A delivery platform takes a share of the price, so its cut grows with every PLN you add. To hold a chosen contribution per item, a restaurant divides the sum of that item's variable costs and its contribution target by one minus the effective take rate, instead of adding the commission on top of the dining-room price.
Why a commission divides your price instead of subtracting from it
A supplier discount is a fixed amount. A platform commission is not: it is a percentage of whatever the guest pays, so the moment you raise the price, the platform's cut rises with it.
Take one dish. Its ingredients cost 14.20 PLN, the container and bag cost 2.30 PLN, and the extra kitchen minutes the delivery channel creates are worth 1.10 PLN. In the dining room the same dish sells for 36.00 PLN and leaves 21.80 PLN behind after its ingredients.
You wanted 21.80 PLN. You are short 6.64 PLN on every plate, and no amount of volume repairs it.
The reason is mechanical.
Prices that must survive a percentage deduction are built by division.
| Way to set the delivery price | What happens to contribution | What happens to the mix | What the guest sees |
|---|---|---|---|
| Dining-room price plus the commission percentage | Falls below target on every item, and the gap widens as the item gets cheaper | Barely moves — the error is proportional, so it hides | A modest, believable uplift |
| One flat percentage across the whole delivery card | Overshoots on some items, undershoots on others | Shifts towards the items that were underpriced relative to their cost | An even, tidy card |
| Each item recomputed by the formula below | Lands on target by construction | Shifts towards the items that genuinely carry the channel | An uneven card that needs a reason |
The third row is the honest one, and it is also the one that needs the most explaining to your own staff. The rest of this page is that explanation, and every number in it is one you can check on your own settlement statement using the finance layer.
The price formula that holds a chosen contribution
Platform price = (Plate cost + Packaging cost + Incremental labour + Target contribution) ÷ (1 − Effective take rate)
Plate cost— the recipe cost of the portion as it leaves the kitchen, PLN. Its calculation belongs to the food cost pillar and this page takes it as given;Packaging cost— the container, lid, bag, cutlery and label that this order needs in order to travel, PLN;Incremental labour— the extra kitchen and pass minutes the delivery channel creates for this item, valued at your loaded hourly cost, PLN;Target contribution— the money you decide this item must leave after all of its variable costs, PLN;Effective take rate— what the platform actually keeps, as a decimal fraction between 0 and 1, read off the settlement statement rather than the rate card.
Target contribution per item — the contribution the operator decides one item must leave after all the variable costs of that channel. It is a decision, not a measurement: two restaurants with identical costs may legitimately set it differently, and the formula holds whichever number is chosen.
Effective take rate — the gap between gross order value and net remittance as a share of gross order value, measured from settlement statements rather than from a price list. The measurement itself belongs to the page on delivery margin after commission; this page only consumes the number it produces.
The arithmetic, forwards and backwards
Worked through on the same dish. Variable costs are 14.20 + 2.30 + 1.10 = 17.60 PLN. You decide the item must leave the same 21.80 PLN it leaves in the room. The numerator is 17.60 + 21.80 = 39.40 PLN. At an effective take rate of 0.30 the platform price is 39.40 ÷ 0.70 = 56.29 PLN.
Check it backwards, because a price formula that cannot be checked backwards is a guess: 56.29 × 0.70 = 39.40 PLN remitted, minus 17.60 PLN of variable cost, leaves exactly 21.80 PLN. The dish now carries the same money into your fixed costs from the platform as it does from a table.
Two conditions the formula does not state
Two conditions sit under that arithmetic and both matter. First, every number above is net of VAT, and the price the guest reads on the platform is a gross price.
art. 41 ust. 12f together with the transitional art. 146ef of the VAT act, and not through the annex that most industry articles cite: the act applies the reduced rate to services classified in the grouping "usługi związane z wyżywieniem (PKWiU 56)" (Ustawa o VAT, consolidated text Dz.U. 2025 poz. 775, art. 41 ust. 12f, opened 26.08.2026), while art. 146ef ust. 1 sets that reduced rate at 8 % and the basic rate at 23 % (same consolidated text, art. 146ef ust. 1, opened 26.08.2026).Convert once, at the end, at the rate that belongs to that item.
Second, this page names no commission rate as a norm, and the 0.30 above is an input taken from a worked example rather than a benchmark. Neither Eurostat nor Poland's statistical office publishes platform take rates: a full pass over the Eurostat catalogue on 26.08.2026 returned two datasets whose titles mention commissions at all, both of them 1997–2001 structural business series for credit and insurance activity, and nothing whatsoever for delivery platforms. Your rate is a fact of your own settlement, and the analytics layer is where it is read.
What goes into the target contribution for a delivery item
Costs the dining room never sees
Two of the four inputs above do not exist at a table. Packaging is the obvious one, and it is rarely a single line: a leak-proof container, a lid that survives a scooter, a bag, cutlery, a sauce pot and a printed label add up to a number most kitchens have never priced per dish. The second is the extra labour the channel creates — the pass, the bagging, the checking, the handover. Both belong to the measurement page for the delivery channel; here they are simply two more cost lines in the numerator, and the formula does not care where they came from as long as they are per item and per channel.
The contribution you decide to keep
There are three defensible ways to pick the target, and picking is unavoidable.
The first is an equal amount per item: the dish must leave the same money on the platform as it does at a table. That is the target used throughout this page, and it is the one that keeps your break-even arithmetic intact, because break-even counts contribution in money and not percentages of anything. Contribution margin — the money left from net sales after variable costs, available to cover fixed costs and, once they are covered, to become profit — is defined on the break-even pillar and quoted here from it.
The second is an equal share of the price, and it does not survive contact with the formula.
Three ratios, one dish, one evening. If the ratio on the guest price has to match as well, the target itself has to rise, and the price rises with it — which is a decision about demand, not about arithmetic.
The third is the number your break-even needs: total fixed costs for the period, divided across the covers and orders you actually expect. That is a planning route rather than a menu route, and the forecast layer is where it lives.
How much higher than the dining-room price — and the condition that formula hides
Required uplift over the dining-room price % = (Platform price ÷ Dining-room price − 1) × 100
Platform price— the price from the formula above for one item, PLN;Dining-room price— the price of the very same item at a table, PLN.
Now the trap.
The two answers differ by 13.49 percentage points on one dish, and the short one is not a rounding of the long one. It is a different statement, true only when the dining-room price was itself built from the same costs and the same contribution target as the delivery price. Ours was not: the delivery version carries 2.30 PLN of packaging and 1.10 PLN of extra labour that the table version never had, and those 3.40 PLN have to be recovered before the commission is even considered.
So the short formula answers "what uplift restores a price after a percentage is deducted from it", and the long one answers "what price leaves me the contribution I decided on". They coincide only in the special case where nothing about the item changes between channels. That case is rare enough that the safest habit is to never use the short form on a menu at all — the what-if layer runs the long one against a changed take rate in the time it takes to type the new number.
Maximum take rate: the commission at which an item stops contributing at all
Maximum take rate that still contributes = 1 − (Plate cost + Packaging cost + Incremental labour) ÷ Platform price
- the numerator of the fraction is the total variable cost of the item in the delivery channel, PLN;
Platform price— the price actually shown to the guest on the platform, PLN;- the result is a decimal fraction; multiply by 100 to read it as a percentage.
Below that rate the item still leaves something; at exactly that rate it leaves nothing; above it, every order destroys money.
This number is the one worth putting on a management screen, because it is the only one that compares an item against its channel rather than against other items. The reporting article is about exactly that habit: few numbers, each of which changes a decision.
The degenerate case is not a footnote. If the maximum take rate comes out at zero or below, the item does not cover its own variable costs at any commission, including a commission of zero.
Nothing about the platform caused that, and no uplift repairs it — the decision is about whether that item belongs on a delivery card at all, in that packaging, at that portion size.
One flat percentage across the delivery menu breaks the mix
The tidy solution is to raise everything by one number. Here is what it does to three ordinary items, each with its own target equal to what it leaves in the room.
| Item | Variable cost, delivery | Target contribution | Required uplift | Maximum take rate |
|---|---|---|---|---|
| Pasta, 36.00 PLN at a table | 17.60 PLN | 21.80 PLN | 56.35 % | 68.7 % |
| Cream soup, 18.00 PLN at a table | 8.30 PLN | 13.40 PLN | 72.22 % | 73.2 % |
| Steak, 62.00 PLN at a table | 31.80 PLN | 34.00 PLN | 51.61 % | 66.2 % |
Three items, three uplifts, spread over more than twenty percentage points.
The pasta lands close to its target and hides the problem.
The damage is not the 3.64 and the 1.54 on their own. It is that the guest now faces a card where the item that needed the smallest uplift got the largest one in absolute terms, and the item that needed the largest got too little. Demand moves towards the relatively cheaper item, which is precisely the item that is now missing its target on every order. The mix has been pushed in the wrong direction by your own pricing.
Menu mix (MM) share — units sold of one dish divided by total units sold in the same menu category. The term belongs to the menu engineering pillar and is quoted from it unchanged; what happens to the matrix after a repricing is that page's subject rather than this one's. This page states only the mechanism: a flat uplift is a change to relative prices, and relative prices are what the mix responds to.
Items you cannot raise, and what to do with them instead
Some lines resist the formula, and it is better to name them than to let an average hide them.
Drinks. A bottled soft drink is the one item on your card whose fair price the guest already knows, and a visible uplift on it reads as a trick rather than as logistics.
A single uplift applied to a gross card therefore lands on two different net bases, and the target contribution you thought you set is not the one you got.
Cheap sides and add-ons. An item at a low absolute price cannot absorb its own packaging: a 2.30 PLN container on a 9.00 PLN portion is a quarter of the price before anything else happens. Raising it far enough to work makes it absurd; leaving it makes it a subsidy paid by the rest of the card.
Anything the guest compares across channels within one screen. If an identical item appears on your own site and on the platform, the difference becomes a statement about you rather than about the platform.
What works instead is changing the item rather than the number: a delivery-only portion size with its own recipe cost, a set that carries its packaging across three components instead of one, a minimum order value that makes the small basket viable, or moving that line to your own channel entirely. All four are menu decisions with a cost consequence, and they are worth modelling before they are announced — the calculator layer exists for that step.
Price parity and contract terms: what to read before you raise anything
Price parity clause — a contract term about how prices on a platform may differ from the restaurant's own prices. Its content is a fact of your contract rather than of this page, and this page makes no claim about what yours says or about what it is allowed to say.
Read your own agreement before touching a single price, and read three things in it: whether platform prices are constrained relative to your own, what base the commission is calculated on, and whether promotional discounts are deducted before or after that base. The second and third change the effective take rate you feed into the formula, which changes every price it produces. The base — gross or net — has its own page in this section, on which denominator a percentage of sales is taken from, because it is the single most common place where two correct numbers describe two different things.
If parity does bind you, the arithmetic does not disappear; it changes shape. Either both channels move together, and the uplift becomes a general price decision with the room's own demand attached to it, or the platform channel accepts a lower contribution and you decide item by item which lines can afford that — which is exactly what the maximum take rate above is for. And whichever price you land on, it has to be the price in the document as well as on the screen: the invoicing article covers what has to reconcile downstream.
What the guest sees: the gap between your own site and the platform
Run the same formula with a take rate of zero, which is what your own ordering channel is, and our dish comes out at 39.40 PLN — the numerator itself.
Beside it, the platform price is 56.29 PLN. The gap the guest can see between the two channels is 16.89 PLN on one dish, and it is not margin: it is the platform's share of a price built to survive that share.
That gap is an argument for your own channel and a risk to your reputation at the same time, and which of the two it becomes depends entirely on whether it is explained. "Prices on delivery platforms include their service costs" is a sentence that costs nothing and prevents a review. What does not work is pretending the two prices are the same when they are one tap apart. The trade behind that decision — what stays with you when the order comes through your own door — is the subject of the article on online orders.
Worth saying plainly: an own ordering channel is not free either. It needs a page that loads, a menu that is current, a payment path and someone answering the phone when it fails. The 16.89 PLN is what buys all of that, and it is a real budget rather than a windfall.
What a price uplift does not fix: the long route and the cold dish
A correct price protects one thing only — the contribution per order. It does not protect anything the guest actually experiences.
A route that takes forty minutes makes a warm dish into a lukewarm one, and no number in the numerator changes that. Packaging that leaks turns a correct margin into a refund and a public review. A menu that is out of date on the platform produces cancellations that cost the plate cost with no revenue at all. These are operational failures, and the arithmetic on this page is completely silent about them: an item can be perfectly priced and still be the wrong item to send twelve kilometres.
Two of those failures are addressable in the same system that holds your prices. Keeping the delivery card, the stock reality and the confirmations in one place is what the automation article describes, and which system layer should own the menu and the price list at all is answered in the CRM or ERP article. A price list that lives in three places will be wrong in at least two of them within a month, and a wrong price is a broken formula no matter how carefully the formula was built.
Where software helps and where it does not. Every calculation on this page is ordinary arithmetic — addition, one division, one subtraction — and it needs no model of any kind. What software adds is that the inputs stop being retyped: the effective take rate comes from the statement, the plate cost from the recipe, the packaging cost from the purchase invoice, and the answer changes when they change. Nothing on this page predicts demand, and nothing on it decides what an item is worth to a guest.
Frequently asked questions
Why can I not simply add the commission to my dining-room price?
Because a commission is a share of the price you charge rather than a fixed amount. When the price goes up, the platform's cut rises in the same proportion, so an addition never catches up. With a table price of 36.00 PLN and a take rate of 30 %, adding 30 % gives 46.80 PLN, of which 32.76 PLN reaches you; after 17.60 PLN of variable cost the item leaves 15.16 PLN instead of the 21.80 PLN it leaves at a table. Prices that have to survive a percentage deduction are built by division, not by addition.
How do I calculate the delivery price that keeps my target margin?
Add up everything that varies with that one item in the delivery channel — plate cost, packaging, incremental labour — add the contribution the item has to leave, and divide the total by one minus your effective take rate expressed as a decimal. With 17.60 PLN of variable cost, a 21.80 PLN target and a take rate of 0.30, the price is 39.40 ÷ 0.70 = 56.29 PLN. Verify it backwards: 56.29 × 0.70 − 17.60 = 21.80 PLN.
Is a flat percentage uplift across the whole delivery menu a good idea?
No, and the reason is measurable rather than aesthetic. Items differ in how much packaging and extra labour they carry relative to their price, so the uplift each one needs differs too — in the worked example, 51.61 %, 56.35 % and 72.22 % for three ordinary dishes. A single flat number overshoots on some and undershoots on others, and the guest reacts to the relative prices you have just changed, moving the mix towards exactly the items that are now missing their target.
At what commission does an item stop being worth selling?
At the rate where the platform keeps everything above your variable costs: one minus the item's variable cost divided by its platform price. The pasta at 56.29 PLN with 17.60 PLN of variable cost stops contributing at 68.7 %. If that figure comes out at zero or below, the item does not cover its variable costs even at a commission of zero, and the decision is about the item, its portion size and its packaging rather than about the price.
Am I allowed to charge more on the platform than in the restaurant?
That is a question about your contract, and this page deliberately does not answer it: what a price parity clause permits is a fact of the agreement that was signed, and it differs between platforms and between contract generations. Read the clause before repricing. If it does constrain you, the arithmetic still applies — either both channels move together, or the delivery channel runs on a lower target and the maximum take rate per item decides which lines can carry that.
What does the guest see when prices differ between channels?
A difference checkable in seconds, because both prices are one tap apart on the same phone. In the worked example the same dish is 39.40 PLN through your own channel and 56.29 PLN on the platform — 16.89 PLN apart, none of it margin. Guests accept the gap when it is explained in one sentence about service costs, and treat it as dishonesty when it is not mentioned at all. The sentence is free; the review is not.
Recompute your three best-selling delivery items with the price formula above, then compare each item's maximum take rate against the effective rate on your last settlement statement. If even one of them comes out below it, the problem is not the size of the uplift, and no percentage will repair it — the rest of this section, starting from the restaurant hub, is about what to do next.