Q&A: Doc Willoughby on Perfecting Recipes in America's Test Kitchen

Twenty years of perfecting recipes. That’s how long America’s Test Kitchen has cooked, baked and obsessed over the results. Based in Brookline, MA, it’s the test kitchen for a PBS television show of the same name, where the staffs of Cook’s Illustrated and Cook’s Country magazines work out their recipes.

How does the staff create a recipe that works every time? I asked John “Doc” Willoughby, the company’s executive editor for magazines. The Harvard grad began his career at Cook’s Illustrated when Chris Kimball founded the magazine in 1993. In 2001, he left to become executive editor at Gourmet magazine, then returned to America’s Test Kitchen in 2010. Willoughby, who writes cookbooks with co-author Chris Schlesinger, a chef, has written nine, including the award-winning The Thrill of the Grill.

Lori Galvin, executive editor of America’s Test Kitchen and a reader of this blog, sent me the company’s latest cookbook, Cook’s Illustrated Cookbook: 2,000 Recipes of 20 Years of America’s Most Trusted Food Magazine, and suggested I talk with Willoughby about the company’s process of developing and testing recipes:

Q. What do you do as executive editor of America’s Test Kitchen?

A. I’m in charge of the two magazines plus 24 special issues for newsstands. I follow along the process for each magazine, starting with ideas like, “Do readers want another roast beef recipe? If so, which kind?” Then we survey readers before doing an article.

Q. How much do you rely on readers for your content?

A. Once we decide what we want to do, we come up with a recipe and send it to readers to test. Eighty percent of those who make it have to want to make it again. Sometimes we end up never publishing the recipe because it’s too complicated or some other reason. Then we do a post publication survey where readers tell us which recipes they like the best and why.

Q. Does the test kitchen staff start with a pre-existing recipe?

A. We use a spectrum of sources. We try to find the early version of a traditional recipe, and we look in major and specialty cookbooks. We look for different approaches. The same approach to apple pie is not helpful.

Q. Then what happens?

A. The standard way to test is to take as many as five different approaches. We make those recipes and the test kitchen staff tastes and rates them, and then gives feedback for each. The cook discusses what she learned from that process and which approaches she liked. She might decide to put three approaches together in one recipe. Then she follows whatever tangents come up. The reason it takes so long is that you have to test one variable at a time.

Q. What are these variables? I read that the average America’s Test Kitchen recipe is prepared an average of 65 times, over a period of six weeks.

A. For a gingerbread cake, you’ll test it at a range of different oven temperatures, with the rack at different levels. The amount of leavener might vary, and then there’s the liquid you might use ’96 water, milk, orange juice, and different kinds of beers. Someone will ask if lager is better so you have to try different beers.

For roast beef, if the goal is super-evenly cooked juicy meat, it’s a different kind of testing. You get into a lot of science, like if you cook it to X degrees and let it sit, then put it back in the oven. Once you get started down that road, there might be 40 tests to do to get those variables right.

One of my favorite discoveries is where we microwave the bananas for the banana bread recipe. It’s a little absurd but it works. The problem is that you want the banana flavor without the banana mushiness. When a discovery like that happens, it’s exciting.

Q. Let’s say you find a technique from someone else and you use it. Do you credit that person’s recipe?

A. You try to. I was making dulce de lece, and I noticed that when people use baking soda, it worked better. It helped the browning and it had a more butterscotch-y flavor. I found a lot of recipes that used the baking soda, but if I had found that only Rick Bayless used it, then I would say so. I saw that he used cinnamon, which I hadn’t seen before, so if I use it I will credit him.

In the gingerbread recipe, the tester used a technique from a colonial recipe, so she said so. She had visited Plimouth Plantation. That also made it a better story.

Q. How did the gingerbread work out?

A. I thought the development process was crazy, just insane. But sometimes the center of the cake sank. The tester was not satisfied and testing went on for weeks. Had it been up to me, at week three, I might have said, “That’s good enough.” But the test cooks get obsessive, which is a good thing.

I made the gingerbread this weekend, for a dinner party. When I poured the batter into the pan I thought it wouldn’t work. The batter was like water. But the tester had solved the problem. (DJ: Here’s why, according to the headnote: “Incorporating the baking soda with the wet ingredients instead of the other dry ones helped to neutralize those acidic ingredients before they get incorporated into the batter and allow the baking powder to do a better job.”)

Q. I have more questions about the testing. You said the editors might taste five versions of a dish and decide what they want. Why have a goal versus the serendipity of creating a great new dish?

A. For the cook, if you leave it wide open, they could do their favorite chocolate chip cookie, and that would be it. But it helps to have a more defined goal. We say we want this kind of cookie, so go make it. It’s a group process so we have to agree at the beginning.

Q. That wasn’t how you did it at Gourmet.

A. A lot of Gourmet was done in menus, where you would have an individual cook who would create all the recipes. There was a lot of tasting as well, but it was more about “Is it good?”

Q. How much of this rigorous process is applicable to people who come up with their own recipes?

A. We have 42 cooks and we can do this. But some of it is applicable. If you’re not satisfied with your dish, the process of looking at all the recipes you can find and asking how do they do it differently and what you could do differently will take care of it.

I did an eggplant recipe for the New York Times and I made it 10 times because I couldn’t get it right. I thought three tests were enough but they weren’t. If you decide “I like this” after two tests, you don’t need to go any further. I hate salting and draining eggplant, and I kept thinking I could get away from it, but I was wrong.

I’d say that if you think of something and it seems silly, try it anyway, like microwaving bananas.

Q. How does this rigorous process of developing and testing influence you as a cookbook author?

A. I write recipes with chef Chris Schlesinger, and most of the recipes come from him. So if I make lamb kebab with Middle Eastern flavors and we both really like it, that’s that. We use recipe testers. We’re just starting a cookbook now, on grilling.

Q. It seems that a lot of your magazine recipes are about teaching technique to readers. There are sidebars that say things like “To keep baked chicken breasts juicy, poke the meat with a fork and salt each piece.” Is your job to help readers understand why a certain step or ingredient is important?

A. Yes. One reason is global and one is more practical. I think people can learn how to make a recipe work better. Two, if you don’t explain why, people won’t do it. Many people don’t actually follow a recipe.

Q. Doesn’t it make the recipe longer, to explain why?

A. Hopefully people read the stories. We put the critical information in the headnotes. We also use a sidebar to explain why the science works and why you should do it.

Q. Cook’s Illustrated puts its recipes behind a pay wall, for $34.95 per year. Why did you decide not to give away recipes, like other magazines do?

A. Because it’s a different business model than other magazines. Traditionally, magazines sell advertising space on their websites. We don’t have advertisers, and the only thing we can make money on is our product, and we can’t give it away. We do have a free website, but it’s not a recipe repository.

Q. How many paid subscriptions do you have?

A. The three company sites have a combined paid circulation of 450,000.

Q. What do you think about food bloggers who give away their recipes for free?

A. Well, they’re not going to make a lot of money! Blogs have their plusses and minuses. You can put work up without a company behind you, put up your own stuff. But not that many people have made money.

Q. Do you read any food blogs?

A. I read Smitten Kitchen, but I don’t read a whole lot regularly. I’ve found that if you follow people on Twitter, other people point out interesting things and then I go read them.

Q. Do you think there’s such a thing as an original recipe?

A. I do. I don’t think there are a million of them. But some people can come up with a combination of flavor and ingredients that no one has seen before, if you search for it and you don’t find it.

Jean-Georges Vongerichten thinks about food in a different way. Maybe the recipes have been made before at some time in history, but it’s the first time you’ve seen it, and it’s not from a culinary tradition but out of his head. Grant Achatz probably has tons of original recipes.

Q. Is it chefs, then, who are more likely to come up with something new?

A. No, cookbook authors can do the same thing. But they’re probably going to get less attention.

Q. Does Cook’s Illustrated go after people who use the recipes as their own?

A. We have gone after people who use 15 or 20 of our recipes on their websites, because then they’re starting to take something we can no longer sell because they are available for free.

Q. What is your advice for someone who wants to be paid to develop recipes?

A. Contact cookbook editors to be recipe testers. My guess is it’s easier to break into.