Case study: how a meditation app cracked Japan without translating the app binary

Most indie devs who think about Japan think about it as the next thing they’ll do “when we have time to localize the app.” Then they never do, because translating UI, testing layouts, handling vertical text and the App Store reviewer’s expectations for the Japanese market all stack up to a multi-week project.

This is a case study about a meditation app that ranked top 50 in Japan Health & Fitness without translating a single string in the app binary. Five weeks, three discrete pushes, ~$150 total spend. Here’s exactly how.

Why Japan is interesting for indie apps

Two reasons:

• It’s the second-largest App Store revenue market in the world. Per-user spend is high.
• The competitive density is much lower than the US storefront for English-first apps. Most US indie devs never localize anything for Japan, so the floor of “minimum effort to rank” is comically low.

The trade-off: Japanese users do expect a polished, locally-appropriate listing. A machine-translated description sitting next to English screenshots reads as spam, not product. So the work isn’t zero — it’s just nothing like rewriting the app itself.

Step 1 (week 1): metadata translation, done with care

We localized the App Store Connect metadata only:

• Name (kept the original English brand mark; added the Japanese category descriptor as a postfix)
• Subtitle (rewritten for the JP storefront, not directly translated — the english subtitle didn’t carry across cleanly)
• Description (translated, then edited by a native speaker for tone)
• Keywords field (entirely re-researched against Japanese-language search behavior, not a transliteration of the English keywords)

The keywords step is the part most people get wrong. The mistake is taking your English keywords and translating each one. The right move is to start over: figure out what Japanese users type when they search for “meditation app” — the actual phrases, including the katakana vs. hiragana vs. kanji choices, plus any English loanwords that have become dominant in the category.

We used a combination of:

• Apple’s Japanese App Store autocomplete suggestions (the same data source as Apple’s own ranking signal).
• Top-3 competitor metadata in the JP storefront, read literally.
• A native Japanese speaker who used the category herself for spot-checks and tone.

Total cost for this step: about 2 hours of the speaker’s time plus a Claude pass for first-draft translation, ~$40 all in.

Step 2 (week 2): localize the screenshots, not the app

This is the part everyone skips because it sounds like a lot of work. It isn’t, if you separate text from image.

The dev had 6 screenshots, each with an English overlay headline and subheadline. Underneath the overlay was a screen capture from the running app — which stayed in English.

We:

• Rendered new versions of the same 6 screenshots with Japanese overlay text.
• Kept the actual app UI behind the overlay in English (Japanese users in the meditation category accept this for indie apps as long as the marketing layer is localized).
• Used the same visual template, same fonts (Japanese fallback), same crop, same backgrounds.

End state: Japanese users see screenshots that look localized, even though the underlying app screens are English. For a product where the daily-use surface is mostly buttons and timers, this is a remarkably effective shortcut.

Total cost: about an evening of canvas work.

Step 3 (week 3): tiny ASA Discovery campaign to prime the index

Apple Search Ads has a campaign type called Discovery that runs against broad-match keywords your team didn’t pre-specify. It’s primarily a research tool — you find out what users are typing — but it has a useful side effect: it drives signal into the App Store ranking system for the keywords that convert.

We ran a Discovery campaign in JP at $5/day for two weeks. Total spend ~$70.

We learned:

• Two long-tail Japanese phrases the dev had never thought to target were converting at 3-4× the cost-per-install of his main terms.
• One English loanword (a brand-name-style term that’s become genericized in Japanese meditation media) was being typed by users in romaji and in katakana — we added both forms to the keywords field.

After two weeks we let the campaign expire. The point wasn’t to keep buying installs; it was to find the phrases worth ranking organically.

What happened

The trajectory across five weeks:

Week	JP daily installs	JP rank (lead keyword)
0 (baseline)	1-2	~280
1 (metadata pushed)	4-5	~140
2 (screenshots)	8-10	~70
3 (ASA running)	14-18	~35
4 (ASA off)	12-15	~30
5	18-22	~12

Top 50 in JP Health & Fitness by week 5. No binary update. The app’s internal copy was still entirely in English.

Two things to watch for if you try this

Reviewer scrutiny. Apple’s reviewers for the JP storefront occasionally flag apps with fully localized metadata + screenshots but an English-only app as misleading. This is rare, but the workaround is small: localize the in-app onboarding screen and the most-visible CTA buttons in code. Two or three days of work, not a full app translation. Worth keeping in mind.

Reviews and support. If you rank, Japanese-language reviews and support requests start coming in. You need a plan for those. The dev set up a forwarding inbox and a Claude-powered first-draft responder; he replies personally to the ones flagged as substantive. Not zero work, but a manageable hour a week.

The generalization

This isn’t a Japan-specific trick. The same playbook works for any large App Store market where the language barrier to a full app translation is high — Korea, Germany, France, Spain (LATAM included), Brazil. In each case:

1. Translate metadata thoughtfully (not literally).
2. Translate screenshot overlays while leaving the binary alone.
3. Run a small ASA Discovery campaign to find phrases worth ranking on.

You’re trading two weeks of careful copy work and a small ad budget for entry into a market that took your competitors a multi-month engineering project to reach. The ROI is rarely close.

---

Related: the English locale trick covers the same logic applied to en-GB / en-AU / en-CA / en-IN / en-SG — free-reach storefronts most indies never enable.

Asomium handles metadata across every locale and renders screenshot overlays in each language from a single canvas template. If the workflow above sounds tedious done manually, it’s literally the workflow Asomium was built for.