As a company that books more than a million guests into rooms, apartments, houses and other accommodations every month, Airbnb has a ton of data to work with about everything from the length of a person’s trip to the kind of house they prefer when they travel. Vice President of Engineering Mike Curtis told the Structure conference in San Francisco that the future is in using that data to help predict what kind of accommodation someone might like, before they have even expressed a preference.
“The really interesting things we’re working on are search and personalization and matching, so that we can make a recommendation that will be so spot on that you don’t even have to go through the process of searching through the listings,” Curtis said Wednesday. Figuring out whether a guest and a host would be a good fit is “actually a really interesting data problem,” he said.
Some of that challenge is just sorting through variables, Curtis said. For example, if a user wants to rent something for just a couple of nights instead of two or three weeks, that’s going to reduce the number of hosts available. And the same applies if a user wants to book a trip tomorrow instead of planning ahead by a few days or weeks: the universe of hosts who will want to make their rooms or homes available will undoubtedly be smaller.
But the really interesting part, the Airbnb vice president said, is using all of the other tiny pieces of data about where a user stayed and the feedback they provided to make smart recommendations about trips or destinations the user might like — all without them having to indicate a preference. In a sense, Airbnb is trying to do something similar to what Google Now does by developing so much intelligence about a user’s behavior that it can recommend something without forcing the user to search for it.
Photo by Jakub Mosur