Extend or replace the Canadian question set when you add a second market, and the four cases that decide it

A Canadian company adding a second market faces a narrower decision than it looks. You are not starting a question set from scratch and you are not translating the Canadian one either. You are deciding, question by question, which of three categories each one falls into: travels unchanged with the country name swapped, travels with the figure or the rule replaced, or does not travel at all and has to be written for the new market. A Canadian seller has an advantage here, because if you sell into Quebec you already run two sets and already know which questions survived a language change. Apply the same sorting to a country change and the work halves.

The three categories, and how to sort a question into one

Category	What it looks like	What to change
Travels unchanged	A question about what something is or how a mechanism works	The country name in the text, nothing else
Travels with a substitution	A question about cost, tax, consent or a regulator	The currency, the tax, the statute and the regulator named
Does not travel	A question about a local platform, a provincial rule or a city market	Write it new. The Canadian version measures nothing in the new market

Sort your twelve or twenty questions into those three buckets before you spend anything, and the size of the third bucket is the real cost of the market. A market where ten of twelve questions travel with a substitution is cheap. A market where seven do not travel at all is a second programme.

The four cases, and the decision in each

• Same language, similar regulation. Extend. Swap the country name and the currency, replace the statute references, keep the structure. The source list still has to be built fresh from the new reading.
• Same language, different regulation. Extend the structure and replace every question that touches consent, claims or tax. Expect about a third of the set to need rewriting.
• Different language, similar regulation. Replace. You already know this from Quebec: a question written in another language is a different question with a different source pool, and translating it measures the translation.
• Different language and different regulation. Replace entirely, and treat it as a new programme with its own baseline, its own source list and its own reviewer.

Why the source list never travels, measured

On 22 October 2026 we ran the same 198 question bank for Canada and for the United States, on the same day, through the same data provider, with each country named inside the text of every question because the provider accepts gl, hl and location parameters and ignores all three. Batch 33 names us in the question and measures a ceiling, excluded from both.

Canada gave 64 blind readings with 2,992 citation URLs across 974 hosts. The United States gave 64 blind readings with 2,866 across 1,143. The Canadian list leaned on torontoseo.com in 13 readings, canadacreate.com in 11, canada.ca in 10, wisemedia.io in 9, competition-bureau.canada.ca in 9, ised-isde.canada.ca in 9, www150.statcan.gc.ca in 8 and ottawaseo.com in 7. The American list leaned on searchengineland.com in 13, clutch.co in 11, semrush.com in 10, ftc.gov in 9, blog.hubspot.com in 9 and consumer.ftc.gov in 6. Google's documentation properties appeared at the top of both. Beyond those, the two lists shared almost nothing.

So a question can travel and its answer's sources cannot. Any plan that extends the question set and reuses the target list is working against last market's evidence. One host in the Canadian raw data, ccn.web.tr, is the data provider's own fetch infrastructure rather than a publisher, excluded from both analyses and never merged into another host. These are two separate measurements with their own denominators, reported side by side and never combined.

What the sorting exercise costs and saves

The sorting itself is two hours for a twenty question set. What it saves is the second market's wasted first quarter, which is what happens when a Canadian set is lifted wholesale: the cost and tax questions return answers about Canadian tax, the local platform questions return nothing, and the reading looks like a visibility problem when it is a question set problem. If you sell in Quebec you have already done this once, so start by pulling out the questions that changed between French and English. Those are the ones most likely to need rewriting for a new country too.

The arithmetic, after the sorting

Tracked buyer questions, times engines read, times refresh frequency, times markets, times language sets. A bilingual Canadian company adding one market is already at four question sets rather than two. Take that number into the pricing conversation, ask which unit the vendor's price is per, a tracked question, an engine, a recorded answer or a domain, and recompute your total. Ask about the fourth set rather than the first, because the per set line is where the bill grows.

What to do before the second market goes live

• Sort the existing set into the three buckets. Two hours, and it is the whole decision.
• Rewrite the third bucket for the new market rather than translating it.
• Build the new market's source target list from its own first reading, never from the Canadian one.
• Replace every currency, tax and statute reference in the second bucket, and have somebody in that market read them.
• Keep the method identical so the two markets' readings stay comparable even though their sources do not overlap.

The wider go to market question, payments, support, references and credibility abroad, sits upstream of this and is costed in what each extra market costs to be named in. For the per set cost arithmetic in full, including the meter question to put to a vendor, see why the unit count multiplies before the dollar figure does.

Questions about extending a Canadian question set abroad

Can I just add the country name to my Canadian questions?

For the first bucket, yes. For anything touching cost, tax, consent or a regulator, no, because those questions will return answers about Canada's rules with another country's name attached.

How much of a set usually travels?

It depends entirely on how many of your questions are about local platforms and local rules. Sort them and count rather than estimating, because the count is the cost.

Does the Quebec experience really help?

It is the closest thing to a rehearsal. The questions that changed between French and English are the ones most likely to need rewriting for a new country.

Should the new market reuse my source targets?

No. Our Canadian and American citation lists shared almost nothing beyond Google's documentation properties, so a reused target list is aimed at the wrong market.

How do I keep two markets comparable?

Keep the method, the question count, the engines and the frequency identical, with separate denominators. The sources will differ and the measurement will still compare.

Reading each market on its own denominator

The one thing that makes two markets comparable is an identical method with separate counts, and the one thing that makes them incomparable is a blended score. AI Knows Us reads ChatGPT, Gemini, Perplexity, Claude and Google AI Overviews per market and per engine with the market named inside the question, states each denominator, records every cited host, and keeps the raw answer text. Outside the United States and the United Kingdom Growth is 199 dollars a month with 100 buyer questions watched; in the United States and the United Kingdom it is 299 dollars. Both read on aiknowsus.com/pricing and aiknowsus.com/llms.txt on 22 October 2026, and the question count is the line to size against your number of sets.