Insights

The numbers that decide whether your restaurant makes money

A former CRA auditor on the handful of numbers that decide whether an owner-run restaurant is profitable: prime cost, the food-cost gap, labour, pour cost, delivery apps, and the January valley.

---

A restaurant is one of the hardest small businesses to own, because a lot of cash moves through the till and very little of it stays. You can be busy every night, have a full parking lot and a line at the door, and still lose money. In Canada today only about 9 percent of operators clear a 10 percent-plus margin, down from roughly a third before the pandemic. Busy does not mean profitable. A few numbers decide that, and most owners cannot see them between visits from the bookkeeper.

I spent years as an auditor at the Canada Revenue Agency reading how businesses actually make and lose money, versus how their owners described it. Now I run an operations firm here in Manitoba and hold a Lean Six Sigma Black Belt. What follows is the short list of numbers that decide whether a restaurant makes money, and how to make each one visible before the month is already lost.

Revenue is vanity, prime cost is sanity

The single number that separates an operator who lives from one who slowly dies is prime cost. Prime cost is the sum of two things: your cost of goods sold, meaning food and beverage, plus your total labour including the burden of payroll taxes and benefits. It is the part of every dollar that walks out the door before rent, insurance, utilities, or anything else. Everything downstream is small by comparison.

The survival ceiling is roughly 60 percent of sales for a quick-service spot and 65 percent for a full-service restaurant. Hold prime cost at or under that line while keeping rent and property costs under about 10 percent of sales, and there is room for a real margin. Let prime cost drift to 70 percent and there is no clever decision downstream that recovers it. That is the whole game in one sentence. The mistake owners make is tracking prime cost monthly, from the accountant, four weeks after they could have done anything about it. Track it weekly. Weekly is the only cadence fast enough to catch a bad week before it becomes a bad month.

The gap between what food should cost and what it did

Here is where profit quietly disappears, and it is almost never where the owner is looking. There are two food-cost numbers, not one. Theoretical food cost is what your recipes say the food should have cost, given what you sold. Actual food cost is what your invoices and your inventory count say it really cost. The space between them is the diagnostic, not the level itself.

Say your recipes tell you the kitchen should run 29 percent and your actual number comes in at 34 percent. That five-point gap is not the price of chicken. It is over-portioning, waste and spoilage, comps and voids, staff meals, an invoice booked to the wrong week, and sometimes theft. On a restaurant doing $700,000 a year, five points is $35,000 of pure profit, gone, and on the profit-and-loss statement every one of those causes looks identical. So the owner blames food prices, which cannot be controlled, instead of portioning, which can. A cook who plates half an ounce of extra protein across two hundred covers is real money, and nobody decided to do it. Portions just creep up over months as the line makes it look generous. You cannot close a gap you cannot see. The first job is almost always to make it visible with a costed recipe, a real closing count, and the variance in front of the owner every week.

Labour, scheduled to a forecast and not a feeling

Labour is the other half of prime cost, and it is the number owners fudge most often, usually to themselves. They will quote a food-cost figure and skip labour entirely, which is exactly where the loss-makers die. The median full-service operator spent around 36.5 percent of sales on labour in a recent year; operators posting losses ran as high as 43 percent, while the profitable ones sat closer to 30 to 34. These are US operations figures and directional for Canada, so confirm against your own payroll, but the shape holds everywhere.

The root cause of a bad labour number is almost always the same. The schedule is built from how busy last week felt, not from a sales forecast. That is over-scheduling by memory, and it guarantees you are paying for hours the sales never needed. The fix is not heroics. It is scheduling to a forecast, watching sales per labour hour, and reading labour as a live percentage of sales during the week instead of as a surprise at month-end. In Manitoba there is a second reason to get ahead of this: minimum wage rises to $16.40 on October 1, 2026, indexed to inflation each year after. Every indexed bump compresses quick-service and casual margins first, because those are the segments running the most minimum-wage hours. Model the next increase into your labour plan before it lands, not after.

Beverage is where the margin actually lives

If food is where margin leaks, beverage is where it is made. A dollar of drink sales is far more profitable than a dollar of food, which is why the fastest lever in a full-service restaurant is often beverage attach, the ratio of drink sales to food sales. For a bar program the number to watch is pour cost by category. As a rough guide, well liquor should run an 18 to 20 percent pour cost, draft beer 22 to 24, and wine by the glass 28 to 32. When the actual numbers sit well above that, the culprit is usually not pricing. It is shrink.

Industry-average bar shrinkage runs a startling 20 to 25 percent of inventory value, while a controlled bar sits in the single digits. The mechanism is simple. A bartender free-pouring two ounces against a one-and-a-half-ounce standard gives away a quarter of every bottle, and does it without any bad intent. Jiggers, a weekly liquor count, a pour-cost target by category, and a quick look at comps and voids keyed in after the guest has already paid will move that number from the twenties toward single digits. For a licensed room, that one discipline is frequently the most valuable thing an operator can do in a quarter.

The delivery-app trap that erases the plate

Third-party delivery is the number-one silent margin killer, and it is worth being blunt about the math. The platforms advertise a commission of 15 to 30 percent, but the true effective cost lands closer to 35 to 45 percent once you count packaging, payment processing, the promotions you feel pressured to run, and refunds. Now put that against a plate built to cost 30 percent in food. A 25 to 30 percent commission does not shrink your margin on that order. It erases it, and then some.

This is what fools people. Delivery shows up as top-line growth, so the sales chart looks great while the bank balance does not move, because you are running negative-margin volume dressed up as success. Here in Manitoba the reflex to just get on Skip is strong, since SkipTheDishes is headquartered in Winnipeg and dominant across the province. The discipline is not to swear off delivery. It is to price the delivery menu separately so the channel pays for itself, cap the promotions, and treat your own pickup and first-party ordering as the margin-protecting channel you actually push.

The January valley and the taxes hiding in your account

Every Manitoba restaurant lives the same yearly swing, and the sharp cold makes it sharper than in milder provinces. December is the peak: parties, gift cards, catering. January and February are the trough, because nobody drives out to dinner at minus thirty. Owners who spend the December cash without holding a reserve start the new year borrowing, and gift cards make it worse, since a card sold in December is a liability you redeem in January against fresh food and labour cost.

There is a scarier version of this that a former auditor cannot help but flag. A restaurant meal in Manitoba carries 7 percent RST and 5 percent GST, so 12 percent of what lands in your account is not yours. It is a trust fund you hold for the province and for CRA. The January valley is exactly when an under-managed operator dips into that money to make payroll, and that is how an operating problem becomes a tax problem with penalties attached. The move is unglamorous and it works: hold back a defined percentage of December cash into a reserve, build a leaner January schedule off the forecast, and remit your sales tax on a weekly cash discipline so the money is gone before you can spend it by accident.

Where to start

You do not need all of this running by next week. You need to see prime cost weekly, and then close the gap between what your food should cost and what it does. From there the early wins are usually the same three: reprice or turn down the delivery apps, catch vendor price creep by checking invoices against a price file, and reserve December cash for the January valley. None of it requires a bigger location or more covers. It requires seeing the numbers that were always there. If you want a straight read on where you stand, our operations health check maps your prime cost, your food-cost gap, and your biggest margin leaks, and shows you where to start. You are always welcome to reach out and talk it through.

Frequently asked questions

What is prime cost and why does it matter more than anything else? Prime cost is your food and beverage cost plus your total labour, including payroll burden. It is the largest and most controllable slice of every sales dollar. Held under about 60 to 65 percent of sales, with rent under 10 percent, there is room for profit. Let it drift past 70 and no other decision recovers it.

What is a good food cost for a restaurant? It depends on segment: roughly 25 to 30 percent for casual, higher for fine dining, lower for quick service. But the level matters less than the gap between your theoretical food cost from recipes and your actual cost from inventory. A large gap is over-portioning, waste, or shrink, and that is where the money is.

Are delivery apps worth it for a small restaurant? Only if you price for them. The true effective cost reaches 35 to 45 percent of the order once packaging, processing, promos, and refunds are counted, which erases the margin on a normally priced plate. Price the delivery menu separately, cap promotions, and push your own pickup and first-party ordering as the profitable channel.

How do I get through the January slowdown? Reserve a defined share of December cash before you spend it, build a leaner January schedule from a real sales forecast, and add a shoulder-season revenue play like private events or a prix-fixe promotion. And keep your RST and GST remitted weekly, because that tax money is a trust fund, not profit to lean on.