Covers Per Labor Hour: Measuring Kitchen Productivity

The problem with measuring productivity in dollars

Most restaurants measure labor productivity in money. Sales per labor hour, labor cost percentage, sales per head — all of them put revenue on one side of the ratio. That works until you try to compare two periods, at which point the measure starts telling you about pricing when you wanted to hear about work.

Consider a kitchen that produced 380 covers on 120 hours in March and 380 covers on 120 hours in June. Identical work. In between, the menu went up 6% and a new dessert list pushed average check from $38 to $42. Sales per labor hour rose from $120 to $133 and the report calls it an 11% productivity gain. Nobody worked any harder. Nobody worked any less.

The reverse is worse, because it hides real problems. A site running a discounted set menu, or a site whose guests have quietly traded down, can show falling sales per labor hour while the kitchen is plating more food than it has all year. Cutting hours in response is exactly the wrong move.

Covers per labor hour removes the distortion by counting guests instead of revenue. It is the same ratio with the pricing variable taken out.

How to calculate it

Covers served divided by total paid labor hours for the same period. That is the whole calculation, and the covers per labor hour calculator will also give you the inverse — minutes of paid labor per cover — which most kitchen teams find easier to reason about.

Three definitional choices matter more than the arithmetic, and the only real requirement is that you make them once and never quietly change them:

• What is a cover. One guest, not one check and not one table. Counting checks understates the work in any site with shared tables or a lot of two-tops.
• Which hours count. Every paid hour in the period, including prep, dish, opening and closing. Excluding the unglamorous hours produces a flattering number that cannot be used to build a schedule.
• Whether salaried staff are in. Include them for a true cost-to-run picture, exclude them if the figure is purely a scheduling tool for hourly staff.

Calculate kitchen and floor separately as well as combined. The combined number is convenient for comparing weeks, but it can stay flat while a stretched kitchen quietly offsets an overstaffed floor, and it never tells you which side to act on.

What drives the number up and down

Menu complexity

The single largest structural driver in the kitchen. Every additional item adds prep, adds a mise en place position, and adds stock that has to be handled. Two sites with identical covers and identical staffing can differ by 40% in kitchen productivity purely on how many dishes are on the card and how much they share components.

Seating pattern

The floor figure is dominated by when guests arrive rather than how many. Two hundred covers spread evenly over four hours needs far fewer paid hours than two hundred arriving in two peaks, because the peaks set your headcount and the troughs still get paid. Anything that spreads demand — earlier reservations, a bar area, a second sitting — improves the ratio without anyone moving faster. This is the overlap with table turnover.

Fixed hours as a share of the total

Opening, closing, prep, deliveries and deep cleans happen at 40 covers and at 140. On quiet days those fixed hours dominate, which is why the ratio always looks worse midweek. Knowing what proportion of your hours are genuinely fixed stops you chasing a number you cannot move and points you at the one you can — the variable hours at the shoulders of service.

Overtime and turnover

Both make hours more expensive without making them more productive. Overtime buys the same output at a premium, and a churning team spends a share of every shift training rather than producing. If your covers per labor hour is acceptable but labor cost still looks wrong, the problem is usually the price of the hours, not the number of them — see reducing overtime and the cost of staff turnover.

Turning it into a schedule

A productivity figure that never reaches the rota is just reporting. The way to make it operational is to invert it: take the forecast covers for a shift, divide by your target rate, and you have the hours the shift should need before anyone is placed.

A Thursday dinner forecast at 160 covers, at a target of 1.6 covers per labor hour, is a 100-hour shift. If the draft rota comes to 118 hours, you have an 18-hour conversation to have before the week starts rather than a variance to explain after it. Build the target from your own best comparable shift at acceptable quality, not from a benchmark table — a target the team recognises as achievable gets worked toward, and an arbitrary one gets ignored.

Set separate targets by shift type. Lunch, dinner and weekend brunch have different arrival patterns, different menus and different service intensity, and a single site-wide target quietly overstaffs one of them while punishing another.

Reading it alongside the money

Covers per labor hour is deliberately blind to price, which is its strength and also its limitation. It tells you how much work each hour produced, not whether that work was worth doing. Read it next to two other numbers and the picture closes.

First, sales per labor hour, which is covers per labor hour multiplied by average check. If covers per hour held steady and sales per hour fell, your problem is check, not scheduling — and no amount of hour-cutting will fix it. The sales per labor hour guide covers that side in detail.

Second, labor cost percentage, which reflects what those hours cost. Productivity can be flat while cost percentage rises, because overtime, wage increases or a heavier senior-staff mix changed the price of an hour. These are different problems with different fixes, and only looking at both distinguishes them.

The diagnostic pattern is straightforward. Covers per hour down and cost percentage up means a scheduling problem. Covers per hour flat and cost percentage up means an hourly rate or overtime problem. Covers per hour up and cost percentage up means volume fell faster than you cut hours. Each points somewhere different.

Where it goes wrong

The most common failure is treating a higher number as automatically better. Push covers per labor hour far enough and you are not gaining productivity, you are forming a queue: ticket times stretch, remakes rise, the floor stops checking back and the guest experience absorbs the saving. Put ticket times and complaint counts on the same weekly report so the trade-off is visible while it is happening rather than in next quarter's reviews.

The second failure is comparing across service models. A tasting menu kitchen at 0.8 covers per labor hour and a fast casual counter at 5.0 are not evidence of anything. The check sizes differ by an order of magnitude and so does what the guest is buying. Compare a site to itself over time, and to sites running the same format, or do not compare at all.

The third is measuring too often. Daily figures are dominated by noise — one large party, one delivery day, one deep clean — and chasing them produces reactive scheduling that unsettles the team without improving anything. Weekly, split by shift type, is the cadence that supports decisions.

Used carefully, this is one of the few operational numbers that means the same thing this year as it did last year. That stability is the point. Everything else in a restaurant P&L moves when prices move; covers per labor hour only moves when the work does.

Frequently asked questions

Is covers per labor hour better than sales per labor hour?

Neither is better; they answer different questions. Covers per labor hour measures work and ignores price. Sales per labor hour measures revenue efficiency and moves with average check. Tracked together, they let you tell a productivity change from a pricing change.

What is a good covers per labor hour figure?

It depends entirely on service model. Full service table service typically runs 1.0 to 2.0 combined, casual dining 2.0 to 3.0, and counter service well above that. Fine dining sits below 1.0 and is not underperforming for it.

Should prep hours be included?

Yes. Prep is labor you pay for that produces covers, just on a different day. Excluding it makes the kitchen look more productive than it is and produces schedule targets that cannot actually be worked.

How do I handle a shift with a large private party?

Count the covers and the hours as normal, but note the event on the record. Large parties are usually served at a much higher rate than the floor average, so an unmarked event week will set an unreachable target for the following month.

Can I use it to compare two of my own sites?

Yes, provided they run the same menu and service style. Where they differ, the comparison tells you about format rather than performance. Comparing each site against its own trend is always safe; comparing them against each other rarely is.

What is the quickest way to improve the number?

Staggered finish times. Most low readings come from the last ninety minutes of service, when arrivals have stopped but the rota has not, and this is usually the largest recoverable block of hours in the week.

Does it work for delivery-heavy operations?

For the kitchen, yes — count each order as a cover and the figure behaves normally. For the floor it breaks down, because delivery consumes almost no service labor, so report the two channels separately rather than blending them.

Run the numbers

Use the free Covers Per Labor Hour Calculator to apply everything above to your own figures.