Reading a restaurant P&L when you did not train in finance

Most operators I have met can run a Friday night better than they can read the statement describing it. That is not a character flaw. Nobody teaches it. You get handed a document written for a lender and are expected to manage from it.

The good news is that a restaurant P&L is a small document pretending to be a large one. Four blocks carry almost all the meaning, and the rest is detail you can read once a quarter.

What are the four blocks and what does each one mean?

Sales. What customers paid you, before anything is taken out. Watch that this matches what you think you sold. If third party delivery is netted rather than shown gross, your sales line is understating volume and hiding a fee.

Cost of goods sold. What the food and beverage you sold actually cost. Not what you bought, what you used. The difference between purchases and usage is inventory, and it is the reason a P&L without a count is a guess.

Labor. Wages, plus payroll taxes, plus benefits if you carry them. Many operators look only at wages and then wonder why the bottom line does not reconcile. The burden is real money and it belongs in the number you manage.

Operating expenses. Rent, utilities, insurance, repairs, software, marketing, the endless small subscriptions. Mostly fixed, mostly boring, occasionally hiding something absurd that has renewed for three years.

Below those sits your profit. Above them sits the only real lever most weeks.

What is each block supposed to be?

There is a published answer, and it is worth knowing before you decide your own numbers are bad.

The National Restaurant Association describes the pre-pandemic shape of a typical independent restaurant as food taking about 33 cents of every sales dollar, labor about 33 cents, and everything else, meaning utilities, occupancy, supplies, general and administrative, repairs and card processing, at about 29% of sales. What was left was a pre-tax margin near 5%.

Five cents. That is the room the entire business operates in. It also means a dollar of margin needs roughly twenty dollars of sales behind it, which is the arithmetic that should govern every decision you make about discounts, delivery commissions, and marketing spend.

Then the inputs moved. In its July 2026 analysis of restaurant profitability, the Association estimates that average hourly earnings of restaurant employees are up 41% since February 2020 and average wholesale food prices are up 35%, with utilities, occupancy, insurance, taxes and swipe fees all up by double digits. Total expenses for an average restaurant jumped 36% between 2019 and 2026. In the same piece, 42% of operators said their restaurant was not profitable in 2025.

The Association publishes the consequence as arithmetic, and it is the clearest illustration in the literature. Take a restaurant that did $1.5M in 2019 at that 5% pre-tax margin, so $75,000 of profit. Hold sales flat at 2019 levels and apply today's costs, and the same restaurant posts a pre-tax loss of $432,600, close to 29% of sales. To get back to break even, not to profit, sales would have to reach $1,932,600, which is 29% above the 2019 volume.

Read your statement with that in mind. If your percentages look worse than the ones you memorized in 2018, part of that is the decade.

Why do percentages matter more than dollars here?

Because dollars move with volume and tell you nothing about control.

Food cost went up nine hundred dollars. Is that bad? If sales went up proportionally, nothing happened. If sales were flat, something happened and you need to find it. The percentage answers that question and the dollar figure does not.

Read the percentage first, then read the dollars to size the problem. Percentage tells you whether there is a problem. Dollars tell you whether it is worth your Tuesday.

There is one trap. A percentage can improve for a bad reason. Food cost as a share of sales drops when you raise prices, even if the kitchen is wasting exactly what it wasted before. Always ask what the denominator did before you take credit for what the numerator did.

The two blocks you can actually move in a given month are food and labor together, which is why prime cost is the number that matters some months more than net profit does.

How do you read one month with actual numbers?

Here is the whole exercise on a made up month. Swap in your own figures as you go.

Net sales for the month: $118,000.

Food and beverage cost: $38,940. Divide by sales and you get 33% of sales, which is the shape the Association describes.

Labor including payroll taxes and benefits: $43,070, or 36.5% of sales. Food plus labor is $82,010, so prime cost is taking 69.5 cents of every dollar that came through the door.

Everything else at 29% of sales is $34,220. Add it up and you have spent $116,230 of $118,000, leaving $1,770. That is a margin of about a point and a half, and it is what most operators actually find when they do this honestly for the first time.

Now do the only comparison that teaches you anything. Last month, sales were $126,000 and food cost was $40,320.

Food dollars went down by $1,380. Food cost as a share of sales went up, from 32% to 33%.

That single point is worth $1,180 in the month and roughly $14,160 over a year at that volume. Nothing on the dollar line would have told you. The month looked cheaper and was worse.

Do the same subtraction on labor, then on the three largest operating expense lines, and you have read the statement. Everything past that is refinement.

What should you compare it against?

Three comparisons, in this order.

Against last period. The most obvious and the least informative on its own, because seasonality moves everything.

Against the same period last year. Better, if the restaurant has not changed materially. Be careful in the current stretch, because the Association's July 2026 outlook has industry sales growing 4.3% in nominal terms and only 0.8% after inflation, with much of the growth coming from higher menu prices rather than more traffic. A year over year sales gain that matches menu price increases is not growth, it is repricing.

Against your own budget or forecast. This is the one that matters, because it compares what happened to what you intended. Variance against intent is a management question. Variance against last March is a weather report.

If you run more than one location, comparing them to each other is the fastest diagnostic you own. Same menu, same prices, same vendor pricing, different result. The gap is almost always operational, and what to actually do when two locations perform differently starts with reading both statements side by side rather than averaging them.

Which lines lie to you?

A few, reliably.

Cost of goods without an inventory count. If nobody counted, the software assumed. The number is arithmetic, not measurement.

Repairs and maintenance. One compressor turns a normal month into a disaster on paper. Note it and move on. Do not redesign the operation around a one time capital event.

Marketing. Often a dumping ground for anything nobody could categorize. Before you conclude marketing is not working, confirm the line only contains marketing.

Delivery revenue and commission. Depending on how it is booked, you can be looking at gross sales with a fat expense below, or net sales with the fee already invisible. It is worth knowing which, because the economics of that channel are different from your dining room, and so is the strategic cost of letting a delivery app outrank your own website for your own name.

Owner compensation. If you do not pay yourself on the statement, your profit is overstated by exactly the value of your labor. That matters the day you try to sell or hire someone to replace you.

Do 90% of restaurants fail in the first year?

No, and the number that gets repeated is one of the most thoroughly debunked statistics in this industry.

The claim has no study behind it. It was checked twenty years ago by H.G. Parsa, John Self, David Njite and Tiffany King in "Why Restaurants Fail," published in the Cornell Hotel and Restaurant Administration Quarterly in August 2005. They went looking for the source of the 90% figure and found that it had become accepted with no data supporting it.

So they measured it, using longitudinal ownership turnover data from 1996 to 1999 alongside Dun and Bradstreet records. The first year failure rate for independent restaurants was 26.16%. Year two was 19%. Year three was 14%.

Two more findings are worth carrying around. Over a longer window, failure rates were roughly comparable between franchise chains at 57.2% and independents at 61.4%, so the brand does not protect you the way folklore says. And the arithmetic disproof is the part I like best: if 90% really failed in year one, a simulated market of 1,500 units would shrink to 254 units in twenty years, an 84% loss. The industry has done the opposite.

Why this belongs in an article about reading a statement: the myth teaches operators that failure is weather. It is not. About one in four independents does not make year one, which is serious enough without folklore, and the ones who survive are usually the ones who could tell you what their food and labor did last month.

Where does reading the statement stop helping?

In three places.

When the statement is not built the same way twice. If your bookkeeper reclassifies an expense, moves manager salaries between labor and general and administrative, or changes when invoices get cut off, your trend breaks and nothing physical changed in the restaurant. Ask for the chart of accounts and freeze it.

When you treat a published benchmark as a target. The Association states plainly that its operating data is "not intended to represent standards or goals for individual restaurants." A steakhouse and a taqueria do not belong at the same food cost. Your own eight month trend is a better instrument than anyone's median.

When the answer is on the sales line and you keep reading the expense lines. Given a 5% margin structure, some months the honest read is that no expense discipline available to you closes the gap and the problem is volume. The statement will tell you that too, if you let it.

What to do this week

Print one month. Paper, pen, no laptop.

Write the percentage of sales next to every line over one percent of the total. Doing this by hand once builds the intuition that reading a formatted report never will.

Circle the three lines that moved most against the prior period. For each, write one sentence explaining the cause. If you cannot explain it, that is the item to chase, and the fact that you cannot explain it is more useful than any number on the page.

Then run the subtraction from the worked example on your own two most recent months, in points rather than dollars, and convert every point into an annual number. Once you have said that figure out loud it stops being an abstraction.

Then hand the same page to your general manager and ask them to do it independently. What they cannot explain tells you where their development sits, which is the real bottleneck in most single owner operations and the subject of developing managers when you cannot be in two rooms at once.

Questions about your statement?

Email me at eric@seod.com with one month's P&L, numbers changed or scrubbed if you prefer, and I will send back the three lines I would look at first and why. Plain language, no accounting vocabulary.

Sixteen years of multi unit operations and more than $54M in annual P&L means I have read a lot of these, including plenty that were formatted badly. Reading raw records to answer a question is the same discipline whether the records are invoices or server logs showing which AI crawlers visit your site. Aggregate, compare, find the cause.

Or keep going with more on running the operation.