Dinner Lab is not your typical restaurant, and the secret ingredient behind its rapid growth may come as a surprise to many foodies: data.

The nearly two-year-old startup, which hosts pop-up dinners in unique sites across 10 cities, creates about a dozen original recipes each week based on feedback provided by its members.

Like the taste of confit duck but not its cured green tomato? Or maybe you felt that turtle swamp chowder should have been a few degrees warmer. Now you can tell someone – and have them actually listen.

Chefs who partner with Dinner Lab get rated on five elements for each of the dishes they prepare across what’s normally a five- or six-course meal. Those reviews combined with psychographic details that individual patrons provide then get analyzed to pinpoint which ingredients New Yorkers might be growing tired of (kale!), what Los Angelenos are eager to try (cauliflower!) and cuisines that cities are seriously lacking (Nashville: what’s up with the dearth of great Asian restaurants?).

All told, a single night of dinner service typically yields 750 to 1,000 data points—and that’s just from one of the 10 cities that Dinner Lab currently operates within, which includes New York, Los Angeles, Chicago and Washington, D.C. Next up: maybe Toronto.

“It’s sort of the Moneyball for food,” says Brian Bordainick, a former food service industry worker who co-founded and runs Dinner Lab as its chief executive officer. “The question we ask ourselves is: What can we do that’s unique, leveraging each community in each city?”

In this era of “big data” and “fast data,” where plenty of companies are harnessing bits of information to boost efficiencies across everything from their product supply chain to customer service, restaurants have been particularly slow to adapt. When they do, it’s usually to track food and beverage sales. Many eateries, of course, track online reviews on sites like Yelp or OpenTable but doing so alone “won’t make or break your business, because there are many other factors that go into running a successful restaurant,” says Cenk Fikri, founder of a boutique hospitality design and consultancy studio based in New York City.

Restaurants that fail are plagued by cash flow issues and too-high overhead. Dinner Lab, by comparison, addresses this with a member-driven model where approximately 15,000 foodies have committed to paying $100 to $175 each year plus an additional $50 to $85 for each dinner (including drinks, tax and gratuity). By offering a mobile kitchen hosted in various locations, its overhead costs are considerably less than a traditional restaurant.

In fact, the underground supper club experience is part of its appeal. Quirkier spots have included a helipad in New Orleans, a motorcycle dealership in Austin and a banjo factory in Nashville—places, which on their own, attract diners who want a different kind of experience each time they eat in addition to an ever-changing assortment of recipes. “People want something unique, and it’s a unique concept,” says Brad Bloom, director of sales at CHD Expert, a data management firm for the food service industry.  

Hatched in New Orleans, where Bordainick felt that good dinner service was hard to find after 9:30pm, Dinner Lab’s original concept was to provide late-night meals. The only problem: Midnight dinners meant patrons often showed up “inebriated,” says Bordainick. “It was a horrible idea.” Chefs also preferred to prepare food much earlier in the evening and didn’t want to cook random dishes, instead hoping to focus on passion projects where they could experiment with, say, finding the perfect green onion croquette accompaniment to alligator. The business model evolved and “through that process, chefs started asking us for feedback,” says Bordainick.

Members who provide feedback prefer to do so through old fashion means—paper and pen, instead of through a smartphone or tablet—though their responses get transferred into computer software. The detailed ratings help Dinner Lab’s network of chefs craft dishes for different locales, and Fikri sees these nuanced reviews as more “unbiased” than what restaurants typically get when a server or manager walks up to a table and asks how the food was. (“Good” or “great” are the typical canned responses, even if these are far from the truth.) “Because you’re not holding dinner at your own restaurant, you’re able to reach out to an audience that’s untried and untested, so Dinner Lab has an advantage over just looking at your regulars everyday,” says Fikri. “It’s great in terms of testing out new markets.”

Yet Fikri warns that the feedback is focused more on the chefs’ food, rather than how well Dinner Lab’s business actually operates, which could limit Dinner Lab’s longevity as it expands into new markets. “You may hold a successful dinner party somewhere on a rooftop in New York but how does that translate to a successful business?” he says. “Of course you get feedback, but how will you incorporate that into running your business day to day?” As with all data, there are limits to the insights it can provide.  

Still, knowing exactly what diners want has helped Dinner Lab expand into smaller markets, such as Miami, Austin, Nashville and Atlanta. Revenues, meanwhile, have grown from $100,000 in 2012 to $1.8 million in 2013—and are projected to hit $15 million by the end of this year, at which point the company hopes to turn a profit. “What’s happened in those markets, and our secret, is that there’s just as big of a propensity to eat new foods but the traditional brick and mortar restaurants haven’t caught up to consumer behavior,” says Bordainick.