Progressive Web Apps in Sri Lanka: Cost, Benefits & PWA vs Native Apps
At a Glance - PWA Development Sri Lanka (2026)
- Best for: Mobile-first retail, news, booking, field tools - when App Store friction is too high
- Requires: HTTPS, web app manifest, service worker, responsive UI
- Offline: Cache shell + catalogue/read content - not full checkout without sync rules
- Push: Strong on Android Chrome; limited on iOS Safari (2026)
- Typical cost: PWA retrofit LKR 150K-400K · new Next.js PWA LKR 500K-1.5M · vs native app LKR 800K-2M+
- Proof: Lighthouse PWA audit + Core Web Vitals before go-live
Introduction
Progressive web app development Sri Lanka buyers ask one question first: do we need an App Store app, or can a fast installable website do the job? A PWA is a secure website with a manifest and service worker - installable from the browser, cacheable offline, and shareable by URL. It is not a magic replacement for every native feature.
This guide covers PWA development Sri Lanka decisions: PWA vs native table, offline/push/install requirements, browser limits (especially iOS), cost bands, Lighthouse evidence we use at delivery, and a real export-retail pattern from our Next.js work. For full native builds, see mobile app services; for the web foundation, see web development services.
What Is a PWA?
A Progressive Web App meets installability and reliability criteria defined by the web platform: served over HTTPS, includes a web app manifest (name, icons, display mode), and registers a service worker for caching and background tasks. Users can add it to the home screen without the Play Store or App Store - but discovery is still via your website and SEO, not store search.
PWA vs Native App: Decision Table
PWA vs mobile app is a product decision, not a technology fashion choice. Use this table in workshops with stakeholders.
| Factor | PWA | Native (RN / Flutter) |
|---|---|---|
| Install friction | Low - browser prompt, no store account | Higher - store download, updates via stores |
| SEO / link sharing | Full URLs indexable | App Store listing only; deep links extra work |
| Build cost (LKR, typical) | 150K-800K retrofit · 500K-1.5M greenfield | 800K-2M+ - mobile CX guide |
| Offline | Cached reads; queued writes with sync API | Full local DB (SQLite, Realm) easier |
| Push notifications | Android: good · iOS: limited (see below) | FCM/APNs full support |
| Bluetooth / NFC / background GPS | Limited or experimental | Full device APIs |
| Payments in-app | Web PayHere/Stripe flows - payment guide | Same gateways + optional Apple/Google Pay SDKs |
| When to choose | Catalogue, content, MVP, budget-sensitive mobile web | Heavy push, device hardware, store presence, offline-first field apps |
Common hybrid: ship PWA first for validation, add React Native when install metrics and iOS push requirements justify it. Compare frameworks: Flutter vs React Native.
Offline, Push & Install Requirements
Install (Add to Home Screen)
manifest.jsonwithname,icons(192px + 512px),start_url,display: standalone- Service worker registered on secure origin (HTTPS)
- Meets Chromium installability heuristics (engagement / manifest validity)
Offline
Define what must work offline before coding. Realistic tiers:
- Tier 1 - App shell: Layout, fonts, navigation load without network
- Tier 2 - Read cache: Product catalogue, articles, last-viewed pages (stale-while-revalidate)
- Tier 3 - Write queue: Form/cart actions stored in IndexedDB; sync when online (Background Sync where supported)
Checkout and payments should not silently queue without user confirmation - require connectivity or explicit "save for later" UX.
Push notifications
- Android (Chrome): Web Push with VAPID keys + service worker
pushhandler - iOS (Safari 16.4+): Web Push only for installed PWAs to home screen; permission UX differs from Android
- Desktop: Supported in Chromium browsers when user grants permission
- Plan fallback: email/SMS for critical alerts if iOS coverage is mandatory
Browser & Platform Limitations (2026)
| Feature | Chrome Android | Safari iOS |
|---|---|---|
| Install to home screen | Yes (prompt + menu) | Yes (Share → Add to Home Screen) |
| Service worker cache | Full | Yes; storage quotas stricter |
| Web Push | Yes | Installed PWA only; no push in browser tab alone |
| Background Sync | Yes | Limited / not equivalent to Android |
| Storage eviction | Under disk pressure | Aggressive if app unused - design re-fetch paths |
If your audience is majority iPhone and marketing depends on push campaigns, budget for a native app or accept SMS/email fallbacks. Do not promise parity with Android on day one.
Lighthouse Evidence (How We Prove It Works)
Marketing claims without measurement are worthless. Before PWA go-live, Hashtag Coders runs Lighthouse in CI or staging on real mobile throttling - not only on a developer laptop.
| Audit | What it checks | Target at delivery |
|---|---|---|
| PWA category | Installable, service worker, manifest, HTTPS | All core checks pass |
| Performance | LCP, INP, CLS (field-aligned lab) | LCP < 2.5s · CLS < 0.1 on key templates |
| Offline test | DevTools offline mode on catalogue + shell routes | Documented routes in UAT sign-off |
| Regression | Re-run on production URL after deploy | Screenshot archived in handover pack |
Deeper performance work: Next.js performance optimization. We use Serwist or @ducanh2912/next-pwa on Next.js projects with Workbox caching strategies - not hand-rolled service workers unless scope demands it.
PWA Development Cost Sri Lanka
| Scope | Deliverables | LKR (indicative) | Timeline |
|---|---|---|---|
| PWA retrofit | Manifest, service worker, install prompt on existing responsive site | 150K-400K | 2-4 weeks |
| New Next.js PWA | Mobile-first UI, offline catalogue, Lighthouse pass | 500K-1.2M | 4-8 weeks |
| PWA + push + sync | Web Push (VAPID), queued actions, admin for campaigns | 600K-1.5M | 6-10 weeks |
| E-commerce PWA | Catalogue offline, PayHere online checkout, install UX | 800K-2M+ | 8-14 weeks |
Broader web budgets: website development cost Jaffna. E-commerce scope: e-commerce development Sri Lanka.
Related reading: Native apps: mobile customer experience · Stack choice: tech stack decision guide · Next.js builds: Next.js Jaffna.
PWA or native - we build both
Hashtag Coders - progressive web app development and mobile apps from Jaffna. Lighthouse-tested delivery, PayHere-ready e-commerce, honest iOS limits explained upfront.
Real Example: Export Retail PWA Pattern
Progressive web app company buyers should ask for a live install flow, not slides. Our pattern for Jaffna export retailers (including Next.js e-commerce builds such as Spices Jaffna) often follows this phased approach:
| Phase | What we ship | Stack |
|---|---|---|
| 1 - Web store | Responsive catalogue, checkout, inventory sync | Next.js, Prisma, MySQL/Postgres |
| 2 - PWA layer | Manifest, install banner, offline product list + images, Lighthouse PWA pass | Serwist / Workbox on Next.js |
| 3 - Engagement (optional) | Web Push for restock/promo (Android-first); native app if iOS push is critical | VAPID + FCM or React Native add-on |
Note: Spices Jaffna launched as a production Next.js store with measurable online sales growth (client-attributed). PWA install and offline catalogue are a natural Phase 2 for similar exporters serving mobile buyers on variable connectivity - scope separately in discovery.
Delivery Process
- Discovery: Offline routes, push needs, iOS/Android audience split
- UX: Install prompt timing, offline empty states, sync feedback
- Build: Manifest, service worker strategies per route type
- Test: Lighthouse mobile, offline walkthrough, push on Android device
- Handover: Lighthouse report, cache version policy, monitoring hooks
Frequently Asked Questions
Can a PWA replace our mobile app?
Often for catalogue, content, and light transactions - yes. For heavy push on iOS, Bluetooth, or App Store discovery, plan native or a hybrid roadmap.
Do PWAs work offline in Sri Lanka?
Cached reads work well on patchy 4G. Writes need explicit sync design. Test on real Dialog/Mobitel connections, not only office Wi-Fi.
How long does PWA development take?
Retrofit on an existing fast site: 2-4 weeks. New Next.js PWA with e-commerce: 8-14 weeks depending on payments and admin scope.
Which stack do you recommend?
Next.js + TypeScript + PostgreSQL for most Sri Lankan PWA projects we deliver - one team, SEO-friendly, PWA plugins mature. Laravel fits if your team is PHP-first.
Conclusion
Progressive web app development Sri Lanka is the right move when you need installable mobile UX, URL discoverability, and lower cost than dual native apps - and you accept iOS push and device API limits. Use the PWA vs native table, define offline tiers honestly, and require Lighthouse evidence before sign-off. When native is the better fit, we deliver through mobile apps; when the web is enough, we ship through web development.