Few things hurt a mobile product faster than instability. A crash on launch loses a user in seconds. A battery that drops 20% while your app sits in the background earns a one-star review and an uninstall. Both problems feed each other: users who see poor stability leave, ratings fall, store visibility drops, and acquisition gets more expensive.
Both stores also watch these numbers. Google Play tracks “bad behaviour” thresholds for crash rates and ANRs (App Not Responding errors) and can reduce your visibility if you exceed them. Apple surfaces similar data in Xcode Organizer.
The good news is that unstable apps are rarely beyond repair. In most React Native and Expo projects I see, a small number of root causes produce the majority of crashes and drain. This guide explains those causes, how to find them with real data, and how to fix them without a rebuild.
The Root Causes of App Crashes and Battery Drain
- Unhandled Promise rejections: A failed network call with no catch handler can leave the app in a broken state or, in some cases, crash it outright. These are common wherever API errors were never anticipated.
- Memory leaks from uncleaned listeners: Event listeners, timers, subscriptions and WebSocket connections that are never removed when a screen unmounts keep objects in memory. Over a long session, memory climbs until the operating system kills the app.
- Excessive background polling: Requesting the server every few seconds keeps the radio awake and the CPU busy. This is one of the biggest silent battery drains.
- Aggressive location tracking: Using high-accuracy GPS continuously, when coarse or periodic updates would do, drains a battery quickly.
- Heavy main-thread work: Large list renders, JSON parsing, image processing or complex calculations on the main or JavaScript thread freeze the interface. On Android, a frozen main thread for several seconds triggers an ANR.
- Oversized images and unbounded lists: Loading full-resolution images into small views, or rendering thousands of rows without virtualisation, wastes memory and CPU.
- Wake locks left on: Keeping the device awake and never releasing it prevents sleep and drains power.
Most unstable apps suffer from several of these at once, which is why guessing rarely works. You need data first.
Step 1: Audit Crash Logs and Performance Metrics
Start by measuring. Without data you will spend weeks fixing the wrong things.
- Install crash reporting: Sentry or Firebase Crashlytics will group crashes by cause, show stack traces and tell you which app versions, devices and OS versions are affected. Make sure source maps and symbol files are uploaded so traces are readable.
- Rank by impact: Sort issues by the number of affected users, not by how interesting they look. Often three or four issues account for most crashes.
- Check store dashboards: Google Play Console’s Android vitals shows crash and ANR rates, plus ANR stack traces. In Xcode, the Organizer shows crashes, energy and hang data.
- Use native profilers: Android Studio Profiler and Xcode Instruments reveal memory growth, CPU spikes and energy impact. Run them on a mid-range or older device, since flagship phones hide problems.
- Reproduce under stress: Leave the app running for a long session, switch screens repeatedly and background it. Watch memory over time.
Step 2: Fix Memory Leaks and Main Thread Bottlenecks
- Clean up useEffect: Every subscription, listener, interval and socket created inside a
useEffectneeds a cleanup function that removes it. Abort in-flight requests withAbortControllerwhen a component unmounts. - Handle every Promise: Add error handling around async calls, and consider a global handler plus an error boundary to catch what slips through.
- Offload heavy computation: Move expensive work off the main thread, defer it with
InteractionManager, or process it on the server when practical. - Virtualise lists: Use
FlatListor FlashList with sensible settings and memoised row components, so only visible rows render. - Optimise images: Serve correctly sized images, use a caching library such as expo-image, and avoid loading full-resolution files into thumbnails.
- Simplify view hierarchies: Deeply nested views and needless re-renders cost CPU on every frame. Profile re-renders and trim them.
Step 3: Optimize Background Processes and Network Requests
Once the app is stable in the foreground, reduce what it does when nobody is looking.
- Replace polling with push: Use push notifications, WebSockets with sensible lifecycle handling, or server-sent updates instead of repeated requests.
- Batch and cache requests: Combine small calls, cache responses and use conditional requests so the radio wakes less often.
- Right-size location tracking: Use the lowest accuracy and update frequency your feature genuinely needs, and stop tracking when it is not required.
- Release wake locks: Acquire them only for the task at hand and release them in all paths, including errors.
- Respect the platform: Use the operating system’s background task scheduling rather than fighting it.
How I Stabilize and Scale Mobile Apps
I am Supratim Chatterjee, founder of Codecreate. I work hands-on in React Native and Expo, diagnosing memory leaks, tuning background processing and refactoring unstable code. I do the work myself, so there is no handoff between a salesperson and an engineer.
A typical engagement looks like this:
- Audit. I review your crash reports, store vitals and profiler data to find the real causes.
- Prioritise. I give you a short, ranked list of issues, what each costs you and what each takes to fix.
- Refactor. I repair the problem areas inside your existing codebase, without rebuilding things that already work.
- Verify. I measure again after release so you can see crash and battery results for yourself.
I cannot promise a specific crash-rate number before seeing your data, but I can show you exactly where the problems are and how to remove them. You can see examples of my work in my web development portfolio and my design portfolio.
Stop Losing Users to Crashes
Send me your crash dashboard, or just describe what your users are reporting. I will tell you honestly what I think is going wrong and what it would take to fix it.
- Contact page: codecreate.net/contact
- Email: supratim@codecreate.net or supratim.dev.87@gmail.com
