Japan Is Not a Localization Problem. I Learned More About Design Here Than From Any Job

I have lived in Tokyo long enough to know I still cannot fully explain it to people who have not been here. That is not modesty. It is the actual state of things after years of trying, and it turns out to be a product problem as much as a personal one.

The clearest version of it I can give you is a receipt. I bought something small at a department store, and the person serving me made a mistake with the wrapping. She noticed before I did. What followed was not an apology script. It was a complete repair of the situation, executed in under a minute, with no drama and no discount, and I left feeling like the mistake had been the interesting part of the transaction rather than the damage.

I have never seen software do that. Not once.

Three countries, three lessons I did not ask for

I am Costa Rican. Latin America taught me that creativity is mostly resourcefulness, and that emotion is a legitimate design material rather than a decoration you add at the end. You learn to make something good with what is actually in the room, and you learn that people decide with their feelings and explain with their reasons.

China taught me speed. Shipping without romance. Version one goes out, the market answers, version two is next week, and nobody holds a memorial service for the idea that did not work. I found it brutal at first and then I found it honest. A lot of what passes for care in design is really just a fear of being told you were wrong.

Tokyo taught me the thing neither of the other two could. Craft as a baseline instead of a differentiator. Consensus as a real constraint rather than an obstacle to route around. Proof before personality. And restraint, which is the hardest of the four, because restraint looks like nothing from the outside and feels like nothing on a slide.

The combination is the point. Not one of those places is better at design than the others, and I am suspicious of anyone who ranks them. What I have is three incompatible instincts arguing in my head every time I open a file, and the argument is the useful part.

Japan is not a localization problem

Here is where this stops being a personal essay and starts costing companies money.

A pattern I have watched several times, with the details changed enough to protect the people involved: a good product, working well in its home market, decides to enter Japan. The team does what every playbook says. They translate the interface. They hire a native speaker to check the copy. They add a yen price. They launch, they wait, and the numbers are so far below the model that people start blaming the channel.

The Japanese was fine. The translation was not the problem. The problem was everything underneath it that nobody thought to question, because it did not look like a design decision. It looked like the product.

The signup flow assumed one person could decide. In a lot of Japanese companies, the person filling in your form is not buying anything. They are collecting enough material to bring to four other people, at least one of whom will ask a question your marketing site does not answer. If your funnel is built around getting one hero to click, you have designed for a decision maker who is not in the room.

The homepage led with personality and put proof at the bottom. That ordering works in markets where a confident voice reads as competence. Here, an unbacked claim near the top reads as a risk, and the reader starts looking for the thing that would make it safe to take you seriously: who else uses this, how long you have existed, what happens when it breaks, whether there is a phone number.

The information density was wrong in both directions. Pages that felt clean to the team felt thin here. And the support documentation, which the team considered a back office concern, was the single artifact a serious buyer was most likely to read before making a decision.

None of that is language. All of it is structure. Translation moves words between languages and leaves the assumptions in place, and the assumptions are what failed.

I have written before about how writing is the weakest English skill in the country that is exporting the software, and about how forty two million visitors a year still reach for their own phone when the interface in front of them gives up. Same underlying failure, different scale.

Restraint is a product advantage, not an aesthetic

The work I am proudest of was made under accessibility constraints, and it won an Apple Design Award. I bring that up here for one reason: designing for people the happy path forgets is the same discipline as designing for a market that will not tell you it is unimpressed.

Both force you to remove things. Both punish the flourish that only works if the user is already convinced. Both make you find out whether the product still stands up when you take the personality away.

Tokyo has almost no tolerance for a sloppy default, and it expresses that tolerance through silence. Nobody writes you an angry review. The number just does not move. That is an unforgiving teacher and it made me much better, because it removed the option of arguing with the feedback.

This matters more now than it did five years ago. AI products over promise by construction. They are demonstrated at their best moment and used at their worst one, and the gap between the demo and the Tuesday is exactly the gap Japanese users are most sensitive to. AI has made the large majority of Japanese companies faster and almost none of them richer, and I do not think that is a technology gap. I think it is a gap in what was designed.

Five Proofs

The portable version of all of this is a checklist I use before saying a product is ready for this market. I call it Five Proofs, because every item is something the user is trying to prove to themselves before they commit.

1. Design for consensus, not for a click hero

Assume the person in front of your interface has to convince other people. Give them the material to do it: a page they can send, a document they can attach, a price they can defend, an answer to the question their finance team will ask. Most funnels are built for a decision that happens alone. That decision often does not exist here.

2. Put proof before personality

Track record, named customers, dates, numbers, an address, a real company entity. Put it high, before the tone of voice does any work. Personality is not wasted here, but it is spent, and you have to earn the budget first.

3. Make refusal and recovery easy

Service culture here is largely a recovery culture. What happens when something goes wrong is more revealing than what happens when it goes right. So design the cancellation, the refund, the failed payment, the wrong order, the answer your model cannot give. If the only path out of your product is embarrassing, people notice before they enter it.

4. Never ship a translation of your structure

Rebuild the information architecture for how the page will actually be read. Order, density, where proof sits, how long the page is, what a section is called. If the Japanese page is the English page with Japanese in it, you have shipped a layout that was optimised for a different reader and hoped.

5. Measure seriousness the way users here measure it

Forms, support, documentation, and the speed and shape of your first reply. These are not hygiene. They are the test. A company that answers a vague enquiry with a careful, specific response has communicated more about its reliability than any amount of homepage copy. Local friction is design research, not a ticket in the localization queue.

What this taught me about building anywhere

None of the five is really about Japan.

Every one of them is a general improvement that this market simply refuses to let you skip. Designing for the people who have to be convinced, putting evidence before voice, treating recovery as a first class flow, rebuilding structure instead of translating it, and taking the unglamorous surfaces seriously. Any product anywhere is better for all five. Most markets are just polite enough about it that you can get away with ignoring them for a few years.

That is the argument I would make to a team debating whether Japan is worth the effort. Not that this is a big market, although it is. That the standard here will find the parts of your product that were never finished, and you will take the fixes home with you.

The part I still cannot explain

I am not going to pretend I have figured this place out. I have lived here for years and I still get things wrong in ways that only become obvious afterwards, and the people around me are usually too kind to tell me at the time.

What I can hand over is the checklist, which is the transferable part of something that is mostly not transferable.

And one question, which is the one I would ask an audience if I had a stage and fifteen minutes. What local standard has your team been translating, when it should have been designing?