The Three Numbers to Check Every Monday Morning

A bad month is four bad weeks you could have caught. By the time your accountant hands you a P&L for March, it is the third week of April, the staffing mistake has been repeated twelve times, and the supplier price that crept up is now the price you accept as normal.

You do not need a full set of books every week. You need three numbers, pulled Monday morning, written in the same place every time: sales, labour as a percentage of sales, and food cost as a percentage of sales. Twenty minutes. No bookkeeper. The point is not precision — it is movement. A number that moves three weeks in a row is telling you something, and you want to hear it on week two, not in a quarterly review.

Number one: sales, split by where they came from

Start with the easy one. Total net sales for the week — Monday through Sunday, same seven days every time, no exceptions. Your POS will print this. If you close out daily, the weekly summary is usually two taps.

Net sales, not gross. Strip out tax and tips. Those are not your money and including them inflates every percentage you calculate afterward.

Then split it. At minimum: dine-in, your own takeout and online ordering, and third-party delivery apps. If you have catering, that is a fourth line. The split matters more than the total, because a flat week can hide a real problem — dine-in down 15 percent while delivery is up 15 percent looks like a steady week on one line and a bleeding dining room on two.

It also matters because a delivery dollar and a dine-in dollar are not the same dollar. Once commission comes out, a $40 app order might leave you less than a $28 walk-in ticket. If your sales line is flat but the mix shifted toward the apps, your actual take-home went down. We walked through the full arithmetic in why your delivery app orders cost more than you think.

One more habit: write down the number of covers or orders next to the sales figure. Sales divided by orders gives you average ticket, which is the single fastest way to tell whether a soft week was fewer people or the same people spending less. Those two problems have completely different fixes.

Number two: labour as a percentage of sales

Take your total payroll hours for the week, multiply by the wages you actually pay, and add anything salaried that runs the restaurant. Divide by net sales. Multiply by 100.

Most scheduling tools will produce the hours total for you. If you schedule on paper or in a group chat, add the hours from the schedule and then adjust for the shifts that ran long — that adjustment is usually where the surprise lives.

Two decisions you have to make once and then stick to:

• Do you include your own labour? If you work 50 hours a week on the line, your real labour cost includes a cook you are not paying. Include a number for yourself, even a modest one, or you will build a business that only works when you never take a day off.
• Do you include payroll taxes and statutory costs? Either way is fine. Just do it the same way every week, or your trend line is meaningless.

The target number depends entirely on your format. A full-service room with servers, bussers, and a prep team runs a very different labour percentage than a counter-service shop with four people. Do not chase a benchmark you read somewhere. Chase your own last eight weeks.

What to watch for: labour percentage climbing while sales stay flat means you are scheduling for a volume you are not doing. Labour percentage dropping while sales climb is usually good — unless service times are going up and reviews are getting sharper, in which case you are harvesting short-term margin from your reputation. The weekly check is where you catch that early. If this is the number that keeps moving on you, your staff schedule is a labour cost decision goes deeper on fixing it at the schedule level.

Number three: food cost as a percentage of sales

This is the one owners skip, because the textbook version requires counting inventory, and nobody is counting inventory at 9 a.m. on a Monday.

So use the shortcut. Add up every food and beverage invoice you received that week. Divide by net sales. That is your purchase-based food cost percentage.

It is not your true food cost. It is distorted by anything you bought ahead, by the week you stocked the freezer, by the long weekend order. Over a single week it can swing wildly. Over four weeks, the swings cancel out and you get something honest.

So track it two ways on the same sheet: this week's number, and the average of the last four weeks. The four-week average is the one you act on. The weekly number is there to show you which week caused a move in the average.

Do a real inventory count once a month if you can manage it, or once a quarter if you cannot. That corrects the drift. The weekly purchase number is your early warning system; the count is your audit.

What a rising food cost percentage actually tells you is one of four things: supplier prices went up, portions got bigger, waste went up, or your menu mix shifted toward low-margin dishes. Those are four different conversations with four different people. Knowing the number moved is what prompts you to figure out which one it was. If it turns out to be supplier prices, how to price a menu when food costs keep moving covers the repricing without scaring your regulars.

Why weekly beats monthly

Monthly numbers are the right unit for your accountant and the wrong unit for running a kitchen.

A month gives you twelve data points a year. Twelve chances to notice something. Fifty-two chances is better, and the reason is not that more data is inherently good — it is that a week is short enough to connect a number to a cause. You remember last week. You remember who was on, what broke, what the weather did, which supplier was late. You do not remember the second Tuesday of seven weeks ago.

Weekly also matches how restaurant costs actually behave. You schedule by the week. You order by the week. The lever and the measurement should operate on the same clock.

And a monthly review gives you one chance to correct. If your labour percentage drifts up in week one, a weekly check means you adjust the week two schedule. A monthly check means the drift runs for four weeks before you see it.

Make it a sheet, not a feeling

One page. Columns: week ending, net sales, dine-in, own takeout, apps, covers, average ticket, labour dollars, labour percent, food purchases, food percent, four-week food average. Google Sheets is fine. A printed grid taped inside the office door is fine.

Fill it in at the same time every Monday, before service thinking starts. Keep every week — the value compounds. By week twelve you can see your seasonality. By week fifty-two you can compare this November to last November, which is the only honest comparison in a seasonal business.

One caution about where the data comes from. If your sales live in four places — the POS, two delivery dashboards, and a notebook for phone orders — Monday takes an hour instead of twenty minutes, and you will stop doing it. Pulling your own ordering onto your own site puts more of the week into one report you control, and it is the same change that gets commission off those orders. If that is the bottleneck, what we build for restaurants covers how ordering and phone orders land in one place.

The three numbers will not fix anything by themselves. What they do is make the next decision obvious — which is more than most owners get from a P&L that arrives six weeks late.