Localization Text Expansion Calculator

How it works

Target length ≈ source × expansion factor, where the factor combines a language baseline (German/Finnish ≈ +30% for long text, French/Spanish +20–25%, CJK −20 to −50% characters but wider glyphs) with the short-string severity (≤ 10 chars up to +200%, > 70 chars ≈ +30%, per IBM guidelines)

English is one of the more compact major languages, so almost every translation runs longer — German compounds, Romance-language articles and prepositions, Finnish case endings — and the shorter the source string, the worse the ratio: a two-letter “OK” can triple, a paragraph grows by a quarter. IBM’s long-standing guideline bands that effect by string length, and this calculator scales each language’s baseline by it. Chinese and Japanese need fewer characters but each glyph is wider and taller, so pixel width shrinks less than the character count suggests and line height must grow; Arabic and Hebrew add right-to-left mirroring. The practical rules that follow are: never fix the width of labels and buttons, keep 30–50% slack (more for short strings), avoid concatenating fragments, and pseudo-localize early so overflow shows up before translators are paid.

Worked example

A 24-character English label into German: the 21–30 band implies about +65%, so plan for 40 characters — roughly 280 px at 7 px per character, overflowing a 160 px control by 120 px. Worst case across languages is similar; Japanese would need only about 14 characters but 155 px because of wider glyphs.

Frequently asked questions