BikeMi
BikeMi’s redesign turns Milan’s public bike-sharing app into a cleaner, station-aware experience, readying it to compete with private micromobility rivals.
Role
UX/UI design, Research, Usability Testing, Prototyping.
Context
CHALLENGE : BikeMi's bike-sharing service had a dated interface that made adoption lag behind newer mobility apps.
OBJECTIVE : Meeting usability goals, faster onboarding and transparent station availability in a single, mobile-first experience.
Platform Evaluation
Methods map onto the Double Diamond framework -
Discover > Define > Develop > Deliver.
PHASE 1
Without Users
TESTING APP &
DOCUMENTATION
INFORMATION
ARCHITECTURE
SERVICE
BLUEPRINT
HEURISTIC
EVALUATION
PREMO
USER JOURNEY
DIGITAL ETHNOGRAPHY
PHASE 2
With Users
SURVEY
USABILITY TESTING
EMOTION RECOGNITION
ANALYSIS
INTERVIEWS
PERSONAS
Key Insights
There were multiple issues with the platform. Mainly, the architecture was too nested and there was a lack of cohesive design language that confused the user.
Consistent design system speeds up recognition and navigation.
Real-time bike counts and type filters cut wasted station visits.
On-demand tutorials with step feedback help new users onboard quickly.
Design Solution
Uniform Design System | Real-Time Availability | Guided Onboarding
IMPROVED ARCHITECTURE
LOFI PROTOTYPE
DESIGN SYSTEM
HIFI PROTOTYPE
TESTING
OLD ARCHITECTURE
Fragmented menus and scattered support links forced riders to dig through multiple subpages for stations, policies, and help.
NEW ARCHITECTURE
A map-centric hub with five clear branches unifies stations, search, account, and help, giving direct one-tap access to every core task.
LoFi Wireframes
Twenty users (18 UX-savvy 18-25, and 2 in their 50s) tested low-fi wireframes via think-aloud sessions in person and on Maze. Issues found in round one were fixed and retested, and with all blockers cleared the final wireframes were ready for high fidelity mockups.
Emotion Recognition Analysis
Users’ emotional responses were recorded and analysed while using the app to see the changes in facial expressions, aiming to identify feelings of frustration, confusion, or satisfaction.
This method, however, yielded limited insights as most users showed little change in expressions, leading to the dismissal of these results and rely on the other modes of feedback for testing results.
Design System
Using Atomic Design principles, the design system was made reflecting the colors of BikeMi while keeping accessibility in mind which was a key issue in the original app.
Tokens
Typography
Font : SF PRO
BODY
BUTTON
HEADING
Colors
PRIMARY
SECONDARY
TERTIARY
Atoms
Icons & Styles
Molecules
Cards, Buttons & Loader
Organisms
Selectors, Bars, Popups & Map
Further testing with the design system applied on the whole app, enabled iterations leading to the final HiFi screens.
The Redesign
I focused on reducing cognitive load, streamlining navigation, and building a consistent design system. By pairing concise text with clear visuals, I made information easier to grasp at a glance.
Key actions became accessible in fewer taps, and the unified design language not only improved usability but also strengthened the overall brand experience.
Original App
Redesigned App
I set the tone with Milan landmarks + clean BikeMi visuals to teach the model fast.
Outcome: faster first-time comprehension, clear next step (Get Started).
Micro-tips show: filter by type > tap station > unlock bike.
Outcome: fewer first-use errors; no need for dense help text.
Persistent ride HUD (route/ETA/cost), clear return receipt, visual tap-to-report.
Outcome: Higher ride confidence, quicker returns, faster damage reports.
Bite-size “How it works” steps. Visuals + one-liners replace long manuals.
Outcome: lower cognitive load and fewer support requests post-onboarding.
Testing (n=18)
I validated the new flow with a mixed group of first-time and returning riders using moderated think-aloud + unmoderated core tasks.
Task success rate
92%
across
Find station, Unlock and Return flows.
Median time-to-unlock
24s
For first-time users (streamlined flow, fewer decision points).
station ACCESS
2 Taps
From map home (reduced steps to lower cognitive load).
IA first-click accuracy
86%
users chose the correct path on the first try for core tasks.
SUS
84/100
Overall usability in the “Excellent” range.