Amir Afroozan
Client:
Snapp
My Role:
Product Design Owner
Year:
2021
Service Provided:
Product Design
gesture
Core Problems
There were no map design guidelines of any kind, and the map itself wasn't built in the tools
designers usually work in — it lived in Mapbox Studio, a cartography tool rather than a design
tool. So the work started twice over: learning an unfamiliar environment, and then working out
what a map should even be evaluated on. Riders struggled to orient themselves and find the
place they were actually looking for.
Approach
We treated this as cartography rather than UI styling. We benchmarked Google Maps, Waze, and
native platform maps, pulling their palettes out and setting them side by side to understand
what map colour actually does — and where the shared conventions are, because a map that
invents its own language is a map people can't read.
From that we built a full POI system rather than a set of screens: categories for the different
kinds of place a rider looks for — green space, government, food, services — and a tiered
palette separating first-class POIs like hospitals and airports from second-class ones, so
importance is legible at a glance. The system was built with headroom, so new POI types could
be added later without breaking the logic. We defined behaviour across zoom levels, built
matching light and dark palettes, and checked every colour choice against colour-blind
accessibility rather than picking what looked good in isolation.
Finding that shaped the design
Parks and green areas turned out to do disproportionate navigational work. Their distinct
shapes and bright colour make them anchors — riders locate themselves against a park, then find
their destination from there. So we gave them prominence they wouldn't have earned on a purely
hierarchical reading of importance. Major shopping centres got similar treatment for the same
reason.
Outcome
We ran navigation testing with real users: given a destination, how fast can you find it? The
new map beat both the old Snapp map and Google Maps on time-to-find. Google Maps is solving a
broader problem than we were, so it isn't a clean like-for-like comparison — but for the task
riders actually come to a ride-hailing map to do, the design measurably won.