Tested: Google Voice Navigation for Cyclists

Staff Writer
Guided by voices through New York's manmade labyrinth

For a couple of years now, cyclists have been able to count on Google to help guide them efficiently from point A to point B, using the 330,000-plus miles of biking lines (which include bike paths, lanes and recommended routes) on Google Maps. Until very recently, though, making use of the big G's data required either a) memorizing the recommended route or b) a lot of looking back and forth between the phone and the road (which can be distracting while riding in/with traffic). That's all changed since Google announced last week that it was adding turn-by-turn, voice-guided navigation for cyclists to Androids and other supported mobile devices.

I decided to take it for a spin in that most daunting of urban landscapes, New York. If all the tall buildings, one-way streets and oddball bridge approaches couldn't fool it, nothing would. So instead of riding my usual to-and-from-work routes, I let Google lead the way. Here's what I found:

• I didn't have a convenient way to mount my phone to my handlebars, so I plugged in earbuds and shoved it in my pocket. It was nice not having the distraction of the screen in front of me, but on a couple of occasions, when Google's robotic voice barked out confusing directions, I had no visual confirmation of what it was telling me.

• It didn't know the streets as well as I do. At one point, on my way home, it tried sending me through the Brooklyn Navy Yard, a sprawling, gated industrial complex that's closed to the public. Then, going into Manhattan, it gave this confusing direction: "Turn left on Delancey Street South, then turn left on Delancey Street South." This on an east-west street. I was baffled and instead started up a route I often take, leaving Google to reroute. That seemed to really throw it, and the rest of my ride into work was a wreck: It sent me down a wide, traffic-heavy cross-town street with no bike lanes, started lagging behind where I was (giving directions to turn onto a street I'd already turned onto) and even tried sending me the wrong way down a one-way street. I guess that's why the bike directions are still in beta.

• The "human algorithm" is missing. Google uses algorithms to calculate its routes—probably millions of them each second—but it doesn't understand shortcuts, choosing different routes based on how heavily trafficked they are and which streets are more potholed or prone to collecting debris. For instance, I live near the end of a one-way street. When I'm coming home from work, I jump off my bike at the corner and walk the 30 yards to my place, or—don't tell anyone—ride the wrong way for five seconds. Google routed me to the other, farther away end of my street first. And that leads to my next point...

• Both rides took longer than my usual routes. The shortest way home I've found is a 4.9-mile route over the Manhattan Bridge that I generally cover in anywhere from 19 to 22 minutes. (What can I say? I'm consistent.) Google's route went over the Williamburg Bridge, was 5.8 miles and took 27 minutes. In the other direction, I usually ride 5.1 miles in 23 minutes. Google took me 5.5 miles in 26 minutes (and it got confused).

• It was kind of fun trying to anticipate where Google would tell me to turn, and it gave me plenty of warning—usually once a couple hundred yards out and again just before an intersection—ahead of time.

• I learned some things, like how far the end of my street is from the nearest 43-tap bar (800 feet, apparently).

Most of my gripes are with Google Maps navigation itself, not with the new, voice-guided feature, which works fine. If you're going to use it in an urban setting, it's best to have it mouted to your handlebars, so you can glance at the map if you're given an unclear direction. Also, as with all technology, it works best when you don't use it blindly. It has things—namely maps—that your brain doesn't, and your brain has things—the ability to reason and weigh other, more variable factors—that it doesn't. Take advantage of that human-technology interaction and you'll find your way wherever you're going.

H/T TreeHugger.