Fred Lalonde likes to get right to the point when talking about the impetus behind his Big Data start-up.
“The problem we’re trying to solve is it actually sucks to plan a trip,” says Lalonde, co-founder of Hopper, a Boston-based start-up that is developing a new approach to the online travel based on Big Data technologies. The company plans to launch its new service later this summer.
Hopper is also the company behind Hack/Reduce, which Lalonde describes as “the best idea we never had,” and is a founding participant in bigdata@CSAIL, the MIT/Commonwealth of Massachusetts initiative to support Big Data startups in the Boston area. Recently he stepped into theCUBE to discuss Hopper with Wikibon Chief Analyst David Vellante.
Trip planning is “a really painful experience, and over the last 10 years there hasn’t been that much innovation there,” says Lalonde, who left Expedia in 2007 to co-found Hopper to find a better answer. “The sites look the same, the experience is the same. We think it’s because the catalog is broken.”
The problem, he says, is that everything is disconnected from everything else. “Obviously you can find hotels and flights, but there’s so much more involved in planning a trip.” Travel, he says, is aspirational, it’s about discovery. Often people are going to places they have never been before, and finding out what is there that they might want to visit or learn about is very difficult.
“Most consumers have no idea of what they are missing. I can get a picture of George Clooney in a balloon shirt wearing sunglasses in under a second on Google. But trying to find a vacation rental near a golf course takes four hours,” says Lalonde
The problem is that if you put a destination such as Dallas into a search engine, you get huge numbers of Web sites but no way to organize them or make a travel plan. The analogy he makes is to music sites. If you put the name of a performer into iTunes you get a list of albums with individual songs listed under them. Imagine if you just got Web sites for each song. Buying the music you wanted would take hours rather than minutes.
Hopper’s vision is to build custom pages of travel information on the fly from a huge catalog of individual travel Web sites in HBase. But instead of just listing sites, Hopper will organize them logically around the consumer’s travel plan and interests, making planning and booking the trip as easy as buying music.
Hopper was actually founded in 2007, before the term “Big Data” existed, before Hadoop. “We started with HBase and Cassandra in 2009. back then, none of this stuff worked. We spent a good two years just trying to get the database to support the catalog.”
In the meantime, Hopper built a catalog of a half billion travel Web sites. It also moved its headquarters from Montreal to Boston, attracted by the large pool of Big Data talent from MIT and other Boston-area universities and the relative lack of competition for that talent, compared to Palo Alto. And the lower priced housing market didn’t hurt, Lalonde admits.
And Hopper is doing its part to encourage that Big Data community with Hack/Reduce, a project Hopper stumbled on almost accidentally while trying to recruit technical staff in Montreal. “We couldn’t get the super geeks to come to events to talk to recruiters because they just don’t do that,” he said. Then someone in the group suggested that they simply make the resources available to play with Big Data and see who showed up. “We figured four guys would show up and one would bring a friend. Back then there were four of us in the company, so if we hired one person, that’s a 20% increase.”
That first Hack/Reduce event actually drew 70 hackers. “The crossover was crazy – people from finance, hipsters, front-end guys, back-end guys, all technologists, and nobody knew anybody else. By the end of the day, people were writing working code.” The trick is that Big Data development requires large servers; student hackers can’t do it in their dorm rooms on their laptops. So this was a big opportunity.
When Hopper moved to Boston it tried the same thing in Cambridge, and 100 people showed up. “Then we realized that we needed a permanent space.” So Hopper’s next step is to create that space and a curriculum to go with it, with backing from several sources, including Microsoft, IBM, and local VCs such as Atlas Venture Development Corp.
“I think all the Big Data companies are going to get into this. We want to make the resources available, not just for startups but also for people with ideas for community projects, analyzing political and social data. Amazing things are going to happen if you give people the tools to do it. We wanted to energize the Big Data community, and this already has.”