- Why Off-Premise Sales Can Look Better Than They Really Are
- Calculate Contribution Margin by Channel
- A Hypothetical $50 Order
- The 10 Off-Premise Metrics Restaurants Should Track
- Revenue by Channel Is Not Enough
- Are Marketplace Orders Incremental?
- Factor Customer Acquisition Into the Math
- Do Not Ignore Operational Costs
- Measure Pickup, Delivery, and Catering Separately
- When a Lower-Margin Channel Can Still Be Worth Keeping
- A Real Orders.co Example of Channel Mix
- How Orders.co Can Help Operators See the Channel Picture
- Frequently Asked Questions
- More Helpful Reads
An off-premise channel is profitable only when the revenue it produces exceeds the variable costs required to fulfill those orders and contributes enough toward your fixed expenses and profit goals. Sales alone will not tell you that. To know whether delivery, pickup, direct ordering, or catering is actually making money, you calculate contribution margin by channel instead of comparing revenue.
Contribution margin is a plain idea. It is what is left from an order after you subtract the costs that exist only because that order happened. If the order goes away, those costs go with it. What remains helps pay for the parts of your business that stay the same, whether the order comes in or not.
Here is the formula for a single order:
Off-premise contribution margin per order = order revenue minus food cost minus packaging minus variable labor minus marketplace, dispatch, or payment fees minus restaurant-funded discounts minus refunds or other variable order costs.
Read that as contribution margin, not final profit. It does not yet cover rent, salaried managers, insurance, utilities, or the other fixed overhead that keeps the doors open. Those still have to be paid out of the contribution your orders generate, and keeping that distinction straight is the whole point of this article.
Why Off-Premise Sales Can Look Better Than They Really Are
Off-premise is no longer a side project. National Restaurant Association research says nearly three-quarters of restaurant traffic now happens off-premises, and its 2025 Off-Premises Restaurant Trends study found that 47% of adults get takeout at least weekly, 42% use a drive-thru weekly, and 37% order delivery weekly. That scale is why a big sales number can mislead you.
You open a dashboard and see “$30,000 in DoorDash sales.” It feels like a win. But that figure says nothing about commission expense, promotion funding, packaging, labor, refunds, chargebacks, remakes, or whether you priced the menu higher on that platform. It also cannot tell you whether the sales were incremental, whether the customer would have ordered directly anyway, or whether the order created a repeat guest. None of that makes marketplace orders automatically unprofitable; it just means the sales figure is where the analysis starts, not where it ends.
Calculate Contribution Margin by Channel
Each off-premise channel earns and spends money differently, so comparing them side by side helps you see where to look.
| Channel | Revenue | Major Variable Costs to Watch |
| Marketplace delivery | Menu revenue | Food, labor, packaging, commission, promotions, refunds |
| Direct delivery | Menu revenue plus delivery fee, where applicable | Food, labor, packaging, payment processing, dispatch, or driver cost |
| Pickup | Menu revenue | Food, labor, packaging, payment processing, or platform fees |
| Catering | Higher-ticket revenue | Food, labor, packaging and equipment, delivery, setup, sales, and admin time |
Do not fill these rows with borrowed industry averages. Your food cost, packaging, labor, and negotiated fees are the only numbers that describe your restaurant. Pull the real figures from your own records.
A Hypothetical $50 Order
Numbers make this concrete. Take a clearly hypothetical $50 marketplace order and subtract the variable costs one by one:
- Start: $50 order
- Food cost: -$15
- Marketplace commission (hypothetical 25%): -$12.50
- Packaging: -$2.50
- Variable labor to fulfill it: -$5
- Restaurant-funded promotion: -$3
- Contribution: $12 before fixed overhead and any other applicable costs
So would you still want that order? Maybe. If it was incremental business filling unused kitchen capacity on a slow Tuesday, $12 of contribution is real money you would not have had otherwise. If the customer was a regular who would have placed the identical order on your own website, the math changes, because you paid a commission to serve someone you already had. That nuance decides whether the order helped you or quietly cost you.
The 10 Off-Premise Metrics Restaurants Should Track
You do not need a finance degree to measure food delivery profitability. You need to track the restaurant delivery metrics that matter, ten of them, measured the same way every time. Vendor analysis from Olo also argues operators should watch guest acquisition, visit frequency, and average order value together, a useful lens if treated as a perspective, not a universal benchmark.
| Metric | Formula | Why It Matters | What to Investigate |
| Contribution margin per order | Revenue minus all variable order costs | The core financial metric for off-premise | Any channel or daypart with a thin or negative result |
| Contribution margin percentage | Contribution per order divided by order revenue | Lets you compare channels of different ticket sizes | Channels with high sales but a low percentage |
| Average order value | Total channel sales divided by order count | Sets the scale, but means little without a margin | High AOV paired with low contribution |
| Effective marketplace fee rate | Total marketplace costs divided by relevant sales | Shows what platforms really cost after promos and fees | A rate well above your plan’s headline commission |
| Packaging cost per order | Total packaging spend divided by orders | A commonly overlooked variable cost | Rising cost per order or waste on specific items |
| Variable labor per off-premise order | Fulfillment labor divided by off-premise orders | Prep, packing, and handoff time add up | Orders that take disproportionate staff time |
| Refund, remake, and chargeback rate | Affected orders divided by total orders | Direct revenue leakage | One channel is leaking more than the others |
| Order accuracy and cancellation rate | Accurate orders divided by total orders | Operational health that eventually hits economics | Modifier errors, missed items, cancellations by cause |
| Channel mix | Each channel’s share of total off-premise orders or sales | Shows how dependent you are on any one channel | Over-reliance on a single marketplace |
| Repeat-order behavior by channel | Returning customers divided by total, where available | Signals that build lasting value in channels | Channels that never produce a second order |
Revenue by Channel Is Not Enough
Sales reporting and profitability reporting are not the same thing, and most restaurant tools show you only the first. A dashboard that reads “DoorDash: $20,000, Direct: $15,000” tells you where orders came from. It does not tell you which channel puts more money toward your rent. For that, you need the cost side of every order.
This is the sentence worth pinning above your desk: the highest-revenue off-premise channel is not necessarily the most profitable. A channel doing $20,000 at a thin margin can contribute less than one doing $15,000 at a healthier one. Until you attach costs to sales, you are ranking your channels blindly. The gap is drawing attention. The second annual Optimizing Off-Premises Strategies study from Restaurant Business and Nation’s Restaurant News surveyed more than 300 operators on channel mix and better ways to measure off-premise ROI. Its August 27, 2026, session was sponsored by Olo, so treat the findings as an informed vendor perspective, not neutral research.
Are Marketplace Orders Incremental?
The hardest question in off-premise economics is also the most important: would this sale have happened without the marketplace? There are three broad scenarios, each with a different conclusion.
True acquisition is a new customer discovering your restaurant through DoorDash. Channel substitution is a loyal regular who already knew you, ordering through DoorDash instead of directly, which means you paid a fee to reach someone you already had. An incremental occasion is a customer who chooses delivery on a night they would otherwise have cooked or gone elsewhere, so the order is genuinely additional.
You often cannot sort every order into the right bucket with certainty, and no honest attribution model will promise you can. Still, naming the three scenarios stops you from treating every marketplace order as pure profit or pure loss.
Factor Customer Acquisition Into the Math
A commission is not only a deduction. It can also buy discovery, delivery logistics, marketplace exposure, and customer acquisition that you would otherwise pay for another way. DoorDash frames its commission partly around those services. Whether or not you accept that fully, it reframes the question: instead of asking how much the platform took, ask what you received for the fee.
Operators feel the weight of these fees. Restaurant Business reported in January 2026 that 50% of surveyed operators named third-party delivery fees as their biggest obstacle to growing off-premise business, ahead of order accuracy at 33% and speed of service at 24%. Marketplace commissions commonly run between roughly 10% and 30%, depending on the service and agreement. As of September 2026, DoorDash’s public U.S. Marketplace pricing lists Basic at 15%, Plus at 25%, and Premier at 30% commission, with pickup at 6% across those plans, and says the plans carry no monthly or signup fee. Never assume every restaurant pays the same rate, because negotiated deals and other platforms differ. If a marketplace regularly brings in new guests who later order directly, part of that commission works as a marketing cost with a return. If it mostly processes orders from people who already know you, the same fee does far less. Same percentage, very different value.
Do Not Ignore Operational Costs
Some restaurant delivery costs never show up as a tidy line item, but they are real. Tablet and device management, manual order entry, missing modifiers, remakes, driver handoff time, delayed orders, refund administration, chargeback disputes, menu updates across platforms, and stitching reports together from disconnected systems all cost something.
Every one traces back to one of four places: labor, accuracy, customer satisfaction, or lost revenue. A missed modifier becomes a remake, which is labor plus food cost plus a slower kitchen. A late order becomes a lower platform ranking, which results in lost future revenue. When you audit off-premise profitability, count the hidden operational drag, not just the obvious fees.
Measure Pickup, Delivery, and Catering Separately
Do not pour all off-premise revenue into one bucket; it hides very different businesses. Pickup usually carries different fulfillment costs from delivery, since no driver or dispatcher is involved. Marketplace and direct ordering run on different fee structures. Catering brings larger tickets but adds production, coordination, delivery, setup, and sales or admin hours that a normal to-go order never touches.
Blend them, and your “off-premise margin” becomes an average that describes none of them. Separate them, and you can see which model earns its keep and which is being carried by the others.
When a Lower-Margin Channel Can Still Be Worth Keeping
A low contribution margin is a reason to look closer, not to shut a channel off. A marketplace that runs a thinner margin can still earn its place if it drives customer discovery, fills spare kitchen capacity, moves off-peak volume, exposes you to a new neighborhood, adds convenience your regulars value, or lets you test a new concept without a full commitment.
The mistake is cutting a channel on its commission percentage alone. The right decision weighs contribution against what the channel does for the rest of the business. Sometimes you keep it and fix the margin; sometimes you shift volume toward a channel you own. You cannot tell which until you measure.
A Real Orders.co Example of Channel Mix
Channel mix is easier to understand with real numbers. In a published Orders.co case study, Spartan’s Grill recorded the following across roughly a 12-month measurement period: direct website orders generated $38,663.75 across 1,138 orders, while Uber Eats generated $35,383.73. Direct orders made up 35% of online orders, with a direct average order value of $33.98 against $34.19 on Uber Eats.
That is a useful picture of channel mix and revenue, and nothing more. The case study does not publish Spartan’s Grill’s food, labor, packaging, or overhead costs, so it cannot tell you which channel produced more contribution margin. Treat it as an example of how direct and marketplace channels can sit close in revenue and order value, not as proof that direct orders delivered a specific profit. Being honest about that limit is the same discipline this article asks you to apply to your own numbers.
How Orders.co Can Help Operators See the Channel Picture
Most operators cannot measure channel profitability because the revenue and order data live in four or five places at once. Olo has reported that 66% of restaurants use more than four technology systems, and fragmented reporting makes performance genuinely hard to read. Orders.co brings the revenue side into one view. It consolidates direct and third-party orders from DoorDash, Uber Eats, Grubhub, and ezCater into one system, and reports sales by channel, order counts, and menu performance, along with customer data from your direct orders and menu management across platforms.
That puts the revenue and order side of the equation in one place instead of scattered across dashboards. It does not, on its own, calculate your true contribution margin because it does not hold your ingredient, labor, and overhead figures. To reach profitability, you combine the centralized channel data with your actual food, labor, packaging, and other costs. The platform gives you the top half of the equation cleanly; you supply the cost half. Together, they answer the question this article opened with.
Frequently Asked Questions
Yes. Catering is off-premise because the food is consumed away from your dining room. But keep it in its own bucket. Catering carries larger tickets, which cost a normal to-go order never has, including extra production, coordination, delivery, setup, and sales or admin time. Folding it into general off-premise numbers hides those costs and distorts every channel’s margin, so measure it on its own.
Yes. Pickup and delivery have different fulfillment costs. Pickup usually skips driver pay, dispatch, and delivery packaging, while delivery adds those and often a higher commission. Combining them produces an off-premise margin that describes neither accurately. Track each channel on its own so you can see which one actually earns its keep and where a fix or a shift in volume would help most.
Packaging is usually treated as an operating cost, not food cost, because it is not an ingredient in the dish. For contribution margin, the label matters less than making sure you count it. Packaging is a variable cost that exists only because the off-premise order happened, so subtract it per order. Many operators overlook it, and on high-volume delivery, it adds up to real money.
Sales tax is collected on behalf of the government and passed through, so it is not your revenue and should stay out of the calculation. Tips generally belong to staff and are not restaurant margin either, unless your local rules and payroll setup say otherwise. Base contribution margin on the money your restaurant actually keeps from the food and any fees you retain, not the gross charge.
Self-delivery can cost less when you have steady delivery volume, drivers you can keep productive, and a delivery area tight enough to keep drive times short. Third-party delivery often wins when volume is unpredictable, distances are long, or paying a per-order fee beats carrying idle driver labor. Compare your true cost per self-delivered order, including driver pay, vehicle costs, insurance, and dispatch, against the marketplace fee for the same order.
