Product7 min

Localisation vs internationalisation: a UX guide for scaling

Localisation vs internationalisation explained for product designers: what each one means, who owns it, and how to design screens that scale to new languages.

Saurabh YadavFounder, Design Studio by Saurabh
ProductI18n vs l10n
Share

Here's a moment most product teams hit sooner or later. The app works, people like it, and someone says, "Let's add Hindi." A week later the translated build arrives and half the buttons are broken. Labels wrap onto three lines, a card overflows, and a date reads backwards.

Nobody did anything wrong in that week. The real mistake happened months earlier, when the screens were designed for one language only.

That's the whole story of localisation vs internationalisation in one example. Internationalisation is the groundwork that lets a product take on new languages and markets. Localisation is the work of fitting it to one specific market. Get the first one right and the second becomes routine. Skip it and every new language turns into a mini redesign.

Localisation vs internationalisation: the plain difference

You'll often see them shortened to i18n and l10n (the numbers count the letters between the first and last). Here's how I explain them to founders.

Internationalisation (i18n) is designing and building the product so it can support any language or region without changing its structure. Flexible layouts, text kept out of images, dates and currency handled by format rather than hard-coded, icons that don't depend on one culture.

Localisation (l10n) is adapting the product for a particular market. Translation is the obvious part, but it also covers tone, examples, units, payment habits, legal text, images and sometimes features.

Internationalisation happens once and keeps paying off. Localisation happens again for every market you add.

Internationalisation (i18n) Localisation (l10n)
Question it answers Can the product handle any language? Does it feel right for this market?
When it happens Early, during design and build For each new language or region
Who leads it Designers and developers together Translators, local reviewers, product and design
Typical work Flexible layouts, string keys, format rules, script support Translation, tone, imagery, units, local payment and legal details
Cost of skipping it Every new language breaks layouts The product works but feels foreign

Some companies call internationalisation "globalisation". The names vary. What matters is the split between building the capacity and doing the adaptation.

Why designers should care, not just developers

It's tempting to treat this as an engineering problem. String files, Unicode, locale settings. Developers handle that, right?

Partly. But most of the expensive failures are design failures. A fixed-width button, a heading that only works at one length, a layout that relies on a short English word. Those decisions are made in Figma, long before anyone writes a translation file.

In India this matters early, not "someday". A product aimed beyond English-speaking metro users will likely need Hindi and other Indian languages much sooner than a typical Western startup needs its second language. I wrote more about this in designing for Indian users.

From my work

Pukaar.ai supports Indian parents in several Indian languages. Writing plain sentences before medical labels, a core decision on that project, also makes the content far easier to translate well, because a short clear sentence survives translation better than jargon does.

Internationalisation in UX design: what to build in from day one

You don't need to support ten languages to design for them. You need a handful of habits.

Let text grow

English is fairly compact. Many translations run longer, and some scripts need more vertical space for characters that sit above or below the line. Devanagari, for example, needs comfortable line height or the top marks get clipped.

  • Use Auto Layout so buttons and cards grow with their content.
  • Avoid fixed heights on anything that holds text.
  • Allow labels to wrap to two lines without breaking the layout.
  • Test with a deliberately long string, not your perfect English copy.

Keep words out of images

Text baked into an illustration or banner can't be translated without redrawing it. Keep text as live text layered over images, so the translation team only touches strings.

Don't build sentences from pieces

"You have" + count + "new messages" works in English and falls apart in languages with different word order or plural rules. Give translators whole sentences with placeholders, and let developers handle plurals properly.

Treat formats as data

Dates, times, numbers, currency and phone numbers vary by region. India writes large numbers as lakh and crore groupings (1,00,000), many countries use 100,000, and date order changes too. Design fields and displays that adapt to a format, rather than hard-coding one.

Check icons and colours

A thumbs up, a hand gesture, a colour meaning "danger" or "luck": these don't carry the same meaning everywhere. When in doubt, pair an icon with a label.

Plan for right-to-left

If Arabic or Urdu is ever on the roadmap, your layout will need to mirror. Back arrows flip, progress bars run the other way. You don't need to design it now, but avoid layouts that can only ever work left to right.

Tip

Make a "stress test" page in your Figma file: every key component filled with a long string, a short string and a second script. It takes an hour and catches most layout breaks before a developer sees them.

Localisation: making it feel local

Once the foundation is solid, localisation is where the product starts to feel like it was made for someone.

Translate meaning, not words. Literal translation produces stiff, sometimes funny copy. Work with translators who understand the product, and give them context: a screenshot, what the screen is for, how the user is feeling at that moment.

Match the register. Many Indian languages have formal and informal ways to say "you". A banking app and a social app may choose differently. Decide this per language, write it down and keep it consistent.

Respect mixed language. Plenty of Indian users mix English and Hindi naturally. A fully "pure" Hindi interface can feel more foreign than a careful mix. Test with real users before deciding.

Adapt examples and imagery. Sample names, festivals, food, payment methods and support hours should reflect the market.

Check legal and trust content. Privacy text, consent wording and disclaimers often need local review, not just translation.

A scaling workflow that doesn't fall apart

Here's the sequence I'd follow for a product planning to add languages.

  • Agree which markets and languages are realistic in the next year or two
  • Build components with flexible sizing and test them with long strings
  • Set up text styles that work for each script you'll support
  • Keep a glossary of key product terms and how each is translated
  • Give translators screenshots and context, not a spreadsheet of loose strings
  • Review every translated flow in the real app, on a real phone
  • Run a few usability sessions with native speakers before launch

That last step catches what reviewers miss: a word that's technically correct but nobody uses. A quick round, like the one in usability testing on a budget, is usually enough.

Watch out

Machine translation is fine for a first draft and terrible as a final answer, especially for health, money or legal content. Always have a fluent human review anything a user will act on.

A shared design system makes all of this far easier, because a fix to one component fixes every screen. If you haven't set one up, my guide to design systems for small teams is a good place to start.

Work with us

Want a team to design it with you? Our UI/UX design agency for startups and SaaS takes products from research to developer handoff.

FAQ

What is the difference between localisation and internationalisation?

Internationalisation is designing and building a product so it can support any language or region without structural changes. Localisation is adapting that product for one specific market, including translation, tone, formats and imagery. You internationalise once and localise for each market.

Which comes first, i18n or l10n?

Internationalisation comes first. Without flexible layouts and proper handling of text and formats, every localisation effort turns into rework. Doing i18n early is cheap compared with retrofitting it later.

Do I need to internationalise if I only support English today?

If there's any realistic chance of adding languages, yes, at least the basics. Flexible components, live text and format-aware fields cost very little when you design them from the start. Retrofitting them across a finished product costs a lot more.

Key takeaways

  • Internationalisation builds the capacity for many languages; localisation adapts the product for one market.
  • Most expensive i18n failures are design decisions, like fixed-width buttons and text inside images.
  • Design for text growth, different scripts, local formats and possible right-to-left layouts.
  • Localise meaning, tone and examples, not just words, and have fluent humans review it.
  • Test translated flows on real phones with native speakers before launch.

What to do next

Open your most important screen, swap in a string twice as long and see what breaks. That ten-minute test tells you how ready you are to scale. If you're planning a multilingual product, or want a second opinion on one, tell me about your project and we can talk through it.

Saurabh Yadav's Memoji

Written by

Saurabh Yadav

Founder of Design Studio by Saurabh, a UI/UX design studio in India. We design apps and websites that feel effortless, from research to shipped pixels. Most recently: Pukaar.ai, as seen on Shark Tank India, and the Demerara Bank app.

Keep reading

All articles →