A hotel with full food and beverage is running several different businesses that happen to share a kitchen, a store, and a purchasing account. There is the restaurant, and alongside it a bar with completely different margins. Banqueting works to its own rhythm, arriving in lumps around the events calendar instead of steadily. And then there is breakfast, which is the strangest of them because most of the time nobody is paying for it directly.
Put them together and you get a single F&B cost percentage that is arithmetically correct and tells you almost nothing. It moves, and nobody can say which of the four moved it.
Why the Blended Number Hides More Than It Shows
The problem is that the four outlets have genuinely different economics, so averaging them buries the signal.
A hotel restaurant might run in the low thirties. The bar, if it is doing its job, runs far lower, because spirits and wine carry margins food cannot approach. Banqueting sits somewhere else again, usually better than the restaurant because volume and a fixed menu remove most of the waste. Breakfast is its own creature and often has no revenue line at all, because the cost of it is buried inside the room rate.
Blend those and you get a number in the mid-twenties that describes no actual outlet. Worse, it moves for reasons that have nothing to do with performance. A month with three big weddings will pull the blended figure down and look like an improvement, when nothing about how the restaurant is run has changed at all. A quiet banqueting month looks like a decline. You end up reacting to your event calendar and calling it cost control.
The same trap catches restaurant groups reading a group average, and it is sharper in a hotel because the mixed outlets sit inside one legal entity and one set of invoices.
Splitting the Number by Outlet
Getting to a per-outlet figure means solving two things: revenue attribution, which is usually easy, and cost attribution, which is not.
Revenue is generally fine, because the POS already separates it. Restaurant covers, bar tabs, and banqueting invoices land in different places, and most hotel systems can report them apart without much argument.
Cost is where it falls down, because a hotel buys once and consumes everywhere. One delivery of butter serves breakfast, the restaurant, and a wedding. Splitting that invoice by outlet at the point of purchase is impossible and not worth attempting.
The workable approach runs from the other end. Attribute cost by consumption, not by purchase. That means each outlet holds its own stock position and records what moves into it, so the store becomes an internal supplier issuing to four consumers instead of one undifferentiated pile.
In practice, that requires a stock location per outlet and transfers recorded when goods move from the main store to the bar, the restaurant line, or the banqueting prep. It is the same discipline a group uses for inter-site transfers, running inside a single property. Once that is running, the theoretical usage for each outlet can be calculated from its own sales and its own recipes, and each outlet gets a variance figure that belongs to it.
Costing a Buffet You Cannot Portion
Breakfast is the part that defeats most hotel F&B reporting, for two reasons that compound.
The first is that consumption is not tied to a transaction. Guests take what they take, and no POS line records it. You cannot calculate theoretical usage from sales because there are no sales in the ordinary sense, so the entire theoretical-versus-actual method has nothing to stand on.
The second is that revenue is often absent too. Breakfast is included in the rate, so it lands in your costs and never in your F&B revenue. Cost with no matching revenue pushes the blended percentage up, and it does so regardless of how well breakfast is actually run.
The fix for both is to change the unit. Stop trying to cost breakfast as a percentage and cost it per cover instead. Take the total cost of breakfast consumption over a period, divide by the number of guests who ate, and you get a euro figure per head. That number can be compared week to week, against your budget, and against other properties.
Getting the consumption figure means counting the breakfast stock separately from everything else, which is the same location discipline as above. Getting the cover count means working from occupancy and a take-up rate, since there is no POS to ask. Neither is difficult; both are usually skipped.
Once you have a cost per cover you can do something with it. A buffet running at four euros twenty a head when it should be three eighty puts production quantity and end-of-service waste on the table, and none of that can be discussed while breakfast is a percentage buried inside a blended number.
What Events Do to a Monthly Average
Banqueting is the other distortion, and it works in the opposite direction to breakfast.
Events run to high volume on a fixed menu with the covers known in advance, which makes them the most efficiently costed food the hotel produces. Waste is low because the quantity was decided from a confirmed number, and there is none of the à la carte unpredictability that makes a normal service hard to buy for. A well-run banqueting operation will beat the restaurant on cost percentage most months.
That is fine until it lands in the blended figure. A month with a large conference will pull the whole hotel’s F&B percentage down by a couple of points. The following month, without it, the percentage rises and someone asks what went wrong in the restaurant. Nothing went wrong in the restaurant. The mix changed.
Which is why event cost belongs in its own line, ideally per event instead of per month. Per event is genuinely achievable, because the covers, the menu, and the timeframe are all fixed and known. A wedding is a small closed operation you can measure end to end, and costing it individually tells you whether your event pricing is right in a way a monthly average never will.
Hotels that separate this often find their event pricing has drifted, because menus get quoted from a cost sheet that was accurate when it was written and has been reused for two years while ingredient prices moved.
Getting to a Number Per Outlet You Can Act On
The end state is four numbers instead of one, each in the unit that suits it.
The restaurant reports a food cost percentage against its own covers and its own recipes, which makes it comparable to any standalone restaurant and to its own history. The bar keeps a separate percentage, split from food, because a blended F&B figure hides both. Banqueting is costed per event and rolls up monthly, while breakfast is measured entirely differently, as a cost per cover.
None of that needs a different accounting system. What it needs is a stock location per outlet, transfers recorded as goods move between them, and a recipe behind everything each outlet sells. Then the same theoretical against actual comparison that works in a single restaurant works four times over.
The payoff is that variance becomes attributable. When the number moves you know which outlet moved it, and usually why, which is the whole point of measuring anything.
Frequently Asked Questions
What is a good F&B cost percentage for a hotel? There is no single useful figure, because a hotel blends outlets whose economics have little in common. Measure the restaurant against normal restaurant benchmarks, the bar separately and much lower, banqueting per event, and breakfast as a cost per cover.
How do you cost a hotel breakfast buffet? Per cover, not as a percentage. Count the breakfast stock separately, take the total consumption over a period, and divide by the number of guests who ate. Occupancy and a take-up rate give you the cover count when there is no POS transaction.
Why does our F&B cost jump around every month? Usually outlet mix, and not performance at all. A heavy banqueting month pulls a blended percentage down and a quiet one pushes it up, independently of anything happening in the restaurant. Splitting by outlet removes the effect.
Do we need separate stock locations for each outlet? If you want per-outlet cost, yes. Attributing purchases by outlet at the point of invoice is not realistic, because one delivery serves several outlets. Attributing by consumption is, and that needs each outlet to hold and count its own stock.
Splitting One Number Into Four
A hotel’s F&B cost only becomes useful when it stops being one figure. Give the restaurant, the bar, banqueting and breakfast a stock location each, the recipes that belong to them, and a unit of measurement that suits what they actually do. The number then starts pointing at something you can change.
Stockifi handles multi-location setups where outlets hold their own stock and transfers are recorded between them, so each part of the operation carries its own consumption and its own variance. See how inventory transfers work.