“To us, it’s a little baffling that it hasn’t happened yet,” says Frederic Lalonde, Hopper CEO, referring to attempts to cross what he terms the “inspiration chasm.”
“If there is an infinite amount of questions, then there is an infinite amount of answers and that is where Big Data comes in,” Lalonde adds.
He hopes Hopper will be able to provide inspirational search results for very broad travel queries such as “skiing,” without any dates or destinations cited, to extremely specific ones.
“You have to be good at the infinitely large and the infinitely precise,” Lalonde says.
So Hopper, which aims to takeÂ natural language queries from consumers and provide destinations and activities with matching flights and hotels, has set itself some very big goals with its Big Data.
A demo of the solution nearly six months ago was characterized by extremely speedy responses, but primarily handled queries about Spain. In the interim, the Hopper database has become global, Lalonde says.
Another Hopper differentiator is that it uses a “push API” rather than Hopper having to query sources, such as a global distribution system, for data, Lalonde says.
Hopper currently is receiving data from an online travel agency through a push API and it could be used with flight and hotel partners, Lalonde says.
“Instead of Hopper searching via a traditional API, they ‘push’ their flight results to us and we add them to our search engine,” Lalonde says. “In turn, when a user finds the flight they are looking for on Hopper, we refer them back to the OTA to complete the booking.”
How successful Hopper will be, with its ambitious goals, remains to be seen, of course.
Founded in 2007, Hopper is currently in private beta and Lalonde declined to specify a launch timetable.
Hopper plans to keep its data processing center in Montreal, but will use part of its new $8 million in funding to relocate its headquarters to Cambridge, Mass., as it seeks tech talent and ramps up its staff from the current 12 to 25 people.
It will also expand its infrastucture to around 600 servers.
Lalonde says Hopper builds its own servers, configuring them to handle Big Data, and closely following open source documentation for the server configuration.
He says Hopper, which uses Machine Learning, and NoSQL databases, as well as Big Data processing, wants to control “its own stack” and can save 24% in data center costs by building its own servers. That sort of efficiency “matters at scale,” Lalonde says.
With the new funding, Atlas Venture partner Jeff FagnanÂ joins Lalonde, fellow cofounder Joost Ouwerkerk, PhoCusWright chairman Philip Wolf and Brightspark managing partner Sophie ForestÂ on the Hopper board.
Hopper’s total funding to date is $10 million. Brightspark participated in the previous round, as well.
Disclosure: Frederic Lalonde is chairman, co-founder and an investor in Tnooz.