Benchmarks, Rules Of Thumb & Other Averages

Thirty percent food cost. Thirty percent labor. Prime cost under sixty. Rent no more than six to eight percent of sales. Price at three times raw cost. Ten percent to the bottom line. Twenty-five covers a server. Keep the menu under thirty items. Six months of cash before you open.

Every operator in this industry can recite most of that list, and almost none of them can say where a single number came from. They did not buy these. They arrived. Somebody said them in a class, in a magazine, at a conference, across a banker’s desk, or in a kitchen twenty years ago, and they installed themselves as the standard the operation gets run to.

They are the most widely adopted set of operating decisions in the restaurant business, and nobody ever made them about your building.

What A Benchmark Actually Is #

A benchmark is a description of a population.

Somebody surveyed a set of operations, added up what they reported, and divided. What comes out the other end is a statement about the middle of that group. It is not a target. It is not a standard. It is arithmetic performed on strangers.

Which produces the inference the whole industry runs on: operations like yours run at this number; therefore your operation should run at this number.

Two problems, and the first one is the small one. Your operation is probably not in that population — different rent, different market, different daypart mix, different model, different labor market, different equipment, different room. The second problem is the one that costs money: even if you were in the population, the middle of a group is not a place worth aiming at. It is the average of everybody in it, including everybody in it who is failing.

A Rule Of Thumb Is A Benchmark That Forgot Where It Came From #

Price at three times raw cost. Labor should never cross thirty. Rent under eight percent. Four times your rent is what you need in monthly sales.

Each of those started as somebody’s number from somebody’s building, or as an average pulled off a survey. Then it went verbal. It got repeated by people who did not know the source, to people who never asked, until the citation fell off and the number was left standing on its own as a thing everybody knows.

That makes a rule of thumb worse than a benchmark, not better. A benchmark at least comes with a population you could examine and reject. A rule of thumb has been stripped of everything that would let you test it. It arrives as common sense, and nothing is harder to question than a number that has no author.

Where The Numbers Actually Come From #

Association and trade surveys. Self-reported, self-selected, and answered by whoever felt like answering. Operators having a bad period tend not to fill out the survey.

Aggregate vendor data. Your point of sale company, your payroll company, your accounting platform. They have real numbers from thousands of operations, which sounds authoritative until you ask who is in the set — it is whoever bought that vendor, which is a business decision, not a sample.

Lender and franchise pro formas. Built to make a deal approvable. The numbers on that sheet exist to get a loan cleared or a franchise sold, and nobody in either transaction is paid to be conservative.

The comparative report your accountant runs. Chart of accounts differences alone will swing food cost several points between two identical operations. One puts paper goods in cost of goods, another puts it in supplies. One books the kitchen manager in labor, another in management. You are comparing two numbers that were never built the same way.

Survivorship. Closed operations do not report. Every benchmark in this industry is calculated off the ones still standing, so the number tells you what survivors report, not what causes survival.

The counsel class. None of the above matters as much as this. The numbers get repeated by consultants, speakers, writers, and course sellers who did not generate them, cannot source them, and have no exposure if they are wrong. Repetition is what makes a number feel true, and repetition is free.

Why Your Numbers Are Supposed To Be Different #

This is the part that never gets said, and it is the whole thing.

Two operations doing the same volume can have correct food costs eight points apart, and both of them are right.

A steakhouse buying protein at center-of-plate prices runs food cost high by every published standard and makes more gross margin dollars per Guest than the pizzeria hitting the benchmark exactly. Percentage says the steakhouse has a problem. The bank account says otherwise. Percentage is a ratio, and you do not deposit ratios.

An operation paying above market to hold a cast that produces a standard nobody on its street can match is running labor high on purpose. That is not a labor problem, it is what its differentiation costs, and the benchmark has no way to see the difference between an expensive cast that is earning it and an expensive cast that is not.

An operation carrying low occupancy cost can afford a food cost that would sink the one paying premium rent for a corner. Same category, same menu, different correct numbers, because the cost structure is different in the one place that matters.

An operation running three dayparts blends four different cost structures into one monthly percentage. The blended number describes nothing that actually happens in the building on any given shift.

The published number cannot account for a single one of those, because to be publishable it had to be stripped of every one of them. What makes your operation yours is exactly what had to be deleted to make the benchmark portable.

Six Ways The Number Costs You Money #

You manage the percentage instead of the money. Food cost climbs a point and the reflex is to fix the point. But a price increase, a mix shift toward a higher-margin item, or a cheaper protein all move that percentage, and only some of them move the dollars. An operator can hit thirty percent food cost with fewer margin dollars per Guest than he had at thirty-three, and celebrate it.

You cut what is carrying load. Labor crosses the line, so hours come out. The hours that come out are the ones that held the standard on a Friday, because those are the ones that look discretionary on a schedule. The percentage lands. The standard slips. The slip shows up two periods later in return interval, by which time nobody connects the two.

You price off cost instead of value. Three times raw cost is the most expensive rule in this industry. It ties what a Guest pays to what a supplier charges you and ignores what the item is worth in your room. It systematically underprices what you are chosen for and overprices what you are not.

You stop reading and start comparing. The number gives you something to check instead of something to understand. Checking takes four seconds. Reading the building takes a shift. So the read stops happening, and the operator can tell you his prime cost but not why tickets stack at seven-forty.

Your competitor is aiming at the same number. Every operation on your street run to the same published targets converges on the same cost structure, which converges on the same menu decisions, the same labor model, the same pricing, and the same operation. You cannot get to a differentiated business by aiming at the middle.

The number outranks you. You decide something about your operation — you are going to pay more to hold this cast, or run higher food cost on the item you are known for — and the benchmark says no. It has no read of your building and it wins anyway, because it feels like a law. A ratio you never chose overrules a decision you made on purpose.

The Only Valid Benchmark Is Your Own Operation #

Your operation last period. Same building, same market, same model, same chart of accounts, same everything except time. It is the only comparison in existence where the population is actually yours.

Trend against yourself, by period, by daypart, by mix. A number moving the wrong way in your own operation is a real signal that something changed and you should go find it. A number differing from a published average is not a signal at all — it is a description of how you differ from strangers, which is the thing you were trying to accomplish.

Setting Your Own Numbers #

Work backward from what the building has to produce, not forward from what the industry reports.

Start with what is fixed and known. Occupancy cost, debt service, insurance, the utilities baseline. These are not percentages, they are dollars, and they are dollars whether you do a hundred covers or four hundred. Write them down as dollars per period.

Name what the operation has to fund on top of that. The cast you need to produce the standard you have ruled — at what it actually takes to hold those people, not at what a survey says. The maintenance and replacement the building genuinely requires. The reinvestment that keeps the operation compounding instead of contracting. What the operation owes you. Dollars, per period, every line.

Now you have the number that matters. Total gross margin dollars the operation must produce per period. That is your target, and it came out of your building. Everything else is derived from it.

Derive your own margin per Guest. Divide the required gross margin by the covers your book actually produces, by daypart. Now you know what each Guest has to contribute. That number, not a percentage, is what you manage.

Read every item against it. Margin dollars per item, not cost percentage. Your highest-percentage items are frequently your best money and the rule of thumb has been telling you to fix them for years.

Derive your cost targets last. Once you know required margin and your actual mix, your food cost and labor targets fall out as arithmetic. They will not match anybody’s published number, and that is the point. They are the numbers your building requires.

Track dollars per hour of stage time. Your room and your hours are the finite thing, not your food cost. What a seat produces per hour, by daypart, tells you what to schedule, what to promote, what daypart to kill, and what to price differently.

Re-derive it every time the load moves. New rent, new daypart, a menu change, a volume shift, a cast change, a vendor spec change. The numbers were drawn against a load, and when the load moves they are describing an operation you no longer have.

What Changes Tomorrow #

Take your last full period’s P&L and write two numbers at the top of a page. Total gross margin dollars the operation produced. Total covers that produced it. Divide. That is margin dollars per Guest, and most operators have never once calculated it.

Then write down what the operation actually has to produce per period — the fixed dollars, the cast at what it really takes, the maintenance, the reinvestment, and what the operation owes you. Divide that by the same cover count. Now you have two numbers side by side: what each Guest contributes and what each Guest has to contribute.

The gap between them is the only operating number you need this week. If it is negative, every published benchmark you have been hitting has been hiding it from you, because you can hit all of them and still not produce what the building requires.

Then take your three highest-volume items and calculate margin dollars on each one, not cost percentage. Rank them by dollars. Compare that ranking to the one you have been managing off. If the order changes, you now know exactly what the rule of thumb has been costing you every shift for as long as you have been running it.

Do that, and you will never look at somebody else’s thirty percent the same way again.

Digging Deeper #

Every term in my framework lives in the Knowledge Base: kb.jeffreysummers.com

Terms used here: [Measurement Lock-In], [Cover Blindness], [Sameness Machine], [Default Architecture], [Designed Architecture], [Default Gravity], [Counsel Class], [Cost Basis Opacity], [Differentiation Economics], [Ordinary By Design], [Safest Mediocre Execution], [People Constraint], [Perspective Constraint], [Constraint Architecture], [Transactional Cost-Plus]

More of my work on this, all of it in the Knowledge Base:

• 077 Rules Of Thumb Are Not A Strategy — the published chapter this entry stands on
• 5.X Food Cost Is Not A Percentage — why the number you are comparing was never a percentage
• 5.X The Four Aspects Of Food Cost — what the benchmark flattens
• 016 A Full Dining Room Is Not A Business Model — the same reversal, read on covers
• Your Sales Are Up Because Fewer People Came — the average hiding the movement