Building a native product without stopping the existing one
App StoreWhen I joined, HelpBnk's mobile app was a WebView shell around the existing web product. I have been moving it toward React Native incrementally: first the shell and navigation, then selected surfaces including the home feed, search, notifications, deep-link handling, and media. Both runtimes still serve users, so each new surface needs a clear owner.
The difficult work lived in the seams: routing between web and native worlds, feed rows that stayed cheap under load, video moving between inline and fullscreen ownership, gesture interruption, startup, and enough instrumentation to distinguish a real improvement from a convincing demo.
What I learned
An incremental migration only works when ownership is legible. The same applies to performance: a benchmark matters when it is anchored to a named user journey and the evidence survives the profiling session.