Building for Bharat: What We Learned Shipping to Tier-2 India Users
The English-language tier-1 Indian user is a fraction of the country. The next 500 million users live with different networks, languages, devices, and expectations. Building for them is a different discipline.
Key takeaways
- Network is unreliable, devices are mid-range, English is a second language at best.
- UX patterns that work in tier-1 (long forms, video-heavy onboarding, dark mode prevalence) don't always translate.
- Vernacular support and voice input aren't nice-to-have, they're often the difference between adoption and abandonment.
- Payment flows must be designed for users where UPI is the only digital payment they trust.
What changes from tier-1 to tier-2/3
Network
Fast 3G or unreliable 4G. Frequent disconnects. Big payloads fail. Design for thin pipes; lazy load; cache aggressively.
Devices
Sub-₹15,000 Android phones with 2-4GB RAM. JavaScript-heavy single-page apps are slow. Native apps with smaller bundles win.
Language
Hindi, Tamil, Marathi, Bengali, Telugu, Kannada, pick the languages your audience reads in. Machine translation is no longer good enough; professional translation by a native speaker is.
UX
Visual-heavy beats text-heavy. Icons with labels beat icon-only. Voice input is a feature, not a luxury.
Trust
App store ratings and reviews matter more (peer trust). Verified-by-government cues (Aadhaar, ABDM) move the needle.
Payment
UPI dominates. Cards are unusual. Cash-on-delivery is still a real expectation for eCommerce.
What works
- Onboarding in <30 seconds, mostly visual
- One-handed, thumb-zone design (most users hold the phone in one hand)
- Offline-first for any core flow
- Vernacular at every layer, not just the title page
- Voice search and voice input as first-class
- WhatsApp-style messaging for support
What doesn't
- Deep settings menus
- English-only error messages
- Video onboarding (data-expensive)
- Anti-patterns from US/EU UX research
What we recommend
Test with real tier-2 users on tier-2 devices on tier-2 networks. Lab testing on a MacBook misses everything important. Recruit users from outside your friend network.
FAQs
Which languages first? Depends on your audience. Hindi covers ~40% of India; Tamil, Telugu, Bengali next.
How big is the localization effort? Significant, typically 20-40% of UI engineering time on a properly localized app.
Do we need a separate app for Bharat? Usually not, but the same app needs to work well for both.