TL;DR
Japanese UI microcopy is the part of localization most often skipped and most consequential to UX. Button labels translated literally from English read as commands rather than invitations, tooltips translated as full sentences feel verbose and condescending, and inconsistent verb endings across screens make the product feel unmaintained. Fixing microcopy is among the highest-ROI Japanese localization investments a SaaS company can make.
Key Takeaways
- Buttons are invitations, not commands — Japanese B2B users respond to "始める" (begin) far better than "開始する" (initiate), which sounds like an order.
- Tooltips should be noun phrases, not sentences — Japanese tooltips that end in 「してください」 feel verbose; short noun-style explanations read more naturally.
- Honorific prefixes carry weight — お and ご prefixes signal politeness and professionalism in B2B SaaS; omitting them can make copy feel curt.
- Verb ending consistency matters — switching between ます-form, dictionary form, and imperatives across a single screen is the most visible sign of machine translation.
- Context-stripped strings produce wrong copy — a translator working from a spreadsheet cannot know if "Save" is on a draft, an invoice, or a setting. Provide UI context, not just strings.
What Microcopy Means in Japanese UI Context
Microcopy is the short text scattered through a SaaS product: button labels, tooltips, placeholder text, validation messages, empty states, confirmation dialogs, and similar fragments. Each string is tiny. But collectively they form the texture of the user experience. In English UX writing, this layer has matured significantly over the past decade — companies invest in dedicated UX writers, style guides, and component-level review.
Most of that maturity does not carry across to the Japanese build. When a SaaS company localizes into Japanese, the carefully designed English microcopy is typically handed to a translator — or an AI tool — as a spreadsheet of strings with no UI context attached. The translator produces Japanese that is grammatically correct and semantically accurate. The product ships. And then Japanese users use the product slightly less, abandon it slightly earlier, and convert slightly worse than in English-speaking markets — without anyone being able to explain why.
The Japanese microcopy is the why. Each individual string passes review. The accumulated effect of small misalignments produces a product that feels translated rather than built for Japan. In SaaS, that perception affects retention, NPS, and renewal — measurably and continuously.
Why Literal Button Translation Damages Japanese UX
English SaaS culture has converged on assertive, action-oriented CTA copy: "Get started," "Start free trial," "Add to cart," "Sign up now." These work in English because the grammar carries an implicit politeness — "Start" reads as opportunity, not command.
Japanese does not work this way. Japanese verbs have explicit forms for command (始めろ), polite request (始めてください), invitation (始めましょう), and neutral action (始める). A direct translation of "Start" requires choosing one of these — and most AI tools default to the form that pattern-matches the English source most closely. The result is buttons that sound rigid or peremptory in Japanese B2B context.
The cumulative effect is a UI that feels like it is barking instructions. Japanese B2B users — procurement managers, IT decision-makers evaluating a foreign SaaS product — read this tone as a signal that the company has not thought carefully about Japanese users. In my experience, they are usually right.
Five button labels that consistently break in Japanese
Why Tooltips Cause More Problems Than Buttons
Tooltips are smaller than buttons but harder to get right. The English convention for tooltips has converged on terse imperative or descriptive sentences — "Click to expand," "Edit the title of this project," "Required field." These compress effectively in English.
Japanese readers respond differently. A tooltip translated as 「クリックして展開してください」 (please click to expand) feels like the product is talking down to them. And 「クリックして展開する」 (click to expand, dictionary form) feels truncated and informal. Both are technically correct. Neither is what experienced Japanese UX writers produce.
The right approach is to convert tooltips from sentences to noun phrases or short descriptors — how native Japanese UI documentation has long been structured. Japanese readers expect tooltips to describe what something is, not to instruct them on what to do with it.
Working rule: If a tooltip is shorter than 12 characters in English, it should be a noun phrase in Japanese. If it is longer, consider whether it belongs in a tooltip at all — Japanese readers prefer dedicated help text over long tooltips.
The Verb-Ending Consistency Trap
Japanese has several common verb endings used in UI: dictionary form (始める), polite ます-form (始めます), polite imperative (始めてください), and noun form (開始). A well-designed Japanese SaaS UI picks one register per component type and applies it consistently — primary buttons in dictionary form, secondary buttons in noun form, confirmation dialogs in polite imperative.
Machine translation does not enforce this consistency, because each string is translated independently. The result is a single screen that mixes endings: "編集する" next to "削除" next to "コピーしてください." Each individually passes a grammar check. Together they signal that no one made a conscious decision about Japanese UI tone. Japanese users register that immediately as low product quality.
This problem is invisible to non-Japanese reviewers, and there is no automated tool that catches it. A glossary helps but only partially — the constraint is at the design system level, not the term level. Fixing it requires either (a) a Japanese UX writer producing the strings from scratch with a register decision in mind, or (b) a Japanese QA reviewer auditing the localized build for consistency.
The Honorific Layer English UX Writing Ignores
Japanese has a layer of grammatical politeness that English does not. The honorific prefixes お (for words of Japanese origin) and ご (for words of Chinese origin) attach to nouns to raise their register — お名前, ご住所, ご連絡, お問い合わせ. In B2B SaaS UIs, omitting these is technically valid but reads as informal in a way English has no equivalent for.
The cost of getting this wrong is asymmetric. Adding the honorific where it does not belong is occasionally awkward but not damaging. Omitting it where it belongs signals an unpolished product. Japanese SaaS products built for the local market apply these prefixes consistently. Foreign products typically omit them, because the decision does not exist in English and no one knows to ask about it.
The touchpoints where honorifics matter most: form labels ("Name" → お名前), confirmation messages ("Your invoice is ready" → ご請求書の準備ができました), customer support copy ("Contact us" → お問い合わせ), and account-related UI ("Your account" → お客様のアカウント). Build the convention once. Apply it consistently.
| English source | Without honorific (less natural) | With honorific (B2B standard) |
|---|---|---|
| Your name | 名前 | お名前 |
| Your address | 住所 | ご住所 |
| Payment method | 支払い方法 | お支払い方法 |
| Contact us | 問い合わせ | お問い合わせ |
| Your invoice | 請求書 | ご請求書 |
| Notification | 通知 | お知らせ |
The Microcopy QA Checklist for Japanese SaaS Builds
Before shipping any Japanese SaaS UI, run this audit against the localized build — not against a string list. Most of these checks are impossible to perform from a spreadsheet, which is why Japanese QA reviewers consistently catch issues that translators miss.
Decide button verb form once and apply consistently
Pick one of: noun form (保存), dictionary form (保存する), or polite form (保存します). Apply across all primary action buttons. Document the decision in your localization brief.
Convert tooltip sentences to noun phrases
Any tooltip ending in 「してください」 or 「です」 longer than 12 characters should be reviewed for conversion to a noun-phrase format. Compactness signals professionalism in Japanese UI.
Apply honorific prefixes systematically
Build a glossary entry for each customer-facing noun and decide whether お or ご should be applied. Common B2B SaaS terms: お名前, お問い合わせ, お支払い方法, ご住所, ご連絡, ご請求書.
Audit confirmation dialogs separately from buttons
Confirmation dialogs ("Are you sure you want to delete?") often need a different register than the underlying action buttons. Polite form (削除しますか？) is standard.
Provide UI context, not just strings, to translators
"Save" on a draft, an invoice, and a billing setting may require different Japanese verbs (保存 vs 確定 vs 適用). The translator cannot make this distinction from a spreadsheet alone — provide screenshots or a staging build.
Review tooltips on hover in the actual UI
Tooltip width, line-break behavior, and content visibility differ between English and Japanese. A tooltip that fits in 60 pixels in English may need 90 pixels in Japanese — review in the live UI, not in a spreadsheet.
Test empty states and error messages with native readers
Short status messages are where machine translation produces the worst tonal mismatches. Native Japanese review catches mismatches that bilingual review consistently misses.
Build a microcopy style guide before scale
Once your Japanese build hits 200+ UI strings, ad-hoc decisions stop scaling. A documented Japanese microcopy style guide saves hours of QA review on every subsequent release and produces visibly more polished output.
Why This Investment Pays Off Faster Than Most Localization Work
Japanese microcopy is one of the highest-ROI areas of localization investment because the changes are small, the impact is broad, and the audit work is fast. A typical SaaS product has 200–500 microcopy strings — buttons, tooltips, validation messages, empty states. A single Japanese QA pass can cover all of them within a week. The product team applies fixes in one sprint, and the localization quality of the entire UI improves measurably.
Compare this to other localization work — full content translation, marketing site rewrites, help center overhauls — which take months and have diffuse impact. Microcopy fixes are concentrated and visible. Every Japanese user on every screen feels them immediately. For SaaS teams looking for the single highest-value Japanese localization improvement, microcopy QA is where I would start.
The teams that ship the best Japanese SaaS UIs treat microcopy as a design concern, not a translation concern. Strings are reviewed in context, verb endings are chosen consistently, honorifics are applied systematically, and the tooltip style is documented in the design system. None of this requires more budget. It requires different attention to the same surface area.
Frequently Asked Questions
What is the single biggest microcopy mistake foreign SaaS products make in Japanese?
Inconsistent verb endings across UI components on the same screen. A product that mixes dictionary form, polite imperative, and noun form on a single screen reads as untouched-by-design Japanese, regardless of whether each individual string is correct. Picking one register per component type and applying it consistently is the single highest-impact microcopy decision.
Should I write button labels as nouns or verbs in Japanese?
Most Japanese SaaS products use noun forms for short action buttons (保存, 削除, 編集) and dictionary form for invitational CTAs (始める, 試してみる). Noun forms feel more like UI labels; dictionary forms feel more like calls to action. Pick the pattern that matches your component intent and apply consistently across the build.
How should I translate tooltips for Japanese users?
Treat them as noun phrases or short descriptors, not full sentences. Tooltips that end in 「してください」 (polite imperative) or 「です」 (declarative) feel verbose and condescending. Replace with a compact noun-phrase format (タイトルの編集, 必須, 詳細を表示). This is the dominant convention in native Japanese SaaS UIs.
Do honorific prefixes (お, ご) really matter in B2B SaaS?
Yes — more than most foreign teams realize. Omitting them is grammatically valid but reads as informal or undermaintained in a B2B context. Customer-facing form labels (お名前, お問い合わせ, お支払い方法) and notification copy (ご請求書, お知らせ) almost always carry the honorific in Japanese-built SaaS products. The pattern is consistent across enterprise tools used by Japanese companies.
Can AI translation tools produce good Japanese UI microcopy?
For first-draft generation, yes — particularly with carefully crafted prompts that specify register, audience, and component type. For final ship-quality output, no. AI tools cannot enforce register consistency across components, cannot apply honorific conventions reliably, and produce contextually plausible but stylistically inconsistent strings. In our QA engagements, a native Japanese review of AI-generated UI microcopy reliably delivers a substantial quality uplift against the internal scoring rubric we apply across projects (register consistency, honorific accuracy, terminology, and component-fit).