Optimising Page Load Times in Flipkart Mobile App
Article Summary
Flipkart shaved 30% off page load times in their React Native app with 500M weekly visits. Here's how they did it without rewriting everything.
With 180M+ downloads and users on flaky networks with low-end devices, Flipkart's engineering team spent a year optimizing their hybrid React Native architecture. They focused on React Native bridge bottlenecks and network layer improvements.
Key Takeaways
- Sending data as strings vs nested objects cut bridge writes 3x, improving loads 15-20%
- Batching bridge calls beat multiple small calls, even with identical data transfer
- Predictive prefetching gave 10% more users instant page loads
- Parallelizing fragment creation with API calls saved another 5-10%
- Architecture shift: moved from native-first to React store-first rendering
The optimizations delivered a 30% improvement in page load times and drove a 0.5% increase in overall revenue.
About This Article
Flipkart's React Native bridge was slow because every write to nested native maps and arrays needed synchronization locks. This became a real bottleneck in their hybrid architecture.
Arun Kumar Reddy's team switched to sending data as serialized strings instead of deeply nested bridge payloads. This cut write operations down to a third of what they were and let them emit payloads earlier.
Page load times improved by 15-20%. The team found that deserializing strings actually performs better than creating objects on the React Native bridge.