Shiageru
(shi-a-ge-ru) — 仕上げる, a Japanese word for the finishing work. The last pass that makes something ready to ship.
Your Japanese users can't find you. The ones who do, bounce.
Both are the same bug: your app was translated, not localized. Japanese users never search the words your listing is written in, and the few who find you land on machine-translated screens. I'm a Japanese engineer — I'll show you exactly where it breaks, in English, then send your .arb, .xcstrings, or strings.xml back fixed.
What you get
Not a translation. A diff.
You don’t read Japanese, so judging the words is a dead end. Judge the format instead: this is what comes back — your file, fixed, ready to merge.
The gap
You have two review counts. You've only ever looked at one.
The App Store and Google Play keep your reviews separately in every storefront. Whichever market you actually watch, that's the count you know. The Japanese one sits on a page you've probably never opened — and for most apps that haven't been properly localized, it's a number you could count on your hands. That isn't a verdict on your app. Nobody in Japan disliked it; they never got to it. The store never surfaced you for the words they type, and the handful who did land bounced off a screen that reads like it came out of translation software.
Check yours in 30 seconds
- Open your App Store page in the Japan storefront — apps.apple.com/jp/app/… , or add ?l=ja&cc=jp to the link you already have. On Google Play, add &gl=JP&hl=ja
- Look at the review count — not the stars
- Compare it with the storefront you actually watch
For most apps that shipped Japanese without localizing it, the ratio lands somewhere past 1 in 500. If yours does too, that's what I'd start with.
If your Japanese star rating is also lower than the rest of your storefronts, that's worth having, but read it carefully: a 3.0 built out of twelve ratings is noise, not a finding. I'll tell you when that's what it is instead of building a pitch on it. The count is the number that holds up.
What a Japanese user actually sees
Discovery breaks in two places, and neither one is your ad spend. First, search: Japanese users type Japanese category words, and a listing built from a machine translation of your English name usually matches none of them. Second, the landing: the small number who do arrive read a page written in vocabulary no Japanese app uses, and leave without installing. You can't see either failure from your dashboard, because neither one produces a review to read.
What they search
家計簿 アプリ 自動連携
"household ledger app, automatic bank sync"
What your listing says
バジェットトラッカー — お金を追跡する
"Budget Tracker — track your money"
バジェットトラッカー is the katakana spelling of "Budget Tracker." It's a real string of Japanese characters that no Japanese person has ever typed into a search box. The word for this category is 家計簿, and if it isn't in your listing you are not in the results.
What they search
タスク管理 アプリ シンプル
"task management app, simple"
What your listing says
ToDoリスト — あなたの生産性を高めます
"ToDo list — raises your productivity"
Two misses at once. The category term Japanese users search is タスク管理, not ToDo. And 生産性を高めます is textbook translationese — grammatical, and immediately recognizable as copy that was written in English first.
What they search
写真編集 アプリ 無料
"photo editing app, free"
What your listing says
フォトエディター — 写真をより良くする
"Photo Editor — makes photos better"
写真編集 is the term with the search volume; フォトエディター is a transliteration competing for nothing. The subtitle is the bigger problem: it makes no claim a user could act on, because the English original was a brand line and the translator kept the words instead of the job.
Composite examples built from patterns that recur across indie apps localized with machine translation — not any specific app's listing. Your audit uses your own listing and your own category's search terms.
Two ways an app loses Japan
From your dashboard both of these look identical — a market that isn't producing anything. They're different failures with different causes, and the work that fixes one does almost nothing for the other. Which is why what follows branches instead of stacking into a bigger package.
Reach
They can't find you.
- How it shows up
- Your Japanese review count is a rounding error next to the rest of your storefronts, and searching your category's Japanese terms doesn't surface you at all.
- What's actually wrong
- Your store listing is written in words Japanese users don't type. A machine translation of an English name produces real Japanese characters that match none of the queries in your category, so the store never puts you in front of anyone.
Fix → Listing — $500
Quality
They find you, then they bounce.
- How it shows up
- Japanese users are arriving — there are enough Japanese ratings to actually read — and the score sits below what you see in your other storefronts.
- What's actually wrong
- The strings inside the app are machine-translated, so the screens read as obviously foreign, and dates, numbers, and currency are still formatted for somewhere else. People install, look, and leave.
Fix → Full — $1,300
Plenty of apps have both, and if you do, reach comes first — there's no point tuning the screens people never reach. If you can't tell which one you're looking at from the outside, that's the entire job of the audit below.
What you actually get
You don't get a spreadsheet of translated strings you have to wire back into your app yourself. You get a diff against your actual localization file — something you review and merge the same way you review any pull request.
.arb.xcstringsstrings.xmlen.json.stringsA diff, not a document
You get a patch against your actual localization file — .arb, .xcstrings, strings.xml, whatever your stack uses — not a spreadsheet of translated strings you have to copy in and wire up yourself.
String-expansion and layout
Japanese and English take up different amounts of horizontal space. I find the strings long enough to break a button or truncate a label and tell you exactly which widget gives, and what to change. Making that change in your codebase is a separate add-on — see pricing.
Plurals, dates, numbers, currency
Japanese pluralization, date order, number grouping, and currency display don't follow English rules. I check that your l10n config actually applies the right formatting, not just that a translated string exists.
Font fallback
If your app doesn't ship or reference a Japanese-capable font, Japanese text renders as tofu boxes or an ugly system fallback. I check for this specifically — it's an easy miss and it kills trust instantly.
Where I stop
Everything above is files: your localization files, your store listing, your policy documents. I don't work inside your codebase as part of any tier — no widget changes, no PRs, whatever your stack is. If you'd rather I did, that's a priced add-on, and keeping it separate is what stops the scope from quietly growing on either of us.
Why Japan is worth the trouble
This isn't the reason to fix your app — your own Japanese review count is. It's the reason the fix is worth more here than in most other markets you could localize into next. If you already shipped German, French, or Spanish, Japan is the locale you skipped, and it's the one where skipping costs the most: it's a top-three market where English isn't a usable fallback.
- $17.9B, third worldwide
- Japanese consumer spending on mobile apps reached $17.9 billion, making Japan the third-largest market in the world behind China and the US.Sensor Tower & Adjust — Japan App Trends report · 2023 data ↗
- 2.5B downloads
- Mobile app downloads in Japan reached 2.5 billion in the same year — a market with volume, not just high spend per head.Sensor Tower & Adjust — Japan App Trends report · 2023 data ↗
- Rank 96 in English
- Japan scores 446 on the EF English Proficiency Index, against a global average of 488 — 96th of the countries measured. Your English listing is not a fallback most of this market can read.EF English Proficiency Index — Japan · 2025 edition ↗
- Local apps win
- Established local titles dominated Japan's top 10 by consumer spend — which is what a market looks like when language and cultural distance keep foreign apps out, and why localization moves more here than in a market that already defaults to English.Sensor Tower & Adjust — Japan App Trends report · Q1 2024 data ↗
Figures checked August 2026, each one linked to the publisher that produced it rather than to a site that re-reports it. Click through and read the source instead of taking the number from me. I don't publish growth forecasts or revenue projections — see the FAQ.
The checklist nobody hands you
Separate from the localization work, there's a set of things Japan's published guidelines and the two store review guidelines ask a paid app to have. Most apps built for a global audience are missing at least one, and store review or a support ticket is usually how the developer finds out. This is a fixed-scope checklist: I compare your listing and documents against what the published sources ask for, show you what isn't there, and link each item to its source so you can read it yourself.
- 01
特定商取引法 (Specified Commercial Transactions Act) notation
Japan's guidelines describe a specific set of seller and transaction disclosures for anything sold to consumers — separate from a generic terms page. Apps built for a global audience often have no equivalent page at all.
Source: Consumer Affairs Agency — 特定商取引法ガイド ↗ - 02
Subscription cancellation display
Both store review guidelines set out how subscription pricing, renewal terms, and cancellation have to be surfaced to the user. This one is a platform rule rather than a law, and it's the one most likely to surface during review.
Source: App Store Review Guidelines / Google Play payments policy ↗ - 03
資金決済法 (Payment Services Act)
Apps that sell in-app currency, points, or gems fall within scope of a statute most developers outside Japan have never encountered, including deposit requirements above certain thresholds. If your app has a coin balance, this is worth knowing exists.
Source: e-Gov — 資金決済に関する法律 ↗ - 04
ステマ規制 (stealth marketing rules)
Paid or sponsored promotion aimed at Japanese users has to be labeled as advertising under a designation published by the Consumer Affairs Agency. Relevant the moment you run influencer or affiliate promotion in Japan.
Source: Consumer Affairs Agency — ステルスマーケティング規制 ↗ - 05
Privacy policy disclosure items
Japan's personal information guidelines list specific things a privacy policy is expected to state, and they don't line up item-for-item with a US or EU policy. I localize your existing policy into Japanese and list the disclosure items the guidelines ask for that yours doesn't currently cover.
Source: Personal Information Protection Commission — 法令・ガイドライン ↗
This is not legal advice. I'm an engineer, not a lawyer. Requirements are checked against publicly available guidelines — Japan's Consumer Affairs Agency, the Personal Information Protection Commission, and the App Store and Google Play review guidelines — and every item links to its source so you can verify it yourself. For anything that needs a legal opinion, take it to a qualified professional in Japan. I can point you to one.
Pricing
Not four options — one diagnosis, two possible fixes, and a way to stay current. Start at the top if you don't know which problem you have; the audit price comes off whichever fix it points you at.
Start here
Audit — Japan Readiness Audit
Which of the two problems do you have — reach or quality? This is the tier that decides it, before you pay for a fix aimed at the wrong one.
$149one-time
Get an audit- Your Japanese review count against the rest of your storefronts, as a ratio and against other apps in your category
- The Japanese search terms in your category, and which ones your listing currently matches — usually none
- 5+ specific mistranslations or unnatural strings, with screenshots
- Any Japanese reviews you do have, translated into English — plus your Japan star rating, read honestly rather than inflated into a finding
- Japan requirements checklist: 特商法 notation, subscription cancellation display, privacy policy disclosure items, ステマ規制 — every item linked to its published source
- A prioritized list of fixes, and a verdict on which prescription you need: Listing, Full, or neither yet
3 business days, delivered as a PDF. Findings are qualitative — no growth projections, no legal advice. Credited in full against Listing or Full if you go on to either one.
Listing — Japanese Store Listing
They can't find you.
$500one-time
The fix for a reach problem. Every field in both stores, rewritten for how Japanese users actually search. Nothing inside your app is touched.
- App Store: name (30), subtitle (30), keyword field (100), description, promotional text (170), and in-app purchase display names
- Google Play: title (30), short description (80), full description
- iOS and Android designed separately. Play has no keyword field and indexes the full description itself, so the same Japanese text can't do both jobs — most apps paste it into both anyway
- Script-variant mapping: 筋トレ, きんとれ, and キントレ are three different queries in Japanese. Deciding which ones go in the visible copy and which get pushed into the keyword field is the actual work
- Keyword field segmentation done for a language that puts no spaces between words, where the English instinct for comma-splitting doesn't apply
- Real search vocabulary extracted from 10 competitors' Japanese listings
- Every field filled to its limit — 30 full-width Japanese characters carry far more than 30 English ones, and a translated English title leaves most of that slot empty
- Promotional text put to work: it's the only field you can change without going through review, and it's usually sitting blank
- Your current rankings recorded as a baseline before anything changes
Delivered as one spreadsheet — current text, proposed text, character count, and the reason, field by field — so you paste and save. Store listing only: no in-app strings, no code, and no screenshot design (caption text yes, the images are a designer's job). Credited in full against Full.
Full — Localization That Ships
They find you, then they bounce.
$1,300one-time
The fix for a quality problem. Everything in Listing, plus your localization files fixed and returned ready to merge.
- Everything in Audit — the full diagnosis, not a summary of it
- Everything in Listing — every field in both stores, designed per platform
- Your localization files fixed and returned (.arb, .xcstrings, strings.xml, en.json, or whatever your stack uses)
- Up to 2,500 strings included
- Every place string expansion will break your layout, located and written up with the change that fixes it
- Date, number, and currency formatting corrected for Japan
- Your existing English privacy policy and terms localized into Japanese, plus the disclosure items the published guidelines ask for that they don't cover
- One round of revisions included
Files in, files out — no tier includes work inside your codebase. Layout fixes are specified here and applied under the add-on below. Covers up to 2,500 strings; $250 per additional 500. One revision round included, further rounds quoted separately. I localize the policy documents you already have — I don't draft new ones from scratch.
Ongoing
$300/month
Keep Japanese in sync as you ship new releases. New strings localized as you release them · Replies to Japanese store reviews · Japanese ASO keyword rank tracking and adjustments · Guideline updates tracked, with the parts that touch your app flagged — what changed, not what it means for you.
Add-on · Code fixes & PR
$400–600
The only place I touch your source code. Layout and l10n-wiring fixes made in your repository and delivered as a pull request, instead of specified in a document for you to apply. Flutter, iOS, Android, React Native — the stack doesn't change whether it's an add-on, because none of the tiers include it.
Priced separately, and honestly so: a clean Flutter codebase is a two-hour job and an unfamiliar Xcode signing setup is a two-day one. Folding that spread into a fixed tier price would mean overcharging most people to cover the worst case. On a first engagement I'll usually bill hourly, and move to a fixed price once I've done enough of these to know the real range.
Start at $149 — it comes off whichever step you take next. Audit, then Listing, then Full costs $149 + $351 + $800 = $1,300: exactly what jumping straight to Full costs. Climbing the ladder is never the more expensive path, so there's no reason to guess up front.
How it works
- 01
Send your files
Send your localization files (.arb, .xcstrings, strings.xml, en.json, whatever your stack uses), or just request the audit if you want the diagnosis first. No repository access needed.
- 02
I run the diagnosis
I pull your Japan storefront numbers, check which Japanese search terms in your category your listing does and doesn't match, read your strings for the mistakes that make people leave without installing, then run the requirements checklist against the published guidelines.
- 03
You get fixed files back
Listing comes back as one spreadsheet you paste into the store console. Full adds corrected localization files, your existing policy documents localized, and a write-up of the layout changes your app needs. You apply them. If you'd rather I made those code changes and opened the PR, that's the add-on.
- 04
Stay current, optionally
If you keep shipping, the ongoing tier covers new strings each release, replies to Japanese reviews, ASO rank tracking, and guideline updates, so this doesn't become a one-time fix that goes stale.
Who I am
Shintaro Akimoto
Software engineer, based in Japan
I'm a software engineer based in Japan. Shiageru is something I built and run myself — not an agency, not a team — so there's no markup for a project manager and no sales call to sit through.
I work directly against your localization files, so what comes back is a diff you review and merge like any other, not a document you have to interpret and re-implement yourself. Work inside your codebase is deliberately kept out of the tiers and priced on its own.
Because it's one person, the entry point is priced like one, and turnaround doesn't depend on a queue of client accounts.
FAQ
How much will my Japanese revenue grow?
I don't know, and anyone who gives you a number is guessing. Being hard to find in Japan is a fact you can check today; what happens after it's fixed is a forecast, and I don't sell forecasts. What you're buying is a diagnosis: the search terms your listing doesn't match, the strings that make people leave, and the requirements your listing doesn't currently meet. All of it is qualitative and checkable. I don't publish growth percentages or download projections, and I won't put one in a proposal either.
Which one do I need — Listing or Full?
It depends which of the two problems you have. If Japanese users never find you, the fix is the store listing, and rewriting your in-app strings won't move anything because nobody is getting to them. If they find you and then leave, the listing is doing its job and the screens aren't, so the files are where the work is. Most people can't tell which one they're looking at from the outside — that's what the $149 audit decides, and it's credited in full against whichever fix it points you at, so having it decided costs you nothing if you go on to buy the fix.
What if my app has more than 2,500 strings?
Full covers up to 2,500 strings, and beyond that it's $250 per additional 500. String count is a rough proxy for effort — a one-word button and a three-sentence onboarding paragraph both count as one — so if your file is unusual in either direction, send it over and I'll quote before you pay anything rather than run the meter on a bad unit.
Why not just keep using AI translation?
AI translation gets the words roughly right and misses the rest: string expansion breaking your layout, plural and date formatting, font fallback, and the fact that a translated string still has to be wired back into your app correctly. I fix the file, not just the words in it.
Do I need to give you access to my repository?
No — and for Audit, Listing, and Full I'd rather not have it. Files only. Send me your .arb, .xcstrings, strings.xml, or en.json and I return them fixed. Repository access only comes into it if you separately buy the code-fix add-on, where the deliverable is a pull request instead of a document.
I don't read Japanese. How do I know the work is good?
You're not being asked to judge Japanese. You're handed a diff against your own file, which you review the same way you review any pull request, line by line, with an English note on each change explaining why it's there. The evidence side is checkable too: your Japanese review count is on a page anyone can open in two taps, search rankings are a thing you can watch move, and every requirements item links to the published source it came from.
Isn't this just ASO? I'm not looking for a marketing agency.
Neither am I, and this isn't that. Nothing here involves buying installs or running campaigns. The reason Japanese users don't find you is that your listing is written in words nobody searches, and the reason the ones who do find you don't install is that your screens read like machine output — both of those are localization defects, and both get fixed in your files and your store listing copy. Keyword work is part of the deliverable because the store listing is a localized string like any other, not because I'm going to sell you a growth retainer.
Are you a lawyer?
No. I'm an engineer. The requirements checklist compares your app and listing against publicly available guidelines — Japan's Consumer Affairs Agency, the Personal Information Protection Commission, and the App Store and Google Play review guidelines — and links every item to its source so you can verify it. It isn't legal advice, and it isn't a judgment about whether you're compliant. For anything that needs a legal opinion, take it to a qualified professional in Japan; I can point you to one.
How do I pay?
Email me and I'll send a Stripe Payment Link. Audit, Listing, and Full are one-time and prepaid; if you've already paid for a lower step, that amount comes off, so ask me for the adjusted link rather than paying the full price twice. Ongoing is a recurring monthly Stripe subscription you can cancel anytime — month to month, no annual commitment.
What does 'merge-ready' actually mean?
It means the file I send back is a drop-in replacement (or an actual PR) for the one you sent me, not a spreadsheet or a list of strings you have to copy in by hand. Open it, diff it, merge it.
How long does it take?
Audit is 3 business days. Turnaround on Listing and Full depends on the size of your app — I'll give you a specific date after I've looked at your files.