Why Google Maps Fails in Amsterdam

TL;DR

Google Maps’ cycling routing in Amsterdam often fails because its algorithms are built around car-centric US assumptions rather than Dutch bicycle-priority route networks.

Briefing Cornell Notes

Briefing

Google Maps often routes cyclists in Amsterdam onto car-dominated streets because its navigation logic is built around American driving assumptions—so it misses the Netherlands’ core design strategy: separate car traffic from bikes not just with lanes, but with different route networks. The result is directions that may still get a rider to the destination, yet frequently choose the least enjoyable and least safe-feeling path for cycling.

The mismatch starts with how US cities are typically planned. In many North American street networks, cars are allowed almost everywhere, and traffic engineering often treats the street grid as a way to keep cars moving even when individual roads get blocked or congested. That mindset feeds directly into navigation apps: real-time rerouting is optimized for shaving minutes for drivers, and “faster” usually means using arterial roads. Cycling in that environment becomes an accommodation problem—riders must fit into a system designed to move as many vehicles as possible.

Amsterdam flips the priorities. Dutch cities aim to move people efficiently while protecting livability: streets are managed to reduce noise, pollution, danger, and space consumption from cars. A key tool is the “hoofdnet” (main network) for through car traffic, with “plusnet” roads prioritizing cars. Crucially, cycling has its own prioritized network with minimal overlap with car plusnets. Safety comes from “ontvlechten” (unbundling/disentangling): driving routes and cycling routes are designed to be fundamentally separate, so cyclists are likely to encounter far less motor-vehicle traffic.

Google Maps doesn’t understand that separation. It generates bike directions using algorithms that assume cyclists will follow routes similar to cars, even when Amsterdam offers better bicycle-optimized paths. The transcript highlights several practical failures: one-way street rules are often not correctly represented for cycling, and routes tend to return riders to major roads quickly—where bike lanes exist—rather than using quieter residential streets that usually require fewer stops.

The comfort-and-safety logic also diverges. In the US and Canada, protected bike lanes often signal the “least dangerous” option because many streets are genuinely hazardous. Google Maps appears to treat those lanes as a universal safety proxy. In Amsterdam, however, once riders leave the major roads, car traffic is already low and streets are designed to keep speeds down (including a city-wide default of 30 km/h and reduced parking). That means protected lanes can be unnecessary or even worse—narrow lanes can become crowded and slow.

Beyond cycling, the transcript argues Google Maps carries a broader car-centric bias in how it visually represents neighborhoods. Streets that are pedestrian-heavy and not meant for through driving can appear smaller or nearly invisible, while car-oriented corridors are emphasized. The piece also criticizes travel-time estimates: Google may use historical data to predict transit and walking times, but not similarly realistic driving factors like parking availability or taxi wait times. In a city where driving into the center is often a poor choice, those optimistic estimates can steer people toward cars, contributing to weekend traffic lines.

The takeaway is not that Google Maps is useless, but that it can be actively misleading in a human-scale city. For Amsterdam cycling, the transcript points to Dutch-specific route planning tools—especially Fietsersbond’s route options and junction-based knooppunten networks—as examples of routing that treats cycling as a distinct mobility experience with different priorities than driving.

Cornell Notes

Amsterdam’s cycling directions often fail because Google Maps is optimized for car networks and American driving assumptions. Dutch cities prioritize livability and safety by separating car and bike routes through systems like hoofdnet/plusnet and “ontvlechten,” which keeps cyclists away from motor-vehicle traffic. Google Maps largely ignores those route networks, misreads cycling access on one-way streets, and tends to push riders onto major roads with bike lanes—sometimes the least comfortable option in a city where car traffic is already reduced. The transcript also argues Google Maps’ broader mapping and time estimates skew toward driving by emphasizing car-relevant street importance and using overly optimistic assumptions for parking and taxi waits. For better results, Dutch tools like Fietsersbond and knooppunten planners align routing with cycling-specific goals.

Why does Google Maps’ cycling routing often feel wrong in Amsterdam even when it still reaches the destination?

Because its bike routing logic is built on the same assumptions used for driving: it treats cycling routes as if they should resemble car routes. Amsterdam’s safety model relies on “ontvlechten” (unbundling/disentangling), where driving and cycling routes are designed to be fundamentally separate, with cycling prioritized on networks that have little overlap with car plusnets. Google Maps doesn’t properly account for plusnets and modal design choices, so it can route cyclists back onto car-heavy roads or onto streets that are inconvenient for bikes (e.g., frequent traffic lights) rather than using the quieter residential streets that typically require fewer stops.

How do Dutch street networks reduce car-bike conflict compared with typical North American approaches?

North American planning often assumes cars can use most streets, so traffic engineering focuses on keeping vehicles moving via rerouting across a grid. Amsterdam instead manages car traffic to protect livability: cars are routed along designated networks (hoofdnet/plusnet) away from residential and shopping streets. Cycling has its own prioritized network with minimal overlap with car plusnets, and “ontvlechten” separates the routes themselves—so cyclists are likely to ride where car traffic is low. Protected lanes exist, but the Netherlands goes further by designing route separation, not just lane separation.

What specific routing behaviors does the transcript criticize in Google Maps for Amsterdam cycling?

Several patterns are called out: (1) Google Maps lacks understanding of plusnets, so it uses US-style algorithms that assume bike and car paths are similar. (2) One-way street handling is often wrong for cyclists—Google can route riders onto roads full of cars instead of using the correct cycling access. (3) Even when street access is configured correctly, Google Maps tends to prefer major roads with bike lanes sooner than residential streets. The transcript also contrasts Amsterdam’s low-stop environment (few stop signs, yields and raised intersections) with North America’s stop-sign-heavy design, arguing Google’s preference for major roads with traffic lights can be worse than Amsterdam’s side-street options.

Why might “protected bike lanes” lead to worse routes in Amsterdam than in the US or Canada?

In the US/Canada, protected lanes often mark the safest available corridors because many streets are dangerous and high-speed. Google Maps appears to treat those lanes as a safety signal. Amsterdam’s context is different: car traffic is already reduced through measures like removing parking space, lowering default speeds to 30 km/h, and restricting through traffic. With fewer cars, cyclists often don’t need protected lanes on most streets. The transcript adds that protected lanes can become crowded and narrow, making them less comfortable than quieter routes without heavy motor traffic.

What is a modal filter, and how does it affect routing and neighborhood life?

A modal filter restricts certain modes while allowing others. In the transcript, Amsterdam uses modal filters to stop cars from using a street as a through route while still permitting walking and cycling (and sometimes access to homes and businesses). This reduces car traffic and improves safety and street usability. The transcript contrasts this with an American approach that keeps through traffic out by making streets winding and inefficient for cars—an approach that can make cycling worse too, because cyclists must follow the same path constraints.

How does the transcript argue Google Maps can mislead drivers in Amsterdam?

It claims Google’s travel-time and destination emphasis can be car-biased. Examples include: driving and taxi times being underestimated because Google uses historical data for transit and walking but not similarly realistic factors for driving such as parking search time or actual taxi wait times. It also notes that Google often shows transit as slower than driving and may claim “plenty of parking” where driving should be avoided. The transcript connects those estimates to real behavior—drivers choosing to drive after typing “Centrum” into navigation and reporting long trips for short distances—plus weekend traffic buildup in the center.

Review Questions

What planning concept in Dutch cities (“ontvlechten”) changes the relationship between car routes and bike routes, and why does that break Google Maps’ assumptions?
List two reasons the transcript gives for why protected bike lanes can be a less optimal choice in Amsterdam than in North America.
How do modal filters differ from winding street designs as a strategy to reduce through traffic, and what does that mean for cyclists?

Key Points

1
Google Maps’ cycling routing in Amsterdam often fails because its algorithms are built around car-centric US assumptions rather than Dutch bicycle-priority route networks.
2
Dutch cities separate car and bike movement through “ontvlechten,” meaning cyclists often ride on routes with far less motor-vehicle traffic by design.
3
Amsterdam’s hoofdnet/plusnet system routes through car traffic away from neighborhoods, while cycling has its own prioritized network with minimal overlap for safety.
4
Google Maps can mis-handle cycling access on one-way streets and tends to steer cyclists back onto major roads with bike lanes, even when residential routes are quieter and require fewer stops.
5
Protected bike lanes are treated as a safety proxy by Google Maps, but in Amsterdam they can be unnecessary or less comfortable due to low car volumes and lane crowding.
6
The transcript argues Google Maps’ neighborhood visualization and travel-time estimates can be car-biased, underestimating driving friction like parking search and taxi wait times.
7
For Amsterdam cycling, Dutch-specific tools such as Fietsersbond route planning and knooppunten junction networks better match cycling priorities like fewer stops, scenic comfort, and car-restricted routing.

Highlights

Amsterdam’s safety model isn’t just “bike lanes next to roads”—it’s route separation (“ontvlechten”), so cyclists are kept away from car traffic by how the city is planned.

Google Maps’ bike directions often mirror car routing logic, which can send riders onto major roads with traffic lights instead of the quieter, low-stop residential routes.

Protected bike lanes can be a worse choice in Amsterdam than in the US because car traffic is already reduced and narrow lanes can become crowded.

Google Maps’ driving time estimates may be overly optimistic for city conditions like parking and taxi waits, nudging people toward cars in places where alternatives are better.

Topics

Cycling Navigation
Amsterdam Street Design
Car Traffic Management
Route Planning
Mapping Bias