistant that summarises rather than by a page a user clicks. Being the source those answers draw from is a different objective from ranking, though it depends on much of the same work, and in Canada it carries a bilingual dimension that most sites ignore entirely. An assistant answering in French draws predominantly on French sources — so a business with an English-only site is absent from the AI answers for half its market, regardless of how strong its English rankings are.

What makes a page usable as a source is fairly consistent. A direct, self-contained answer near the top, phrased so it can be quoted without the surrounding paragraph. Specific facts with attribution rather than general claims. Clear structure with headings that match the questions people actually ask. Structured data that states unambiguously what the page is, who published it and what it is about. And an entity that is coherent across the web — the same business name, the same address, the same description in your markup, your profiles and your citations — so the model can tell that all these mentions are the same organisation.

The French version needs the same treatment independently, which means writing French answer content rather than translating the English answers, marking the language correctly in the markup, and making sure the French page carries its own structured data. We build to this alongside conventional SEO because the underlying discipline — be specific, be well structured, be quotable, be consistent about who you are — improves ordinary rankings too. What we will not do is promise placement in AI answers, because no one controls that, and any agency selling guaranteed AI citation is selling something they cannot deliver.

Questions worth asking any vendor

These apply to us as much as to anyone else bidding for your work.

• Check the canonical tag on a French page. If it points at the English URL, your French site is being consolidated away and no amount of content will fix it.
• Check whether switching language changes the URL. If it does not, only one language is being crawled and the other is invisible to search.
• Ask whether the French keyword list was researched or translated. A translated list targets terms that may have no local search demand at all.
• Ask what a proposed provincial or city page will say that the national page cannot. If there is no answer, it is a doorway page and it is a liability.
• Be sceptical of guaranteed rankings or guaranteed AI citations. Nobody controls either, and the guarantee tells you what kind of work is being sold.