RevPASH: The Metric Hotels Use That Restaurants Should

A seat hour is perishable inventory

Ask a hotel general manager how last night went and they will not tell you how many rooms they sold. They will tell you revenue per available room, because the industry worked out long ago that a room-night which goes unsold is not stock sitting in a warehouse waiting for tomorrow — it is gone. Occupancy alone can be bought with cheap rates, and average rate alone can be protected by leaving the hotel half empty. Only the combined figure resists both forms of self-deception.

Restaurants own exactly the same kind of inventory and mostly refuse to count it. Your dining room does not produce covers; it produces seat hours. An 80-seat room open six hours has 480 of them to sell tonight, and at eleven o'clock every unsold one has expired. Rent was charged on all 480. The manager's salary covered all 480. The lights were on for all 480.

RevPASH — revenue per available seat hour — is the restaurant version of the hotel metric, and it is calculated the same way: total revenue for a period, divided by seats multiplied by hours open. An 80-seat dining room taking $9,600 over a six-hour dinner earns $20.00 per available seat hour. That single figure carries both how full you were and how much people spent, which is why neither can be used to hide the other.

Why covers and average check both mislead

Every restaurant already tracks two numbers, and each of them is systematically misleading on its own.

Cover counts ignore money and time. Two hundred covers is a good night if they arrive at $45 across two turns and a poor night if they arrive at $22 across five hours. The cover count is identical, the P&L outcome is not remotely similar, and nothing in the covers report distinguishes them.

Average check ignores capacity. A restaurant can lift its average check by 15% while its revenue falls, if the guests it lost were the ones who would have filled a second turn. Operators do this to themselves regularly: they reposition upmarket, celebrate the check figure, and cannot work out why the bank balance is worse. RevPASH would have shown it on day one, because the check went up and the covers per seat hour went down by more.

The two combine into one identity worth remembering: RevPASH equals average check multiplied by covers per seat hour. That means there are exactly two ways to earn more from the same room — charge more per guest, or move more guests through each seat. Every plan you will ever be pitched is one of those two, and the discipline of naming which one keeps you from spending marketing money on what is actually a menu problem.

Calculating it without buying anything

You need three numbers you already have and one you can estimate: seats, hours open, dine-in revenue, and average dining duration. Nothing else. The RevPASH calculator will do the arithmetic, but the sequence matters more than the tool.

Start by defining the periods honestly. Lunch, dinner and brunch are separate businesses that happen to share a lease, and averaging them together destroys the only insight the metric offers. If your point of sale can export sales by hour, use hourly buckets instead — the shape of the day is where the money is hiding.

Then be strict about the numerator. Dine-in revenue only, net of tax and service charge. Delivery and takeaway do not sit in your chairs, so including them makes RevPASH rise whenever your off-premise mix rises, which tells you nothing about the dining room and quietly breaks every comparison you make afterwards. Off-premise belongs in its own analysis where the commission arithmetic is visible.

Finally, record the seat count next to every figure. Rooms change — a banquette comes out, a service station goes in, four covers vanish to widen an aisle. If the denominator moves without a note, a year of trend data becomes unreadable.

The shape of the curve is the point

Calculate RevPASH by hour for a week and you will get a chart with two or three sharp peaks and a lot of flat ground between them. That flat ground is where restaurants lose money without noticing, because nothing about it feels like a loss. Nobody is standing idle in a panic. The kitchen is calm. It simply is not earning, while every fixed cost continues to accrue at exactly the rate it does during the peak.

Three patterns show up repeatedly. The first is the long tail: a service that peaks at eight and drags on until eleven at a fraction of the rate, kept open by habit rather than by demand. The second is the dead shoulder: five to six-thirty produces almost nothing, then the room fills, which usually means booking policy and not demand is doing the shaping. The third is the false peak: a period with high covers and a collapsed check, typically a discounted or set-menu session that fills the room and earns less per seat hour than the quiet service either side of it.

Each pattern has a different remedy, and the remedy for one makes the others worse. Cut hours to fix a long tail and you damage nothing. Cut hours to fix a dead shoulder and you have removed the capacity you needed to spread the peak. This is precisely why a single daily figure is useless: it averages the three into a number that recommends nothing.

Fixing a weak period

Once you know which period is weak, split its RevPASH into occupancy and check before you act. Multiply covers by average dining duration in hours, divide by the seat hours available, and you have the proportion of your inventory that actually sold. Everything follows from that split.

• Empty room, decent check: the problem is demand, and discounting is the wrong instrument — it lowers the one thing that is working. Shape demand into the empty hours instead, or shorten the trading window so you stop paying for hours nobody wants. The arithmetic of discounting is brutal at low occupancy.
• Full room, weak check: the problem is the menu and the sequence of service. Attachment on starters, sides and wine by the glass moves the number faster than anything else, and at high occupancy nearly every extra dollar of check converts directly into seat hour revenue. Raising average check is the whole job here.
• Full room, decent check, still constrained: you are out of inventory and the only lever left is speed. Ninety seconds saved at payment across a full service is free capacity, which is why turnover discipline pays even in restaurants that do not want to feel rushed.
• Empty room, weak check: the period is not viable in its current form. Before closing it, check what it costs to run — a thinly staffed quiet session can still contribute, and break-even analysis will tell you whether closing actually saves anything, since the rent continues either way.

The pairing that matters most is RevPASH against labour. A period with poor seat hour revenue and heavy staffing is the one that is genuinely bleeding; a period with poor seat hour revenue and light staffing may be doing no harm at all. Read it alongside sales per labour hour and the picture resolves quickly.

Where RevPASH will mislead you

It is a diagnostic, not an objective, and the difference is not academic. The fastest way to raise revenue per available seat hour is to get people out of the door sooner, and a restaurant run purely to that number becomes somewhere nobody chooses for an anniversary. Guest experience is not a rounding error in this business; it is the demand curve.

It also says nothing about cost. A period can have excellent RevPASH and terrible profitability if it runs on overtime, agency cover and a menu of low-margin dishes. Revenue metrics describe the top line only, which is why they belong next to prime cost rather than instead of it.

And it does not travel. Comparing your figure with another restaurant's is meaningless unless you use the same seat definition, the same trading hours and a similar price point, which you almost certainly do not. Anyone quoting an industry RevPASH benchmark is selling something. The comparison that matters is Tuesday against Saturday in your own room, and this quarter against the same quarter last year.

Used that way — per period, next to a cost metric, over time — it is the shortest route to the question most operators never quite ask: which hours of the week am I paying to stay open for, and what would happen if I stopped?

Frequently asked questions

What does RevPASH stand for?

Revenue per available seat hour. It is the restaurant adaptation of RevPAR, revenue per available room, which hotels have used as their headline performance measure for decades. Both work on the same principle: capacity that goes unsold expires and cannot be recovered.

How do I calculate RevPASH?

Divide the dine-in revenue for a service period by the number of seats multiplied by the hours you were open. Eighty seats open for six hours gives 480 seat hours; $9,600 of revenue across that period gives a RevPASH of $20.00. Do it per service period rather than per day.

Is a higher RevPASH always better?

No. It rises when you rush guests, and it rises when you cut trading hours to peak only, neither of which is automatically good for the business. Treat a low figure as a question to investigate rather than a fault to eliminate, and always check what happened to costs and to guest experience alongside it.

Should takeaway and delivery revenue be included?

No. Those sales consume no seat hours, so including them inflates the figure in proportion to your off-premise mix and makes period-to-period comparison unreliable the moment that mix changes. Keep dine-in in the numerator and analyse off-premise separately.

How do I get average dining duration if my POS does not report it?

Sample it manually. Note table open and payment times for every table across one busy service and one quiet one, and average them. Two services of observation gets you within a few minutes, which is accurate enough for a metric you are using to compare periods rather than to file accounts.

Does RevPASH work for bars and cafes?

Yes, with one adjustment. Where a substantial share of trade is standing or takeaway, the seat count understates capacity and the figure will read low. Either restrict the calculation to the seated area and its revenue, or accept it as an internal trend line and never compare it with a full-service dining room.

How does RevPASH relate to table turnover?

Turnover is one of its two components. Covers per seat hour is essentially turnover expressed against capacity rather than against tables, and multiplying it by average check gives you RevPASH exactly. Improving turnover raises RevPASH only if the check holds — which is why the two should always be read together.

Run the numbers

Use the free Revenue Per Available Seat Hour Calculator to apply everything above to your own figures.