Hopper, With $8M in New VC Bucks, Looks to Leapfrog Online Travel Search Via Big Data
In Baltimore, a “hopper” is a young, street-level drug dealer (as devotees of The Wire can tell you). In Montreal, Hopper is a young travel search company. In Boston, well, we’ll see what happens in Boston.
Hopper is announcing today an $8 million financing round led by Atlas Venture, with previous investor Brightspark Ventures also participating. The company started in Montreal in 2007 but says it is moving its headquarters to Cambridge, MA, soon. The reason?
“We’re making a big bet on the talent pool,” says co-founder and CEO Frederic Lalonde. The company is currently scouting office spaces around Kendall Square and Central Square, and is looking to hire about 15 people, mostly engineers, he says. With a local ecosystem that includes online travel companies such as Kayak, TripAdvisor, Goby, and Google/ITA Software, and “big data” firms like IBM/Netezza, HP/Vertica, and EMC, (my examples, not his), Lalonde hopes to find a “particular kind of programming geek” well-suited for Hopper’s technology challenges.
The idea behind Hopper is to take natural language travel-search queries—things like “best beaches in Spain” or “scuba diving in the Caribbean”—and return a list of places, as well as flight and hotel options for each place, ranked according to measures of quality, convenience, and cost. It’s a more open-ended, “discovery” type of search than what has become standard on itinerary comparison sites like Kayak, Bing Travel, Orbitz, and newer sites like Hipmunk, InsideTrip, WaySavvy, and Yapta.
In other words, if you’re looking for a flight from A to B (and a hotel to stay in), there are plenty of other sites to help you do that. “Travel is a complex discovery process,” says Lalonde. “We’re working on the depth, quality, and intelligence of the search.”
He certainly knows the sector. Lalonde and co-founder Joost Ouwerkerk came from travel firm Expedia (NASDAQ: EXPE), which bought Lalonde’s previous company, Newtrade Technologies, in 2002. Together with co-founder Sebastien Rainville, the Hopper team plans to move to Boston, but they aren’t saying exactly when yet. They will also keep some operations in Montreal—notably the technical infrastructure and servers that crunch the firm’s travel data—because it’s much cheaper to do it there, Lalonde says.
Here’s a little more about how Hopper works. Say you type in “scuba diving Caribbean.” The site will access a “giant statistical grid of user information” that takes into account all mentions of relevant scuba spots—from articles, blogs, forums, reviews, social media, and so forth—and returns a list that’s ranked according to those mentions, but also things like distance, flight costs, and time of year, Lalonde says. The goal is to do all of that in less than a second, he says.
The whole approach requires some serious computing power. Hopper’s database includes information from 100 million Web pages. The finer technical details elude me, but the company is applying cluster computing techniques to build and analyze what it calls the world’s largest database of travel information. Throw in some machine learning and advanced database technologies to make things fast, and you have a pretty sophisticated machine on your hands. (Interestingly, Lalonde says understanding the sentiment of consumer reviews plays a relatively small role in the ranking algorithm; it’s more important that the New York Times reviewed a particular hotel, for example.)
Much of this sounds a lot like what Google already does, or is trying to do—especially with its recent acquisition of ITA. So what’s Hopper’s plan to beat the search giant in travel?
“Google will always be about search, and ITA is good at flight [search],” Lalonde says. “We’re not in this to take away from the general search process. But we believe the discovery aspect [of travel] is entirely unique. We’re trying to inspire the user and guide them through the discovery process.” What’s more, he adds, “There are tons of companies that compete well against Google. We think you stand a good chance of making an impression on the consumer side if you find a place where Google’s strengths are actually weaknesses.”
That’s much easier said than done, of course. Plenty of startups say they do better than Google for particular kinds of search—usually because of a computing advantage plus specializing in those types of queries—yet they all face big marketing challenges. For now, Hopper is “very focused on building the engineering team and scaling up the data,” Lalonde says. That will be challenging enough, without even worrying about making money just yet. (He says the business model is based on referral fees.) The company is currently running private beta trials and getting feedback from a few thousand users.
And in case you’re wondering, yes, Lalonde has heard about the ongoing engineering talent crunch in Boston, especially for startups—though it’s certainly not unique to Boston. But he’s confident he’ll be able to recruit some top talent once people see what Hopper is building.
“The problems we’re solving are unique,” he says.