Minimum viable localization: the smallest scope that is still useful
Localization does not have to mean translating everything or nothing. Between those extremes are several partial scopes that players genuinely tolerate, as long as they are applied consistently. The goal of this article is to name those scopes and the one rule that determines whether players accept a partial localization or feel cheated by it.
Store page only
Translating the store listing — title, description, screenshots with localized captions — while the game itself stays in its original language is a common and reasonable minimum. It costs a fraction of full localization and gives a player enough information to decide whether to buy, even if they will play in a language they do not natively speak. Players who buy under this scope generally understand what they are getting, provided the store page does not imply more localization than actually exists.
This scope works best when your gameplay itself is not heavily dependent on text — a player who does not speak the game's language can still learn mechanics visually or through trial and error.
UI-only localization
Translating menus, settings, buttons, and system messages while leaving in-game dialogue or narrative text untranslated is another workable minimum, common in games where the UI is what a player interacts with constantly but the story is optional flavor. This scope lets a player navigate and operate the game confidently even if they miss narrative nuance.
It works less well the more the gameplay itself is carried by text — a narrative-heavy game with an English-only story but a translated menu still leaves the core experience inaccessible.
Subtitle-only for voiced content
For games with voice acting, translating subtitles while keeping the original voice audio is a well-established middle ground. It is dramatically cheaper than recording a full localized voice track, and many players accept or even prefer subtitles over dubbing, particularly when the original performances are part of what they are buying.
The requirement for this to work is that subtitle timing and line length are checked against the actual voice pacing — a subtitle that scrolls past before a player can read it is not meaningfully localized, even if the translation itself is accurate.
What players tolerate and what breaks trust
The pattern across all three of these scopes is consistency: a player can tell the difference between a game with a deliberately limited scope and a game with a broken one. What players tolerate:
- A clearly-scoped partial localization — a store page or UI translated, with no untranslated text appearing where translated text is expected
- Original-language voice with well-timed, accurate subtitles
- A store page that honestly represents what is and is not localized, so expectations match reality
Where trust breaks
What breaks trust is different in kind, not degree: mixed languages appearing inside the same screen or the same sentence, especially inside the core gameplay loop the player interacts with most. A menu that is translated except for one untranslated button, or a quest log where half the entries are in a different language than the other half, reads as broken rather than intentionally scoped — even though the underlying work (partial translation) might be identical to a well-received UI-only release. The difference is entirely in whether the boundary of what is translated is clean and predictable.
Structuring for growth from day one
The reason minimum viable localization works as a strategy rather than a permanent limitation is that it can expand later — but only if the underlying text was structured to allow it. Keep every piece of text, regardless of current translation status, behind the same key-based lookup system rather than mixing hardcoded strings with translated ones. That way, adding a language later means filling in a translation table, not restructuring code — and expanding scope from UI-only to full dialogue localization is a translation task, not an engineering rewrite.