Plate costing: the before and after of putting the recipe card to work

What does plate costing cost to implement in 2026?

As of August 2026 the range runs from 0 to 18,000 USD, and your monthly revenue sets the bracket, not your appetite for a pretty dashboard.

Below 30,000 USD in monthly sales, a disciplined spreadsheet costs 0 USD and roughly 40 hours of your own time spread over six weeks, with a return that almost always beats software. Between 30,000 and 120,000 USD of monthly sales, assisted implementation runs 2,500 to 7,000 USD per location, with an annual license of 900 to 2,400 USD. Above 120,000 USD a month, or with three sites or more, a serious project —POS integration, master recipe book, yield control— lands between 9,000 and 18,000 USD. Full-service food cost closed 2024 at a median 32.0% of sales (National Restaurant Association), and that benchmark decides whether the spend pays for itself.

What each bracket actually includes?

The 0 USD tier buys three things:

technical sheets for your twenty highest-rotation dishes, prices actually paid pulled from real invoices of the last ninety days, and a thirty-minute monthly review that you either sustain or the whole thing is worthless. The 2,500 to 7,000 USD tier covers the build-out of 60 to 120 recipes, yield tests on the eight raw materials that weigh most in your purchasing, loading into software connected to the POS, and two training days for the kitchen team. Above 9,000 USD you get what genuinely separates the tiers: a versioned master recipe book across sites, theoretical against physical inventory with weekly counts, alerts for supplier price drift, and a menu engineering model that crosses contribution margin with popularity. Nobody sells you weekly counting for 3,000 USD, and whoever promises it is quoting something else. The first one is your head chef's time: 60 to 100 hours to build recipes and weigh yields, which at a 2,200 USD monthly salary means 750 to 1,250 USD of payroll pulled off the line.

Three costs no proposal will ever declare

The second is data maintenance, because a 2023 technical sheet with frozen prices lies more than having none at all: budgeting 8 to 12 hours a month just to refresh prices paid is the floor, and that line never shows up in the quote. The third is the gap between list and invoice. Between the supplier's reference price and what actually gets billed sit freight, order minimums, presentation changes and seasonal adjustments, and across protein purchasing in 2025 that spread averaged 6.8%: enough to turn a dish costed at 30% into a real 34%. Diego F. Parra puts it without diplomacy in Masterestaurant audits: 32% is a CEILING, never a target. Assuming 4% loss on everything coming through the back door is comfortable and it drains money every service. A whole salmon yields around 52% after scaling, filleting and pulling pin bones; if you cost it at 96% yield, you are giving away nearly half the fish without it appearing in any account.

Flat waste is the most expensive lie on the sheet

Beef tenderloin behaves the same way: real trim loss settles at 11.5% while the sheet claims 4%, and that delta explains how a menu's star dish gets reported at 29% while it actually runs at 38%. Do the multiplication: 120 portions a week, a 14 USD selling price, nine mismeasured points of food cost equal 15,700 USD a year walking out of the till with no accounting trace. Food waste costs the U.S. restaurant industry roughly 162 billion dollars annually (The Restaurant HQ, 2025), and part of that figure is simply bad arithmetic. A percentage is a ratio, and a ratio improves when you worsen the denominator. That error has sustained entire managerial careers: the salad at 18% food cost leaves 9 USD of contribution margin per unit, the ribeye at 36% leaves 22 USD, and the manager rewards the salad because the dashboard glows green.

Low food cost, broke restaurant: the percentage paradox

Push the menu toward the low percentage and your location's food cost drops three points while profit falls, because payroll and rent are not paid in percentages, they are paid in dollars. Here is the bridge: the percentage governs purchasing and supplier negotiation, contribution margin governs the menu and the server's script. And that is why at Masterestaurant payroll, rent and utilities are NOT loaded onto the plate —they belong to the location's break-even— because the moment rent enters the technical sheet you lose the ability to compare two dishes against each other. Active recipe count: every block of 50 sheets adds 1,100 to 1,800 USD, because the build-out is paid by kitchen hour, not by license. Sites: a second location raises the project by roughly 45%, not 100%, since the master recipe book gets reused and only training is replicated.

Five factors that move the implementation price

POS integration: if your point of sale exposes an open API the work costs 600 to 1,200 USD; if the connection has to be solved through manual file exports, add 2,000 USD and a monthly headache. Third-party channels: operating delivery forces you to cost two prices per dish, because DoorDash and Uber Eats charge 15% to 30% per order and Grubhub 15% to 25% (Rezku, 2026), and that margin has to live inside the model. Card processing, averaging 2.35% per transaction according to the Texas Restaurant Association in 2025, belongs in there too. Ask for the project split into two deliveries and pay the second against measured results: first delivery, the 25 highest-rotation recipes with yields weighed in your own kitchen; second delivery, the rest of the recipe book, invoiced only if period food cost dropped at least two points against baseline. That single change takes 15% to 25% off list price, because the vendor discounts the risk it is no longer carrying.

How to negotiate the implementation down

Demand as well that yields be measured with YOUR product and YOUR cook, never with generic imported tables, and that the master file stays exportable under your name —if the data lives trapped inside the license, in eighteen months you will be paying rent on your own numbers. Negotiate the annual license upfront at a 20% discount only once implementation is verified, never before. And write into the contract who refreshes supplier prices each month. If you have less than 2,500 USD available, do not buy software: spend six weeks costing your twenty highest-rotation dishes by hand using paid invoices, not price lists, and that exercise alone gives back two to four points of food cost. If you have between 2,500 and 7,000 USD and a single site billing over 30,000 USD a month, hire assisted implementation and demand yield tests on the eight raw materials that concentrate your purchasing.

The rule for deciding with what you have this month

With three locations or more, the 9,000 to 18,000 USD project stops being an expense and becomes control: without a versioned master recipe book, each site invents its own recipe and you lose the comparison entirely. The sector does not forgive improvisation —1,600 restaurants closed in Colombia between August 2023 and 2024 according to Acodrés, and 348 full-service chain locations disappeared through bankruptcy in the United States in 2024 (Technomic). Start this week: pull the protein invoices from the last ninety days and compare the price paid against the one written on your technical sheet. Reference price versus price paid. Most plate costing starts from the supplier's list and is already bent at birth, because between that list and the invoice came freight, order minimums, a pack-size change and a seasonal adjustment. Measured on protein purchases in 2025, the average spread was 6.8% — enough to turn a 30% plate into a 34% one.

The four differences that move cash

Flat waste versus measured yield. Assuming 4% loss across the board is comfortable and it is the most expensive lie on the recipe card: a whole salmon yields roughly 52% after scaling, filleting and pin-boning, so if you cost it at 96% yield you are giving away almost half the fish on every plate and it shows up nowhere. Food cost versus contribution margin. A 34% plate leaving 14 USD clean pays more payroll than a 24% plate leaving 5 USD, and ordering a menu by percentage alone is the mistake I have had to unwind in more board meetings than any other. Both numbers together, never one on its own. Static costing versus alerted costing. If nothing tells you avocado jumped 22% in three weeks, you learned it from the managerial P&L and by then you sold six thousand plates with the margin bitten. That lag runs about 3,450 USD per month of delay in a 75,000 USD location.