Google Experiment: Route Coordination Reduces Congestion and Emissions in Cities

Google Research has published in Nature Cities the results of the first large-scale real-world study of using navigation platforms to improve traffic. During the experiment in 10 major US cities, the Google Maps algorithm was modified to redirect trips from congested segments to alternative routes with similar travel times. The changes affected less than 2% of all trips—only those passing through preselected problem segments (about 100 per city). The experiment lasted six months using a switchback design: days with the modified algorithm alternated with control days. The results showed a statistically significant improvement: median speed increase on target segments was about 2%, and fuel consumption reduction ranged from 0.5% to 1.0%. Across the entire network of affected segments, speed increased by an average of 0.35%, and during peak hours up to 0.5%. At the city scale, this equates to thousands of tons of CO₂e emission reductions per year per city. The effect was achieved by dispersing traffic from major bottlenecks: peripheral roads handled the additional flow while maintaining higher average speed and low emissions. The researchers emphasize that all road users benefit—both navigation app users and non-users. The work also lays the foundation for future experimentation with dynamic traffic signal control and coordination of infrastructure, vehicles, and network routing.