How the Knowledge of Locals Is Helping Google Build Better Maps

The company is recruiting millions of users to update its maps and add more detailed—and more subjective—information.

Remember when maps just showed you what a place looked like? It’s hard to imagine now, but in the mid-2000s that’s basically how it was, before smartphone map apps came along to tell you where to turn, where traffic was backing up, or where to grab a latte en route to your destination.

People’s expectations of what information a map can provide have become incredibly detailed, says Jack Menzel, product management director for local search at Google, one of the companies driving the evolution of maps. Now, when people search for a pharmacy or pull it up on Google Maps, they expect to see its hours, and not just its regular hours, but the holiday hours too. “People have this expectation that ‘I want to know that one day a year when their hours are different, and I want to know how they are different.’”

The next step, according to Menzel, is to add even more detailed and subjective information—not just which Thai restaurant in the neighborhood gets the best reviews, but whether it has a romantic vibe or is a good place to take the kids. And to do this, the company is asking for users’ help.

View Images

Clicking on a business in the Google Maps app brings up reviews, photos, and a graph showing times when it’s busy.

Detailed queries are driving Google, a company built on algorithms and big data, to try a different approach and lean more on users’ local expertise. In recent weeks, the company has rolled out a few software changes to make it easier for users to submit information and vet edits suggested by others. Another recently launched program at Google has recruited millions of people to contribute photos, reviews, and other information about the places where they live. Of course, as those user submissions grow, they’ll likely provide tons of new data that Google’s algorithms can go to work on.

It’s not the company’s first foray into crowdsourcing. Map Maker, a program launched in 2008, lets users trace roads on top of satellite imagery and add missing features. And Google didn’t invent collaborative mapping—more than two million people have contributed to OpenStreetMap, a project launched in 2004 to create maps anyone can edit and use for free, making it a popular option for humanitarian groups and other nonprofit organizations.

Like OpenStreetMap, Map Maker has been helpful for getting factual information, especially about less developed parts of the world. But Google’s new initiatives are aimed at getting a different kind of information, says Laura Slabin, director for local content and community. “We’re wanting to better understand some of the more subjective picture,” Slabin says. “What’s the vibe of the place? What’s your opinion of the place?”

View Images

These screenshots show several ways users can contribute information to Google Maps.

To that end, the company launched a program called Local Guides back in November to recruit users to submit reviews, answer questions about things like the noise level inside a restaurant, and upload photos of businesses and other places near where they live. To encourage participation, Google offers benefits like increased cloud storage and the opportunity to beta test new products. The most active contributors get a chance to participate in virtual and real-life meet-ups (something the program has in common with Yelp’s events for Elite reviewers). “We felt that we really needed to create a community to help inspire people,” says Slabin, who directs the program. So far more than four million people have signed up; 20 percent have contributed within the last week.

View Images

Participants in Google’s Local Guide program get points, and potentially rewards, for their contributions.

Uploading photos is a big part of the Local Guides program. You might see some of these user-submitted photos if you tap on a business in the Google Maps app. For now it’s up to you to decide if the place looks rowdy or romantic, but in the future you might get an algorithmic assist. “One of the things we’re really excited about as computer vision gets better is how can we make it more assistive to our users,” says Menzel. Google already uses computer vision to extract addresses, street names, and information about which turns are permitted from a given lane from the photos collected by its Street View cameras. In the future, that might extend to interior photos submitted by users, Menzel and Slabin say. “One could imagine if we’re getting a lot of internal photos of a place, and let’s say there’s candlelight; is that a signal that it’s a romantic place or maybe not good for children?” Slabin says.

Likewise, Google uses a combination of algorithms and data generated by users to create graphs of when businesses like bars and restaurants are busiest. “We compute popular times just like we compute traffic—based on the anonymized aggregated movement of people that we receive through location reporting,” says Elizabeth Davidoff, communications manager for Google Maps. “We've built up an anonymized aggregated history of what foot and car traffic is usually like on specific roads or at specific places at certain times that enables us to predict how traffic will change over time.” (Users concerned about privacy can opt out of location tracking).

Getting more data from users is the only way to meet the growing expectation that Google can answer any question, no matter how trivial, and no matter where you are, says Menzel. “The only way to answer everybody’s questions is to get people engaged in sharing what they know about their community, their part of the world.”