How to compare the hotel restaurant against itself month over month without being fooled by occupancy

In a hotel restaurant, sales rise and fall with occupancy long before they rise and fall with the kitchen. If you compare one month’s total against the previous one without correcting for how many rooms the front desk sold, you will congratulate one team for a high season and scold another for a low one. The honest comparison is against yourself, but with the same rules every time.

Why gross sales lie in a hotel

A street restaurant has demand that depends on the neighborhood, the weather and its own reputation. Your hotel’s restaurant has, on top of that, a captive demand that arrives through the front desk and that it does not control: occupancy. When occupancy climbs twenty points, sales climb even if the team did nothing different. When it drops, sales fall even if the team did everything right.

That is why the controller does not compare total sales. The controller compares normalized metrics: how much the restaurant sold per occupied night, how many guests it captured, how much each check spent. Those numbers describe the restaurant, not the room market.

There is a second, quieter trap: the calendar. July can have five Saturdays and August four. If your Saturday dinner is the strongest of the week, August starts out a whole Saturday behind and nobody is to blame.

Step one: normalize by occupied nights

Take net food and beverage revenue for the month, without tax or tips, with comps deducted and with room charges included. Divide it by the occupied room nights the front desk reports for the same period and the same cut-off. That is your revenue per occupied room, and it is the first figure compared month over month. The full definition is in the article on how it is calculated (How food and beverage revenue per occupied room is calculated, and what a good number looks like).

If total sales fell and revenue per occupied room rose, the restaurant did a better job with fewer guests. If total sales rose and the metric fell, occupancy saved the month and the restaurant let people walk past. That reading alone changes the conversation in the meeting.

Step two: separate the guest from the street

Walk-in sales do not depend on occupancy. If you mix them with guest sales and then normalize by occupied nights, you punish the restaurant in high season (more nights in the denominator, the same street in the numerator) and reward it in low season. Separate the two sources.

To separate them you need to know which check belonged to a guest. The cleanest way is for the room charge to be tied to the folio (Room charge) and for everything else to be flagged as walk-in when the check is closed. That gives you two series: guest revenue, normalized by occupied nights, and walk-in revenue, normalized by days open.

• Guest revenue ÷ occupied nights: how your capture and your check are doing with the people already sleeping in the building.
• Walk-in revenue ÷ days open: how your restaurant is doing as the neighborhood restaurant.
• Capture: folio nights with consumption ÷ occupied nights. How many come in, not how much they spend.
• Spend per consuming night: guest revenue ÷ folio nights with consumption. How much each room that did come in left behind.

Step three: compare Tuesday with Tuesday

The Tuesday guest is the business traveler who eats quickly or goes out with clients. The Saturday guest is the family or the couple who do not want to leave the hotel. Their average check, their timing and their capture are different. A month with more weekends has a better mix without the restaurant having changed anything.

The practical way: calculate revenue per occupied room by day type (weekday and weekend) and compare each type with its equivalent in the previous month. If you prefer a single figure, weight it: apply this month’s metrics to last month’s mix of days, and the two months now share a calendar.

A calendar example

Invented figures to show the effect. July had 5 Saturdays and August 4. If your Saturday leaves 220 per occupied night and the other days leave 120, July starts with an advantage of 100 per occupied night across one full Saturday. With 60 rooms at 80 % occupancy, one Saturday is 48 nights: 48 × 100 = 4,800 of difference that has nothing to do with the team.

Step four: season is corrected with the same month last year

Neither night normalization nor the calendar corrects for season. In August the guest is on vacation and in October the guest is at a conference. They behave differently at the table. The clean comparison against season is the same month of the previous year, with the same normalized metrics.

If you do not have twelve months of clean data, start today. The first year you compare against the previous month with the corrections from steps one to three; from the second year on, against the same month of the previous year. The discipline of not changing the definition along the way is what makes the series useful.

Step five: price versus volume

If revenue per occupied room went up, you still need to know why. Break it into three factors: capture (how many came in), checks per consuming night (how many times they consumed) and average check (how much they spent each time). The product of the three is the metric. A menu price increase raises the average check and can lower capture. A new bar schedule raises capture without touching the check.

Without this breakdown, a good month can hide a bad decision. Raising the menu 10 % and losing 10 % of capture leaves the metric almost unchanged and the guest more annoyed.

A complete illustrative example

Invented figures to show the calculation. They are not from any hotel. To keep it simple, both months have 30 days.

Total sales fell 12 %. With that figure alone, the meeting would have been uncomfortable. Normalized, August was the better month in almost everything: more guest revenue per occupied night, more capture, more walk-in per day open. The only thing that fell was spend per consuming night, from 400 to 380, and that is the one legitimate question for the chef and the manager: what changed in the check?

Notice that walk-in revenue rose in the month with lower occupancy. That is common: with fewer guests there are free tables for the neighborhood. If you do not separate it, that revenue disguises the occupancy drop and makes you believe the guest consumed more than they did.

The one-page sheet

All of the above fits on one page: four normalized metrics, each with this month, last month on the same calendar, the same month last year and the difference. If your point of sale knows which check was a guest check and the front desk gives you occupied nights with the same cut-off, the sheet fills itself. If not, someone builds it by hand on the eighth and nobody reads it anymore.

One rule for the meeting: nobody explains a normalized figure with occupancy. Occupancy is already inside it. Valid explanations are about the menu, price, hours, service, room charge friction or corporate agreements. Everything else is talking about the room market, and there is another meeting for that.

What to do this week

1. Ask the front desk for occupied nights for the last three months with the cut-off the restaurant uses. If the cut-offs do not match, agree on one and put it in writing.
2. Split the revenue of those three months into guest and walk-in. If you cannot, write down how many checks cannot be classified: that is your first problem to solve.
3. Calculate the four metrics on the sheet for the three months and place them side by side.
4. Count the Saturdays in each month and write down the difference before reading any number.
5. Bring the sheet to the next meeting and agree that this, not total sales, is the figure under discussion.

Inn Restaurant separates guest revenue from walk-in revenue because every room charge is born tied to the folio, and it builds the normalized metrics the controller uses (Controller) in the reports section (Reports). If you want to see the sheet with your hotel’s data, you can request the 15-minute demo on the contact page (contact).