Foursquare has made it no secret that it has a pool of users who aren't all that interested in checking in. Instead of telling Foursquare where they go, they're asking Foursquare to tell them where they should go, through recommendations powered by the service's Explore search tool feature.
Starting today, Foursquare is opening up Explore to people who aren't registered users, meaning anyone can use the tool to discover local recommendations right off its redesigned homepage. It also means Foursquare is about to make a direct play to win over a new audience that extends much further than its current 25 million users, who are currently making one million search queries a day. That number's nothing to scoff at, but making Explore available to anyone, regardless of if they've ever used Foursquare, helps extend Foursquare's potential reach and, ultimately, helps it posit itself as less of a social-based check-in service and more of an all-purpose local search tool to rival efforts from Google, Bing, and Yelp.
But how do you make really good recommendations for someone you effectively know nothing about? Foursquare's head of search Andrew Hogue tells Fast Company that Explore starts by looking at universal signals from current users. So in order to determine what to suggest to a brand-new visitor to Foursquare.com, the Explore algorithm will, naturally, consider whether a place is popular. But it will also consider subtler signals, such as whether a place has a lot of repeat customers and whether users leave positive or negative tips. Explore also tries to determine which users are true connoisseurs—for example, someone who checks into a lot of gourmet coffee shops and writes popular tips—then suggests the venues that tend to attract those experts.
"What’s nice about these signals is they’re independent of any given user ever having checked in somewhere," Hogue says. "We can still do a good job of recommending things even if we don’t know anything about an individual user."
Hogue estimates just 10% to 15% of searches on Foursquare are for specific places. Much more often, users are searching within broader categories, such as "sushi" or "free Wi-Fi." That low percentage is good, he says, because it means people are actually finding Foursquare's recommendations useful. In fact, they're working pretty well: Foursquare says 20% of people who search for places to go using Explore end up checking in at a spot it suggests within three days.
Hogue says Foursquare is experimenting with other, non-social signals to figure out how to become just as useful a resource to people who don't have friends or activity histories on the service. It already takes certain factors such as the time of day into account—it won't recommend a bar to anyone at 8 in the morning. But it's also looking at weather data, so in the future recommendations will also consider where users are checking in when it's nice out, or cold out.
"I think it’s good that we started out as service that was about logging in and going out and being at a place, because we really do get these really amazing, visceral signals from people who are out having experiences in the real world," he says. "If we didn’t have that, we’d have a much sparser idea of what was going on in the world, and we wouldn’t be able to turn around and build this website that's really great for everybody else."
[Image: Flickr user Jan Kromer]