By Matt Wampler, CEO of ClearCOGS
Walk into the revenue office of almost any hotel and ask how many guests will be sleeping in the building eleven days from now. You will get a number. Not a guess, a number, with a confidence range around it, adjusted for the season, the day of the week, the conference in town, and what the same week did last year. Somebody is paid to own that number, and their performance is measured against it.
Now walk downstairs and ask the kitchen how many portions of the signature dish they will sell that same night.
In most hotels, you will get a shrug and a story about how Thursdays are usually decent.
The same building. The same guests. The same eleven days of warning. One side of the property runs on a forecasting discipline refined over forty years. The other runs on the memory of whoever is on shift.
How the hallway got so wide
This is not because hotel operators are careless. It is an accident of where the discipline grew up.
Revenue management was built for rooms, and rooms were the perfect place to build it. Fixed capacity, perishable inventory, advance reservations, and an obvious cost to being wrong in either direction. Airlines got there first, hotels followed, and by now the practice is deeply institutionalized. There are dedicated teams, dedicated software, and a shared professional language across the industry.
Restaurants sit in the same category on paper. Fixed capacity, perishable inventory, demand that varies by the hour. But the discipline never made the jump with the same force. A recent systematic review of the research literature found that restaurant revenue management, despite the proven track record of these methods in airlines and hotels, has seen comparatively limited adoption in the restaurant sector. The review covered 108 studies published between 1997 and 2023 (Bujalance-Lopez, Gonzalez-Serrano, Lechuga Sancho, and Talon-Ballestero, British Food Journal, 2025).
The gap is not a technology gap. The math that predicts occupancy is not fundamentally different from the math that predicts how much protein to thaw. It is an organizational gap, and inside hotels it has a specific cause.
F&B was treated as an amenity, not a business
For a long stretch, the restaurant inside a hotel existed to make the hotel more attractive. It was an anchor, a convenience, sometimes an outright loss leader. Nobody expected it to carry a margin, so nobody built the machinery to manage one.
That assumption has quietly stopped being true. Food and beverage now represents a meaningful share of total property revenue, and at properties that run F&B as a real business rather than a guest service, that share climbs substantially higher. The most successful hotel restaurants tend to be the ones run the way a restaurateur would run them, with someone accountable for cost of goods, prep discipline, and waste, rather than the way a hotel department gets run.
Which creates an awkward position for the executive who owns F&B across a portfolio. The revenue is now material enough that variance shows up in the group’s numbers. The operating discipline is often still set up as though it does not.
The problem multiplies with geography
Ask a group F&B leader what keeps them up and you will usually hear a version of the same answer: every property does it differently.
One property follows the standard. Another interpreted the standard loosely. A third built its own process years ago and nobody has revisited it. The reporting formats differ. The recipes have drifted. The variance between theoretical and actual food cost sits in a range that would be alarming in a restaurant group and is often simply unknown here, because the reporting was never built to surface it consistently.
The instinct is to go see for yourself. Get on a plane, walk the property, talk to the cost controller, find out what is actually happening. That works, and it works well, and it does not scale. A leader with fifty properties across several countries is looking at six months of travel to complete one pass, by which point the first properties visited have drifted again. If the portfolio is still growing, the loop never closes.
You cannot audit your way to consistency at that size. The audit finds the problem after the money is gone.
The signal the kitchen already has and mostly ignores
Here is the part that should be frustrating, and also hopeful.
A standalone restaurant forecasting tomorrow’s demand is working from historical sales, seasonality, weather, local events, and whatever the manager knows. It would love to have a reliable advance count of how many people will be within walking distance tomorrow night. It cannot get one.
A hotel restaurant can. The occupancy forecast is sitting in a system upstairs, produced by professionals, updated continuously. It is one of the strongest forward-looking demand signals a food operation could ask for, and in a great many properties it never reaches the kitchen in a usable form. The chef finds out the hotel is full when the dining room fills up.
Occupancy is not a perfect predictor of covers. Guests eat elsewhere, groups behave differently from transient travelers, and a conference block with catered meals changes the math completely. But those are modeling problems, and they are tractable ones. They are far better problems to have than no signal at all.
The opportunity in hotel F&B is not to invent forecasting. It is to connect a forecasting capability the property already owns to the decisions that currently run on instinct: what to prep this morning, what to order this week, how much to have ready before the dinner rush, when to stop producing and start cleaning.
What this looks like in practice
For a group leader trying to move a portfolio rather than one property, a few things matter more than they might seem.
Standardize the decision, not the person. Trying to make fifty F&B managers think alike is a training program that never ends. Giving all fifty the same number, produced the same way, before their shift starts, is a system. The second one survives turnover.
Start where the spread is widest. You do not need every property live to learn something. Pick a handful that differ from each other in the ways your portfolio differs, prove the approach, and use the results as the internal case rather than the vendor’s case.
Treat the forecast as the consistency mechanism. This is the part that tends to click for multi-property leaders. A shared forecast is not just a prep tool. It is the only practical way to compare properties on equal footing, because it establishes what each one should have done before you look at what it did. Variance stops being a mystery and becomes a short list of places to look.
Let the number travel instead of you. The goal is not to eliminate site visits. It is to make them targeted, so the trip happens because the data flagged something rather than because it has been a while.
The wider point
None of this is really about hotels. It applies anywhere food and beverage operates inside a larger business that measures itself on something else: stadiums, casinos, hospitals, campuses, theaters, grocery. In each case there is usually a sophisticated forecasting function somewhere in the organization and a kitchen that never received its output.
The discipline already exists in the building. It was simply built for a different department, in a different decade, for a different kind of inventory. Moving it down the hall is a smaller project than most operators assume, and the return shows up on the line where the margin has been leaking the whole time.
This is the work we spend our days on at ClearCOGS: taking demand signals a business already produces and turning them into what to prep, order, and staff before the shift begins. If you run food and beverage across multiple properties, we would be glad to talk through what that looks like.
Sources
- Bujalance-Lopez, Lourdes, Gonzalez-Serrano, Lydia, Lechuga Sancho, Maria Paula, and Talon-Ballestero, Pilar. Restaurant Revenue Management: A Systematic Literature Review and Future Challenges. British Food Journal, 127(6), 2169–2196. April 2025. sciencedirect.com
