Nara Pockets $4M for Neuroscience-driven, Personalized Web Discovery
Tom Copeman says he moved to Boston from Southern California in 2010 to be near the best and brightest in neuroscience and computer science. He wanted to tap East Coast PhDs to tackle a tech problem that’s decidedly not as associated with Boston: consumer Web search.
“There’s so much inundation with information on the Web and we spend so much time trying to figure out what we’re looking for,” Copeman says. “A lot of good intentions end up getting abandoned. People fall back into old habits, what they know. But I’m really passionate about discovery.”
He enlisted Nathan Wilson—-with an MIT doctorate in brain and cognitive sciences and a Cornell master’s in computer scientist in artificial intelligence—as the chief technology officer of his company, Nara Logics. Now, the Cambridge, MA-based startup is announcing that its technology is live in the form of its Nara discovery platform.
It’s also revealing that it has closed $4 million in Series A funding to support its mission to make the Web more relevant and easily navigable for consumers. The money comes from Peter de Roetth and other angels.
“Over the last two years, we’ve been hunkered down in stealth mode, building and growing the team, and developing a personalization algorithm that’s quite complex and quite exciting,” says Copeman.
Nara offers a service that’s Pandora meets Yelp—in Copeman’s words—to help consumers discover things on the Web based on their existing tastes, starting with restaurants.
It starts by analyzing structured data on the Web for a given restaurant—like hours, location, and cuisine type—and confirms the information to form a node. As more restaurant properties tracked by the algorithm show similar properties, they develop “nodal networks,” Copeman says. The engine can also overlay that with inferences from other less structured data, like what people are saying about a restaurant in a review, and abstracts it all to make completely new recommendations to consumers based on information they’ve supplied about restaurants they already like. “That’s where the personalization engine starts to work its magic,” he says.
For each city they search, users are directed to a page with dozens of restaurant listings, each with information on food type, price range, location, style, and more. If they’ve already been to a restaurant they can give each restaurant their approval or disapproval or dial down more deeply to give ratings on facets of the restaurant like its food, vibe, or service. Like Pandora, the engine becomes smarter the more consumers interact with it on this level. (I gave it a try, liking three restaurants that I’ve been to, and the engine instructed me to reload the page with fresher results based on the feedback I provided). This also plays into Nara’s function as a “neural network,” which is “basically a system that computes and thinks like the brain does,” in Copeman’s words.
“When you give it thumbs up or thumbs down it sends a sharper signal back to it,” he says. (Now you see why a neuroscience background would be important in all of this)
Beyond the neuroscience bent, Nara is also … Next Page »