Restaurant KPIs worth tracking every week

Restaurant KPIs are the handful of numbers that tell an owner or manager whether the business is on course, early enough to do something about it. KPI stands for key performance indicator, and the important word is key: not every number the till can produce, but the few that move profit and that someone in the building can change. A venue with forty KPIs has none; a venue with eight that are read every week, with an owner and a target for each, is run on facts.

Most venues already have the data. The point of sale records sales, covers and checks; the booking system records bookings, no-shows and party sizes; payroll and supplier invoices hold labour and food costs. What is often missing is the discipline of choosing the few numbers that matter, putting them on one page, and deciding in advance what happens when one of them moves.

This guide sets out eight numbers worth tracking every week, what each one is for and how to read it, and the three that most operators track faithfully without ever acting on. It ends with a one-page KPI sheet, a worked example and the mistakes that turn a good list into a ritual.

A performance indicator is a measure chosen because it shows how well something important is going. In a restaurant, that means a number connected to profit, guests or the team, measured the same way every time, and compared with a target or with the past.

A KPI is not a report. A report can hold dozens of figures; a KPI is a single number with a purpose. It is also not a goal on its own. A food cost percentage of 30 per cent is a target; the KPI is the measured figure each week, read against that target. Keeping the two apart stops a list of hopes from being mistaken for a list of measures.

Every extra number on the list takes attention from the others. A manager who reads twenty figures on a Monday morning will skim most of them and act on none. A manager who reads eight can hold all of them in mind, notice which one moved, and ask why.

A short list also forces choices. Deciding which eight numbers matter most is itself a useful conversation, because it makes the owner and the team agree on what the business is trying to do this year: grow covers, raise spend, protect margin or reduce waste. The list should follow from that, not from whatever the software happens to show.

Each number on the list should pass four tests:

A number that fails the second test, such as the weather or the rent, belongs in the background. A number that fails the fourth is a statistic, not a KPI. The rest of this guide applies those four tests to eight numbers.

Covers, the number of guests served, are the base of almost every other figure. Track them by service, lunch and dinner at least, and compare each with the same week last year. A change in covers tells you whether more or fewer people chose the venue, before any question of how much they spent.

Covers belong to the general manager or the front of house manager, because they are shaped by bookings, walk-ins, table management and how quickly the waitlist is worked. Our guide to the covers report, read on a Monday morning covers how to read them in a few minutes.

Average spend per head is sales divided by covers: what each guest spent on average. It shows whether the menu, the prices and the way the team sells are working. A rise in covers with a fall in spend may mean a promotion pulled in guests who spend less; a rise in spend with flat covers may mean a price change or better selling of drinks and desserts.

Never read it as a single figure across the week. Lunch and dinner, weekdays and weekends, food and drink all behave differently. Our guide to average spend per head, split properly covers the splits that make it useful.

Net sales, after discounts and before VAT or sales tax, are the headline. On their own they say little, because sales rise and fall with the season. Against the same week last year, adjusted for holidays that moved, they show whether the business is growing.

Sales are the product of the first two KPIs, covers multiplied by spend per head, so read them together. When sales move, ask which of the two moved them. That single question usually points to the right action: a covers problem calls for bookings and marketing; a spend problem calls for the menu and the floor.

Prime cost is the cost of goods sold, food and drink, plus the full cost of labour, including taxes and benefits. As a share of sales, it is the most important cost figure in the restaurant, because it is the largest controllable cost and it moves quickly. A venue that keeps it in its target range rarely gets into trouble; one that lets it drift for a quarter usually does.

Track prime cost weekly, not monthly. It needs a reasonable estimate of food and drink cost each week, from purchases and a quick count of the most valuable stock, and the week's wage bill. Our guide to prime cost, the one number to watch weekly covers how to calculate it in under an hour.

Labour cost percentage is the wage bill, with taxes and benefits, as a share of sales. It is part of prime cost but deserves its own line, because the decisions that shape it, the rota and the hours worked, are made by different people from those who buy food.

Read it by service rather than by week alone. A week can look fine overall while Tuesday lunch runs at a labour cost far above the target and Saturday dinner runs below it. Our guide to labour cost percentage by service covers how to split it.

Food cost percentage is the cost of food used as a share of food sales; drink cost percentage is the same for drinks. Together with labour, they make up prime cost. Separately, they show whether the kitchen and the bar are buying, storing, portioning and pricing as planned.

The weekly figure is an estimate, from purchases adjusted for changes in stock. A full stocktake once a month gives the exact figure. Track both, and look at the gap between the weekly estimate and the monthly figure: a gap that grows suggests waste, portion drift or recording problems. Our guide to calculating food cost percentage covers the method.

Table turnover rate is the number of times each table is used during a service. It shows how hard the room is working. A higher rate on busy services means more guests served from the same seats, provided the guests do not feel rushed.

It matters most on the services that sell out. On a quiet Tuesday, turnover hardly matters, because there are empty tables anyway. On a full Saturday, a few minutes saved between seatings can add a full sitting of covers across the room. Our guide to table turnover rate covers how to raise it without hurrying anyone.

No-show rate is the share of booked covers who did not arrive and did not cancel. It costs the venue twice: in preparation for guests who never came, and in tables that could have been sold to someone else. A rising no-show rate on busy services is one of the clearest signals that the booking process needs attention.

Track it by service and by party size, and measure it the same way every week. Our guide to your no-show rate covers how to count it fairly before deciding what to do.

Some venues add one or two numbers to the eight. A busy kitchen may track ticket times, the minutes from order to pass, because slow tickets lose turns and goodwill. Our guide to ticket times covers how to read them. A neighbourhood restaurant built on regulars may track the share of covers from returning guests.

Add them only if someone will own them and act on them. A tenth number that is read but never discussed is worse than none, because it trains the team to skim the list.

In most venues, at least three numbers are faithfully collected, printed and read, yet nothing ever changes because of them. They are not bad KPIs. They are good KPIs used badly: measured too late, averaged too far, or owned by nobody.

The three that most often fall into this trap are food cost percentage from the monthly accounts, a single table turnover figure for the whole week, and one average spend per head for every service. Each can be rescued with a small change in how it is measured and who owns it.

Many venues learn their food cost percentage from the accountant, weeks after the month ends. By then, the menu has moved on, the supplier price that caused the problem has been paid for several weeks, and the stock that was wasted is long gone. The number is accurate and useless.

The fix is a weekly estimate owned by the head chef: purchases for the week, adjusted for the stock of the five or ten most expensive items, against food sales. It will not match the accounts exactly. It does not need to. It needs to show, within a week, that something has moved, so the chef can find out what.

A single turnover figure for the week mixes services that sell out with services that are half empty. The average looks fine, so nobody acts, while Saturday dinner loses a sitting to slow table changes and Tuesday lunch has turnover that does not matter.

Track table turnover rate only for the services that sell out, and give it to the front of house manager for those services. Then look at the parts that shape it: how long a table waits for the bill, how long it waits to be reset, and how long the first course takes. Each has an owner and an action.

A single spend figure for the week is the most common of the three. It moves a little each week, everyone nods, and nothing happens, because nobody can tell whether the change came from lunch, dinner, food, drinks or the mix of guests.

Split it: by service, and into food and drink at least. Then give each split an owner. The bar manager owns drinks spend at dinner; the head chef owns how often desserts are ordered; the floor manager owns spend at weekend lunch. Split and owned, average spend per head becomes one of the most useful numbers on the sheet.

Some KPIs report what has already happened: net sales, prime cost, food cost. These are lagging indicators. Others give warning of what is coming: covers booked for next week, the share of tables booked online, how far ahead guests are booking. These are leading indicators.

A good sheet has both. The lagging numbers tell you whether last week went to plan; the leading numbers tell you whether next week will. A venue that only reads lagging numbers is always reacting. One that adds a leading number or two, such as covers already booked against the same point last year, has time to act.

A KPI without a target is a number to admire. For each one, agree what good looks like: a range for prime cost and labour cost, a minimum for covers against last year, a ceiling for no-show rate. Base the targets on the venue's own history and budget rather than industry rules of thumb, which vary widely between formats and countries.

Review the targets once or twice a year, not every week. A target that moves every time the number misses it stops being a target.

Weekly figures move about. A food cost of 29.5 per cent one week and 30.8 the next may mean nothing more than a delivery that landed on a Saturday. Treat each target as a range, with a green band where no action is needed, an amber band that calls for a question, and a red band that calls for action.

Ranges stop the weekly review from chasing noise. They also make the sheet easier to read: anything in the green band can be passed over in seconds, leaving time for the numbers that moved.

Every number needs a name beside it: the person who explains it when it moves and proposes what to do. Covers and no-show rate usually belong to the front of house manager; food cost to the head chef; drink cost to the bar manager; labour cost to whoever writes the rota; prime cost and sales to the general manager or owner.

Ownership is not blame. The owner is the person closest to the decisions that move the number, so they are best placed to explain it. When a number moves for reasons outside their control, such as a supplier price rise, the owner's job is to say so and suggest a response.

Daily figures are useful for running service; weekly figures are better for running the business. A single day is too noisy for most KPIs: one large party, a storm or a broken oven moves it more than any decision. A week smooths that noise while still being recent enough to act on.

Some numbers deserve a daily glance: covers booked for tonight and tomorrow, and sales against the same day last week. Keep those to a simple screen. Our guide to a restaurant dashboard you look at daily covers what to put on it.

Put all eight numbers on a single page or screen, in the same order every week. For each, show this week's figure, last week's, the same week last year, the target range, the owner, and a colour for where it sits in the range. Add a line at the bottom for notes: events, holidays, problems.

The page should take two minutes to read. If it takes longer, it has too much on it.

The KPI sheet sits on top of the venue's restaurant reporting. Each number on the sheet has a report behind it: the sales report, the covers report, the labour report, the food cost report and so on. When a KPI moves, the owner goes to the report behind it to find out why.

Keep the reports few and the KPIs fewer. Our guide to restaurant reporting: six reports, no more sets out the reports that support a sheet like this one.

Against last week, a KPI shows short-term movement. Against the same week last year, it shows whether the business is genuinely better or worse, with the season taken out. Show both on the sheet. Align weeks by weekday and mark moving holidays, so that a festival that fell in a different week does not create a false alarm.

In the first year of trading, there is no last year. Use the last four weeks as the comparison until a full year has passed.

In a group, give each venue the same eight KPIs, measured the same way, so the sheets can be compared. Add a single group sheet that shows each venue's figures side by side. The comparison is useful, but it should prompt questions rather than league tables: a venue with a high labour cost may simply be the one with the longest opening hours.

Agree definitions centrally: what counts as a cover, which labour costs are included, how food cost is estimated. Without that, a group sheet compares different measures and the conclusions are wrong.

A small venue with one owner on the floor may not have the time for eight numbers every week. Start with four: covers, average spend per head, prime cost and no-show rate. Those four show whether guests are coming, what they spend, whether the main costs are under control and whether the book can be trusted.

Add the others as the venue grows or as a particular problem appears. A small venue struggling with labour hours, for example, may add labour cost by service for a few months.

When a number falls into the red band, the owner explains it at the weekly review and proposes one action, with a date to check the result. The action should be specific: re-cost three dishes, move one cook's start time, call large bookings the day before. A vague promise to watch it more closely is not an action.

Check the result at the agreed date. If the number has returned to the green band, note what worked. If it has not, try the next action. Over a year, the notes become a record of what works in this venue. Our guide to a weekly review that takes thirty minutes sets out the meeting where this happens.

Share the KPIs with the people who move them. A kitchen that sees the weekly food cost estimate takes portioning more seriously; a floor team that sees spend per head by service understands why the specials matter. Keep the version for the team simple: the numbers they influence, the target, and how last week went.

Avoid using KPIs to rank individuals in public. Celebrate the team's results instead, and deal with individual performance privately.

Review the list once a year, or when the business changes: a new format, a second venue, a new service. A number that has sat in the green band for a year may be ready to drop to a monthly check; a new problem may need a number of its own for a while.

Change the list deliberately and rarely. A list that changes every few weeks never builds the history that makes a KPI useful.

Some numbers feel good but rarely lead to a decision: followers on social media, page views on the website, the number of five-star reviews this month. They can be worth watching in a marketing report, but they do not belong on the weekly KPI sheet unless a clear action follows from them.

The test is simple: if this number doubled or halved next week, would anyone do anything differently? If not, leave it off.

Here is an illustrative KPI sheet for a 70-seat venue in one week. The figures are examples, not benchmarks.

Actions: the front of house manager calls large bookings on Thursdays; the chef re-costs the two dishes whose supplier prices rose; the rota trims one lunch shift midweek; the floor team tries a dessert round at weekend lunch. All are checked at next week's review.

In servearound, the dashboard shows revenue, covers, spend per cover and no-show rate against the previous period, with what is happening right now: covers and bookings today, guests seated and waiting, tables in service and open orders. Analytics goes deeper by area. Table analytics shows utilisation, median seating length, revenue per seated hour and the gap between seatings; order analytics shows average and median check with void and comp rates; reservation analytics shows covers, no-show and cancellation rates by meal period, source and party size.

The menu report shows gross margin once dish costs are set, and kitchen analytics shows ticket times and on-time rate. Labour and purchase costs live in payroll and supplier invoices rather than in servearound, so prime cost is worked out from those alongside the sales figures. See Analytics.

Eight is a good number for most venues; four is enough for a small one. More than ten, and the list stops being read.

Prime cost is the most important cost figure, and covers with spend per head explain sales. If you could track only one, prime cost as a share of sales would be the one.

Weekly for most. A daily glance at covers booked and sales against the same day last week is useful for running service.

It depends entirely on the format: a café turns tables many more times than a tasting menu restaurant. Compare with your own busy services over time rather than with other venues.

From the venue's budget and its own history. Industry rules of thumb are a starting point at most, because they vary so much between formats and countries.

The owner and the managers who own the numbers see it in full; the wider team sees a simple version of the numbers they influence.

Either the target is wrong, or the action is not working. Revisit the target with the budget, and if it stands, try a different action.

They can be, if someone owns the response and acts on themes. A star rating watched without action is a vanity number.

Restaurant KPIs are the few numbers that show whether the business is on course early enough to act. Eight are worth tracking every week: covers, average spend per head, net sales against last year, prime cost, labour cost percentage, food and drink cost percentage, table turnover rate on the busy services, and no-show rate. Each needs an owner, a target range and an agreed action when it moves.

Watch for the three that are tracked without being acted on: food cost from the monthly accounts, one turnover figure for the whole week, and one spend figure for every service. Fix them by measuring weekly, splitting by service and giving each split an owner, and keep the whole sheet on one page, read at a short weekly review.