With her iconic blonde top-knot wobbling on her head and the skirt of her floral dress protected by a white apron, senior associate food editor Alison Roman cut the roast pork shoulder into thin, tender slices. A little juice trickled onto the wooden cutting board—artfully, it seemed to me. Was presentation important, I asked her, even though this pork was for an editor tasting, not a photo shoot? She looked at me over her round glasses and answered, "Incredibly."
Roman, you see, was about to present the pork shoulder to Bon Appétit's editors at a recipe tasting, an almost-daily event that determines which dishes appear each month in our magazine and on our website. And though she'd begun cooking it that morning, the pork shoulder had originated, as all our recipes do, long before. Those recipes start out as ideas, get approved by the senior editors (including editor-in-chief Adam Rapoport), get fleshed out into actual dishes, get tested and tasted and tested again before being finalized, styled, and shot—they do not, as executive editor Christine Muhlke recently put it in a meeting, "wake up like this."
The process may be lengthy, and expensive, and maddening, but it guarantees that our recipes meet high standards—that they taste good and can be reproduced by regular home cooks in regular home kitchens. And at that moment, Roman's pork shoulder—which was being developed for a fall issue—was in the middle of this process.
The Idea A lot of our recipes, like our Fast, Easy, Fresh recipes or some of our straight-up recipe features, come right from the brains of our test kitchen staff: Alison Roman, senior food editor Dawn Perry, assistant food editor Claire Saffitz, test kitchen contributor Alfia Muzio, and test kitchen manager Brad Leone. (Overseeing all of our food editors is the food editor, Carla Lalli Music.) Other times, we look to chefs or other cool people in the industry, like David Tanis or April Bloomfield, for recipes. As you can imagine, the recipes we get from these talented chefs are great; but sometimes they need a little work to adapt them for home cooks. We have a very specific recipe style, and typically tailor recipes to serve 4–8 people—while chefs are in the habit of cooking for way more people at one time.
We also cut the esoteric ingredients and equipment that people are unlikely to have in their kitchens, or we recommend substitutions. "We test with regular blenders and hand-held mixers," explains Roman.
When the dish is an original recipe, like the pork shoulder, the food editors have a very clear idea of how they want something to taste: They've eaten it before, have been inspired by something at a restaurant, or have a stroke of genius. In the middle of development, the editors taste each others' dishes and discuss them. Does it need crunch? A little acid? More fat? They play around with flavors, producing different iterations of the same dish, until landing on something that's great.
In the case of the pork shoulder, Roman says, "I wanted to do something different. Usually, it's pulled and put into sandwiches. But a pork shoulder you can slice is better for a dinner party, and a lot more interesting." So she started playing around with roasting a pork shoulder on low heat until it was tender and sliceable.
The Tasting
After the ideas have been tweaked to our food editors' content, they will, as they say, "put up" their dishes at our (almost) daily 3 p.m. tasting. Editors who sit in on the tasting include Christine Muhlke, Carla Lalli Music, and any other editors (including photo editors) who are working on the story. On any given day, they'll have to taste anywhere from 1 to 20 dishes, ranging from breakfast to dessert to bar snacks, and Thai to Turkish to Italian.
The tasting is similar to a restaurant service: The table is properly set. The food has to come out as it's intended to be served. It has to be served in a timely fashion. It has to look great for the editors to get an idea of how it will look on the page. And before serving the food, the test kitchen snaps a photo of it: first, to be sure the recipe fits into the issue (more on that below), and second, to make sure the look of the dish matches with the professional photo that will run in our magazine.
Whoever puts the dish up has to sell it—hard. He or she will explain the dish, why the flavors work, a cool trick or technique they may have used, and what it's best served with. The editors take a bite or two, then weigh in. They muse about what the recipes need, or what should be taken out. For instance, our July issue's Charred and Raw Corn with Chile and Cheese recipe went through multiple iterations. The first was simply dressed with lime; the second added a sprinkling of oil and cheese; the third combined the charred corn with raw kernels. With criticism, each recipe evolves into the best version of itself.
"The hallmark of a good recipe is when we say, 'I can't wait to make this at home,' or go in for another bite after the tasting's over," says Dawn Perry. "The worst is when you serve a dish that you love and put your heart into, and they say, 'It's...fine.'"
Once the editors give the test kitchen the go-ahead, the kitchen team puts all of the reference pictures up on a wall to see if there are any holes or overlaps in the issue's recipes. If there are, the kitchen nixes the offending dish and tries something else. As for Roman's pork shoulder, it made the cut. "They just wanted me to make the sauce a little less salad-dressing-y, so instead of puréeing the vegetables, I cut them up into a chunky sauce."
Time to print the recipe in the magazine, right? Nope, first it's time to cross-test.
The Cross-Testing
Cross-testing is the process by which we make sure our recipes are easy to follow, complete, and—most important—correct. Cross-testers are given a full recipe (written by one of our food editors), and then they have to cook the recipes straight through, as it's written—no using instinct, here. If the recipe says to put a cup of salt in the cake batter, chances are that's a typo. Cross-testers are there to catch it. And those cross-testers? They could be anyone from another food editor to a print editor to a hired freelancer.
A good cross-tester can see where something is unclear or missing and note it in the recipe, just like an editor would in a story with an obvious hole in the reporting. "Ideally, a recipe is cross-tested once. That means the recipe is tight and well-written," says Perry. "This happens about 50 percent of the time." About 25 percent of the recipes need to be cross-tested twice—this is particularly true for baking recipes, because of their precise measurements—and the rest go several rounds. "There’s always a funny recipe that doesn’t quite work for unknown reasons. Could be an oven temp that's off, or a tester forgetting a step." And if something just totally does not work, the recipe gets axed.
Roman's pork shoulder, in fact, was a multiple-time test. "It kept coming out differently," she says. "We figured out that each pork shoulder was different: Even though we'd prepared it the same way every time, some would come out fatty and hard to cut, and others would come out on the drier side. All of them were delicious, but we have to test for consistency."
In the end, Roman and her test kitchen colleagues figured out a way to keep the results consistent (and consistently delicious). If you'd like to try making it yourself, just cli—oh, wait! It's not coming out until the fall, and plus the recipe needs to be edited, and copy-edited, and proofread, and we have to take a nice picture of it, too. Sound like a lot of work for one recipe? It is, but when you taste the results, you'll be glad you waited.