A translated listing can look polished while sending people into an app they cannot use. The problem might be a misleading caption, a missing language or a sign-in error that was never translated. Review the full experience in one language before expanding to more markets.
Use this guide to review one locale at a time. If the work is limited to a store page, start with Listing updates and retain the exact version being reviewed. In-app strings and behavior need a separate implementation and test pass.
1. Translating words without preserving meaning
A literal translation can follow the grammar and still suggest the wrong action or tone. Short copy is especially sensitive because there is little context. Give reviewers the complete user task and identify whether each phrase is a caption, button, description, or error.
Fix: ask a fluent reviewer what the reader would understand and do. Replace idioms that do not transfer. Do not assume that every person in a country prefers the same tone or visual style. Preserve important limitations when rewriting for clarity.
2. Treating English keyword translations as local research
Several terms can describe the same task. A dictionary can suggest equivalents, but it does not establish which phrase people use in a store search. You still need to check how people describe that task in the target market.
Fix: collect local task vocabulary from relevant listings, reviews, and available search data. Record the source, country, date, and whether a metric is an estimate. Track selected keywords in the correct storefront once you have chosen relevant terms.
3. Changing captions while leaving misleading UI
Translated captions around an English interface can imply that the app itself works in the target language. A screenshot should show what a customer will actually experience. The same applies to currencies, maps, example content, and subscription labels.
Fix: capture the localized app when available and use the screenshot editor to arrange the campaign. If a feature or language is unavailable, say so. Review the export at a realistic display size rather than assume that legibility on a desktop means legibility in the store.
4. Ignoring field limits and layout
Text can become longer or shorter depending on the phrase. A fixed “German expansion percentage” cannot tell you whether your title fits. A title can also meet the character limit and still be cut short in a search result.
Fix: rewrite to preserve meaning, then count and validate every field. Use the field reference and inspect the actual store preview. Do not cut a translation in the middle of a word or assume that a generator guarantees acceptance.
5. Using irrelevant or repetitive keywords
Terms that do not describe the app can attract the wrong audience and conflict with store guidance. Repeating a phrase to hit a density score also makes a description harder to read.
Fix: connect every important term to an actual feature or task. Apple's keyword guidance explains formatting and terms to avoid. Google's listing guidance warns against excessive or irrelevant keywords. A full field is not inherently a stronger field.
6. Confusing formatting with value conversion
Changing a dollar symbol to a euro symbol does not convert money. A locale formats the value; the currency code identifies its unit. Dates, times, decimals, measurement units, and names also need explicit handling.
Fix: use platform number and date formatters with the correct underlying values and currency. Test sample amounts, time zones, and longer names. String translation APIs such as NSLocalizedString are not currency formatters.
7. Testing only when everything goes right
An app may look translated in a screenshot while its sign-in error, empty state, permissions prompt, or payment recovery remains confusing. Font coverage and text wrapping can vary with device and accessibility settings.
Fix: test a complete task on supported devices. Make something go wrong too: enter an incorrect password, lose the connection or retry a payment. Keep screenshots and a short record of what was observed. The implementation and QA guide separates structural, visual, linguistic, and functional checks.
8. Mirroring everything for RTL languages
Right-to-left layout is more than aligning text to the right. Navigation order and reading direction matter, but not every icon, photograph, chart, or media control should be reversed. Mixed-language strings and numbers need testing.
Fix: use platform direction-aware layouts and review the actual interaction with a fluent user. Test both the product and the store campaign. Android's pseudolocales can reveal some layout assumptions before human review.
9. Letting translations become stale
A source description can change while its translations still advertise a removed feature. Approval of the old text does not cover a later change.
Fix: associate each locale with a source version, owner, and review status. Recheck changed text and screenshots with every relevant release. In AppDrift, approved-wording reuse is a paid capability for exact reviewed matches in the same app and context; it does not make changed content automatically approved.
10. Expanding faster than you can support
Publishing many locales can create a larger maintenance obligation without proving demand. A country's total app spending does not establish whether your particular service is available or useful there.
Fix: choose languages using existing customers, requests, category relevance, availability, and review capacity. The language selection guide provides a market comparison. Set a budget for both launch and ongoing updates.
What to give your reviewer
| Before review | Before sending to the store |
|---|---|
| Source version, locale, task, glossary, screenshots | Approved fields, remaining issues, target app/store, correct draft state |
| Clear feature and language limits | Validated text, matching assets, retained baseline |
Metadata translation prepares AI drafts, with token usage quoted for the selected work. Have a fluent reviewer check the draft before use. Sending supported text to a connected store is a separate paid action and follows the store's review process. Changes inside the app still require development work.
Frequently asked questions
How many downloads do localization mistakes cost?
There is no universal percentage. Inspect the specific problem, traffic, user feedback, and task completion before estimating its effect.
Should I translate screenshot captions?
Translate captions for the intended audience and make them match the actual app experience. Use real localized UI captures where available; do not imply unsupported product languages.
How do I handle longer translations?
Rewrite for meaning and readability, use flexible layouts, and validate each final field. A fixed expansion percentage is not a substitute for testing the actual text.
Does AI translation eliminate these mistakes?
No. AI can create a draft, but keyword relevance, meaning, product scope, formatting, layout, and publishing still need review.
