Which dishes to remove from the menu to gain profitability: +4.1 EBITDA points by cutting 11 items with the Masterestaurant Standard Recipe Generator

The first Monday's inventory: 11,800 USD sitting in the walk-in

Eleven thousand eight hundred dollars parked in the walk-in, inside a trattoria billing 620,000 a year, works out to 2.4 inventory turns a month where a healthy 14-table, 38-seat house holds four or five. We took that number on the first Monday, before the weekly order came in, and it explained the distance between paper and bank in one stroke: the costing sheets said 33.1% food cost, reality read 39.9%. Almost seven points, and not one dollar stolen. Nine years the owner had run the place with a 21.40 USD average check, a full room Thursday through Sunday, 78% of sales on the floor, 22% in owned delivery, and a P&L reporting 9% operating profit built without closing inventory, kitchen waste dissolved inside purchase cost. Median limited-service food cost closed 2024 at 32.4% according to the National Restaurant Association (2025).

Twenty-three dishes that never sold a single unit a day

Twenty-three of the 46 dishes sold at least one unit daily; the other twenty-three split 11% of the sales mix between all of them. That dead half dragged 34 exclusive ingredients along with it, items appearing in no other recipe, bought in the supplier's minimum pack and thrown out later at expiry. There sits the cost no menu declares: tied-up capital plus scheduled waste. Casual dining runs 30% to 34% food cost according to the National Restaurant Association, and this kitchen operated nearly six points above its own concept without the chef touching a single recipe. When you carry a menu longer than your cold storage, the problem does not live in purchasing, it lives in DESIGN. And design gets fixed in one afternoon; purchasing, never.

The cut criterion is not food cost, it is contribution margin in dollars

A dish dies when its contribution margin in dollars falls below the menu's weighted average AND its share of the mix misses 70% of what its position should earn. Cutting by high food cost is the most repeated mistake in this trade, and I got it wrong for years before I learned to measure it: the osso buco here carried 38% food cost, inside the fine dining band the National Restaurant Association places between 34% and 40%, and it left 9.80 USD of margin per unit against 5.10 from a pesto pasta costing 26%. Killing it on percentage alone would have stripped 4.70 USD of contribution from every plate sold. Percentages pay the chef's ego; dollars pay Tuesday's payroll.

How the Masterestaurant menu engineering matrix was applied?

Twelve weeks of sales by dish and the costing rebuilt at real purchase prices, not list prices: that is what loaded the Masterestaurant method's menu engineering matrix.

Each dish landed by two axes, contribution margin in dollars and relative share of the mix, and the cut fell on the low-low quadrant under one condition Diego F. Parra imposes in every menu audit: no dish leaves if pulling it orphans an ingredient three other recipes need. Nineteen of the 46 dishes went, along with 4 of the 14 wines by the glass. Two margin losers stayed for strategic reasons, the lasagna as the room's recognition anchor and a kids' plate that holds the table of six. The menu closed at 27 dishes, and 34 exclusive ingredients came down to nine.

What happens if you cut the menu but never touch purchasing?

Suppose you make the same cut and never rebuild the weekly order.

The menu drops to 27 dishes, the kitchen fires faster, the guest waits less, and sixty days later food cost still reads 39%, because the supplier keeps the old lines on the recurring order and the walk-in refills with the same product. We forced that scenario deliberately in week three, order frozen, and inventory barely moved from 11,800 to 10,400 USD. Only after rebuilding the order from the 27 new spec sheets did it fall to 6,900 USD, with turns at 4.1 a month. At six months the case closed like this: food cost from 39.9% to 32.6%, real operating profit from 9% declared to 14.8% verified with closing inventory, average check from 21.40 to 23.15 USD after repricing the survivors.

The paradox of the shorter menu that sells more

The owner fought this tension across three meetings: fewer options should sell less, and they sold more. Sales rose 6.4% in the following half-year with 19 dishes gone, and the mechanism holds no mystery, it is kitchen arithmetic. Every dish removed handed seconds of decision back to the guest and minutes of mise en place back to the line, and those minutes became one more table per turn on Friday and Saturday. Market context pushed the same way, since US consumer spending in restaurants grew 2% in 2024 with traffic flat according to Circana (2024), while food and beverage spending added 3% year over year in the first half of 2025 according to Circana (2025). The money sits in the check, not at the door. A long menu asks the kitchen for what the market stopped paying.

Transferable lessons by annual revenue band

Under 500,000 USD a year the job is this week's: count how many dishes sold zero units yesterday and kill the ones sharing no ingredient with another recipe, no matrix and no software, Saturday's ticket rail in hand. In this trattoria's band, 500,000 to 1 million, take physical inventory on a Monday before the order lands and divide monthly cost of sales by that value; under 3.5 turns, your menu is longer than your cold storage. Past the million, recost your twenty highest-volume dishes at real purchase price before touching anything else. Above 5 million, audit which exclusive ingredient props up each signature dish, because menu ego there gets paid through central purchasing. And the multi-unit group over 10 million? The celebrity-chef archetype with six themed rooms usually carries one common 60-dish menu where 40% sells in only two locations. Start with the mix per location.

Limits of this case

I would not expect this result in three contexts, and it is worth saying before somebody copies the cut blind. First comes the high-volume operation with a short menu by design, a QSR running 25% to 30% food cost according to the National Restaurant Association, where the fat is gone and margin gets fought on service speed and labor cost. Second, the destination restaurant with a chef-driven menu, where variety IS the product sold and pulling dishes destroys the reason for the trip; there you trim portions and renegotiate suppliers instead. Third are kitchens holding under four months of clean per-dish sales data, since a cut made on six weeks of low season kills the dishes that bill in December. This trattoria had nine years of history and twelve recent weeks loaded. Without that floor you are not doing menu engineering, you are guessing with a spreadsheet.

Root cause diagnosis: every symptom and the number that exposed it

Symptom: actual food cost hit 39.9% while the costing sheets promised 33.1%. Behind it sat 34 exclusive ingredients belonging to dishes that never sold a unit a day, bought in minimum supplier packs and binned at expiry. The first Monday's physical inventory gave it away: 11,800 USD in the walk-in for a house billing 620,000 a year, which works out to 2.4 turns a month against the four or five a healthy trattoria of that size holds. External benchmarks size the gap, since median limited-service food cost closed 2024 at 32.4% of sales according to the National Restaurant Association (2025) and casual dining runs between 30% and 34% (National Restaurant Association). This kitchen operated nearly six points above its own band. Four cooks for 38 seats, and Friday tickets still left the pass twenty-five minutes late.

Root cause diagnosis: every symptom and the number that exposed it — in practice

A 46-item menu demanded 61 separate mise en place preparations, plenty of them for dishes nobody ordered, which bought the kitchen an extra six-hour prep shift every single day. That 36.2% Labor Cost was never a wage problem. It was a menu problem. Once the count dropped to 35 dishes, mise en place fell to 38 preparations and the extra shift vanished without a layoff, because that cook moved onto hot line. Three years with the average check nailed at 21 USD while ingredient costs climbed, and the cause was printed on the menu itself: an alphabetical list per category with a dollar sign on every line and no pricing architecture anywhere. There is menu-design research documenting that small changes in how price is presented can move spend per person, and here nobody had even tried it. Worse, the highest contribution margin dish lived on the last line of its section, exactly where the guest's visual memory has already given up.

Root cause diagnosis: every symptom and the number that exposed it — key points

Kitchen turnover brushed 94% annualized because nobody memorizes 46 costing sheets that change without notice. The Skills Gap here was structural rather than attitudinal: with no written standard recipes, every new cook needed eleven weeks to produce at the incumbent's pace and quit around week nine. A long menu manufactures turnover. And turnover is disguised CapEx in training that never reaches the P&L, though you pay it monthly in someone's hours teaching what should have been written down. The P&L claimed 9% operating profit and cash flow said 4%, so here symptom and cause almost touch. Closing without ending inventory and without accruing last-cut purchases sent each month's food cost traveling deferred into the next one. That is no minor accounting slip. It is why an owner decides which dishes to remove from the menu against a margin that does not exist, and why we rebuilt fourteen months backward before touching a single plate.