Real world examples of localization as infrastructure. How localization systems enabled global product growth at Notion, Lyft, and Medium.
Notion
Problem
Localization was ad hoc, manual, and limited to a few languages. No dedicated team.
System Built
- Implemented TMS integrated with CMS
- Established continuous localization workflows
- Automated translation across all product touchpoints
Outcomes
- Expanded from 2 → 12 languages
- ~70% of users (and majority of revenue) now international
- Lean team (3 people) supporting global scale
Key Lessons
- Leadership alignment accelerates global expansion
- Automation enables small teams to operate at global scale
- Continuous localization is critical beyond 5–10 languages
- This is what localization as infrastructure looks like: global scale supported by a lean team.
Lyft
Problem
Significant technical debt from US-only architecture; no multilingual support.
System Built
- Refactored core systems for internationalization
- Built multilingual infrastructure across rider/driver apps
Outcomes
- First international launch (Toronto)
- 6 languages supported
- Avoidable $1M+ engineering cost due to late i18n. Even worse was the 18 months it took to retire tech debt, all while Uber was already operating internationally in many languages.
- This highlights the cost of treating localization as an afterthought instead of infrastructure.
Key Lessons
- Localization alone is not sufficient for international success
- Product-market fit and regulatory alignment matter
- Late-stage internationalization creates significant cost and delay
Medium
Problem
Strong global demand, but no localization support and tech debt.
System Built
- Refactored platform for multilingual operation
- Enabled multilingual publishing features
Outcomes
- Growth in high-demand international markets such as Brazil
- Platform usable globally despite limited UI localization
Key Lessons
- It is important to lay the technical foundation for global accessibility early on. It becomes expensive and time consuming to retire tech debt.
- This shows how early infrastructure decisions shape global growth potential.
Some Common Themes
- Localization can unlock 10–100x ROI
- Most cost is upfront engineering, not translation
- Tech debt is the biggest scaling risk
- Lean teams (1–3 people) can support global scale with automation