The New Epicurious Infuses Recipes With Google And Tinder

Leave it to the recipe sites to cook up bold experiments in design.

It’s been a year of big change for food sites. NYT Cooking is making a century and a half of New York Times recipes searchable and beautiful. IBM’s Chef Watson powered by Bon Appétit is recalculating your tired pantry items into experimental recipes. And now, one of the original recipe websites, Epicurious, has launched its first major renovation in eight years.


“Based upon where we came from, you’ll understand why it’s a mic drop,” says Eric Gillin, executive director of Epicurious, when we speak on the phone about a month or so before launch. For the last year, Gillin has been spearheading a rebrand and redesign, backed by a new editorial team of writers and chefs. Together, they are rebuilding Epicurious as both a smarter recipe engine that can learn your tastes, and a snappily headlined blog that will whet your palate for more generalized culinary knowledge.

That’s the vision at least. Today, Epicurious has transitioned from a jumbled wall of recipes and advertisements to a cleanly presented grid of delectable recipes and food stories. But I’d say Epicurious has realized about 70% of Gillin’s gleeful hyperbole in a site that is trending for the better. Here’s what’s notable and new:


Search Is Front And Center

Search used to be a term reserved for Google–now it’s the core of the new Epicurious, which greets you with the words “I want to cook” and an empty text field that you fill in like a culinary Mad Lib.

In the old Epicurious, the search box was tucked away in the upper right hand corner of the screen, 1990s-web style. “Somehow 40% of users found it in all the clutter to use it,” Gillin says, “so now we put it front and center.”

But looking like Google would be lame. So the Epicurious staff photographed a series of “moments”–food porn spreads that surround the search box to greet you when you arrive at the page.



There’s A Sprinkling Of Machine Learning

“The problem with recipes are there are too many of them,” Gillin says. “You don’t need a million recipes. You need one recipe for tonight.”

Epicurious’s efforts in promoting and improving search are part of that. But the team is also investing in machine learning, algorithms that learn what you like over time so recipes are better suited to your tastes (and may even be based on your family’s dietary restrictions).

The first manifestation of this idea comes in Cook It–what Gillin calls a Tinder for recipes. The system takes you through five recipes, like for a Curried Pea Frittata with Fresh Tomato Chutney. You answer if it’s something you would make or not. When you’re finished, the system offers up five recipes that it thinks you’ll like.

“We may find some sort of weird connections, like people who love tomatoes love eggs,” Gillin says, “or people who cook a lot of breakfast don’t like recipes that take a lot of time.”

Eventually, however, Gillin would like to personalize Epicurious’s morning newsletter, and even the homepage you visit. A single guy is going to cook for himself differently compared with a father of three children. Food allergies and other dietary restrictions could inform this personalization, too. Someone who doesn’t eat gluten would see gluten-free recipes on the page more often, and when gluten-free recipes appear, they could be marked with a special icon. Currently, the site doesn’t do these things, but they’re pieces in development for the future.


They’re good ideas that are similar to the way supermarkets work today, with aisles dedicated to special foods that are marked by extra tags and logos. But to be implemented, it will take more time than the past year of Epicurious’s development allowed.


Epicurious Recommends Food Based Upon The Weather

Some of Epicurious’s updates are–pun intended–just cheesy. A new Weather Channel widget called the Food Forecast on the lefthand side of the page recommends a dish based upon the weather.

“[The Weather Channel] has all this weird information like, when it’s hot in Dallas, string cheese sales go up,” Gillin says. “Don’t ask me why or how.”

After 19 inches of snow had fallen in Chicago, it recommended a “lighter” beet salad with goat cheese for me. Was that the right call for my next meal? And was that what I was supposed to be craving or what I was supposed to eat? I like beets as much as the next guy, but while this is a fun experiment, it’s the sort of cross-promotional gimmick that undercuts the neater innovations that Epicurious is after.



Food Is Getting Smarter

Last Thanksgiving, 16 million people–5% of the U.S. population–flocked to Epicurious to hunt for recipes. For those people, Epicurious’s latest updates will offer a faster, more beautiful experience. But when you look at the site’s updates in light of what’s been going on with NYT Cooking and Chef Watson, you see a bigger trend in the works: Some of the most practically grounded experimentation in web design today are in cooking websites.

Epicurious and its peers are giving us a taste of how better search can streamline our menus, and how machines can learn–and design customized experiences around–our preferences. Happy eating!


About the author

Mark Wilson is a senior writer at Fast Company who has written about design, technology, and culture for almost 15 years. His work has appeared at Gizmodo, Kotaku, PopMech, PopSci, Esquire, American Photo and Lucky Peach