Keigo Mistakes That Make Your Japanese Business Website Look Unprofessional

One keigo slip and the product no longer matters

Here is the uncomfortable mechanic. A Japanese B2B reader is not grading your honorifics in isolation — they are using them as a proxy for everything they can't yet verify. If the keigo is off, the reasoning runs roughly like this: "If they didn't get the basics of polite Japanese right, did anyone Japanese actually review this? And if not, what else didn't get reviewed — the contract terms, the security claims, the support promises?"

The honorific mistake becomes a competence signal. It doesn't matter that your product is excellent; the reader never gets far enough to evaluate it, because the copy has already told them you're operating without a native check. In a market where trust and 安心 (peace of mind) drive the purchase, that's not a cosmetic flaw. It's a disqualifier.

The good news: the failures cluster into a small number of recognizable patterns. You don't need to master keigo to defend against them. You need to know the three that do the most damage and build a check that catches them every time.

Sonkeigo vs kenjougo: the reversal that signals you don't know the rules

Japanese keigo splits politeness into two directions, and getting the direction backwards is the single most damaging mistake.

• 尊敬語 (sonkeigo) — respectful language — elevates the other person's actions. You use it when describing what your customer, client, or reader does.
• 謙譲語 (kenjougo) — humble language — lowers your own actions. You use it when describing what you or your company does.

The reversal happens when copy uses humble forms for the customer ("we humbly receive your distinguished self") or respectful forms for itself ("our esteemed company graciously deigns to provide"). To a native reader this is jarring in a specific way: it doesn't read as merely awkward, it reads as not understanding who is above whom in the conversation — which in a sales context is close to the worst possible impression.

Why machine translation produces this so reliably: sonkeigo and kenjougo often map to the same English politeness ("please," "we are glad to"), so a translation engine with no model of social direction picks whichever conjugation is statistically common and frequently lands on the wrong one. The English looks identical. The Japanese is backwards.

This is also why the failure is so detectable to a native and so invisible to a non-native reviewer: the politeness is technically present, just pointed the wrong way. Nothing in the English flags it.

二重敬語 (double keigo): trying too hard reads as trying too hard

If the reversal is the error of not knowing the rules, 二重敬語 (double keigo, nijuu keigo) is the error of overcorrecting. It means stacking two honorific markers on a verb that only needs one — applying respect twice to the same action.

The instinct behind it is understandable, even sympathetic: more polite must be safer. It isn't. To a Japanese reader, double keigo reads as strained and slightly unnatural — the linguistic equivalent of someone bowing too many times. It often shows up in copy that has been "polished" by a non-native who knew enough to add honorifics but not enough to know when to stop, or by an over-eager AI pass instructed to "make it more formal."

The signal it sends is subtler than the reversal but still costly: it tells the reader the writer is performing politeness rather than speaking it. Native business Japanese is confident and economical with its keigo. Copy that piles it on reads as anxious — and an anxious vendor is not a reassuring one.

Too casual: when です・ます isn't enough for B2B

The opposite failure is just as common, and it has gotten worse as teams lean on a friendly, conversational English brand voice and translate it straight across.

です・ます (the desu/masu polite form) is the floor for any business-facing Japanese — but for B2B website copy, the floor is not always enough. A homepage hero, a pricing page, a sales email to an enterprise prospect — these often call for register above plain です・ます, with appropriate sonkeigo when addressing the reader's actions and a controlled, professional tone throughout.

Two specific ways the "too casual" failure shows up:

• Casual sentence-final particles and contractions carried over from a breezy English voice (the kind of "hey, let's do this together!" energy that works in English SaaS copy and reads as unserious in Japanese B2B).
• Plain です・ます where the situation expects elevation — for example, addressing a prospective enterprise client's actions with flat polite forms instead of the respectful forms the context calls for.

The result reads as a vendor who doesn't grasp the formality of the relationship they're asking to enter. In a market where the buyer is often building a quiet internal case to bring you to their boss, casual copy makes that internal case harder to win. (This is a recurring theme — overseas brand voices that work at home routinely fail to survive the trip. We go deeper on it in Translated, not built for Japan.)

How a non-Japanese manager can catch these without reading Japanese

You can't audit keigo line by line if you don't read Japanese. But you don't have to. You can run a short set of structural questions that surface whether anyone qualified actually reviewed the copy — without parsing a single character.

Ask, about every piece of Japanese content before it ships:

1. Who did the final native pass — and can you name them? Not "the agency handled it." A specific person whose first language is Japanese, who read the final copy in context. If no name exists, no native pass happened.
2. Was it reviewed in context, or string by string? Keigo is relational — it depends on who is being addressed and in what situation. Copy translated as disconnected strings (UI labels, table cells, a spreadsheet of segments) loses the relational context that keigo depends on. Ask whether the reviewer saw the page, not just the strings.
3. Did a Japanese B2B reader — not a translator — sanity-check the tone? Translators optimize for accuracy. Whether the register lands as appropriately senior for your buyer is a marketing-and-relationship judgment, not a translation one.
4. Can your reviewer point to a specific change they made and explain why? "It looked fine" is not a review. A genuine native QA pass produces edits and a reason for each one. If the review came back with zero changes on machine-translated copy, it wasn't a review.

These four questions are deliberately about process, not language. They're the non-speaker's version of reading the keigo — and they're a reliable filter for whether your Japanese site is quietly working against you. We map out the wider set of failures these questions surface in The Japanese localization mistakes foreign companies make most.

Teardown: real copy, the keigo flagged and corrected

To make this concrete, here is an illustrative composite — anonymized, not any real client's copy — of the kind of Japanese a B2B SaaS company's site and outreach often ships after a translation-only pass. Each example shows the original, what's wrong, and a corrected version.

Notice the through-line: none of these are grammar errors a spellchecker or a fluency-focused MT engine would flag. They're register errors — and register is precisely what an English-side reviewer can't see and a translation-only workflow can't catch. This is the same evidence-first approach we apply across our published teardowns: real (anonymized) copy, the specific problem named, the fix shown.

Building a QA pass that catches this every time

The reason keigo errors keep shipping isn't a lack of skill — it's a missing step. Most overseas Japan content pipelines look like this: English source → translation (human or machine) → publish. There's no stage where a Japanese B2B reader looks at the final copy in context and edits for register.

That missing stage is the whole job. A QA pass built to catch keigo failures every time has three properties:

• It's a native edit, in context, on the final copy — not a translation, not a string review, not a fluency proofread. The reviewer sees the page or the email as the buyer will, and edits the register to match the relationship.
• It produces named changes with reasons — so the value is auditable even by a manager who doesn't read Japanese, and so the same mistakes get learned, not just patched.
• It treats register as a marketing decision, not a language one — because how senior, how warm, how formal your Japanese should be is a function of who your buyer is and how Japan buys. That's a positioning call.

This is the layer Foothold Japan is built to be. The work is planned, edited, and shipped with a native B2B reader's eye on register from the start — so the keigo never becomes the thing your buyer remembers about your brand. And because the standard is the same regardless of who you are, our pricing is published, month-to-month, and free of the multi-month minimums that lock you in before you've seen whether the work lands.

A correct honorific won't win you the deal. But a wrong one can lose it before the product ever gets a hearing — and that's a risk worth removing from your Japan launch entirely.

Where to go next

• Translated, not built for Japan — why brand voices that work at home fail the trip, and what "built for Japan" actually means.
• The Japanese localization mistakes foreign companies make most — the wider pattern of failures that keigo errors are one symptom of.
• What Japan content actually costs — real numbers for agency retainers, in-house hires, and a transparent alternative.

See it in practice: We publish teardowns of real (anonymized) Japanese B2B copy — the keigo and register flagged, the fixes shown. It's the clearest way to see what a native QA pass actually catches.