A training tracker that had collected six years of features. The work was mostly subtraction — deciding what the app is for, then designing the interface around that single answer.
The problem
The home screen carried eleven competing elements. People opened the app, saw all of them, and closed it again without starting a session.
How it went
- 01
Find the one job
Mapped every flow and asked what each screen was actually for. Almost everything pointed back to the same need: tell me what today is.
- 02
Design the answer, hide the rest
Today became the whole first screen. History, plans and settings moved one level down — still there, no longer shouting.
- 03
A system, not screens
Built the UI on tokens and shared components so new features land consistently instead of adding another visual dialect.
- 04
Front-end build
Implemented the interface against the design system, with motion used only where it explains a state change.
Along the way
What shipped
- User flows and wireframes
- Product design, end to end
- Component library and tokens
- Front-end implementation
- Handover documentation
After
The app now opens on a decision instead of a dashboard. Any engagement figures would have to come from the client — none are invented here.
No invented numbers. Ever.Next project