A concept pitched to Google Maps — adding pre-attentive colour intelligence to the route line so drivers are warned of approaching highway exits before it's too late, without adding a single new UI element.
This concept uses color cues to help drivers anticipate and prepare for exits. The color system is configurable and can also be represented through variations in line thickness and other subtle visual indicators.
Before touching any design tool, I observed real drivers using Google Maps on live highway journeys. The goal was to understand the moment of a missed exit — not from data, but from the driver's lived experience on US interstates.
Drivers in our diary study reported missing at least one GPS-guided highway exit per week. The highest frequency was on unfamiliar routes during rush hour.
Average number of times a driver glances at the phone per exit approach. Each glance takes eyes off the road for 0.3–0.8 seconds at highway speed.
Typical distance at which drivers first notice a lane change is needed. At 65 mph that is only 17 seconds — far too late for a safe multi-lane merge.
"By the time I saw the exit sign, I was already in the wrong lane with a truck on my left. I had to just… keep going and reroute."
"The blue line looks the same whether I'm ten miles from the exit or ten seconds. There's nothing to warn you it's about to happen."
Research revealed the problem isn't a lack of information — Google Maps has all the data. The failure is in the signal: the route line communicates nothing about proximity to the next exit manoeuvre. Every participant was staring at the right element. It just wasn't telling them anything.
Highway drivers using Google Maps need to perceive their exit approach earlier and without consciously reading the screen, because the uniformly-blue route line provides no advance spatial signal — causing late, unsafe lane changes and missed exits.
The instinctive solution — orange for "prepare", red for "exit now" — creates a direct semantic collision with Google Maps' existing traffic encoding on the same route line. Orange = moderate traffic. Red = heavy traffic. Using these colours for exit proximity would introduce a dangerous double-meaning. A driver seeing a red segment could not distinguish "heavy traffic" from "take this exit now". This is a safety failure. All colour selection must start by eliminating the entire reserved spectrum.
Ideation here was a systematic elimination process. Rather than open-ended brainstorming, I mapped the entire colour spectrum against three filters in order: reserved by Google, WCAG contrast, colour-blind distinguishability. What survived was the answer.
Teal (190°) and Purple (275°) are separated by 85° from the nearest existing Google colour (Green at 145°), and 85° from each other — ensuring perceptual distinction for all users including the most common forms of colour vision deficiency.
Teal occupies the 185–195° hue range — entirely absent from Google's palette. Psychologically it signals calm readiness and transition — between the passive blue of normal navigation and something requiring action. On the map's warm-grey road surface it achieves a 5.8:1 contrast ratio (WCAG AA). Under deuteranopia simulation, it reads as a distinctive dark blue-green — clearly different from the normal route blue.
Purple sits at 275° — as far from red-orange (15–35°) as possible on the hue wheel, yet unambiguously distinct from teal. It triggers immediate attention without the aggression of red, and is perceivable across all three major colour vision deficiency types. On the map road surface: 7.5:1 ratio — WCAG AAA. The strong luminance drop from teal to purple creates a clear sequential signal even in greyscale.
The prototype simulates the three-state journey on I-95 Southbound in New York — the exact stretch of highway where exit-miss events are well-documented due to multi-lane complexity and a short merge window at the Cross Bronx Expressway interchange.
Route line stays Google Blue · 6px weight. No change from today's experience. Zero cognitive load added. Driver focuses on traffic conditions and general routing.
Route transitions to Deep Teal · weight increases to 8px · Navigation bar shifts to teal-tinted background · Bottom banner: "PREPARE: Move toward exit lane" · A single haptic pulse fires. At 65 mph, 1.5 miles = ~90 seconds — enough time to move safely across multiple lanes.
Route transitions to Deep Purple · weight 10px · 2-second breathing animation · Banner: "EXIT NOW · Stay right" · Second haptic pulse (2 taps). 0.4 miles at 65 mph = ~22 seconds. Urgent but not panic. Driver confirms position and holds the exit lane.
After GPS detects the exit manoeuvre is complete, the route line returns to Blue #4285F4 at 6px. No animation. Recovers gracefully whether the driver takes the exit or misses it (standard reroute fires as normal).
The Google Maps Navigation SDK already provides RouteSegment and StepInfo objects with manoeuvre type and distance-to-next-step data. The route polyline is already being drawn. This feature requires splitting the polyline into colour zones at calculated distance thresholds. No new data pipeline. No SDK research. Estimated development scope: a single rendering module change.
Testing covered two tracks simultaneously: WCAG 2.2 compliance including colour blindness simulation, and a proposed A/B testing framework for live validation. Both tracks used the I-95 New York scenario as the benchmark context.
Target reduction in exit-miss events (GPS-detected reroutes after missed exit) vs. control cohort.
Target avg. distance at which driver moves to exit lane — up from current 0.3 mi.
Target % of users who correctly explain teal/purple meaning after 3 navigation sessions.
Target increase in post-session confidence rating for highway navigation sessions.
The final design adds zero new UI components, requires no new data infrastructure, works on every phone running Google Maps globally, and passes WCAG 2.2 AA on all tested backgrounds. Delivery scope is a single rendering module change in the Navigation SDK.
| Dimension | Before | After |
|---|---|---|
| Route line | Uniformly Google Blue · single weight · no state change | 3-state: Blue → Teal → Purple · progressive weight 6→8→10px |
| Driver attention | Must consciously read text prompt to know exit is near | Peripheral colour change perceived in <200ms — pre-attentive |
| Advance warning | ~0.3 mi (voice prompt, often too late for safe lane change) | 1.5 mi / ~90 seconds at highway speed (Teal prepare trigger) |
| New UI | None required or added | None — only existing banner text content changes |
| WCAG compliance | Blue on road: 3.4:1 (AA UI only) | Teal: 5.8:1 · Purple: 7.5:1 → both AA to AAA |
| Colour-blind safe | N/A — no differential colouring | Yes — teal/purple avoids red-green spectrum entirely |
| Hardware required | None (phone-only) | None — works on every phone, worldwide, from day one |
Pre-attentive first — colour registers before reading
Systems thinking — zero semantic collision with existing palette
Accessibility-first — CVD safe by design, not by addition
Progressive enhancement — adds no new UI surface
Redundancy — weight + haptic + text label as backup channels
Low surface area — change as little as possible to fix the problem
The design decision wasn't "which colours look good" — it was "which colours survive a rigorous elimination process". Starting by mapping what's already reserved, then filtering candidates through WCAG and CVD, left exactly two answers. The process found the solution; taste only chose between them.
WCAG 1.4.1 requires colour not be the only channel. But beyond compliance — in a car at speed, in bright sunlight, on a cheap screen — even good colours fail. Line weight + haptic + text ensures the signal survives all conditions. Redundancy is robustness.
User-customisable colour pairs for personal CVD preferences. Extension to roundabouts, toll plazas, and complex interchanges. CarPlay / Android Auto large-screen adaptation where the signal would be even more impactful. A controlled user study on a closed road to measure actual lane-change timing.
Google Maps has ~2 billion users. Even a 1% improvement in exit-miss rate across global highway navigation sessions represents millions of safer journeys per day. The feature requires no hardware, no SDK research, and ships to all existing users immediately upon release — unlike Google's own Live Lane Guidance, which requires car-embedded cameras.