ASO localization: why translating your metadata is killing your rankings

When indie devs decide to "go international", the workflow is almost always the same:

1. Open Google Translate.
2. Paste your English title and description.
3. Copy the translated text into App Store Connect.
4. Ship.
5. Wait for the install lift that never comes.

Translation is not localization. They're related but different operations, and the difference is the entire reason 80% of indie apps that "support multiple languages" get zero organic traffic outside their home market.

Here's why, and what to do instead.

What translation does

Translation gives you a grammatically correct rendering of your source text in another language. It is concerned with:

• Vocabulary equivalence (what's the Spanish word for "poster")
• Grammatical correctness (does the article agree with the noun)
• Tonal preservation (is "amazing" rendered as "asombroso" or "increíble")

Translation gets you a string that a native speaker would accept as not-broken. That's the bar.

For ASO, that bar is too low.

What localization does

Localization gives you copy that real users in that market actually type into the search bar. It is concerned with:

• The word for the concept that has the highest search volume in that country
• Local connotations and emotional weight that translation flattens
• Format conventions (do users in this country search with accents? With diacritics?)
• Script choice (is the canonical search term in romaji or katakana?)
• Cultural framing of the use case

A translation can be linguistically perfect and an ASO disaster. Conversely, a localized title can be a literal mistranslation by translation standards but rank brilliantly because it uses the word real users actually type.

A worked example: posters in Spanish

You're shipping a poster customization app. Your English title is "Posterly — Star Map Posters". You want to localize for Spain (es-ES) and Mexico (es-MX).

The translation answer:

• es-ES: "Posterly — Pósteres de Mapas Estelares"
• es-MX: "Posterly — Pósteres de Mapas Estelares"

Same string for both. Both are grammatically correct Spanish. Both will fail to rank for the queries real users actually search.

The localization answer:

• es-ES: "Posterly — Láminas Mapas Estelares"
• es-MX: "Posterly — Cuadros Mapa Estelar"

Different strings. Different keyword anchors. Different markets, different real-user vocabulary.

In Spain, the word for a hangable wall print is overwhelmingly "lámina". In Mexico, the same physical product is most often searched as "cuadro" or "póster". Brazilians (different language, different country) search "quadro". We covered this in detail in ASO for Brazil: lámina vs póster vs quadro.

The translation approach gives the same word ("póster") to all three. It's the word an ASO-naive translator picks because it's the closest cognate to the English word. It's also the word with the lowest search volume in two of the three markets.

Why this happens (and why translators don't fix it)

Translators are not ASO specialists. Their training is to preserve meaning, register, and tone — not to optimise for store search behaviour. When you ask a translator to render "wall art for the bedroom", they produce a phrase a Spanish speaker would accept as well-written. They don't tell you that no Spanish speaker would ever type that phrase into the App Store.

The information a translator can't give you:

• Which of three valid synonyms has the highest search volume in this country
• Whether users in this country search with full diacritics or stripped (does "constelaciones" or "constelaciones" rank better?)
• Whether the cultural framing of the use case differs (a "gift" for an anniversary in Brazil is framed differently from a "gift" in Mexico)
• Whether your category is dominated by English loanwords or local terms

A real localization process needs all four of those, and translators aren't paid to surface them.

The cost of translating instead of localizing

Some honest numbers from working with apps shipping into multiple Spanish-speaking markets:

• Translated metadata (same string for all es- locales): ~30% of the organic traffic of properly localized metadata
• Translated metadata (different strings per locale, but using translation logic): ~55% of the organic traffic
• Properly localized metadata: 100% baseline

The gap is starkest in markets with strong local-vs-cognate splits — Brazil, Spain, Korea, Japan, Germany. It's narrower in markets where English loanwords dominate (Netherlands, Sweden, Israel).

How to actually localize

The honest process, simplified:

›1. Identify the target country, not just the language

"Spanish" is not a market. Spanish-speaking countries each have different ASO behaviour. es-ES (Spain), es-MX (Mexico), es-AR (Argentina), es-CO (Colombia) all need separate consideration. Picking the right primary localization for your audience matters more than ticking every locale box.

›2. Find the local search vocabulary

The free way: Apple's autocomplete API and Google Play's search suggestions. Type your category seed into both stores in the target country and look at what surfaces. The autocomplete only surfaces terms users actually type — it's the most honest demand signal you have without paying for SDK data.

For a poster app, autocomplete in es-ES surfaces "lámina pared", "lámina decorativa", "lámina mapa" but rarely "póster". That's signal.

›3. Validate with native speakers, but ask the right question

If you ask a native speaker "is this word correct?" you'll get yes for any of three valid synonyms. Wrong question.

The right question: "if you were searching the App Store for an app that does X, what word would you type first?" Then: "what's the second word you might try?"

Optimise for the first answer. Use the second answer in the keywords field.

›4. Apply locally relevant cultural framing

Even with the right word, the framing matters. In Brazil, a poster gift for an anniversary is often framed around "lembrança" (memory/keepsake) more than "presente" (gift). In Spain, "regalo personalizado" (personalised gift) is the cultural anchor. In Korea, gift-giving for couples leans toward "기념일 선물" (anniversary gift) much more heavily than the equivalent phrase would in English.

These aren't translations. They're cultural anchors that shift the keyword strategy.

›5. Don't share metadata across locales unless you have to

App Store Connect lets you share a default metadata across all locales of a language. Don't. Even within Spanish, the per-country differences are large enough that the default-share approach leaves install volume on the table in every market except the one you happened to write the default for.

The exception: if you're a 1-person team and you have to choose one Spanish localization, pick the country with the largest installed base in your target audience. For most consumer apps, that's es-MX. For business apps, it's often es-ES.

Where most tools fail you

Most ASO tools that "support multiple languages" do one of two things:

• Translate your metadata using a generic LLM or DeepL pipeline. Better than Google Translate, still not localization.
• Give you a per-locale UI but generate metadata using English keyword candidates that are then translated.

Neither of these gives you the local-vocabulary-first workflow that actually ranks.

We built ASO Pilot around the opposite premise. The system prompt encodes per-country search behaviour for the markets we support — not as translations, but as locale-native vocabulary preferences. When the engine generates Mexican-Spanish metadata, it's not translating "poster" to "póster"; it's choosing between póster, cuadro, and other locally-valid terms based on which one has higher search behaviour in Mexico specifically.

The other half: real autocomplete-verified demand signals. Every keyword candidate is checked against Apple and Google's autocomplete in the target country, and the demand score reflects whether the term actually appears in real-user search behaviour there.

What to do this week

If you're already shipping in multiple locales:

1. Pick your second-largest market (after the home country).
2. Open Apple's App Store in that country (you can do this in browser by changing the URL country code). Type the seed term for your category. Look at the autocomplete suggestions.
3. Compare the suggestions to your current localized title. If your title doesn't contain at least one of the top 5 autocomplete suggestions, you're translating, not localizing.
4. Rewrite the title and subtitle for that market using the actual local vocabulary. Ship it.
5. Wait 2–3 weeks. Measure.

The lift from a single market localization done properly is usually larger than any other ASO change you can make in a quarter.

If you want a tool that does this with you instead of for you, give ASO Pilot a try. You'll get per-market metadata that's localized, not translated, with the autocomplete signals to back it up. One free analysis a month.