My Restaurant Isn't Making Money. Where Do I Look?

By Marji · Published August 10, 2026 · Updated September 30, 2026 · 6 min read

First check whether it is profit or cash that is missing: loan principal and owner draws drain cash without appearing as P&L expenses. Then work down one representative month — sales, prime cost, occupancy, everything else — against your own history, find the line that moved, and size each fix in monthly dollars before pulling it. For scale, the National Restaurant Association's 2025 data abstract (2024 figures) put median income before taxes at 2.8% of sales for full-service restaurants and 4.0% for limited-service.

The instinct when a restaurant stops making money is to cut something. The obvious cuts — portion sizes, staff hours, quality — are the ones guests can see, and they can make the problem worse. Find the actual leak first; with a month's P&L and your sales report, it takes an afternoon.

Step 1 — Profit or cash?

A P&L can show a profit while the bank account shrinks. Loan principal, owner distributions (as distinct from a salary run through payroll), equipment you capitalize rather than expense, and a bigger stock of inventory when you count it, all use cash without being expenses on the P&L; interest, depreciation and an owner's payroll salary are the parts that are. If the P&L is positive and the account is not, the answer is in those items or in timing, not in the kitchen. If the P&L itself is negative, keep going.

Step 2 — Get the numbers for one month

Use a representative month — not your best, not your worst — and put the same months from last year beside it.

The numbers to pull, and where each comes from
Number	How to get it	Compare with
Total sales	POS or the P&L, after discounts and comps	Same month last year
Prime cost %	(COGS + fully loaded labor) ÷ sales	Your own trend; the 65% rule of thumb for table service
Occupancy %	(Rent + property taxes + property insurance) ÷ sales	Your own trend; NRA 2024 medians 5.7% full-service, 5.2% limited-service
Everything else %	All other operating expenses ÷ sales	Your own trend, line by line
Owner pay and loan payments	Bank statements	What is left after the four lines above

What is left after the first four lines is profit before owner pay, interest, depreciation and tax. National medians are thin reference points, not targets: the Association's 2025 Restaurant Operations Data Abstract found 2024 median income before taxes of 2.8% of sales for full-service and 4.0% for limited-service restaurants, and a median prime cost of "65 cents of every sales dollar in the limited-service segment." A median means half the restaurants surveyed did worse.

Step 3 — Find the line that is out of range

Worked example · Hypothetical diagnosis

Monthly sales: $104,000
Prime cost: $73,840 · 71.0%
Occupancy: $9,400 · 9.0%
Other operating: $9,900 · 9.5%
Left before owner pay, interest and tax: $10,860 · 10.4%
Owner draw $7,500 + loan payments $4,000: $11,500
Cash after owner and loan: −$640 a month

Occupancy and other costs are in line with last year. Prime cost at 71.0% is six points above the 65% rule of thumb, and six points of $104,000 is $6,240 a month — more than the whole shortfall, and it is the half of the P&L you can change fastest.

Step 4 — Size each lever before pulling it

Compare fixes in the same unit: monthly contribution — the dollars that reach the bottom half of the P&L. Extra sales are not extra profit; a price increase keeps the whole increase if volume holds, while add-ons bring their own food cost and extra covers bring food, card fees and labor.

What each move is worth per month on $104,000 of sales; the assumptions are in the note below
Move	Extra sales	Extra contribution
Prices +3% across the menu, same volume	$3,120	$3,120
Average check +3% from add-ons ($1.14 a cover at a $38 check)	$3,120	$2,184
Covers +3%	$3,120	$1,872
Food cost −1 point	—	$1,040
Labor −1 point	—	$1,040
Rent −1 point	—	$1,040

Assumptions: add-ons keep 70 cents of each dollar after a 30% food cost; an extra cover keeps 60 cents after food, card fees and the hourly labor it needs; card fees on a price increase are ignored. Use your own figures. On these, a modest price increase is the biggest single lever if volume holds — test that on one dish first with the price increase calculator, which shows how much volume the new price can lose and still pay. Rent is the one line that rarely moves without a lease negotiation.

Step 5 — The order to act in

Reprice the plowhorses. High volume, below-average contribution, proven demand — the cheapest price test on the menu. See menu engineering.
Recost everything and fix what drifted. Recipes costed at launch miss every price change since; see why is my food cost so high.
Work on attachment. Sides, drinks, dessert. It costs training and a server script, not capital.
Then labor. Schedule against covers by hour, and cut only what service does not need; RevPASH by the hour shows where the room is empty.
Structural last. Rent, concept, hours, footprint. Slow, expensive, sometimes necessary — but not before the operating levers are tested.

When the answer is structural

Sometimes the numbers do not work at this volume in this space, and no amount of operating discipline fixes it. The signals: a break-even cover count you have never once hit, occupancy that stays high as a share of sales even in your best months, or a prime cost you have already worked down with little left to trim.

That is worth knowing early rather than after another year of subsidising it. Run your own numbers through the break-even calculator: if the covers-per-day figure is above what your room can turn in the hours you are open, the problem is not your food cost. The break-even guide shows how to split labor into its fixed and variable parts and what delivery commission does to the number.

What profit margin should a restaurant make?

There is no standard. In the National Restaurant Association's 2025 data abstract, 2024 median income before taxes was 2.8% of sales for full-service restaurants and 4.0% for limited-service — medians, so half did worse. Among full-service restaurants it was 1.1% for those with sales under $2 million and 4.3% at $2 million or more. Set your own target from your debt, the owner pay the business must cover and the return you need on the money invested.

How long should it take to see a difference?

A price change shows in the next week's sales and contribution. Recosting and portioning fixes show in the next inventory period. Labor changes need a few weeks of comparable covers to read. Marketing takes longer. Sequence your effort accordingly.

Sometimes, for prep and inventory reasons rather than direct cost: fewer items can mean fewer ingredients to stock, less spoilage and faster service. Start with the dogs — low volume and low contribution — and check what shares their ingredients first.

What if I can't pay myself?

Then the business isn't profitable, whatever the P&L says. Put a market-rate salary for the job you do into your fixed costs and recompute break-even. A restaurant that only works because the owner works free isn't a restaurant that works.

The cost side, weekly.

What ingredient prices did, what it does to a typical dish, and one thing worth repricing. One email, Monday.