Case Study 2 — Aroon's Duck and the Saturday Question
The Saturday Problem
Mae Som restaurant on Bloor Street in Toronto's Annex neighborhood serves a Saturday-night menu that is different from the rest of the week's. Saturday is the night Chef Aroon Sornprasit reserves for a tasting of seven plates, no choices, $145 per person, by reservation only. He runs it because he likes it. His regulars come back every six to eight weeks. Saturdays are usually full.
Aroon was 47 years old when he first put a duck breast on the Saturday menu. The dish was bpet yang prik thai — duck breast with a Thai black-pepper-and-garlic glaze that his grandmother had made every year for the Songkran festival in the Chiang Mai household where Aroon grew up. He had served it at home for years. Now he wanted it on his menu.
The problem was that he could not consistently cook it on a Saturday night.
A duck breast is a particular cut. The skin is thick and fatty; the meat underneath is dense and dark. To cook it well, you need two things: the skin to render and crisp, and the meat to come to medium-rare without overshooting. Traditionally, this is done by low-and-slow rendering — placing the duck skin-side down in a cold pan, slowly bringing the heat up so that the fat under the skin melts out and the skin crisps before the meat overcooks. It works. It is the technique he learned in Chiang Mai and the technique he taught his sous chef Aki when she joined the kitchen.
The problem with Saturday was the rhythm. On Saturday, eight tables of two-tops or four-tops, all eating the seven-course tasting, all needed their fourth course (the duck) at roughly the same window. Aki could not stand at the stove monitoring eight pans of slowly-rendering duck breast for 25 minutes each. The kitchen had two burners free in the relevant window and three other dishes plating during it. When she tried, two breasts came out perfect, two came out medium-well (skin on too long, meat overcooked), one came out underdone (pulled too early, skin not crisp enough), and three came out somewhere in between.
For a six-week stretch, Aroon pulled the duck breast off the menu and replaced it with a less time-sensitive dish. He thought about how to fix the rhythm.
The Conversation
He told the story over coffee one morning to Diana, his old culinary school classmate who now consulted for restaurants on operational efficiency. Diana had been doing a lot of work with Toronto restaurants on prep-ahead techniques. She listened. She thought for a minute. Then she said: "Aroon, you've been to Modernist Cuisine. You know sous vide. Why aren't you using it for the duck?"
Aroon had used sous vide, occasionally, for short ribs and oxtail. He did not use it for duck. He had a specific aesthetic problem with sous vide: he thought it produced food that tasted, sometimes, thin — without the rendered-fat character that separates a great pan-cooked duck from a merely-edible one.
Diana said: "Run a comparison."
The Comparison
The next Sunday, when the restaurant was closed, Aroon ran the experiment with Aki and his other sous chef, Marcos. He brought in six duck breasts of equal size from his usual supplier (a small farm in Ontario) and split them into three pairs.
Pair 1: Traditional pan-rendered. Skin scored, salted, started skin-side down in a cold pan, heat brought up gradually, skin rendered for 15 minutes until deeply golden, then flipped for 4 minutes to finish. Rested 5 minutes. Sliced.
Pair 2: Sous vide only. Salted, vacuum-sealed, cooked at 57°C / 135°F for 90 minutes. Removed from bag, dried, sliced without searing.
Pair 3: Sous vide then sear. Salted, vacuum-sealed, cooked at 57°C / 135°F for 90 minutes. Removed from bag, surface dried thoroughly, then seared skin-side down only in a screaming-hot cast-iron pan with no added fat — the skin's own rendered fat became the cooking medium — for 90 seconds, until the skin was deeply crisp and golden. Rested 2 minutes. Sliced.
The three of them tasted blind. Aki and Marcos had not been told what was what. Aroon did the cutting and plating; Diana scored.
Pair 1 (traditional): Aki said this was the dish she knew. "This tastes like duck breast." The skin was rendered and crisp. The meat was medium-rare in the center, slightly more cooked at the edges. The fat had carried flavor.
Pair 2 (sous vide only, no sear): Marcos said the meat was "perfect" but "naked." The skin was rubbery and uncooked. The meat was uniformly medium-rare from edge to edge, juicy, but without the contrast and the rendered-fat character. Aki said "you can't serve this in a restaurant."
Pair 3 (sous vide + sear): All three of them stopped talking when they tasted it. The skin was — possibly — more crisp than Pair 1, because the surface had been pre-dried by 90 minutes in the bag and the sear was therefore more efficient. The meat was perfectly medium-rare from edge to edge — as it was in Pair 2 — but now with the rendered-fat carry that Pair 2 had been missing. Aroon scored Pair 3 a hair above Pair 1 on technique. Aki scored them tied. Marcos scored Pair 3 above Pair 1.
Aroon ran the experiment again the next Sunday with different breasts. Same result. Pair 3 was reproducibly at least as good as the traditional method, and the technique was hands-off until the final 90-second sear.
The Saturday Solution
Two months later, the Mae Som Saturday tasting menu had a new procedure for the duck breast.
Friday night: Aki bagged eight duck breasts and dropped them in the bath at 57°C / 135°F. She started a 14-hour timer (longer than necessary, but the duck doesn't suffer from extended bath time at this temperature, and Friday-overnight is the most convenient window). A timer alerted her in the morning to check the bath water level and confirm temperature. When Saturday-morning prep began, the bag had been in the bath for 12 hours; Aki pulled four breasts out, reserved them in the walk-in fridge bagged, and left four in the bath running until service.
Saturday service, 8 p.m.: Aki pulled each pre-cooked breast from the bath as the table progressed through earlier courses. She unbagged, dried thoroughly, and seared skin-side down in a screaming-hot pan for 90 seconds — a procedure that took less than two minutes per plate, including the rest. The skin was crisp. The meat was medium-rare edge to edge. Eight tables ate the duck within a 20-minute service window, with no failures.
Aroon, who tracked every plate sent out, recorded zero send-backs of the duck breast over the next six months of Saturdays. He had previously run roughly a 1-in-15 send-back rate on the duck breast when cooked traditionally during a busy Saturday rhythm.
The Cultural Question
Aroon's grandmother, when he called her in Chiang Mai to tell her about the new technique, was unimpressed at first. "Why do you need a machine," she asked. "I cooked duck for fifty years without a machine."
He said: "You cooked one duck at a time."
She thought about this. "How does it taste?"
"It tastes like the duck you taught me to cook," he said. "But I can do it for eight tables on the same Saturday and they all taste the same."
She thought about this for longer. Then she said: "If the duck tastes right, you can use whatever you need to use to make it." She paused. "Don't tell my sister."
Aroon told the story at his next industry panel in Toronto: that the technique was a tool, not a betrayal of the food, because what mattered was whether the duck on the plate was the duck his grandmother had taught him to cook. His grandmother's recipe was not "the technique." His grandmother's recipe was the duck on the plate. The technique was what got you there, and one technique was as legitimate as the next if the result honored the food.
This is, in a sentence, the deep argument for sous vide as a technique. It does not replace traditional cooking. It is one of the traditional cooking methods now, alongside the others, with its own strengths and its own limitations. The cook who uses it well uses it where it serves the food, and uses something else everywhere else.
Analyze This
Aroon's procedure resolves a throughput problem — the kitchen could not serve duck breast for eight tables in the same 20-minute window using traditional pan-rendering. It resolves it by uncoupling the slow protein-cooking step from the fast skin-finishing step, allowing the slow step to be done hours in advance for any number of breasts at once.
Identify two other restaurant dishes where this same uncoupling principle could be applied. For each, describe:
- The slow step that's hard to scale on the line.
- The fast finishing step that adds the textural or flavor element traditionally produced by the slow step (e.g., crisp skin, Maillard sear, char marks).
- The sous vide protocol — temperature and time — that you would propose for the slow step.
Bonus prompt: Aroon's grandmother said "if the duck tastes right, you can use whatever you need to use." Identify one cooking tradition or dish where a modern technique has been used in a similar way to enable scale or precision without changing the fundamental flavor profile of the original. Identify another where the modern technique did change the flavor profile, and discuss whether that change was a loss, a gain, or a different food entirely.