Restaurant KPIs: the metrics worth tracking every week (and the ones to ignore)

Search "restaurant KPIs" and you will find lists of 15, 22, even 30 metrics, usually written by software vendors whose dashboard happens to display all 30. For a multi-site group with a finance team, fine. For an independent restaurant owner who does the ordering, the schedule and half the service, a 30-metric dashboard is not a management tool, it is homework that never gets done.

The reality of the trade is harsher and simpler. Margins are thin: in the National Restaurant Association's 2026 State of the Restaurant Industry, 42% of US operators reported their restaurant was not profitable in 2025, and UKHospitality's Q2 2025 survey found one third of UK hospitality businesses trading at a loss. When the game is decided over a few percentage points, what saves you is not measuring everything; it is measuring the right handful of things every single week, early enough to react.

This guide covers the nine KPIs that earn a place on an independent's weekly scorecard, how to compute each one from data you already have (your POS and your reservation book, nothing more), the vanity metrics that waste your attention, and a one-page scorecard template you can copy tonight.

In short:

• nine KPIs are enough: covers, average check, food cost %, labor %, prime cost, table turnover, no-show rate, RevPASH and your review score;
• every one of them can be computed from two sources you already have: the POS and the reservation book;
• weekly beats monthly: cost drift caught in seven days is a correction, drift caught at the accountant's is a bad year;
• revenue, social media followers and total reservations are vanity metrics: they can all go up while the restaurant loses money;
• a paper scorecard filled in every Monday beats a beautiful dashboard nobody opens.

Why weekly, and why so few?

A KPI (key performance indicator) is only useful if it passes two tests: you can act on it this week, and you will actually look at it. Most metrics on the long vendor lists fail one or both. Employee turnover rate matters, but you do not act on it weekly. Website traffic is interesting, but no decision follows from it on a Tuesday morning.

The nine metrics below pass both tests. They split into three families:

Notice what makes the list work: none of it requires new software. If you take reservations on paper, the reservation-book column still works; it is just slower to add up. A free online reservation book like ViteUneTable does the counting for you (covers, no-shows, party sizes are logged automatically), but the method comes first and the tooling second.

The money KPIs: where the margin lives

Food cost percentage

Food cost % = (opening inventory + purchases − closing inventory) ÷ food sales × 100

The share of every dollar (or pound) of food revenue that goes back out the door to suppliers. The National Restaurant Association's pre-pandemic baseline for an average US restaurant puts food and beverage costs around 33 cents of every sales dollar, with wholesale food prices up 35% between 2019 and 2026. Your own target depends on your concept; the weekly job is not hitting a magic number, it is catching movement. A two-point jump in one week means a supplier price moved, portions crept, or something is walking out the back door.

Compute it weekly from purchases (this week's invoices) against the POS food sales, with a quick inventory adjustment if you can manage it. A monthly deep version with a full count keeps the weekly quick version honest.

Labor percentage

Labor % = fully loaded payroll ÷ net sales × 100

Fully loaded means wages plus payroll taxes and benefits, not just the hourly rates. The same NRA baseline puts labor at roughly another third of sales, with average hourly earnings up 41% since 2019. Read this one as a trend across comparable weeks: a rainy quiet week mechanically inflates the ratio without anyone having done anything wrong. The actionable question is whether the schedule matched the actual covers, service by service.

Prime cost

Prime cost % = food cost % + labor %

The single most important number in the restaurant business: your two big controllable costs added together. In the NRA baseline they absorb about two thirds of revenue, and since occupancy and other operating expenses take close to 29% more, a prime cost drifting above 70% leaves essentially nothing. We break down prime cost, break-even covers and the full cost structure in our guide to restaurant profit margins; for the weekly scorecard, the discipline is simply to write the number down and watch the direction.

Average check (or spend per head)

Average check = revenue ÷ covers served

What Americans call the average check, UK operators call average spend per head or the average bill; the math is identical. It is the quality of your revenue where covers are the quantity. At equal traffic, every extra dollar of average check falls almost entirely to margin, because the guest is already seated and the fixed costs already paid. Track it separately for lunch and dinner: blending a $18 lunch and a $46 dinner into one number hides everything. If it stagnates while your costs rise, that is a menu and selling problem, and there are proven, guest-friendly ways to work on it.

The seat KPIs: your reservation book is a data source

Most KPI lists stop at the POS. That misses half the picture, because the POS only sees guests who showed up. Your reservation book sees the ones who booked, the ones who cancelled, the ones who vanished, and the empty chairs in between.

Covers

The raw count of guests served, per service. It is the denominator of half your other ratios, so record it first: Monday lunch, Monday dinner, and so on. Two weeks of per-service cover counts will teach you more about your restaurant than a year of monthly revenue totals, because they show exactly which services are weak. Weak services are fixable; a weak month is just a fog.

Table turnover

Table turnover = parties served ÷ available tables (per service)

How many times each table earns its keep in a service. TheFork Manager's guide gives the same formula and notes that a common average is around 3 turns per service, while stressing it depends entirely on the concept: a lunch brasserie lives on turns, a tasting-menu room deliberately sells one long seating. The metric is not "more is better"; it is "know your number and design your booking slots around it", then raise it through pacing and floor-plan work rather than by rushing guests.

No-show rate

No-show rate = no-show reservations ÷ total reservations × 100

The KPI that only exists if you log it, and the one with the most direct money attached: a no-show is a cover you counted on, at a table you may have refused to someone else, with fixed costs already paid. The scale is documented on both sides of the Atlantic: a YouGov survey commissioned by OpenTable found 28% of US adults admitting to skipping a reservation without cancelling in the previous year, and Zonal and CGA's UK research measured a no-show rate of 14% in 2024, up from 12% a year earlier. Tag every booking as honored, cancelled or no-show and the rate computes itself. If yours runs high, confirmations, reminders and deposits all work, and we detail them in our guide to reducing restaurant no-shows.

RevPASH: one number for the whole room

RevPASH = revenue for a period ÷ (available seats × opening hours in that period)

Revenue per available seat-hour, the metric introduced for restaurants by Cornell professor Sheryl Kimes in her work on restaurant revenue management published in the Cornell Hotel and Restaurant Administration Quarterly in 1998. Where covers measure traffic and average check measures spend, RevPASH multiplies everything together: how hard is each seat working per hour of opening?

A worked example: 40 seats, open 4 hours for dinner service, $2,400 of dinner revenue. RevPASH = 2,400 ÷ (40 × 4) = $15 per seat-hour. Say Saturday runs at $15 and Tuesday at $4: that gap, expressed in one comparable unit, tells you where the upside is. You raise RevPASH by filling more seats, turning tables at the right pace, or lifting spend, which is precisely why it makes a good headline number: any real improvement anywhere shows up in it, and it cannot be gamed by discounting (that lifts covers but drops the numerator).

There is no honest universal benchmark for RevPASH; concepts differ too much. Compare yourself to yourself, service by service, week by week.

The reputation KPI: your review score

One number from outside your walls belongs on the scorecard: your Google rating (and its count of new reviews this week). The link to revenue is measured, not folklore: Harvard Business School economist Michael Luca matched Yelp ratings with restaurant revenue reported to tax authorities and found a one-star increase corresponds to a 5 to 9 percent revenue increase, with the effect concentrated on independent restaurants. Weekly, the job takes five minutes: note the rating, read the new reviews, reply to them. A sudden dip in a specific week usually points to a specific service, and your cover log tells you which one.

Vanity metrics: the numbers that flatter and teach nothing

A vanity metric is any number that can go up while the business gets worse. The usual suspects:

• Revenue on its own. The most dangerous one, because it feels like the whole point. A record month with a prime cost at 74% is a record loss. Revenue only means something next to its costs.
• Social media followers and likes. Nice to have, impossible to bank. Ten thousand followers who never book are worth less than one regular who comes twice a month. If you want a marketing number, count bookings and covers, which is where marketing either shows up or did not happen.
• Total reservations taken. Reservations are a promise, covers are revenue. Counting bookings without tracking the no-show rate is counting chickens before they hatch, precisely 14% of which, on UK figures above, never do.
• Website visits. A means, not an end. The number that matters is how many visits end in a booking.
• Number of menu items. Often worn as a badge of pride, usually a cost driver: more prep, more waste, more inventory, slower tickets.

The pattern: vanity metrics count inputs and applause; real KPIs count money and seats.

A simple weekly scorecard (copy this)

One page, filled in every Monday morning for the week just ended, next to the same week last month and, once you have the history, the same week last year. On paper, in a spreadsheet, wherever you will actually keep doing it.

Three rules make it work:

1. Same day, same definitions, every week. A KPI measured differently each week measures nothing.
2. Write down one action. The scorecard's output is a decision: rework Thursday's schedule, call the fish supplier, add a reminder email for Saturday bookings. No action, no point.
3. Watch directions, not decimals. One week is weather; three weeks is a trend.

Where does the data come from? Food cost and labor come from invoices and payroll. Everything else comes from the POS and the reservation book. Let's be honest: no reservation system computes your prime cost, ViteUneTable included; that side of the scorecard stays with you and your accountant. What a reservation system does is make the seat-side KPIs automatic instead of a counting chore. ViteUneTable's free plan logs every booking, party size and status with 0% commission, no per-cover fees and no time limit; the Standard pack (€29 excl. VAT per month) adds automatic email reminders and Reserve with Google to push the no-show line down; and the Standard + Anti No-Show pack (€49 excl. VAT per month) adds credit card holds for the bookings that hurt most.

Frequently asked questions

What are the most important KPIs for a restaurant?

For an independent restaurant, nine cover the essentials: covers served, average check, food cost percentage, labor percentage, prime cost, table turnover, no-show rate, RevPASH and your Google review score. Prime cost (food plus labor as a share of sales) is the single most important one: in the National Restaurant Association's baseline it absorbs about two thirds of revenue, and above roughly 70% there is nothing left for profit.

How often should I check my restaurant's KPIs?

Weekly. Monthly is too slow: a cost drift discovered after four weeks has already cost you four weeks, and the trail (which supplier, which service, which schedule) has gone cold. A fixed ritual works best: every Monday, one page, the week just ended against the previous one, plus one concrete action.

What is RevPASH and how do I calculate it?

RevPASH is revenue per available seat-hour: revenue for a period divided by (seats × opening hours). The metric was introduced for restaurants by Cornell professor Sheryl Kimes in 1998. A 40-seat room taking $2,400 over a 4-hour dinner service runs at $15 per seat-hour. Its strength is combining traffic, turnover and spend in one comparable number; there is no universal benchmark, so compare your own services against each other.

How do I calculate my no-show rate?

Divide no-show reservations by total reservations for the period, times 100. It requires tagging every booking as honored, cancelled or no-show, which a digital reservation book does automatically. For scale: UK research by Zonal and CGA measured a 14% no-show rate in 2024, and 28% of US adults told a YouGov survey they had skipped a reservation without cancelling in the previous year.

Which restaurant metrics are vanity metrics?

Any number that can rise while the business loses money: revenue without its costs, social media followers, total reservations taken (without tracking no-shows), website visits, menu size. They count applause and inputs. Useful KPIs count money (prime cost, average check) and seats (covers, turnover, no-shows, RevPASH).

Can I track restaurant KPIs without buying software?

Yes. Food cost and labor come from invoices and payroll, covers and average check from any POS, and a paper reservation book still yields covers, no-shows and turnover if you count them up weekly. Software just removes the counting: ViteUneTable's free plan, for example, logs bookings, party sizes and no-shows automatically, with no commission and no time limit, so the seat-side of the scorecard fills itself.