Prime cost: traditional method vs the Masterestaurant method
Prime cost is food and beverage cost plus total labor cost divided by net sales for the same period, and in a healthy full-service restaurant it lives between 55% and 60%; above 65% the operation burns cash even when the accounting P&L shows a profit. What separates the two methods is not the formula, which is identical, but FREQUENCY and depth: the traditional method closes it monthly with an estimated inventory and delivers it somewhere between the 15th and the 25th of the following month, when nothing can be fixed; the Masterestaurant method closes it every Monday against a physical count of the 20 SKUs that carry 80% of spend, breaks it open by cost center and produces an actionable variance in under 48 hours. If you must govern your restaurant with a single number, govern it with this one.
A restaurant billing 780,000 USD a year at 63% prime cost keeps 16,200 USD of operating margin before rent; the very same restaurant at 57% keeps 63,000 USD. Six percentage points, almost four times the profit. That is the whole business, which is why I insist prime cost is not one more tile on the dashboard: it is THE number, the one that puts every other metric in order.
In one-to-five-unit operations the problem is rarely that the owner does not know the formula, since any blog carries it. The problem is that it gets calculated late, from data that does not match the shelf, on a P&L built for the accountant instead of for whoever places Tuesday's order. With that picture you get the right diagnosis six weeks after the leak opened.
Side-by-side comparison
| Traditional (accounting) method | Masterestaurant method | |
|---|---|---|
| Calculation frequency | ✕Once a month, at accounting close | ✓Weekly, closed Mondays at 10:00 |
| Latency to the number | ✕15 to 25 days after period end | ✓Under 48 hours from week close |
| Inventory basis | ✕Estimated, or one full count per month | ✓Physical count of 20 SKUs (80% of spend) weekly |
| Cost breakdown | ✕2 lines: total food and total payroll | ✓7 centers: cold line, hot line, bar, pastry, floor, kitchen, admin |
| CapEx and OpEx handling | ✕Blended: a new fryer wrecks the month | ✓Separated: CapEx sits outside prime cost with its own payback |
| Alarm threshold | ✕None formal; mentioned when it hurts | ✓+2 points over budget triggers a 24-hour review |
| Labor included | ✕Gross wages of operating staff only | ✓Wages, payroll taxes, overtime, turnover cost and staff meals |
| Owner time per month | ✕3 to 4 hours reading a PDF nothing can be done about | ✓40 minutes a week ending in a written decision |
Step 1. Close out net sales for the period before you touch a single cost line
The denominator rules everything: prime cost is calculated on NET SALES, meaning gross sales minus consumption taxes, minus comps, minus discounts, and minus delivery marketplace commissions whenever the order comes through a platform. Half of all diagnoses die right here, because a DoorDash or Uber Eats ticket posts at menu value while the restaurant actually receives between 70 and 85 cents on the dollar, with fees running from 15% to 30% depending on the plan (Rezku, Third-Party Delivery Fees 2026). Take out card processing too, averaging 2,35% per transaction (Texas Restaurant Association, 2025). Deliverable for this step: one sheet showing gross sales, six deduction lines and net sales for the period, reconciled against the bank deposit within 0,5%. If it doesn't reconcile, everything downstream is decoration. Real food cost comes from a single formula: opening inventory, plus purchases for the period, minus closing inventory, divided by net sales.
Step 2. Physical opening and closing inventory, counted by hand, same day and same hour
It does not come from the purchasing system or from adding up invoices, which is the shortcut showing up in eight of every ten operations that reach Masterestaurant convinced their food cost is under control. Count physically, two people, printed tally sheet, Sunday night and the following Sunday night, always after close and always with identical counting units. A restaurant doing 15.000 USD in weekly sales usually holds between 9.000 and 14.000 USD of inventory on the floor: counting it takes ninety minutes and protects a six-figure number every year. The deliverable is the signed tally sheet, valued at last purchase cost. Labor inside prime cost means total cost of personnel: kitchen and front-of-house wages, tips when they form part of pay, overtime, benefits, employer contributions, social security, uniforms, and the temp who came in Saturday for the event. Standard accounting payroll leaves out somewhere between 18% and 32% of that total, and that gap explains most of the prime costs reading 54% when they should read 63%.
Step 3. Add up FULL labor, not the payroll your accountant hands you
Include the market salary of an owner who works shifts, even if you aren't paying yourself today, because the day you hire someone to do what you do, that cost shows up and the margin disappears. Deliverable: one single loaded labor line, reconciled against actual cash outflow for the period, with variance under 2%. A global prime cost of 61% tells you nothing about what to do on Tuesday. Break it apart: hot kitchen, cold kitchen, bar, bakery, delivery, each with its attributed sales and its direct cost. The pattern surfacing again and again is a kitchen sitting at 27% food cost, inside threshold, alongside a bar at 34%, well above the 32% ceiling the Masterestaurant method sets. With that view the decision stops being generic cost-cutting and turns concrete: recalibrate spirit pours, close the comps register, check the jigger. Diego F. Parra has run prime cost weekly for twenty years for an arithmetic reason, not an ideological one: a 900 USD variance caught on day seven costs 900 USD, and the same variance caught on day thirty costs 3.800 USD.
Step 5. Separate CapEx from OpEx before new equipment distorts the indicator
A 12.000 USD extraction hood, the combi oven, the restroom remodel and the point-of-sale software do NOT belong in prime cost, not even in the month you pay for them. They are capital investment and get depreciated across their useful life, three to ten years depending on the asset. Dropping them into the period destroys comparability, which is precisely the only value weekly measurement has. Rent, utilities, marketing, insurance and bank fees stay out as well: those live below the break-even line and get handled with other levers. Watch also for the small recurring expense that genuinely is OpEx yet still isn't prime cost, such as delivery packaging, which in high-dispatch operations runs between 1,5% and 3% of sales. Deliverable: a chart of accounts with every line flagged CapEx or OpEx. Five failures repeat across almost every one-to-five unit operation. First, calculating from invoices instead of physical inventory, which ignores waste and theft entirely.
Common mistakes that wreck the calculation, and how to shield yourself from each
Second, measuring monthly: month-end delivers the right diagnosis six weeks after the leak started, and food waste already costs the US restaurant industry roughly 162 billion dollars a year (The Restaurant HQ, 2025). Third, forgetting inter-unit transfers, which inflate one inventory and deflate another. Fourth, leaving delivery sales at menu value, with Grubhub commissions of 15% to 25% never deducted (Rezku, 2026). Fifth, changing the counting criteria mid-quarter. Write the procedure down, assign it to a named person at a fixed hour, and protect it from Saturday improvisation. Here sits the paradox the craft has to resolve. Cutting labor lowers prime cost this week and sinks average ticket within three, because a dining room one server short serves worse, sells less dessert and turns tables slower; buying the cheaper ingredient lowers food cost and sends plates back to the kitchen. The way out isn't cutting, it's reallocating: move hours from the dead Tuesday shift to the Friday peak, negotiate volume pricing by consolidating vendors, adjust portion where the guest won't notice, and raise price where elasticity allows it.
The real tension: lowering prime cost without burning down the operation that produces it
If Colombian restaurant sales fell 44% during 2024 (Acodrés, 2025) and the US full-service segment now runs about 18% smaller than in 2019 (Technomic, 2024), margin no longer arrives through volume. It arrives through cost discipline. Your prime cost is properly built when it clears these six verifications. One: net sales reconciles against the bank deposit within 0,5%. Two: a signed physical inventory sheet exists, dated and timed, at both ends of the period. Three: loaded labor cost reconciles with cash outflow inside 2%. Four: no CapEx line appears in the period. Five: the indicator is broken out across at least four cost centers, each with its own threshold and its own owner. Six: the result was published Monday before noon covering the week that closed Sunday. A restaurant with 780.000 USD in annual sales and a 63% prime cost leaves 16.200 USD of operating margin before rent; at 57% it leaves 63.000 USD.
How to know everything landed right: closing checklist for the cycle?
Six points, nearly four times the profit. Count your inventory this Sunday. FREQUENCY. Food cost moves weekly, not monthly. A Monday count catches a 900 USD leak seven days in;
a monthly close catches it at day thirty, when it already costs 3,800 USD. Diego F. Parra has run weekly prime cost for twenty years for that single arithmetic reason: capital leakage is a function of how long you leave the tap open. GRANULARITY. A blended 61% prime cost tells you nothing useful. Broken out by cost center it usually turns out the hot line sits at 27% food cost, inside the threshold, while the bar runs 34%, well past the 32% hard ceiling the Masterestaurant method sets. The decision then is not 'cut costs': it is to check spirit pours and how comps get logged. CapEx / OpEx SEPARATION. Dropping a 6,400 USD extraction hood into the month's cost sinks the metric and destroys the historical comparison.
The four differences that change the outcome
CapEx leaves prime cost, gets amortized and is judged on payback; OpEx goes in whole, with no exceptions and no 'this month was odd'. FULL LABOR COST. I got this wrong for years: I counted gross wages and left out payroll taxes, staff meals and the cost of turnover. That omission hides between 4 and 7 points of prime cost, which happens to be the exact distance between a restaurant that distributes profit and one that merely survives. MANAGERIAL P&L. Your accountant's statement satisfies the law; it cannot drive purchasing. A managerial P&L reorders the same figures by cost center and by week, using net sales after delivery commission, which is the only honest base for calculating food cost.
Criterion-by-criterion comparison
When the traditional method is enoughAccounting
- A single unit under 220,000 USD in annual sales with fewer than 25 menu items.
- Suppliers on a six-month fixed contract, with no price swing above 3% between orders.
- Stable payroll, no overtime, annual turnover under 20%.
- An owner on the floor seven days a week who sees every delivery and signs every invoice.
- The first six months after opening, when there is no historical series to compare against.
When the Masterestaurant method pays for itselfMasterestaurant
- Two or more units, or a single one above 400,000 USD in annual sales.
- More than 30 menu references and rotating fresh-product suppliers.
- Payroll with split shifts, overtime and weekend casual staff.
- Any operation where the owner no longer counts inventory with their own hands.
- Restaurants planning a new location within 18 months that need a clean benchmark.
- Businesses with owned delivery or marketplace channels, where an 18% to 30% commission distorts net sales.
Side-by-side comparison
| Traditional (accounting) method | Masterestaurant method | |
|---|---|---|
| Calculation frequency | ✕Once a month, at accounting close | ✓Weekly, closed Mondays at 10:00 |
| Latency to the number | ✕15 to 25 days after period end | ✓Under 48 hours from week close |
| Inventory basis | ✕Estimated, or one full count per month | ✓Physical count of 20 SKUs (80% of spend) weekly |
| Cost breakdown | ✕2 lines: total food and total payroll | ✓7 centers: cold line, hot line, bar, pastry, floor, kitchen, admin |
| CapEx and OpEx handling | ✕Blended: a new fryer wrecks the month | ✓Separated: CapEx sits outside prime cost with its own payback |
| Alarm threshold | ✕None formal; mentioned when it hurts | ✓+2 points over budget triggers a 24-hour review |
| Labor included | ✕Gross wages of operating staff only | ✓Wages, payroll taxes, overtime, turnover cost and staff meals |
| Owner time per month | ✕3 to 4 hours reading a PDF nothing can be done about | ✓40 minutes a week ending in a written decision |
The numbers that frame the decision
“We were measuring off the accountant's close and the number reached me on the 22nd. Once we moved to Monday counts on the 20 key SKUs, week one exposed the bar running 36% cost from unlogged comps. Prime cost went from 64.8% to 57.9% in eleven weeks, and the next quarter operating margin climbed from 9,100 to 41,400 USD on almost identical sales, 612,000 USD annualized.”
How to calculate and govern prime cost in seven steps
Five things go on the table and without them do not start: POS net sales with delivery commissions already deducted, the full folder of purchase invoices for the period, opening inventory at value, payroll including taxes and overtime for the same date range, and the SKU list ranked by cumulative spend. DELIVERABLE: one sheet with those five sources and their totals. CHECKPOINT: invoices must reconcile to the bank statement within 2%; if they do not, you have invoices outside the system and any prime cost you compute is fiction. TYPICAL MISTAKE: using gross instead of net sales, which flatters the metric by 3 to 6 points in delivery-heavy operations.
Pick the range and never move it again. A Monday-to-Sunday week aligns the count with supplier deliveries and the payroll cutoff, which is what makes one data point comparable to the next. Calendar months put four weekends into some periods and five into others, swinging the metric two points with nothing happening on the floor. DELIVERABLE: a 52-week calendar with count date and owner. CHECKPOINT: every window holds exactly seven days and no day appears twice. TYPICAL MISTAKE: shifting the window mid-year 'to match the accountant' and losing the entire historical series.
Do not count everything. Rank last quarter's purchases high to low and you will find that 18 to 24 references explain roughly 80% of input spend: protein, dairy, spirits, oil, coffee. Those get counted physically every Monday before opening, on a scale, in the same unit you buy them in. Everything else is counted monthly. DELIVERABLE: a signed count sheet with quantity and value. CHECKPOINT: counting the 20 SKUs must take two people under 45 minutes; longer than that means the list is badly built. TYPICAL MISTAKE: counting in units other than purchase units, portions where the invoice says kilos.
Opening inventory plus period purchases minus closing inventory equals consumption. Divide that consumption by net food and beverage sales and you get food cost, which under the Masterestaurant method carries a hard 32% ceiling per dish and should sit between 26% and 30% on a weighted average. Pull staff meals and comps out of consumption; they belong on a separate managerial P&L line. DELIVERABLE: period food cost, blended and by cost center. CHECKPOINT: the gap between theoretical food cost from recipe cards and actual must stay under 2 points. TYPICAL MISTAKE: ignoring bar-to-kitchen transfers, which inflate one center and deflate the other.
Half the badly calculated prime costs I review die right here. Labor means gross wages, payroll taxes and benefits, overtime, casual staff, guaranteed tips where the law requires them, staff meals valued at cost, and what it costs to recruit and train whoever replaces the person who left. Divide that total by net sales and you get labor cost, which in healthy full service runs 28% to 33%. DELIVERABLE: period labor cost with all seven lines open. CHECKPOINT: the sum of those lines must match actual payroll disbursement within 1%. TYPICAL MISTAKE: leaving out the owner who works shifts, which understates the metric and makes benchmarking impossible.
Food cost plus labor cost over net sales: that is prime cost. Now comes the part almost nobody does, which is comparing it against a budget set at the start of the year rather than against the prior period, because last month only tells you whether you beat a number that may have been wrong to begin with. If budget says 57% and you print 59.4%, that 2.4-point variance on 60,000 USD of monthly sales is 1,440 USD gone. DELIVERABLE: a three-column table, budget, actual, variance in points and dollars. CHECKPOINT: cumulative quarterly variance stays under 1.5 points. TYPICAL MISTAKE: celebrating an improvement that actually came from one freak sales week.
A metric that does not end in an action with someone's name on it is expensive decoration. Every Monday, right after the calculation, write a single line: what gets fixed, who does it, by when, and which number should move. Check the whisky pour at the bar, owner is the head bartender, deadline Friday, target bar cost from 34% down to 30%. One decision a week, fifty-two a year, beats what most operations we review manage. DELIVERABLE: a weekly decision log with a target figure. CHECKPOINT: at least 8 of every 10 decisions carry a measured result two weeks later. TYPICAL MISTAKE: opening five fronts at once and closing none.
Once a month, somebody who does not do the counting recounts three random SKUs and compares against the sheet. It sounds like mistrust and it is not: it is the control that stops the count from decaying into copying last week's figure, which is how 90% of the costing systems I have helped install eventually die. DELIVERABLE: a cross-audit record with the three SKUs and the gap found. CHECKPOINT: under 3% difference on each recounted SKU; above 5% the whole inventory gets recounted that same week. TYPICAL MISTAKE: assigning the cross audit to the direct supervisor of whoever counted, which turns the control into a signature.
And with AI?
Project your food cost, spot margin leaks and simulate pricing scenarios in minutes. Diego F. Parra is an expert in AI applied to restaurants.
Free tools to apply this now
Which ecosystem tools hold it up
No spreadsheet saves an operation on its own, yet three Masterestaurant ecosystem tools shorten the distance between data and decision, which is exactly where prime cost is won or lost. The first orders the business model before you measure it, the second projects what each recovered point does to growth, and the third watches cash, which is what actually kills restaurants.
Diego F. Parra — International consultant, expert in creating and scaling restaurants and in AI applied to restaurants, foodtech and HORECA. Methodology applied in 8.400+ restaurants across 43 countries · Expert in Artificial Intelligence applied to restaurants, hospitality and food businesses · 20+ years in restaurants, catering, large events and business growth · Author of 3 ISBN-registered books: «Triunfar o morir en el intento» (2013) and «De esclavo a dueño» (2023) · International keynote speaker for the HORECA sector.
Frequently asked questions about prime cost
What is a good prime cost percentage for a restaurant in 2026?
What is a good prime cost percentage for a restaurant in 2026?
Between 55% and 60% of net sales in full service, and 60% to 65% in quick service, where labor weighs less but food cost climbs. Above 65% the operation burns cash even if the income statement reports accounting profit, because that profit includes items you have not paid out yet.
How often should I calculate my restaurant's prime cost?
How often should I calculate my restaurant's prime cost?
Weekly, closing Monday morning. Food cost moves week to week, and a leak caught at day seven costs roughly a quarter of the same leak caught at day thirty. Your accountant's monthly close exists to satisfy the law, not to drive purchasing decisions.
Do rent and utilities belong in prime cost?
Do rent and utilities belong in prime cost?
No. Prime cost covers food and beverage plus full labor, nothing else. Rent, utilities and insurance are fixed costs judged against your break-even, not against a plate. Blending them makes every dish look unviable and pushes owners to raise prices where no increase was needed.
How does delivery affect the prime cost calculation?
How does delivery affect the prime cost calculation?
It distorts the base if you use gross sales. With marketplace commissions between 18% and 30%, calculate on NET sales after commission and open a separate cost center for the digital channel. A restaurant doing 25% of volume through delivery can be three points worse off than it believes.
Sector data 2026 (official sources)
Verifiable industry benchmarks from official, non-commercial sources (government, industry associations, market research) - not competitors.
| Metric | Benchmark 2026 | Source |
|---|---|---|
| Ventas de servicios de comida y bebida en Canadá | CAD 96.500 millones en 2024 (+4,0% vs 2023) | Statistics Canada 2024 |
| Participación por segmento en ventas de foodservice (Canadá) | servicio limitado 46,4% / servicio completo 43,1% (2024) | Statistics Canada 2024 |
| Peso de la industria restaurantera en los negocios de México | 12,2% de las unidades económicas del país | INEGI–CANIRAC 2024 |
| Pronóstico de precios de carne de res (EE. UU.) | +7,5% en 2026 (hato ganadero en mínimo de 75 años) | USDA ERS (Food Price Outlook) 2026 |
| Pronóstico de precio mayorista de carne de res (EE. UU.) | +9,4% en 2026 | USDA ERS (Food Price Outlook) 2026 |
| Pronóstico de precios de bebidas no alcohólicas y café (EE. UU.) | +5,7% en 2026 | USDA ERS (Food Price Outlook) 2026 |
Related content
Grow your restaurant with the Masterestaurant method
Applied in +8.400 restaurants across 43 countries.