How we shipped the Shark app in 14 weeks

Shark had a problem: 5+ separate apps for different robot-vacuum models. Each with unique UX, separate onboarding, different icons. Support tickets ran 40% higher than they should.

We proposed unifying everything into one flagship app. Deadline — 14 weeks. Here's how we pulled it off.

Weeks 1-2: Audit and architecture

We walked through all 5 legacy apps. Recorded every flow, every popup, every error. Found 12 duplications and 4 patterns worth keeping.

Weeks 3-4: Design system

Instead of 5 different designs — one system with 60+ components in Figma. Atomic tokens: colors, typography, spacing, shadows. Components work in light/dark themes.

Weeks 5-10: Development

React Native + TypeScript. BLE bridge in native Swift/Kotlin for fast device links. MQTT layer for real-time commands. Every 2 weeks — client demo.

Weeks 11-12: Beta

Internal testing at Shark. 40+ insiders. Found 23 bugs, 3 critical (BLE reconnect after iOS 17 update).

Weeks 13-14: Launch

App Store submission → approved in 3 days. Google Play → 5 days. Phased country rollout: US → UK → CA → AU.

Results after 6 months:

- 4.7★ on App Store (was 3.9★) - -38% support tickets - +52% engagement (sessions per user/week) - Successful launch in 4 countries

The key lesson: unifying the design system BEFORE the code saves weeks. Don't skip fundamentals — invest in them.