Cross-platform mobile app testing means checking one shared codebase on every operating system, OS version and device type your users actually run. In the US that list is longer than most teams expect: in September 2026, six different Android versions each held more than 5% of US Android traffic, while three iOS versions covered almost every iPhone. BugsQA builds and runs these device matrices for React Native, Flutter and hybrid apps through our cross-platform app testing service.
The pain is familiar. The app was tested on the developer’s iPhone and one Pixel, it shipped, and the first one-star review came from someone on a Samsung running Android 13.
This guide gives you the numbers behind a realistic device list, the matrix we start from with new clients, and what that coverage costs per release.
Why does a cross-platform app still need separate iOS and Android coverage?
A cross-platform app needs separate iOS and Android coverage because a shared codebase only shares the app logic. Navigation, permissions, keyboards, notifications, fonts and the back gesture all come from the operating system, and each OS version changes some of them.
React Native’s own documentation is direct about this. It says component tests “do not take into account any iOS, Android, or other platform code” and that “if there is a bug in the iOS or Android code, they will not find it” (React Native testing overview).
The audience split makes both platforms non-negotiable for North American products:
| Market (Sept 2026) | iOS share of mobile traffic | Android share |
|---|---|---|
| United States | 53.59% | 46.39% |
| Canada | 64.87% | 35.13% |
Source: StatCounter, mobile OS share by page views, September 2026.
Takeaway: in the US, skipping Android testing means shipping untested code to almost half of your mobile users.
Which Android versions should you test a cross-platform app on in 2026?
You should test a cross-platform app on Android 16, Android 15 and at least one older version between Android 12 and 14, because the US Android audience is split across six versions. Android 16 is the most common version, but it covers fewer than half of US Android users.
| Android version | US share (Sept 2026) | Canada share |
|---|---|---|
| Android 16 | 46.85% | 55.16% |
| Android 13 | 11.04% | 9.41% |
| Android 15 | 10.74% | n/a in top list |
| Android 14 | 9.37% | n/a in top list |
| Android 12 | 6.51% | n/a in top list |
| Android 17 | 5.73% | 12.02% |
Source: StatCounter Android version share, United States, September 2026.
Those six versions add up to about 90% of US Android traffic. No single phone covers that spread.
Why do Android 12 to 17 all still matter in the US?
Android 12 to 17 all still matter because Android users upgrade slowly and unevenly. Updates depend on the phone maker and the carrier, and many budget phones stop receiving major versions after two or three years.
Older versions also behave differently in ways a cross-platform framework can’t hide. In our test plans the riskiest areas on older Android are:
- Runtime permissions: notification permission prompts only exist from Android 13, so the same flow has two paths.
- Edge-to-edge layouts: apps that target Android 16 can no longer opt out of edge-to-edge, so headers and bottom buttons can slide under system bars (Android 16 behavior changes).
- Back navigation: Android 16 turns on predictive back by default for apps that target it, and
onBackPressedis no longer called.
That last change is now mandatory for anyone shipping updates: from August 31, 2026, Google Play requires new apps and updates to target Android 16 (API level 36) (Google Play target API requirements).
Which Android brands should you cover?
You should cover Samsung first, then one other brand such as Google Pixel or Motorola. Samsung alone is 28.11% of US mobile devices, more than every other Android maker combined.
| Vendor (US, Sept 2026) | Share of mobile devices |
|---|---|
| Apple | 53.58% |
| Samsung | 28.11% |
| Xiaomi | 5.89% |
| Motorola | 3.49% |
| 3.16% |
Samsung ships its own One UI skin, keyboard and browser, so a layout that passes on a Pixel can still break on a Galaxy. In Canada, Google (5.29%) overtakes Xiaomi (1.25%), so a Pixel is the better second device for Canadian audiences.
Which iOS versions should you test on?
You should test on the current iOS release, the previous major version and iOS 18. In the US in September 2026, iOS 26 releases held at least 76% of iPhone traffic, iOS 18 still held about 11%, and iOS 27 had just started at 2.02% after its September release (StatCounter iOS version share).
| iOS version (US, Sept 2026) | Share |
|---|---|
| iOS 26 (26.6, 26.5, 26.3) | 76.07% |
| iOS 18 (18.7, 18.6) | 11.41% |
| iOS 27 | 2.02% (first weeks after launch) |
iOS is easier than Android, but not one-version easy. iOS 18 users are often on iPhones that can’t install iOS 26, such as the iPhone XR and XS, so they also have older screen sizes and slower chips.
How fast do iPhone users move to a new iOS release?
iPhone users move to a new iOS release within months, not weeks. Apple reported that by June 7, 2026, 86% of iPhones introduced in the last four years were running iOS 26, and 79% of all iPhones were (Apple App Store support data). iOS 26 had launched the previous September.
For your test plan, that means the newest iOS starts small and becomes the majority within one to two quarters. We add a new iOS major to the matrix on day one and keep the previous major until its share falls below roughly 5%.
What does a minimum cross-platform device matrix look like?
A minimum cross-platform device matrix covers one current iPhone, one older iPhone, one Samsung on the latest Android, one older Android phone and one tablet. This is the starting matrix BugsQA uses for US-focused apps, built from the StatCounter data above and reviewed every quarter.
| Tier | Device and OS | Why it’s on the list |
|---|---|---|
| 1. Always | iPhone on current iOS (27 or 26), 393x852 | iOS is 54% of US mobile traffic |
| 1. Always | Samsung Galaxy on Android 16, 384x832 | Samsung is 28% of US devices; Android 16 is the top version |
| 2. Recommended | Older iPhone on iOS 18, 375x812 | iOS 18 still has about 11% of US iPhone traffic |
| 2. Recommended | Android 15 phone (Pixel or Motorola) | Second Android brand, second-largest recent version |
| 2. Recommended | iPad in portrait and landscape | Most common tablet in the US |
| 3. If your audience needs it | Android 12 or 13 budget phone, 360x800 | About 17% of US Android users |
| 3. If your audience needs it | Android tablet, 800x1280 | Most common US tablet resolution |
Every build in the matrix also gets checked in landscape, with large system text, in dark mode and on a throttled connection. Those four settings expose layout bugs that no single device shows on its default settings.
How does the matrix change between the US and Canada?
The matrix for Canada shifts toward iOS because iPhones carry 64.87% of Canadian mobile traffic against 53.59% in the US. Canadian Android users are also newer: Android 16 and 17 together reach 67% of Canadian Android traffic.
For a Canada-first app, we swap the Android 12 or 13 phone for a second iPhone screen size (375x812 class) and use a Pixel as the second Android brand. For apps launching in both countries, the US matrix covers Canada as well.
Should you test cross-platform apps on real devices or emulators?
You should use emulators and simulators during development and real devices before every release. Simulators are fast for checking layouts, but they don’t reproduce real push notifications, camera and biometrics, Samsung’s One UI, battery saving, or the timing of a mid-range phone.
| Check | Emulator or simulator | Real device |
|---|---|---|
| Layout on many screen sizes | Good | Good |
| Push notifications | Limited | Required |
| Camera, Face ID, fingerprint | Limited | Required |
| Vendor skins (Samsung One UI) | No | Required |
| Performance on mid-range hardware | No | Required |
| Release build signed for the store | Partly | Required |
React Native’s documentation recommends end-to-end tests that build the app “in the release configuration” and run on a device or simulator, which is the same principle: test what users install, not the debug build.
How often should you refresh your device and OS list?
You should refresh your device and OS list every quarter and every September. September matters because Apple ships a new iOS major every fall, and Google Play’s yearly target API deadline (August 31 in 2026) changes how your Android build behaves.
Our quarterly refresh takes four steps:
- Pull the latest StatCounter OS, version and vendor shares for your markets.
- Compare them with your own analytics, which beat any public dataset.
- Add any OS version above 5% share and drop any below 2%.
- Re-run the full regression suite on any device you add.
If your analytics show a very different audience, such as mostly Android in Latin America or mostly iPads in schools, your own data wins.
What does cross-platform mobile application testing cost per release?
Cross-platform mobile application testing costs depend on how many devices are in the matrix and how often you release. At BugsQA, a one-off release test of an iOS and Android app is a fixed $2,800 (from $1,500 for one platform), recurring sprint testing starts at $1,700 a month for 40 hours, and smaller jobs are billed at $45 per hour for manual QA (how we work and pricing).
| Situation | BugsQA plan | Price |
|---|---|---|
| Launch, major update or redesign | Release QA | $2,800 fixed for iOS and Android |
| Releases every sprint | Ongoing QA | From $1,700 / month (40 h) to $5,900 full-time |
| Stable flows that repeat every release | QA automation | $6,500 setup, maintenance quoted with it |
| Small or irregular work | Hourly | $45/hr manual, $55/hr automation |
Release QA includes a test plan and test cases, manual testing across the agreed devices, bug reports with video and severity, one retest round and a sign-off report. Device count is the main driver: every device you add multiplies the regression time.
Frequently asked questions
How many devices do you need for cross-platform mobile app testing?
Most US-focused apps need five to seven devices for cross-platform mobile app testing: a current iPhone, an older iPhone on iOS 18, a Samsung on Android 16, a second Android brand on Android 15, an older Android 12 or 13 phone, and a tablet if your app supports one. Based on StatCounter’s September 2026 data, that set covers the OS versions used by about 90% of US Android and iPhone traffic.
Is testing on Android 16 enough for a new app?
Testing only on Android 16 is not enough, because Android 16 covered 46.85% of US Android traffic in September 2026. More than half of Android users ran Android 12, 13, 14, 15 or 17, and older versions handle permissions, notifications and back navigation differently, so each needs at least one test device.
Do React Native and Flutter apps need separate iOS and Android testing?
React Native and Flutter apps need separate iOS and Android testing because the framework shares app logic, not operating system behavior. Permissions, notifications, keyboards, gestures and store builds are native to each platform, and React Native’s own documentation says JavaScript component tests will not find bugs in the iOS or Android code.
What is cross-platform app testing?
Cross-platform app testing is the process of checking that an app built from one shared codebase, such as React Native, Flutter or a hybrid framework, works correctly on both iOS and Android across the devices and OS versions its users run. It combines shared functional testing with platform-specific checks for layout, permissions, navigation and release builds.
Get a cross-platform device matrix for your next release
BugsQA tests React Native, Flutter and hybrid apps on the iOS and Android versions your users actually run, from nearshore teams in Colombia working in North American hours. Send us your app, your release date and your analytics split by OS, and we will reply with a device matrix and a fixed quote for your next release.
- cross-platform
- react native
- flutter
- device matrix
