Most ways of getting restaurant recommendations rely on the assumption that you have the same taste in food as other people. Whether you seek advice from your food-obsessed friend, a Zagat guidebook or the restaurant critic who writes for your local newspaper, you're essentially hoping and trusting that you will like what someone else has liked in the past. Obviously, this is far from a given.
A new-ish website called Nara is trying to change this, using a concept from neuroscience to generate restaurant recommendations tailored to each user's idiosyncratic preferences. The site's bots have trawled the internet to build a database of the restaurants in the country, affixing publicly-available information on cuisine, price and quality to each restaurant's profile. Nara links similar restaurants to each other in a "neural network" that functions much like the connections in your brain. When you first sign up for the site, you're asked what kinds of restaurants you like. Nara then uses those preferences to suggest other restaurants close to the ones you like in the neural network.
"One thing we're really proud of is that the site worked just as well on day one as on day 100. Nara was ready out of the box, because of its neural network technology," CEO Thomas Copeman told The Huffington Post.
Copeman and his co-founders hope that the company's innovative approach to finding great restaurants can eventually make it a more useful alternative to sites like Yelp. They aim to eventually expand outside restaurants to hotels, shopping, music and beyond, challenging the hegemony of traditional search engines like Google. Their unofficial motto? "Stop searching, start doing."
The site has found some success since it was founded in 2010. Nara has received millions of dollars in venture capital funding, which it has used to expand to mobile and to cities across the country. The site partnered with OpenTable and GrubHub to add more functionality to the searches. And on June 4, Nara announced that it was making its recommendations national for the first time, so users can now search for restaurants anywhere in the country, not just in major metropolitan areas.
"The biggest feedback we've been getting has been, 'Nara looks great, but it's not in my city,'" Copeman said. "But now it is. No matter where you go, Nara's got your back."
At this point, though, the recommendations themselves leave something to be desired.
When I searched for cheap restaurants in my neighborhood, Brooklyn's Clinton Hill, the top several choices were spot-on: the best slice joint, a good burrito spot, an anonymous-looking deli with killer sandwiches -- plus an interesting-looking wine bar I've never heard of. But the site suggests McDonald's and White Castle a bit further down. And there's no clear demarcation between strong recommendations and faint ones. Nara, unlike something like Netflix, doesn't display the number of stars it thinks I'll rate a given restaurant; it just lists it, without explanation, as a recommended eatery.
The problem seems to be that Nara straddles the line too ambiguously between search engine and recommendation generator. Right now, it's less comprehensive than Google, but less transparent in its evaluations of restaurants than a review site. At this point, I'm certainly less likely to trust it than a knowledgeable friend or critic.
That said, Tuesday's news that the site was going truly national adds serious value to the product -- now, Nara's recommendations will transcend geography. The company's neural network understands New Yorkers just as well in San Francisco as in Manhattan. So if I travel somewhere I've never been, where I know no food lovers, I might just log on to Nara and see what it recommends.