Planning for Localization: Typical Expansion by Language
The same message doesn’t take up the same space in every language. Planning for text expansion can reduce rework and the risk of delays.
I make my coffee from roasted coffee beans every morning. Recently, I started buying a different coffee brand in cans. The pull-tab lid comes with multilingual safety instructions arranged in concentric arcs. I realized I had a Localization 101 lesson in my hands. The sentences had to follow a particular order: the more concise languages had to be placed closer to the center of the lid. The person laying out this text may not be a translator, but they understand something fundamental about localization.
When you work with some creative types of translation (e.g., transcreation), text expansion causes an issue that you have to fix, project after project. Even with technical translations, you come across tight table cells, tight text boxes sized so the source text would fit, and one-line headers that would need two lines after translation. Package inserts, box labeling, and GMP translations must be adjusted for text length. Software localization presents challenges, such as when the developer limits the number of characters. Sometimes you see a button sized exactly to display “Save”, but then it has to be translated into many languages.
How Much “Text Real Estate” Do Different Languages Need?
Here are some average expansion (or contraction) rates when we translate from English.
- German: +30%
- Portuguese: +15%
- Spanish: +15%
- Russian: +10%
- Japanese: –10%
- Chinese: –20%
These are rough estimates, and the actual expansion rate depends on the specific document and writing system (e.g., Japanese, which may also need more vertical space). Short strings may expand proportionally more than longer strings.
Typical Practices at Translation Companies
When we come across these situations, technical translation companies usually have to adjust the formatting of the translated document. Some typical actions:
- Reduce line spacing (vertical). This may be the least aggressive layout adjustment. Still, if compacted too much (e.g., by more than 5%), the translation layout starts feeling different from the source layout.
- Reduce character spacing (horizontal). This should be used very carefully, and in difficult situations, e.g., in tight table header cells. In critical situations, use of a condensed font may be required, subject to client approval.
- Reduce font size (proportional). This is usually not recommended, and frequently not even allowed, although this may be a good solution for spreadsheets.
What Can Authors Do?
The author of the source document to be translated can take some measures to prevent the layout of translations from feeling “different” from the source.
- Remember that almost always the length of the translation will be different from the source text. Prepare for that.
- Even if larger text boxes are not needed, consider making them larger to accommodate translated text. Don’t leave the decision of box size to the translation company.
- In software localization, ideally the client would decide which languages the text will be translated into, right at the beginning of the development process.
- After the source document is ready considering all of the above, run a pseudo-translation test to the intended target language(s).
Good translation and localization is teamwork that starts before the translation company receives the files. It starts right at the author’s desk. Remember: planning ahead can reduce revision cycles and avoid project delays.
I almost forgot to finish making my coffee! Time to go back and finish brewing while I admire this small piece of “text real estate” optimization art.
About Varendi
Varendi Translation is a technical and certified translation agency founded in 1989. We specialize in pharmaceuticals, life sciences, regulatory, and industrial/engineering content in English, Brazilian Portuguese, and Spanish. Our ISO-certified processes and terminology management support eCTD submissions, GMP documentation, labeling, and software/UI localization. Workflows are human-led, using CAT tools and, where appropriate, human-supervised AI to meet accuracy, compliance, and turnaround requirements.
