America’s Test Kitchen has long taken a highly scientific approach to developing recipes for home cooks. Every dish outlined in the brand's magazines, books, and TV show has been methodically and obsessively tested by a staff of cooks and tasters, who sometimes try up to 100 iterations. The method isn't cheap; each recipe costs about $11,000 to produce, explained Dan Souza, a food science expert with ATK.
The premise on which the ATK business was built is this: If home cooks follow the recipes carefully — using the exact ingredients, techniques, and equipment outlined — they'll get the "perfect," "best," "foolproof" results promised in many ATK headlines.
But taste, of course, is subjective, and not everyone believes ATK's recipes are, in fact, the best. I, for one, have found their instructions sometimes too finicky to follow. (What happens if you don't use a fifth of a teaspoon of baking powder as outlined, or you can't find the exact type of mustard or wine called for?) As New York Times food writer Melissa Clark has pointed out, they stifle the home cook "because they leave no room for options."
But the brand has managed to build a loyal following and gain millions of subscribers over the years by selling certainty in a domain where many feel lost and intimidated — the kitchen. And, for the record, I read the magazine regularly because I'm impressed by their rigor and often surprised by the counterintuitive tricks their method unearths.
Now ATK is leaving the kitchen for a new venture with the launch of the website, Cook's Science. Unlike the company's recipe-centric legacy brands (Cook’s Illustrated, Cook’s Country), which were started under ATK's recently departed and controversial founder, Christopher Kimball, the new venture will focus on telling stories about food science. "We will report in the field," said Souza, "talking to scientists, chefs, farmers and professors at universities about what they're doing — then coming back to the test kitchen."
I talked to Souza about the new site, some food science tricks he's picked up over the years, and why busting the conventional wisdom about food can lose friends and alienate people. Our conversation was edited for length and clarity.
Julia Belluz: At ATK, you take an obsessively empirical approach to cooking. Can you describe the method you typically use to optimize a recipe?
Dan Souza: We try to become experts on whatever dish or ingredient we're interested in covering. Say we're looking at doing a recipe for roasted chicken. We have lots of recipe books here, so we will pull what we think demonstrates the spectrum of recipes that a home cook might find.
We might make five of those recipes at the test kitchen and then we'll have a big group of people — over 55 test cooks, plus the editorial staff — taste and comment on them. This gives us an idea of what's out there and what a home cook could run into if they looked up a recipe for roast chicken. There will be some things in that test that we gravitate toward — say a certain brining method that produced moist meat — and so when we put together our first working recipe, we'll include that brining method. We'll also include some test kitchen common practices.
Then we go full gear with the scientific method. We start systematically testing variables using that working recipe as the base. The first test might be brine strength. We keep everything about the recipe identical except for how we brine the bird. We taste, pick a winner, and then change the working recipe to reflect that choice. Then on to the next test.
We do that 40, 60, sometimes 100 times for eight or more weeks until we find a strong recipe. Then we send the recipe out to panels of volunteers who try it out. We look for 80 percent of people to say they'd make the recipe again. If they say no, we go back to the kitchen. We have to spend roughly $11,000 on each recipe we do.
JB: Do you have any tips for home cooks who want to be more scientific in their approach to recipes?
DS: Let's say [you] make a recipe and want to play with it. The most important thing is taking really good notes. If you're going to make a change, don't make too many at one time, and keep track of all the changes. For taste testing, it's best when you can try things side by side.
So if you're looking at apples, try a few varieties, cut them up, and don't tell people which ones they are. Ask them to rate the apples based on acidity, sweetness, and firmness. We do that all the time with products, and I'm always surprised. I was certain I liked a certain brand of Mexican beer, but when I tasted a few brands blinded, I actually put [the Mexican beer] at the bottom of the pack.
JB: How does Cook's Science build on all this method?
DS: We are taking our journalistic curiosity and combining with our curiosity in the kitchen. We are going chase down stories by traveling and talking to experts in the field.
For instance, we attended a nine-day course at Penn State on the science of ice cream. We covered everything from ice cream physics to how different sweeteners affect texture, and interviewed dairy scientists. We are using our new knowledge to build better ice cream recipes for home cooks, and we'll publish those along with the story of our experience at Penn State.
We've also been reporting on the difference between cheese made from cow and goat milk, traveling to two local farms, picking cows and goats, and using their milk to make identical cheese.
JB: By using science in cooking, you're often challenging the conventional wisdom about food, like when you suggested it's actually better to cook steak when it's frozen compared with defrosting the meat. Do you ever get any blowback after these revelations?
DS: When we do, I feel like that applies a lot to really simple classic American foods, like doing scrambled eggs. A lot of people said they liked our recipe but it wasn't the way their grandmother or mother did scrambled eggs. When we published an article around the fact that you don't need boiling water to cook pasta — you can cook it at a lower temperature or start the pasta in cold water — plenty of Italians ignored that.
We try to explain as much as possible in our stories about the "whys," so that if a reader wants a different result than the one we chose, they'd have the knowledge to make that adjustment.
JB: Over the years, you've discovered some pretty cool tricks for picking the choicest versions of various ingredients. I understand you recently discovered how to identify the best hunks of Parmesan?
DS: We did some cool tests where we found that because Parmesan wheels are so large, there's a big disparity in the quality of the cheese from the exterior to the interior. You get more advanced aging at the exterior than you do in the interior. It's like heating a big roast in the oven — it heats from the outside in. So if you're going to the store, pick cheese that was cut closer to the rind. It'll be crumbly, more densely flavored, with more of those crunching, piercing crystals. Cheese at center of the wheel, which wouldn't have any of the rind on it, will be more plastic-y, cleaner, and mild.
JB: You've often noted that marinating meat doesn't do quite what you'd expect. How does science explain that?
DS: Marinating is useful, but it really only goes skin deep. So it can only go a few millimeters into your food. The only things that move into meat and make a difference in its flavor are salt and sugar. So if you have plenty of salt in your marinade, that's great. But don't expect the flavors of your marinade to penetrate inside your meat. That's magical thinking.
Most of the flavors we use in marinades are fat-soluble — as opposed to water-soluble — and meat is about 75 percent water. We know that oil and water don't mix, and it's the same thing here: Fat-soluble flavor compounds don't readily move into a water-packed piece of meat. Garlic is both fat- and water-soluble, and we've found that if a marinade has a lot of garlic in it you can taste it subtly inside the meat. But again, that shouldn't be your goal with marinating. You get great flavor on the exterior, and the salt moves in for better seasoning and moister meat.