Sales Per Labor Hour Calculator
Enter the sales and the hours worked for a shift, a day or a week to see how much revenue each staffed hour actually produced.
What sales per labor hour measures
Sales per labor hour is the simplest productivity ratio in a restaurant: net sales divided by every hour somebody was paid to be on site. It answers one question — how much revenue did an hour of staffed time actually produce — and it answers it in a unit a manager can hold in their head while building next week's rota.
The figure is only honest if the hours are complete. Count front of house, back of house, and the salaried management time genuinely spent on site. Managers are the line most often left out, and leaving them out inflates the result by ten to twenty per cent — which quietly makes the most over-managed site look like the most productive one.
A worked example. A Tuesday dinner service takes $4,200 in net sales with 46 front of house hours, 34 kitchen hours and 8 management hours: 88 hours in total, or $47.73 per labor hour. The same restaurant on Saturday takes $11,800 across 148 hours, which is $79.73. Same menu, same wage rates, same building — and a 67% difference in how hard each staffed hour worked. That gap is the entire scheduling conversation, and no weekly labor cost percentage will ever show it to you.
Why it beats labor cost percentage for scheduling
Labor cost percentage is the right number for a period review and the wrong number for a Tuesday. It moves for two independent reasons — how many hours you used, and what those hours cost — and it cannot tell you which one moved. A shift landing at 34% might be badly overstaffed, or correctly staffed with your three most expensive people on at once. The percentage looks identical either way.
The two metrics are joined by an identity worth committing to memory, because it converts a percentage target into a scheduling target you can act on before the shift rather than after it:
Work it backwards. If your fully loaded average hourly cost is $19 and you want labor to land at 28% of sales, the sales per labor hour you need is $19 ÷ 0.28 = $68. From there the rota builds itself: hours you can afford equals forecast sales divided by 68. A $9,500 Friday supports roughly 140 hours. A $3,600 Monday supports 53. That is a decision a shift manager can make in ten seconds, without any percentage arithmetic during service.
It also explains why the two numbers sometimes disagree. Push wage rates up 8% and hold hours constant, and your labor cost percentage rises while sales per labor hour does not move at all. The scheduler did nothing wrong; the market did something. Sales per labor hour isolates the part a manager controls. Percentage blends it with the part they do not.
Target ranges by service model
| Service model | Typical sales per labor hour | What drives it |
|---|---|---|
| Coffee shop & bakery counter | $40 – $60 | Low average check, very high transaction count |
| Quick service & fast casual | $50 – $75 | Short service cycle, limited menu, tight prep |
| Pizza with delivery | $45 – $65 | Driver hours drag the denominator down |
| Casual full service | $50 – $70 | Check average against server section size |
| Bar-led venue | $65 – $95 | High drink mix, minimal kitchen hours |
| Fine dining | $70 – $110 | High check average carrying a high staff ratio |
Treat these as orientation, not as goals. They assume a broadly US-style wage structure, and the moment your average hourly cost differs from the norm the range shifts with it. The reliable way to set your own target is the identity above: divide your fully loaded average hourly cost by the labor percentage you are willing to run. Everything else is somebody else's cost base.
One structural point the table makes clearly: a delivery-heavy pizza operation and a bar-led venue are not doing the same job with an hour. Driver hours produce revenue slowly, bar hours produce it quickly, and comparing the two across sites tells you nothing except which format you chose. Compare a site to itself, on the same day of the week, before you compare it to anyone else.
Read it by shift, not by week
A weekly figure is an average of good decisions and bad ones, and it hides both. The value of this metric appears when you calculate it per shift and line the shifts up against each other.
- Split it by daypart. Lunch and dinner have different check averages and different staffing needs. Blending them produces a number that describes neither.
- Split front of house from back of house. Kitchen hours are far less flexible than service hours — a station has to be manned whether it sells six covers or sixty. If your figure is falling, knowing which side moved tells you whether you have a rota problem or a volume problem.
- Use hours worked, not hours scheduled. The gap between the two is where most labor overruns actually live, and the rota will never show it.
- Track the same day of the week over time. Tuesday against Tuesday over eight weeks is a trend. Tuesday against Saturday is just a description of the calendar.
- Post it where the team can see it. A single dollar figure per shift is understood immediately, in a way that a labor percentage never quite is.
Once you have eight weeks of per-shift figures, the weak shifts identify themselves. The usual finding is not that the restaurant is overstaffed but that it is overstaffed for two hours at each end of the day — a shape problem rather than a headcount problem, and a much cheaper one to fix.
What actually moves the number
Sales per labor hour has exactly two levers, and it is worth being explicit about which one you are pulling.
- Stagger start and finish times. The single largest gain in most operations. Bringing four people in at once because the shift starts at five is how you lose fifteen hours a week to an empty room.
- Raise check average rather than cutting hours. A dollar added to the average check flows straight into the numerator without touching service quality. Cutting hours flows into the denominator and eventually into the guest experience.
- Turn tables faster at peak. More covers through the same staffed hours is the purest form of improvement here, which is why table turnover and this metric move together.
- Cut the shape, not the depth. Trim the first and last hour of each shift before you trim the number of people on the floor at eight o'clock.
- Move prep to the quiet hours. Prep done on a Monday afternoon costs the same wage as prep done during a Friday rush, but only one of those hours could have been selling something.
Every one of these raises the number without reducing the number of people a guest can see. That distinction matters, because the fastest way to improve any labor metric is also the most expensive: cut until service degrades, watch sales fall, and discover the ratio has not improved at all because both halves moved together.
Where this metric will mislead you
Sales per labor hour is blind to menu mix and blind to cost. A site that pushes high-priced, low-margin items will post an excellent figure while making less money than the site next door, because revenue per hour says nothing about what that revenue cost to produce. Read it alongside prime cost or you will optimise the wrong half of the P&L.
It is also insensitive to wage inflation, which is a feature for scheduling and a defect for budgeting. A year of pay rises will leave the metric flat while your margin quietly compresses, which is precisely why the percentage still belongs in the period review even though it is useless on a Tuesday.
Finally, it can be gamed. Cut hours far enough and the number rises right up until the moment service collapses and sales follow it down. If the figure improves while covers, check average and review scores all fall, you have not found productivity — you have found the edge of your capacity, and you are already over it.
Frequently asked questions
What is a good sales per labor hour figure?
Most full-service restaurants sit between $50 and $70, bar-led venues between $65 and $95, and fine dining between $70 and $110. Rather than borrowing a benchmark, divide your fully loaded average hourly labor cost by the labor percentage you want to run — that gives you the target your own cost base actually supports.
Should I include salaried managers in the hours?
Yes, for any hours genuinely spent on site. Excluding them typically flatters the figure by ten to twenty per cent, and it makes cross-site comparison meaningless because the site with the deepest management structure will appear to be the most productive one.
Do I use gross or net sales?
Net sales, excluding sales tax and service charges, so that the numerator matches what appears on the top line of your P&L. Whichever you choose, use the same definition every period — a metric that changes its own definition is worse than no metric.
How is this different from labor cost percentage?
Labor cost percentage blends hours and wage rates into one figure, so it cannot tell you which of the two caused a change. Sales per labor hour holds wage rates out of it entirely and measures only what a scheduler controls: how many hours you deployed against the sales you took.
Can I use it to build a schedule directly?
That is its best use. Divide forecast sales for a shift by your target figure and you get the number of hours that shift can afford. A $9,500 Friday at a $68 target supports about 140 hours. Distribute those hours across stations and start times, and the rota is built before anyone opens a spreadsheet.
Should front of house and back of house be measured separately?
Ideally both together and apart. The combined figure is what you schedule against; the split tells you where a change came from. Kitchen hours are much less elastic than service hours, so a falling combined number caused by the kitchen is a volume problem, while the same fall caused by front of house is usually a rota problem.
How often should I calculate it?
Per shift if you can, and at minimum per day. A weekly figure averages away exactly the variation you are trying to see, and the whole advantage of this metric over a percentage is that it works at the level where scheduling decisions are actually made.