How to price a tasting menu: a real fine dining case that lost EBITDA at every table

The diagnosis: a 28% food cost that was really 41%

The root error was measuring food cost on three signature dishes instead of the full cover. The owner swore it was 28%; the cover audit —seven courses, complimentary bread, plating waste, and suggested pairing— came in at a real 41% on a 78 USD ticket. That 13-point gap is what evaporated in production: «packed every weekend and the bank never grows». It is not a rare case. An average restaurant wastes 4% to 10% of what it buys (The Restaurant HQ, 2025), and in a multi-course tasting that share weighs more because each plate adds its own plating waste. On top of that, full-service foodservice concentrates over 43% of the sector's total food surplus (ReFED, 2024). Without measuring the whole cover, the price starts from a false number, and no amount of full tables fixes a wrong base cost. You set it from the theoretical cost of the full cover, not from plate-by-plate food cost.

How do you set the price of a tasting menu?

In this case we broke the cost down by course, added the complimentary bread and suggested pairing, and measured real plating waste: the cover cost 32 USD (41% of 78 USD), not the 22 USD the owner believed.

The rule we applied was to anchor each cover to a 30% target food cost on the menu price, which required raising the ticket to 92 USD and not a cent less. That adjustment did not come from watching competitors: it came from the arithmetic of the seat. Worth remembering that U.S. restaurant menu-price inflation peaked at 8.8% in March 2023, the highest in over two decades (National Restaurant Association, 2023); pricing by perception in that context leaves you short by default. Courtesy has a cost, and in a tasting menu it costs far more than it seems.

Pairing and courtesies: the cost almost nobody books

Here the house bread, whipped butter, and two welcome snacks added 4.10 USD per cover that nobody charged to the plate; the five-glass suggested pairing, served at cost with no margin «so as not to scare the guest», drained another 6 USD of contribution per seat. Between courtesies and mispriced pairing, over 10 USD per diner leaked out —13% of the ticket. Input volatility makes this worse: arabica coffee hit a record 4.41 USD per pound in February 2025 (Bellwether Coffee, 2025) and farm-level eggs rose 43.1% in 2024 in the U.S. (USDA ERS, 2024). A tasting menu that does not re-cost courtesies every quarter bleeds margin without changing a single line of the menu. Table time is the cost that perception pricing ignores, and in a tasting menu it is decisive. A seven-course menu occupies the seat for three hours; with 32 covers and 11 kitchen and floor staff, each seat-hour had to cover its share of payroll, rent, and utilities before yielding any EBITDA.

The seat-hour: why three hours at the table change the price

We calculated the structural cost per seat-hour at 9.50 USD; at three hours per diner, each cover had to contribute 28.50 USD of margin just to pay for structure. With the 78 USD ticket and a 32 USD cover, only 46 USD of gross margin remained: it grazed break-even and left no return. That is why the price anchors to contribution margin per seat-hour, not to isolated food cost. It is the metric that separates «full» from «profitable», and the one almost no fine-dining operator calculates. The tool was the Masterestaurant method's Full-Cover Costing template, applied course by course. With Diego F. Parra we loaded the seven dishes, the bread, the snacks, and the pairing into a single sheet that computes real food cost, plating waste, and contribution margin per seat-hour at once. The sheet revealed that two courses concentrated 60% of the overcost: a seafood plate with 22% plating waste and a pre-dessert served in tableware that kept breaking.

The Masterestaurant tool we used: full-cover costing

We redesigned the plating of both, swapped the at-cost pairing for one with 35% margin, and raised the ticket to 92 USD. Case result: the cover's food cost dropped from 41% to 31% and the margin per seat-hour went from grazing break-even to a positive 14 USD, without losing a single reservation (per the case data). The case result was that the same full room started leaving cash. After raising the ticket to 92 USD, re-costing courtesies and pairing, and redesigning the plating of the two courses that bled, the full-cover food cost fell from 41% to 31% and the contribution margin per seat-hour went from near zero to 14 USD (per the case data). Occupancy did not move: it stayed full on weekends with advance reservations, because the tasting-menu guest pays for experience, not cents. In six months the venue's monthly EBITDA multiplied by 2.3 (per the case data).

The measurable result: same full room, different bank

The value context of these adjustments is worth noting: the median sale price of a small U.S. restaurant reached 773,000 USD in 2025, up 24% from 2021 (BizBuySell, 2025); a venue with healthy, documented margin sells, a full and broke one does not. The transferable lesson is that price starts from the full cover, and the first step depends on your size. Small independent (one room, owner in the kitchen): this week cost ONE whole cover by hand —the seven courses, bread, snacks, and pairing— and compare it with what you think you charge; there are almost always 10 to 15 hidden food-cost points. Mid-size (two shifts, station chef): this week measure the real plating waste of your three priciest dishes by weighing discards across three services; that is where the overcost lives. Multi-unit group: this week standardize a single per-cover costing sheet for all locations and audit the most profitable against the least; the gap tells you what to replicate.

Transferable lessons: your first step this week by size

In all three, the anchor is contribution margin per seat-hour, not isolated food cost. The restaurant industry carries weight: in Mexico it is 12.2% of the country's economic units (INEGI–CANIRAC, 2024). This result is not universal, and saying so avoids survivorship bias. First, I would not expect it in a tasting menu within a saturated premium zone where direct competitors already set the price ceiling: raising the ticket to 92 USD worked because the venue had demand with advance reservations and little nearby substitution; in a corridor with ten equivalent options, the same jump empties tables. Second, I would not expect it in a high-turnover format —bistro or short 60-minute menu— because there the seat-hour is optimized with volume, not ticket, and the margin analysis shifts its axis.

Limits of this case: where I would NOT expect the same result

Third, I would not expect it if the real problem is volatile input costs with no menu power: with retail ground beef at 5.63 USD per pound by mid-2026 versus 4.56 in 2025 (USDA, 2026), a rigid menu without quarterly re-costing loses the margin this method recovers. 'By perception' pricing asks 'what does everyone else charge?'; correct costing asks 'what does the full cover cost me and what seat-hour margin do I need?'. The first leaves you full and broke; the second leaves you full and with EBITDA. Perception pricing treats food cost as an imprecise global figure; the right method breaks it down by course, measures real plating waste (heavier in tasting menus because of the number of plates) and adds the cost of courtesies and pairing almost nobody counts. Perception pricing ignores table time; the right method recognizes that a tasting menu occupies a seat for three hours and that seat-hour carries an opportunity cost —so price is anchored to contribution margin per seat-hour, not to isolated per-dish margin.