RevPASH: How to Calculate Revenue per Available Seat Hour, Hour by Hour

The measure comes from restaurant revenue-management research at Cornell. Kimes and colleagues argued in 1998 that "instead of counting table turns or revenue for a given day part, restaurant operators should measure revenue per available seat hour (RevPASH). This measure captures the time factor involved in restaurant seating" (Kimes et al., 1998). Turns count guests; revenue counts money; RevPASH counts money against the one thing a dining room cannot make more of in the moment — seats multiplied by time.

The formula, and what goes into it

Use revenue from seated guests only. Thompson and Sohn, who tested how RevPASH should be measured, say it "should not include tips, service charges, or any taxes", and takeout belongs outside it too, because takeout does not use a seat (Thompson & Sohn, 2009). Available seats are the seats you open and staff for that period; a closed section is not available.

Worked example: one Friday, hour by hour

The evening average of $10 describes none of the hours. The early hour has guests who spend well — $25 per occupied seat-hour — but only a fifth of the seats are in use, so it is an occupancy problem. The middle hours run at $20 per occupied seat-hour with most seats full; there, a walk-out or a table left empty for twenty minutes costs real money. The last hour has both low occupancy and lower spend. Each hour needs a different decision, which is the point of calculating it by the hour.

Spread each check across the time the table was occupied

The common shortcut credits each check to the hour it was opened. Thompson and Sohn found that, in their data from one restaurant, this was off by more than 40% for any analysis period shorter than 150 minutes, while spreading each check's revenue from open to close kept the error under 6%. Hourly RevPASH needs the second method.

The studies above used the check open and close times recorded by the restaurant's POS; if yours exports those, a check-level export is enough to do this in a spreadsheet. Marji does not import POS sales or check times, so this calculation sits with your POS reports; the plate costs Marji keeps come in when you turn revenue into contribution, below.

Where seat turnover fits

Seat turnover is covers divided by seats for a service period. It counts guests, not money, and it cannot see time: two turns of short, cheap meals and two turns of long, expensive ones read the same. Kimes described RevPASH as "closely related to the number of turns and the length of the meal, or service cycle" — turnover is one ingredient, and RevPASH is the result that includes spend and duration too.

What the research found when restaurants tried to raise it

The best-documented case is a 230-seat Chevys in Glendale, Arizona (Kimes, 2004). Hourly RevPASH there ranged from $0.27 to $7.03 on an average check of $12.70, and peak occupancy was 61%. The restaurant replaced 56 tables — mostly four-tops — with 76 tables of mixed sizes, keeping 230 seats, and changed service practices, at a remodel cost of $49,000 plus about $15,000 a year in labor. About three months later, occupancy in its busiest hours had risen from 50% to 59%, and RevPASH in those hours from $5.85 to $6.32. Because sales can move for reasons outside a restaurant, the author compared it with two other Chevys and attributed a 5.1 percentage-point difference in year-over-year sales growth to the interventions — while noting that without that comparison "it was unclear whether that increase resulted from the managers' interventions or from other market conditions." One restaurant, with changes designed and measured by the researcher, is a case, not a benchmark.

Shortening meals is the tempting lever, and it is weaker than simple arithmetic suggests. Kimes calculated in 1999 that a one-minute reduction in meal time during high demand could raise RevPASH by 1.5 to 2 percent. When Thompson simulated more than 1,200 scenarios, the average revenue gain from shorter dining was "less than one-quarter of the amount predicted", and even the best case was "less than one-half" (Thompson, 2009). A shorter meal only earns more when there are guests waiting for the seat; in a half-empty hour, it just empties the room sooner.

If you use price to move demand between hours, guests judge the methods differently. In a survey of 157 hotel guests rating hypothetical restaurant practices, two-for-one coupons, time-of-day pricing and lunch-versus-dinner pricing were seen as fair, weekday-versus-weekend pricing as neutral to slightly unfair, and charging more for a better table as somewhat unfair (Kimes & Wirtz, 2002). Framing mattered: the same time-of-day price gap scored 2.29 as a discount and 3.11 as a surcharge, on a scale where 1 is extremely fair and 7 extremely unfair.

Add costs: contribution per available seat hour

RevPASH measures revenue, and a busy hour of low-margin dishes can look better than it is. Kimes's formula already allows profit in place of revenue. The practical version for most kitchens is contribution: revenue minus the food and beverage cost of what was sold, per available seat hour.

This is where reviewed plate costs matter: two hours with equal RevPASH can earn very different amounts if one sells mostly steaks and the other mostly pasta. The contribution margin guide covers the dish-level half of the calculation; labor, rent and other costs still sit outside it, so it is not profit.