Skip to content

Mobile Development

How to Choose Between Native, Cross-Platform and PWA

Score your app against 8 factors for Native, React Native, Flutter and PWA, then check five if-X-pick-Y scenarios and a worked one vs two platform cost.

Shayma Parween
Lead Developer
Published
Jun 22, 2026
Reading time
11 min read
How to Choose Between Native, Cross-Platform and PWA

Pick the cheapest option that clears your hardest requirement. Start with a PWA. Move to React Native or Flutter when you need app-store presence, reliable push or real device APIs. Go fully native (Swift and Kotlin) only when the app depends on background execution, heavy graphics or new OS features on day one. Most business apps stop at the second step.

Key takeaways

  • Run three knockout questions before you compare anything else: deep background work, performance-critical UI, and app-store distribution. Most decisions are settled there.
  • A PWA is a real option for internal tools and companion apps, but iOS limits push, background work and hardware access. Check those limits before you commit.
  • React Native suits teams that already write React. Flutter suits teams that want pixel-identical UI on both platforms and are happy to adopt Dart.
  • Shipping on two platforms does not double the budget. In our estimating model it adds about 50%, because the backend, design and QA are shared.
  • You pay for maintenance every year, not once. Two native codebases mean two release trains, two sets of OS upgrades and two hiring pipelines.

First, the three knockout questions

Most native-vs-cross-platform debates go round in circles because they compare averages. Your app isn't average. It has one or two requirements it can't do without, and those decide the stack. Answer these first.

1. Does the app need to do serious work while it's closed? That means continuous location tracking, long-running Bluetooth sessions with a peripheral, background audio processing or large background syncs. If yes, go native, or cross-platform with a planned native module for that part. A PWA is out, because browsers on iOS give you almost no background execution.

2. Is the core experience a performance-critical UI? Think camera pipelines, real-time video effects, 3D, complex gesture-driven canvases, or anything where a dropped frame is the product failing. If yes, go native. Flutter is the strongest cross-platform option here because it draws its own UI, but if the hard part is the camera or GPU rather than the widgets, you'll still be writing native code.

3. Must users find and install it from the App Store or Google Play? For consumer apps the answer is usually yes, and that rules out a pure PWA. For internal tools, B2B dashboards and companion apps to a web product, the answer is often no. Skipping the stores saves review cycles, store fees and release friction.

If none of the three applies, you're choosing on cost, team skills and maintenance. The table below covers that.

The decision table

Score your app against each row, then weigh the rows that matter to you. ● = strong fit, ◐ = workable with effort, ○ = weak fit or blocker.

FactorNative (Swift / Kotlin)React NativeFlutterPWA
Device APIs and background work● Full access, day-one OS features◐ Most via libraries; edge cases need native modules◐ Most via plugins; edge cases need platform channels○ Limited, especially on iOS
Performance-critical UI● Best ceiling◐ Fine for business UI; heavy animation needs care● Consistent frame rates; own rendering engine○ Browser-bound
Fits an existing web/React team○ New languages, two toolchains● Same mental model as React◐ Dart is easy to learn but new● Same stack as your website
One platform only● No overhead◐ Framework overhead with no sharing benefit◐ Same● Works everywhere by default
Two platforms on a tight budget/timeline○ Two codebases● One codebase, mostly shared● One codebase, mostly shared● One codebase, plus desktop
Long-term maintenance◐ Stable, but everything happens twice◐ Framework and library upgrades take ongoing effort◐ Upgrades fairly smooth; plugin quality varies● Deploy like a website, no store review
Offline use● Full local database and sync control● Mature local storage and sync options● Mature local storage and sync options◐ Service worker and IndexedDB; fine for read-mostly, harder for complex sync
App size● Smallest◐ Adds a JS runtime◐ Adds a rendering engine● Nothing to download from a store

Two things the table doesn't show:

  • Most real apps are hybrids. A React Native or Flutter app with one or two native modules, say for payments, biometrics or a Bluetooth peripheral, is normal. Choosing cross-platform doesn't mean promising never to write Swift. It means writing Swift only where it pays off.
  • A ◐ is a cost, not a blocker. "Workable with effort" means you budget extra hours for it. Too many ◐ marks in the rows you care about tells you to move one column left.

If your app is X, pick Y

A companion app for an existing SaaS product, used by staff and customers. Pick React Native if your web team writes React, Flutter if they don't. You need push notifications, in-app payments and role-based screens, and you want store presence so parents, patients or field staff can find it. This is what we built for Esparks Edu: a React Native app with more than 21 screens and role-based navigation (a five-tab dashboard for admins and staff, two tabs for parents and students), Firebase push for attendance, fees and exam results, and Razorpay checkout inside the app. One codebase serves both Android and iOS.

An internal tool or back-office dashboard that people use on their phones. Pick a PWA. Nobody searches the App Store for your inventory screen. Our Sparks Business platform ships as an installable Next.js PWA with a dedicated offline view and home-screen shortcuts that open straight to New Sale or Stock. There's no store review, and every deploy reaches every device straight away.

A consumer app where the experience is the product: fitness tracking, media editing, AR, anything camera-heavy. Pick native, or Flutter if you need both platforms on one budget and the hard part is UI rather than hardware. Background sensors and camera pipelines are where cross-platform abstractions leak first.

An MVP to test demand on both platforms within a quarter. Pick cross-platform. The whole point of an MVP is to learn cheaply, and two native codebases double the work of every pivot. Keep the architecture modular so you can rewrite one hot screen natively later if the data justifies it.

An app that talks to custom hardware over Bluetooth, NFC or USB. Pick native, or cross-platform with a native module that you own and test on real devices from week one. Don't rely on a community plugin you haven't run against your specific peripheral. A PWA won't work here, because Web Bluetooth and Web NFC aren't available in Safari on iOS.

What one vs two platforms costs: a worked example

Illustrative example. These figures come from the formula behind our public cost calculator. They are indicative, not a quote.

Assumptions: a standard-complexity mobile app with user accounts, push notifications and in-app payments. The model uses a blended rate of USD 45/hour and a team delivering about 60 productive hours a week (roughly two developers).

StepOne platformTwo platforms
Mobile base hours × standard complexity (1.0)460 h460 h
+ Accounts & auth (40), notifications (35), payments (70)605 h605 h
× Platform factor× 1.0 = 605 h× 1.5 ≈ 908 h
Cost at USD 45/h, shown as a band (−15% / +25%)~USD 23,000–34,000~USD 34,500–51,000
Timeline at ~60 h/week~8–13 weeks~12–19 weeks

How to read it:

  • The second platform adds about 50%, not 100%. The backend, API, data model, design system and much of the QA are built once. The extra effort covers platform UI behaviour, device testing, store submissions and platform-specific bugs.
  • The ×1.5 assumes a shared codebase. That's how the calculator models selecting both iOS and Android. With two fully native apps, the client-side work is genuinely written twice, so plan for the top of the band or beyond. The backend still doesn't double.
  • A PWA falls below the one-platform column. You skip store submission and most device-specific testing, though you still build and test for both mobile browsers.
  • Every add-on scales with the platform factor. Payments cost 70 hours on one platform and about 105 on two. Cut features before you cut platforms if budget is tight.

For a breakdown of what drives cost beyond platform choice, see our guide to mobile app development cost in 2026.

Failure modes we see when teams pick the wrong stack

  • Choosing native "to be safe" for a forms-and-lists app. You pay double on every feature for performance headroom you never use.
  • Choosing a PWA, then discovering iOS push requires installation. On iOS, web push only works once the user has added the PWA to their home screen (iOS 16.4 and later). If notifications drive your engagement, test that install flow with real users before you commit.
  • Choosing cross-platform, then building half the app in native modules. If more than a handful of screens or services need native code, the shared codebase is costing you rather than saving you. Revisit question 1.
  • Ignoring the upgrade treadmill. Every year Apple and Google change permissions, background rules and store requirements. Whatever you pick, budget recurring time for OS upgrades and framework updates. Without it, an app that worked at launch stops passing review 18 months later.
  • Letting the agency's favourite stack decide. Ask any vendor, including us, to fill in the decision table for your app and explain each ◐ and ○. If they can't, the recommendation is a habit, not an analysis.

If you've settled on cross-platform and want to know what a good delivery partner looks like, our cross-platform mobile development guide covers vendor selection. If you'd like a second opinion on the decision itself, our mobile development team will score your requirements against the table with you.

Frequently Asked Questions

Is React Native or Flutter better for a business app in 2026?

Both are mature, production-ready choices for business apps. React Native is usually the better fit if your team already writes React or TypeScript, because skills and some code carry over from the web. Flutter is a strong choice if you want identical UI on both platforms, and the Dart learning curve is modest.

Can a PWA replace a native app?

For internal tools, dashboards and companion apps that don't need app-store distribution, often yes. It installs to the home screen, works offline for read-mostly use, and updates on every deploy. It falls short on iOS background work, hardware access such as Bluetooth and NFC, and push notifications for users who haven't installed it.

How much more does it cost to build for both iOS and Android?

In our estimating model, adding the second platform multiplies effort by about 1.5 with a shared cross-platform codebase, because the backend, design and much of the testing are shared. Two fully native apps cost more than that, since the client-side code is written twice. Use the cost calculator to see the band for your feature set.

When should you rewrite a cross-platform app in native?

Rewrite when a measured problem can't be solved with a native module. Examples are a performance issue in the core experience that profiling traces to the framework, or a background requirement the framework can't support. Rewriting one critical screen or service natively is usually cheaper than rewriting the whole app.

Do cross-platform apps get rejected from the App Store?

Not because of the framework. Apple and Google review the app's behaviour, content and policy compliance, not the language it was written in. Apps that are only a thin wrapper around a website risk rejection for minimal functionality, which is a reason to use a real cross-platform framework rather than a web view.


Work with Esparks IT Solutions

Planning a project around this? We work with businesses in the United States, Canada and the UAE, as well as the UK and Australia. Explore our Mobile Development services and portfolio, estimate your project cost, or book a free call.

Written by

Shayma Parween

Lead Developer

View team

Lead Developer at Esparks IT Solutions. Writes from the team's own delivery work across software, cloud, DevOps and security.

Share this article

Work with us

Have a system to build, fix or scale?

Tell us what you are working on and where it hurts. We will tell you how we would approach it.