App store localization strategy in short

Ship 2 or 3 complete locales first, picked from markets already sending you downloads, then expand in funded waves. Translate the name, subtitle, keyword field, description, and screenshots together per locale, because half-filled locales convert worse than fewer finished ones. Measure installs per page view in each new storefront, and let growth pay for the next wave.

That's the whole strategy in one paragraph. The rest of this guide turns it into a plan you can run this week: picking markets, choosing depth, sequencing fields, shipping waves, and reading results.

Start from revenue, not from language size

Open App Store Connect sales and trends and sort the last 90 days of downloads by territory. Any non-English market already inside your top 10 earns the first slot even when its volume looks small, because those shoppers found you despite an English-only page. Add one high-upside East Asian storefront next, usually Japanese or Simplified Chinese, where localized pages across indie catalogs consistently outpull English ones by multiples. Take German or French third: large spending markets where translation is straightforward and screenshots rarely need cultural rework. Cap wave one at 3 locales and translate every field per locale, name through screenshots, before enabling a fourth. After release, compare installs per product page view per storefront for 4 to 6 weeks. When a new locale beats your primary-language baseline, fund wave 2 from that lift. When it trails, fix the page before buying more languages.

Notice what this order ignores: speaker counts. Spanish has more speakers than Japanese, but a paid utility app often earns more per Japanese install than per Spanish one. Speaker counts describe people, not buyers. Your own download territories describe buyers, so they outrank every generic top-10-languages list. Where your data is thin (a brand-new app with 40 downloads), borrow a proxy: the storefronts where your closest competitors run localized screenshots. If 3 rival apps all ship Japanese galleries, that market pays for your category.

One data pull beats ten blog opinions. If a market already sends you downloads, it earned slot one before any language ranking did.

Full, partial, or metadata-only: pick a depth per locale

Not every locale deserves the same spend. Agencies sell 3 depths, and the labels are worth keeping because they price the work honestly. Full localization means every field plus screenshots plus culturally adapted captions. Partial means name, subtitle, keywords, and description translated while screenshots stay in English or get captions only. Metadata-only means the index fields (name, subtitle, keywords) without description or art changes. Each depth has a market where it's the right call.

  • Full depth for wave-one locales and any market where English penetration is low. Japan and Korea punish English screenshots, so partial work there reads as broken.
  • Partial depth for test markets where you want signal before art spend. Translated words plus English art still convert better than a fully English page.
  • Metadata-only for cheap coverage of secondary variants. Once Brazilian Portuguese is done, European Portuguese reuses most of it, and en-GB rides on en-US with spelling tweaks.

Match depth to evidence, then write it down per locale before translating anything. A plan that says Japanese: full, Spanish: partial, en-GB: metadata-only is a budget you can hold. A plan that says translate everything everywhere is a wish that stalls at 40 percent. Revisit depth each wave: partial locales that convert earn full art, and metadata-only locales that send nothing stay cheap.

What to translate first inside each locale

Inside one locale, order matters as much as market choice. App Store Connect splits each locale across 2 levels: app level holds the name and subtitle, version level holds the description, keyword field, release notes, and screenshots. A locale goes live only when both levels are filled, so translate in an order that fills one level per sitting.

  1. Lock a glossary first: brand spelling, feature names, and banned competitor words. Every later string reuses these tokens.
  2. Draft name, subtitle, and keyword field together. They share a 30 plus 30 plus 100 character budget of meaning, so write them as one row, not 3 separate jobs.
  3. Write the description next, reusing glossary terms and the subtitle's benefit language so the page reads like one product.
  4. Build screenshots last: translated UI plus translated captions at each size Connect asks for, checked against the words above.
  5. Paste app-level fields, then version-level fields, preview the storefront top to bottom, and submit with the version.

Why this sequence works: steps 1 through 3 produce the submittable text pack, and step 4 produces the art pack. Teams that translate first and click second finish a locale in one sitting. Teams that enable languages inside Connect and translate between clicks end up with half-English pages that stall in review. Our setup guide to App Store Connect localization walks the 2-level model screen by screen, and our locale list shows which languages you can add.

Roll the strategy out in waves of 2 or 3

Waves are the mechanism that keeps the strategy funded. A wave is 2 or 3 locales shipped together at one depth each, then measured before the next wave starts. Wave 1 is your revenue picks from the download data above, all at full or partial depth. Wave 2 clones the cheap wins: language twins of wave 1 (es-MX off es-ES, pt-PT off pt-BR, en-GB off en-US) plus one new bet. Wave 3 spends wave 1's lift on the next high-upside market.

  • Freeze English source text before wave 1. Editing source mid-wave invalidates translations already in flight.
  • Keep a one-row-per-locale sheet: market, depth, text status, art status, ship date. You'll read this sheet more than any dashboard.
  • Time waves with releases. Version-level fields need an editable version open, so locale work rides free on a planned update.
  • Stop adding after each wave and measure. The pause is the strategy, not a delay in it.

Indie teams usually reach 5 to 8 locales across 3 waves and then slow down deliberately, because each later locale serves a smaller market. That's success, not stall: the strategy did its job when extra languages cost more than they return.

Measure each wave before funding the next

Each wave ends with a read, not a feeling. In App Store Connect analytics, split product page views and installs by territory for the 4 to 6 weeks after a wave ships, and compare each new locale against your primary-language baseline. A locale that lifts installs per view earned its cost and argues for a twin or a neighbor. A locale that trails needs page surgery (usually screenshots or subtitle) before you buy more languages.

  • Installs per page view per territory: the headline metric. It isolates page quality from traffic swings.
  • Search impressions per locale: shows whether the translated keyword field opened discovery you lacked.
  • Screenshot-first markets: if views rise but installs lag, the art mismatches the words.

Give each wave 4 to 6 weeks before judging it. Storefront data lags, review adds days, and early numbers wobble. One common trap is reading week-one noise as failure and pausing a strategy that needed a month. The second trap is never reading at all and funding wave 4 on hope. The sheet from the waves section plus a calendar reminder is the whole measurement stack an indie team needs.

How App Store Localizer fits this workflow

App Store Localizer (https://asolocalization.com) builds the locale packs this strategy assumes. Paste a public App Store URL and it imports your title, description, icon, and screenshots. It translates the title, subtitle, description, and the 100-character keyword field per target locale, generates localized keywords from free heuristics over your listing, and re-renders screenshots with translated captions in iPhone 14 Plus 1284x2778 and iPad Pro 12.9-inch 2048x2732 sizes. Then download the assets or publish toward App Store Connect.

Coverage runs to 20-plus target locales, from Japanese, Korean, and both Chinese scripts through German, French, Spanish, and Brazilian Portuguese to Hindi, Thai, and Vietnamese, so waves 1 through 3 stay inside one tool. Pricing is one-time per listing: Starter at $19.99 covers 3 languages with every available screenshot, and Global at $99.99 covers all supported languages. Two honest limits: it handles the store listing only, not interface strings or code, and keyword ideas come from listing heuristics rather than paid search-volume data.

If the strategy above ends with wave 1 funded, start it today: generate the 3-locale pack this afternoon at https://asolocalization.com, then walk the field order with every string and screenshot ready to paste.

Frequently asked questions

  • How many locales should an indie app ship first? Three complete ones. Pick them from your own download territories, not from speaker counts, and translate every field plus screenshots per locale. Three finished locales that convert will fund the fourth, while 10 half-filled ones convert nothing and teach you nothing.
  • Should every market get full localization? No. Give full depth to wave-one markets and places where English art kills trust, partial depth to test markets, and metadata-only coverage to cheap twins like en-GB or pt-PT. Promote a locale to full depth when its numbers earn it.
  • Does the description affect search, or only the keyword field? Shoppers read the description and search reads the keyword field, but both shape conversion, and conversion shapes rank. Write the description for humans with the same vocabulary as your subtitle and captions so the page feels like one product.
  • How long before I know a new locale works? Wait 4 to 6 weeks after release, then compare installs per page view in that territory against your primary-language baseline. Data lags and review adds days, so week-one numbers wobble. Set a calendar reminder per wave and judge once.
  • Can I localize only the listing and leave my app in English? Yes. App Store Connect treats each localization on its own, so a Japanese listing can point at an English interface. Many indie teams ship this way to test a market before paying for full interface work.