We crossed from Belgium into France by road recently and never noticed the language change. Not because the languages are similar — they are not — but because English kept working on both sides of the line. The border was administrative. The floor underneath it did not move.
That experience is the reason European companies file Japan in the wrong category, and it is worth explaining carefully, because the usual telling of this story is both condescending and wrong.
European companies are not lazy about language
The standard version of this argument is that English-speaking Europeans assume English works everywhere and cannot be bothered to localise. We checked, and that is not what the evidence shows.
On 7 August 2026 we requested the home page of seventeen Finnish companies with a browser asking for English (Accept-Language: en-US) and recorded which language each site actually served.
| Group | Language served | Result |
|---|---|---|
| Domestic-facing Kesko, S-Etukortti, Elisa, OP, Posti, VR, Yle, Alko | Finnish | 8 of 8. None switched to English when asked. |
| Global industrial Nokia, KONE, Wärtsilä, Neste, Fortum | English | 5 of 5 |
| Technology Wolt, Supercell, Reaktor, Futurice | English | 4 of 4 |
The split is clean, and it is not about nationality. It is about audience. Finnish companies serving Finns publish in Finnish. Finnish companies serving the world publish in English. They already make exactly the distinction that Japan requires, and they make it correctly, every day, at home.
So the failure is not ignorance about localisation. It is a filing error.
Japan gets filed under “world”, not under “market”
A European company has two mental buckets. One is home, which gets the local language and everything that goes with it. The other is international, which gets English and has been quietly reliable for decades — across the EU, across the Nordics, across most of the places these companies sell.
Japan arrives and gets dropped into the second bucket, because that is where every previous new market went and nothing bad happened. The English site carried Belgium. It carried France. It will carry Japan.
It does not, and the reason is more specific than “Japanese people don’t speak English”.
In Japan, English is not unintelligible. It is uncitable.
This is the part that changes the decision, so it is worth being precise.
Plenty of Japanese B2B buyers read English perfectly well. That is not the constraint. The constraint is what happens after they read it.
Japanese B2B purchases are written up internally before a vendor is ever contacted. A staff member assembles a proposal, circulates it, and it is reviewed by people who did not do the research. That document is in Japanese. Every source quoted in it has to survive being pasted into a Japanese document and read by a manager who will not click through.
An English page cannot do that job. It can be understood, and then it is set aside in favour of a Japanese page that can be quoted directly. The English page loses not because it failed to communicate, but because it could not be used. We wrote about the mechanics of this in BtoB vs BtoC SEO in Japan.
The same logic now applies to machines. Japanese AI assistants answering a Japanese question cite Japanese sources. An English-only presence is invisible to that entire layer, regardless of how good the English is.
What an English-only Japan page actually does
- It ranks for almost nothing. Japanese buyers search in Japanese, including buyers who could have searched in English.
- It converts the traffic it does get at a lower rate, because the reader has to do the translation work before they can act internally.
- It signals something you did not intend. A company with no Japanese-language presence reads as not seriously present, which is a problem in a market where buyers check company details carefully.
- It produces a misleading result. Low traffic gets read as “no demand in Japan” when the correct reading is “no visibility in the language demand is expressed in”.
That last one is the expensive failure. Companies cancel Japan on the strength of a number their own setup guaranteed.
A note on accommodation
Something we noticed on that trip: English worked in both countries, but Belgians met us further along the road. Same language, different amount of effort extended toward the visitor.
Japan extends a great deal of effort toward visitors, and almost none of it in writing. The accommodation is in person, in service, in the room. It does not appear on the web, where a Japanese buyer expects to be met in Japanese and does not consider this a concession.
So the reciprocity has to be returned somewhere else. Not by speaking Japanese in meetings — an interpreter handles that — but by publishing in Japanese, sourced and dated, so the person building the internal case can use your material without doing extra work. That is what accommodation looks like in this market.
What to do instead
- Refile Japan. Move it out of the international bucket and into the one where your own home market sits. That single decision determines the rest.
- Japanese on a subdirectory of your main domain, in most cases — you keep the authority you have already built. A separate
.jpstarts from zero. - Do not translate your English pages. Research in Japanese first; the questions differ before the words do. See Japanese keyword research.
- Start with ten to fifteen pages covering distinct intents, not a full mirror of your English site. Depth per page is what ranks here.
- Record a baseline first, so that two quarters from now you are reading a real result rather than the artefact of an English-only setup.
Common questions
Our Japanese partners all speak English. Isn’t that enough?
For the relationship, often yes. For search and for the internal document that decides the purchase, no. Those are different audiences that happen to be the same people.
We run six European languages already. Isn’t Japanese just a seventh?
Operationally it looks like one more locale in the CMS. Editorially it is not. The other six can often share a structure and an argument; Japanese usually needs a different structure because the questions and the accepted forms of proof differ.
Can we start with machine translation and improve later?
You can, and it usually costs more than starting properly. Pages that answer the wrong question rank poorly, and replacing them later means re-earning the ground twice.
How much does the Japanese-language section need to cover?
Less than your English site and to a higher standard. Ten to fifteen pages done properly beats a hundred translated ones, and it is cheaper.
Working on Japan?
We are a small Japanese SEO practice working with companies entering the Japanese market. We spend time in the markets our clients come from, because the assumptions people arrive with are as much a part of the work as the search data. Everything is produced natively in Japanese. The full sequence is in our Japan market-entry SEO guide. More about how we work.
Method
The language table above is our own measurement, taken on 7 August 2026. Each home page was requested once over HTTPS with a desktop Chrome user agent and Accept-Language: en-US,en;q=0.9, following redirects, and the served <html lang> attribute was recorded. Seventeen companies, grouped by who they primarily sell to. It is a snapshot, not a survey — but it is dated, and it is repeatable.
Last reviewed: August 2026.
