PWA vs native app: when a web app is the right choice for a business
Every client who says "we need an app" usually means one of three things: a website, a PWA, or a native app. The cheapest wrong decision is to build a native app before validating demand. Here is how we decide, and what it costs.
The short version
A PWA (progressive web app) is a website with a manifest and a service worker. It installs to the home screen, works offline, and feels like an app — without app store review.
- Website — from $500, 3–10 days;
- PWA — from $800, 2–4 weeks;
- Native app — from $1,500, 4–10 weeks, plus store accounts and moderation.
Figures are ballpark for a typical scope; the real price depends on screens, integrations and design.
When a PWA wins
- Validating a hypothesis. You don't know yet whether users return. A PWA ships in 2–4 weeks and shows real demand before you invest in native development.
- Internal tools. Inspections, checklists, field reports: employees don't need an app store — they need an icon, a camera and a form that works in a basement.
- Repeat visits. If users come back regularly (bookings, orders, cabinets), installing to the home screen gives an app-like experience without store overhead.
- Event launches. Conferences, seasons, campaigns: PWAs are fast to build and update instantly for every user.
- Limited budget and deadlines. No store review, no platform-specific codebases.
When a PWA is not enough
- Heavy background work — continuous recording, sensors, long-running computations: browsers throttle background processes.
- Complex hardware — Bluetooth devices, specific drivers, deep camera control.
- Maximum performance — games, heavy graphics, real-time video processing.
- "Store-only" requirements — if the client or a partner insists on RuStore/Google Play.
- iOS limitations — some APIs are restricted, push notifications only work on installed PWAs from iOS 16.4+.
In these cases we say "native" before the project starts, not in the middle of it.
What a PWA project includes
- Manifest: icons, name, theme color, display mode;
- Service worker: app shell cache, offline page, update strategy;
- Offline data: local storage, send queue, conflict handling;
- Installation: install prompt, iOS instructions;
- Push notifications (optional);
- Integrations: payments, forms, Telegram/email, analytics;
- QA: Lighthouse, real phones, "no network" scenarios.
A real example
A car dealer needed vehicle inspections: the driver photographs 8 angles, the manager approves or marks a defect with a comment. Instead of a native app we built a PWA: form, camera, client-side compression, offline queue, simple PHP admin panel.
- 1–2 weeks instead of 4–8 weeks for a native app;
- an order of magnitude cheaper, no store accounts or moderation;
- hypothesis confirmed: 80% of drivers submit a report on the first try.
SEO and PWA
A PWA is still a website, so pages index normally — and often rank better thanks to speed and caching. Keep canonical URLs, sitemaps and meta tags in order; do not block the offline shell in robots.txt; measure installation as a separate analytics event.
Checklist: is a PWA right for you?
- Do users return more than once or twice a month?
- Is there a "on the go" scenario — inspection, request, booking, report?
- Does it need to work with poor or no connectivity?
- Are camera/files/forms enough, without exotic hardware?
- Can you start without app stores?
- Can you iterate every 1–2 weeks?
- Is 2–4 weeks to a first release acceptable?
Mostly "yes" → PWA. Heavy background work or a store requirement → native.
What we learned building them
- Installation must be explained on iOS. Half the audience won't find "Add to Home Screen" on their own.
- Don't cache everything. Stale data is worse than a spinner.
- Offline queue needs conflict rules before launch, not after.
- Analytics from day one — installed PWA is a separate event worth tracking.
Written by the team at TenLux (tenlux.tech). We build websites, PWAs and native apps for small businesses — and we will happily tell you when you don't need the expensive option.
Top comments (0)