Shipping Echo: Lessons From Building a Native Finance App in 10 Weeks
We built Echo (iOS + Android) in 10 weeks with a team of three. Here is what we got right, what we got wrong, and what we would do differently next time.
Echo is the mobile client for trade. We built it in 10 weeks with three people: one iOS (Swift/SwiftUI), one Android (Kotlin/Compose), and one backend engineer handling the API and push notification infrastructure.
What we got right: starting with a shared API contract before writing any mobile code. Both platforms implemented against the same OpenAPI spec, which meant the iOS and Android apps were feature-equivalent from day one. We also shipped to TestFlight/internal testing in week 3, which gave us 7 weeks of real user feedback.
What we got wrong: we underestimated push notification complexity. Reliable delivery of time-sensitive trading signals across both platforms, with proper deduplication and ordering, took almost 2 weeks — 20% of our timeline. We should have started this in week 1, not week 5.
We also made a mistake with our charting library on Android. We chose a popular open-source option that looked great in demos but had performance issues with real-time data updates. We ended up writing a custom chart renderer in week 8, which delayed our Android launch by 4 days.
If we did it again: push notifications first, shared API contract second, UI third. And we would build custom chart rendering from the start rather than hoping a third-party library meets our performance requirements for real-time financial data.