Revenue Per Available Seat Hour Calculator
Enter your seat count, the length of the service and what you took, and see what every seat hour actually earned — plus how much of the gap is occupancy and how much is the check.
What RevPASH actually measures
Covers and average check each tell you half a story, and the half they leave out is usually the expensive one. A hundred covers at $30 sounds better than sixty at $45 until you learn that the hundred took five hours to serve in a room built for eighty, while the sixty went through in two. RevPASH — revenue per available seat hour — puts the two halves back together and asks one question: for every hour that a seat existed in your dining room, how much money did it make?
The idea comes from hotels, where revenue per available room has been the standard measure for decades because a room-night is perishable inventory. A seat hour is exactly the same kind of inventory. A seat that sits empty between six and seven o'clock has not been saved for later — that hour is gone, and a full room at nine does not recover it. Counting only the seats you filled hides every seat you did not.
Because the denominator counts every seat you own for every hour you were trading, RevPASH is unforgiving in exactly the right way. Add ten seats you cannot fill and it falls. Open an extra hour that takes $80 and it falls. Turn the same room twice instead of once at the same check and it doubles. It is the only widely used restaurant metric that treats the dining room as capacity you are paying for rather than as scenery.
Working through a real service
Take an 80-seat restaurant trading dinner from 5pm to 11pm. That is six hours × 80 seats = 480 seat hours of inventory, whether anyone sits in them or not. The service takes $9,600 across 240 covers, so the average check is $40 and RevPASH is $20.00.
Now lunch in the same room: three hours, 240 seat hours, 120 covers at $26, revenue $3,120. RevPASH is $13.00. Lunch feels busy, and its covers per seat hour is actually identical to dinner's, but each seat hour earns roughly two thirds of what dinner earns while carrying the same rent, the same salaried manager and the same lights.
The optional dining duration field converts covers into the seat hours you genuinely sold, which is where the diagnosis lives:
At 240 covers holding a table for 95 minutes, dinner consumed 380 of the 480 available seat hours — 79% occupancy. Divide RevPASH by that occupancy and you get the ceiling: $25.25 per seat hour if every seat were occupied every minute. Nobody reaches 100% with real arrival patterns, so the ceiling is not a target. What it tells you is how much of your upside is still sitting in empty chairs and how much will have to come out of the check instead.
That third identity is the one to memorise. RevPASH moves for two reasons only: guests spent more, or seats worked harder. Every intervention you can name — a new menu, a shorter turn, a booking policy, a wine list rebuild, a two-hour table limit — lands in one of those two columns, and knowing which one keeps you from applying a check fix to an occupancy problem.
Read the week, not the night
A single RevPASH figure is close to meaningless. The value appears the moment you calculate it for every service period separately and line them up, because the spread between periods is almost always wider than operators expect. Here is the same 80-seat room across five periods:
| Service period | Seat hours | Revenue | RevPASH |
|---|---|---|---|
| Tuesday dinner (6 hrs) | 480 | $6,240 | $13.00 |
| Friday lunch (3 hrs) | 240 | $3,120 | $13.00 |
| Friday dinner (6 hrs) | 480 | $12,480 | $26.00 |
| Saturday dinner (6 hrs) | 480 | $14,400 | $30.00 |
| Sunday brunch (5 hrs) | 400 | $6,000 | $15.00 |
Saturday dinner earns 2.3 times what Tuesday dinner earns from identical fixed capacity. Friday lunch and Tuesday dinner produce the same RevPASH from very different-looking services — one short and full, one long and thin. That equivalence is invisible in a covers report and invisible in an average check report, and it is the sort of thing that decides whether you should be open on Tuesday at all.
Do not compare your RevPASH to anybody else's. A 40-seat neighbourhood bistro and a 200-seat hotel dining room produce numbers that cannot be reconciled, and published benchmarks are worse than useless because seat definitions differ. The only comparisons that mean anything are period against period in your own room, and this month against the same month last year.
Which lever to pull
Once you have split RevPASH into occupancy and check, the diagnosis is usually obvious and the two cases call for opposite behaviour.
Low occupancy, healthy check
The room is not full and the people in it are spending properly. Discounting here is close to self-harm — you would be reducing the check on guests who were already coming in order to attract guests who may not exist. The work is demand shaping: moving bookings into the empty hours rather than the peak, shortening the trading window so you stop paying for hours nobody wants, or adding a distinct offer that suits the quiet period. Check the arithmetic on closing a weak session against your break-even point before you cut it, because fixed costs do not fall when the door stays shut.
High occupancy, weak check
The room is full and each seat hour still earns little. Now the answer sits in the menu and the sequence of service, not in marketing. This is average check work: attachment rates on starters and sides, wine by the glass, a dessert offer that actually gets asked for, and pricing that reflects what the dish is worth. A dollar added to the check at 79% occupancy adds roughly 79 cents of seat hour revenue at no additional capacity cost, which is the cheapest revenue in the building.
High occupancy, healthy check, still constrained
You are out of inventory. The only remaining moves are faster table turnover, better table mix so two-tops are not sitting at four-tops, or physically more seats. Turn speed is the one to try first, because ninety seconds saved on payment across a full service is free capacity.
Decisions RevPASH changes
- Whether to open a session at all. A lunch running at a third of dinner's RevPASH is a candidate for closure, but only after you check the labour. Compare it against sales per labour hour for the same period — a weak session that is thinly staffed can still contribute.
- Whether to extend hours. Adding two hours adds seat hours to the denominator immediately and revenue only slowly. If the new hours earn less per seat hour than your variable cost of being open, you have bought yourself a busier, poorer restaurant.
- Whether to accept a large booking. A party of twenty at a modest per-head rate that occupies a quarter of the room for four hours can easily earn less per seat hour than the walk-in trade it displaces. Price group business per seat hour, not per head.
- Whether the seat count is right. Squeezing in eight extra covers raises capacity, and if you cannot fill them, RevPASH falls while service quality falls with it. Fewer seats turning faster frequently beats more seats sitting empty.
- Which menu changes to prioritise. Cross-read RevPASH with menu engineering. A high-margin dish that adds fifteen minutes to the table is not the win it looks like once the seat hour cost is charged against it.
Mistakes that make RevPASH useless
The first is measuring the whole day as one block. Averaging a dead Tuesday lunch with a heaving Saturday dinner produces a number that describes no service you actually run and hides the only insight the metric offers. Measure by period, and by hour if your point of sale will export it.
The second is mixing delivery, takeaway or private-room revenue into the dine-in figure. Those sales do not consume seat hours, so including them inflates RevPASH exactly in proportion to how much off-premise business you do — and makes the trend meaningless the moment that mix shifts. Keep delivery in its own analysis where the commission arithmetic is visible.
The third is changing your seat count without noting it. If you take out four covers to widen an aisle in March, every comparison across March breaks silently. Record the seat count alongside the figure every time.
The fourth is treating RevPASH as a target to maximise. It is a diagnostic, not a goal. The fastest route to a high RevPASH is to rush people out of the door, and a restaurant optimised purely for seat hour yield stops being somewhere people want to spend an evening. Use it to find the periods that are quietly losing money, then fix those with judgement.
Frequently asked questions
What is a good RevPASH for a restaurant?
There is no transferable figure, because it scales with your price point and your seat count. A fine-dining room and a canteen can both be well run at RevPASH numbers that differ by a factor of five. Judge it against your own other service periods and against the same period last year — those are the only comparisons that carry information.
Should bar seats and the terrace be included?
Include them only if their revenue is in the figure you enter. The rule is that the numerator and denominator must describe the same room. If your terrace is seasonal, run it as a separate calculation rather than adding forty seats to the denominator for eight months of the year when nobody sits outside.
How is RevPASH different from sales per square foot?
Sales per square foot measures how productively you use space over a long period, usually a year, and it is a property decision tool. RevPASH measures how productively you use space over a single service, and it is an operating decision tool. Square footage tells you whether the site works; RevPASH tells you whether Tuesday does.
Do I need average dining duration to use this?
No — RevPASH itself only needs seats, hours and revenue. Duration is what splits the result into occupancy and check, which is where the actionable part lives. If your point of sale timestamps table open and close, you already have it; otherwise a manual sample over one busy and one quiet service is close enough.
Can RevPASH go up while profit goes down?
Yes, and this is the main trap. Pushing turns harder often means more staff on the floor, more comps from rushed service and more remakes. If RevPASH rises while your prime cost rises faster, you have bought revenue with margin. Always read the two together.
What if a service period has almost no covers?
Then RevPASH will be very low and the calculator will flag it, which is the correct answer rather than a fault. The follow-up question is whether that period covers its own variable costs. A near-empty session that costs you three staff and a full kitchen brigade is a different problem from one run by two people who would otherwise be prepping.
How often should I calculate it?
Weekly, by service period, is enough for most operations. Monthly hides too much and daily encourages reacting to weather and one-off events. What you are looking for is which periods sit persistently at the bottom of the table, not which single Tuesday was quiet.