A contract caterer never chooses a napkin in a kitchen. It is chosen during mobilisation, in a document, by someone pricing a bid against a specification that lists the napkin as a line with a unit of a thousand. By the time the first case arrives on site, the decision has been locked for the length of the contract.
That makes catering unlike every other part of hospitality. A restaurateur holds a sample before buying. A contract caterer often commits to a product across twenty sites without anyone having pulled a sheet out of a stack, on the strength of a single number that says nothing about how many sheets one cover will consume.
The second difference is operational rather than commercial. In a canteen the operator decides how many napkins a person gets, because the operator is the one laying them out: on the tray, at the cutlery station, or in a dispenser at the end of the line. Consumption stops being something you observe and becomes something you set. What follows is how to work in cost per cover instead of price per thousand, why one ply holds up on merit where cutlery is used, and how to write the napkin line so the lowest bid cannot win with the wrong product. We write as the manufacturer who converts this tissue.
Table of contents
- Catering is the only setting where the operator sets the count
- Why a specification picks the wrong napkin
- Cost per cover, not price per thousand
- Cutlery changes what the napkin has to do
- Three ways to hand it out and what each does to the count
- Size for a canteen: 33×21 cm and when 33×33 cm
- The hospital tray and the patient with one usable hand
- One ply or two in contract catering
- White or Eco on the tray
- Documents the specification will ask for
- Term time, not calendar year
- How much to order and how to find your own figure
- Most common mistakes
- Checklist before you order
- Key takeaways
Catering is the only setting where the operator sets the count
In every other type of venue, napkin consumption is the result of guest behaviour. The guest takes as many as they feel they need and the venue influences that only indirectly, through quality and placement. Catering inverts it.
Meals are served in an organised flow, at one point, in a fixed order. The napkin is either already on the tray, or in a holder at the cutlery station, or in a dispenser at the end of the servery. In all three cases the number one person receives is the result of an operating decision rather than chance.
This has a consequence that exists nowhere else: consumption in a canteen can be planned in advance and hit. You do not need to estimate it by observing a few weeks of trade, because it follows from how the service point is set up. It can equally be broken by a single decision, such as laying two sheets on every tray where one was doing the job.
Planning therefore starts from a different question than in a restaurant. Not how many a guest will use, but how many we intend to issue and whether that is enough. The reference point for a canteen is two to three napkins per meal served, the lowest figure anywhere in hospitality.
> In short: in catering, consumption is a parameter the operator sets rather than a variable you measure. Before calculating volumes, decide how many sheets one cover receives and at which point it receives them.
Why a specification picks the wrong napkin
A specification compares bids in a column whose content is one number: price per thousand. That criterion is objective, easy to verify, and entirely disconnected from what it costs to serve one meal.
Napkins are a product where unit price and consumption move against each other. A thinner sheet is cheaper per piece and is taken a second time more often, because the first one did not do the job. If the price difference is a fifth and consumption rises by half, the cheaper product costs more per cover. The column cannot show that, because it measures pieces rather than meals served.
The second source of error is an item description with no grammage. The line "white paper napkins, 2 ply, 33×33 cm" looks precise, and admits products that differ enough in grammage to be two different articles in the hand. Grammage per ply in hospitality sits in the 16–19 g/m² band, and that, not the ply count on its own, determines how the sheet behaves.
The third is a silent pack format. The count per case and the way sheets are folded decide how much time it takes to restock a service point mid service. When a site puts out several hundred covers inside an hour, that is labour, not a packaging detail.
| What the specification states | What it produces in service |
|---|---|
| Price per 1000 only | The sheet people take twice wins the bid |
| Ply count with no grammage | Two articles under one name, different in the hand |
| Size with no service attached | 33×33 cm where 33×21 cm would do |
| No pack format | Slower restocking at the peak of service |
| Quality wording with no parameter | No grounds to reject a non-conforming delivery |
Cost per cover, not price per thousand
The unit of comparison in catering is cost per cover: unit price multiplied by actual consumption per meal served. That is all that is needed to compare two offers honestly.
A practical method that avoids pricing arguments altogether: take both samples and run one service on each, at the same point and the same time of day. Count how many sheets went against the same number of covers. Only then set that ratio against the price difference. If the cheaper offer consumes more per cover than its price advantage, it is the more expensive product, and you can show it with a number from your own servery.
In a tendered contract this test has a second value: it is repeatable and documented. A result from two days of service is a concrete argument when the specification is rewritten for next year, rather than the opinion of whoever runs the line.
Keep the distinction between fixed and variable cost in view while doing it. A napkin laid on a tray is consumed at the moment it is laid, whether or not anyone touches it. A napkin taken by hand is consumed only when someone reaches for it. Those are two different cost models in the same building, and averaging them hides the whole effect.
Cutlery changes what the napkin has to do
Meals in catering are eaten with a knife and fork, or a spoon. The hand is in contact with food for a fraction of the time it is in a pizzeria or over a sandwich. That is not a remark about table manners, it is a statement about the work the sheet performs.
A napkin alongside cutlery wipes the mouth, collects crumbs and occasionally touches fingers. It does not have to take sauce off a whole palm or hold grease for several minutes. One ply in a canteen is therefore justified on merit rather than on price alone, and that is the sharpest difference between this segment and the rest of hospitality.
The exceptions are predictable and worth listing, because they decide whether a second line is needed at all. Finger food, bread with toppings, fruit, meals issued in takeaway packaging and anything that drips all push the requirement towards higher grammage. On those days the simpler move is issuing two sheets rather than changing the product.
The arithmetic behind that is straightforward. One ply is 16–19 g/m², and two one-ply sheets give 32–38 g/m² in the hand, the same as a single two-ply. It lets you raise grammage on selected days without carrying a second article and without reopening the contract. We compare both options in the guide to 1 ply and 2 ply napkins.
Three ways to hand it out and what each does to the count
How the napkin reaches the person eating governs consumption more strongly than the product does. Canteens use three arrangements and each has its own cost behaviour.
Laid on the tray. Fixed cost, exactly one sheet per cover. Consumption is fully predictable and needs no restocking during service, because the tray is dressed with the cover. You also pay for everyone who does not use it. This suits sites where service speed matters and where every person has to be guaranteed one.
At the cutlery station. Partly variable. People take a napkin along with a fork, so the pickup is natural, but nobody is counting sheets. This is where handfuls happen, especially when the sheets are thin and hard to separate singly. A holder with a restricted opening works better than an open basket.
Dispenser at the end of the line. Variable cost. A 1/8D dispenser unit issues sheets one at a time thanks to the offset, the last panel sitting shifted past the flush edge of the stack so fingers can grip exactly one napkin. In the right situation that cuts the take by 25–40%, but the saving is a property of the situation rather than of the equipment. It disappears when the unit is overpacked, when it is running on the last third of its load, and when it stands on the table instead of at the pickup point.
These three often coexist in one building: tray service for pupils or patients, and a dispenser on the tables for external customers. If consumption is tracked as one average for the site, none of the three effects is visible.
Size for a canteen: 33×21 cm and when 33×33 cm
The baseline format for catering is 33×21 cm, the lunch sheet. It is designed for exactly this application: bistro and lunch service, canteens and food courts.
The reason is the one above. With a meal eaten using cutlery the sheet does not have to wrap a whole hand, so the full gastro sheet is more than the job requires. The lunch format sits on the tray beside the cover, does not overhang it and does not slide off while the tray is carried.
The 33×33 cm sheet has two defensible uses in a canteen. The first is table service to external customers, where the napkin does the same job it does in a restaurant. The second is a menu with food eaten by hand, where a smaller sheet forces people to take two at once and the format saving cancels itself out. The 1/8D dispenser sheet is also a full 33×33 cm, so choosing a dispenser does not mean dropping down a size.
Smaller formats, 24×24 cm and 15×15 cm, are misplaced in this category. The first belongs to bar service, the second to points where the napkin accompanies a drink rather than a meal. The full map of formats and their applications is in the guide to paper napkin sizes.
The hospital tray and the patient with one usable hand
Hospital catering is the one place in this segment where the napkin is not a consumable at all. It is a counted component of a sealed tray, assembled in a production kitchen, trolleyed to a ward and placed in front of someone who cannot go and fetch another one. There is no service point, no cutlery station and no dispenser on a wall. Whatever is on the tray is the entire supply for that meal.
That changes which property of the sheet matters. In a restaurant the question is absorbency. On a ward it is tensile behaviour, because a significant share of patients eat one handed: an arm in a cast, a cannula in the back of a hand, a stroke affecting one side, or simply the weakness that comes with being ill for a fortnight. A one handed person does not lift a napkin, they drag it across the tray surface and pin it with a thumb. A sheet that tears under that movement has failed, however well it would have performed a wipe. This is the one place in catering where it is worth testing a sample by pulling it flat across a wet tray rather than dropping water on it.
The second consequence is that the count is a line in the tray build specification rather than an operating habit. If the tray spec says one, that is what the patient gets, every meal, across every ward. Changing it from one to two is not a kitchen decision, it is a budget decision multiplied by bed days, and that is exactly how it should be presented when it comes up. The honest version of that calculation uses meals served, which hospital kitchens already count precisely.
The third is placement on the tray itself. A tray leaves the kitchen with a hot plated meal under a cover, and the sheet riding under or beside a hot item picks up steam in the trolley on the way to the ward. By the time the cover comes off, part of its capacity has been spent without anyone touching it. Put the napkin on the cool side of the tray, away from the plate and away from any hot beverage, and that loss stops happening.
For a contract caterer running wards alongside a staff restaurant in the same building, those are genuinely two different products working to two different criteria, and one site average will describe neither. Count them as separate lines from the first week of mobilisation.
One ply or two in contract catering
The default answer in this segment is one ply, and it is one of very few hospitality categories where that is true.
One ply is 16–19 g/m². With a meal eaten using cutlery such a sheet does its job, and the perceived difference against a two-ply is smaller here than it is with food eaten by hand. A two-ply napkin in a canteen is not a mistake, but it is a decision about standard rather than about function, and it should be justified in those terms when the budget is questioned.
There are situations that justify going up. Table service to external customers, meals issued in takeaway packaging, dishes with sauce eaten without a knife, and hospitality events held in the same room. In all four, issuing two one-ply sheets is a simpler move than changing the contracted article, because it delivers the same combined grammage.
One point is worth stating separately, because it causes confusion when requirements are drafted: a cellulose tissue napkin is not a grease barrier. Grease passes through it and leaves a translucent mark, whatever the ply count. If the specification needs an insert separating food from packaging, that is greaseproof paper, a separate product category, and it should be written as such.
White or Eco on the tray
Both versions are the same cellulose tissue and they do not differ in absorbency. The choice is about appearance and about the buyer's criteria, not about performance.
White is neutral and safe wherever the napkin sits on a tray next to a plate. With pale sauces and dusted products a mark shows less on white than on unbleached paper, which matters when trays come back through a public dining room.
Eco in unbleached paper is often scored or explicitly required where the buyer applies environmental criteria. Where a site already separates waste at the return point, unbleached paper and cardboard packaging go into the same stream. We cover the category in the guide to eco napkins.
What not to plan as an everyday standard here: coloured napkins and print on every cover. The range for catering is white or Eco, and print is a separate product with a longer lead time, 7–14 days against 3–5 days for plain. It belongs to scheduled hospitality events in the same room, not to daily service.
Documents the specification will ask for
A catering contract almost always carries a clause about documentation, and it is often worded loosely enough that the two parties read it differently.
One rule covers most of it: ask for documents on the exact product to be supplied, not a general supplier folder. Check that the product reference on the document matches the line in the offer, because that is the difference between a document and a document about something else.
What to ask for explicitly:
- Written confirmation of suitability for food contact for the offered article, stating which reference it covers.
- A technical specification with grammage per ply, ply count, size, composition and pack format.
- Documentation on fibre origin, where the buyer applies environmental criteria. It is a separate matter from food contact and does not replace it.
- A declaration that parameters stay constant for the life of the contract, meaning the supplied product will not change grammage mid term.
The last point matters most in practice, because the usual complaint in this segment is not about the first delivery but about the third. The product wins on price, the opening cases match the sample, and the difference appears later. A constancy clause gives you grounds to reject a load, which the price column on its own never does.
Term time, not calendar year
A canteen does not trade twelve months a year, and that is one of the most frequently skipped inputs when volumes are agreed.
A school canteen serves for roughly ten months, with two long closures and a hard restart in September when consumption returns to full volume overnight. A staff restaurant has shallower but regular dips through holiday periods and around public holidays. A hospital kitchen runs continuously, but its occupancy moves seasonally for entirely different reasons.
A twelve month average understates the working months and overstates the closures. The consequence is always the same: a stockout in the first week back and pallets sitting in a store through the summer.
The practical calculation is simple. Work out demand for one full month of service, meaning covers per day multiplied by the napkins per cover you have decided to issue, multiplied by service days. Only then multiply by the number of operating months and add a buffer. Set the reorder point before the restart rather than after it, allowing for the 3–5 day lead time on plain stock.
Storage conditions deserve a line too, because canteen stores usually sit next to the kitchen. A damp environment softens the paper and, in dispenser loads, flattens the offset, which is precisely the feature that makes sheets come out one at a time. A case standing next to the wash up or under a window loses its parameters before it is opened.
How much to order and how to find your own figure
The starting point for a canteen is two to three napkins per meal served, but your own figure is easier to establish here than anywhere else, because covers are counted exactly.
Take consumption from one full month of service and divide it by the covers served that month. The result is your real figure rather than a planning assumption. If it comes out above three on a cutlery menu, the cause is almost always the service point rather than appetite for napkins: an open basket instead of a restricted holder, sheets that are hard to separate singly, or two laid on a tray where one was enough.
Count each service mode you run separately. Trays, the cutlery station and table service to external customers carry different figures, and only separated do they show where consumption is generated. The method for stock cover, reorder points and safety buffers is set out in the guide to how many napkins to order.
On annual contracts there is one more thing worth checking before signing: how much space a year of stock occupies. A canteen store is sized for food, not for cases of tissue. More frequent, smaller deliveries often work better in this segment than one annual drop at a better price with nowhere to put it. Terms on larger volumes are covered in the guide to wholesale napkins.
Most common mistakes
- Scoring offers on price per thousand alone. It omits consumption per cover, the only number that decides cost in this segment.
- An item description with no grammage. Two products with the same name and size can differ enough to be different articles in the hand.
- One average for the whole site. Trays, cutlery station and external table service carry different figures and cancel each other out in the mean.
- 33×33 cm everywhere. On a cutlery menu the 33×21 cm lunch sheet does the same job.
- An open basket at the cutlery station. It invites handfuls, especially with thin sheets that are hard to separate.
- Planning stock across twelve months. A school canteen runs about ten and restarts abruptly in September.
- A store next to the wash up. Damp softens the paper and flattens the offset that dispenser loads depend on.
- No constancy clause. The complaint usually concerns the third delivery, and the price column gives you no grounds for it.
- Treating a hospital tray as a consumable. The count is fixed at assembly and the patient cannot top it up on the ward.
- Testing a sample only for absorbency. Where people eat one handed, how the sheet behaves when dragged across a tray matters more.
Checklist before you order
- Decide how many sheets one cover receives and record it as an operating decision.
- List every service mode separately: trays, cutlery station, dispenser, external table service.
- Work out your own figure: one full month of consumption divided by covers served.
- Write requirements as parameters checkable on a sample: size, ply count, grammage per ply in g/m², white or Eco, pack format.
- Run two samples through one service and count sheets rather than opinions.
- Convert both offers to cost per cover before comparing prices per thousand.
- Ask for documents on the exact offered article and check the reference against the offer.
- Include a constancy clause for the life of the contract.
- Plan stock on operating months and set the reorder point before the restart.
- Check where a year of stock will physically stand and whether the store is dry.
- If you run wards, specify the tray napkin as a counted component of the tray build and test it by dragging, not only by dropping water on it.
Key takeaways
In catering the napkin is chosen by a document rather than by the person who hands it out. Most of the errors in this segment start there, and so should the fix.
The price per thousand column does not describe cost. Cost per cover does, meaning unit price multiplied by actual consumption per meal served, and that consumption can be measured on your own servery in two days.
This is the only hospitality segment where the count per person is an operating decision. Trays, cutlery stations and dispensers give three different cost models, and a single site average shows you none of them.
A meal eaten with cutlery gives the napkin less work, so one ply holds up here on merit rather than on price. Where more is needed, two one-ply sheets deliver the same combined grammage as one two-ply without a second contracted article.
The baseline format is the 33×21 cm lunch sheet. Keep the full 33×33 cm for external table service and for food eaten by hand.
On wards, treat the napkin as a counted part of the tray build and judge it on how it behaves for someone eating one handed, which is not what an absorbency test tells you.
Gastronomic napkins · Eco napkins · Dispenser napkins · B2B for hospitality
