Can I translate my App Store keywords with Google Translate?

As a first draft, yes. As the final field, no. Machine translation gives the dictionary word, and App Store searches use the spoken word: shorter, often borrowed from English, sometimes a brand name used as a category. Run each translated term through the storefront's search suggestions before it goes in the field; a term with no suggestions is a term nobody types.

Do I have to translate the whole keyword field for every language?

Only for storefronts where you already get impressions or expect to. Apple indexes several localizations per storefront, so the US store reads both English (US) and Spanish (Mexico) fields, and many European storefronts read English (UK) alongside the local language. Fill the local field with local terms and let the English field carry the English ones; never repeat a term across the two.

Does a translated keyword need plural and singular forms?

Usually not. Apple matches most inflected forms for the major languages, so invoice covers invoices and Rechnung covers Rechnungen. Spend the characters on a different word instead. Where a language changes the stem in the plural, check the suggestions for both forms and keep the one that appears.

FunnelHound / Guides / Keyword translation

App Store keyword translation: how to translate keywords without losing rankings

By Emin, indie iOS developer building FunnelHound · Published 3 Nov 2026

Short version: translate the intent, not the word. Take each English keyword back to the problem it stands for, type that problem into the target storefront's search box, and keep the terms Apple suggests. A keyword field built that way ranks; a field run through a translator ranks for words that exist in dictionaries and nowhere else.

"Fahrtenbuch" is not a translation of "mileage log"; it is what a German driver calls the thing. That gap is the whole job.

Why the dictionary loses

Searches are spoken, not written. People type the short word they say out loud. "Belege scannen" beats "Quittungsscanner" because nobody says Quittungsscanner.
English leaks into every storefront. "App", "Tracker", "Scanner", "Budget" are typed in English in Germany, Brazil and Japan. A translation that replaces them loses the term people actually use.
Grammar moves the search term. Verbs ("Rechnung schreiben", "escanear recibos") outrank nouns in some languages; compound nouns split or merge; Japanese flips between katakana and kanji for the same object.
Categories have local names. A "budget app" is a "Haushaltsbuch" in Germany and a "gestor de gastos" in Spain, neither of which contains the word budget.
A brand can be the category. In some markets the dominant app's name is what people type for the whole category. You cannot use it, but you need to know that the generic term underneath it is small.

The difference between translation and localization as concepts, and which storefront reads which language field, is in the keyword localization guide. This page is the hands-on part: you have an English field and you need a German one by Friday.

The four-step method

Steps 2 and 3 are the dictionary. Step 1 makes sure you look up the right thing.

Rewrite each keyword as an intent. Not "mileage log" but "record my trips for the tax office". Not "invoice" but "write a bill to a customer". The intent survives translation; the noun does not. Ten keywords become ten short sentences.
Type each intent into the target storefront. Switch the App Store app's storefront to the target country, or use a device set to that region, and type the first two or three words of the intent in the local language. The suggestions that drop down are Apple telling you what people type. Write down every suggestion that matches your app; discard your own guess if it never appears.
Read the top competitors in that storefront. Search the two or three biggest suggestions, open the top five results, and copy their titles and subtitles. The words that repeat across them are the vocabulary of that market. If three of five say "Fahrtenbuch", that is your word, whatever the dictionary says.
Assemble the 100 characters. Commas without spaces, single words rather than phrases (Apple combines them), nothing that is already in the localized title or subtitle, no plurals, no category name, no app name. Rank by how often each term appeared in steps 2 and 3, fill from the top until the counter hits 100. The keyword field mistakes guide has the full list of wasted characters.

Where machine translation fits

Use it for step 1, to get the intent sentence into the target language so you can type it. Do not use it for the field. An AI model with a good prompt gets closer than a dictionary, because it knows "Fahrtenbuch", but it still cannot see which of three plausible terms Apple actually suggests this month. Thirty minutes with a native speaker who uses apps, not a professional translator, beats both: show them the intents, ask what they would type, and watch them do it on their own phone.

Which storefronts to translate first

Where you already get impressions. App Store Connect → Analytics → Territory. A storefront sending you 300 impressions a month with an English-only field is the cheapest win on the list.
Where the field is shared. English (UK) is indexed in many European and Asian storefronts, and Spanish (Mexico) in the US. One translated field can serve several countries; the extra US locales guide maps which.
Where your category is big. Germany and Japan for productivity, Brazil for utilities, France for anything with a strong local brand. Guesses, until the Territory report says otherwise.

Judge each storefront on its own

A translated field changes rankings in one storefront and nowhere else, so the only honest measurement is impressions and downloads for that territory, seven days before against seven days after, with the two-day lag in mind. A global impressions chart buries a German win under US noise. FunnelHound filters the funnel by territory for exactly this reason: pick Germany, pin the keyword change as a milestone, and read whether impressions moved and whether the page views and installs that followed moved with them. The change-measurement guide covers the before-and-after arithmetic.

Signs the translation went wrong

Impressions up, page views flat. You rank for a term with the wrong intent; people see you and scroll past. Back to step 1 for that term.
Nothing moved after four weeks. The terms exist but nobody types them. Usually a dictionary noun that no suggestion ever showed.
Impressions from the wrong country. A shared language field is being read by a storefront you did not aim at. Fine if it converts; otherwise move the term to the territory-specific field.

Translate once, judge per storefront

FunnelHound draws impressions, page views and installs per territory on your iPhone, with your keyword change pinned as a milestone. Germany gets its own verdict; so does Japan.

Get FunnelHound

Share this on X

Data notes: keyword field limit 100 characters per localization, per Apple's App Store Connect Help. Which localizations each storefront indexes follows Apple's published list as of 2026 and changes occasionally. German examples were checked against the German storefront's search suggestions in October 2026; suggestions rotate, so repeat step 2 before each submission.