On a per-order basis, take-out wins. The customer does the last mile for free, there is no commission, no driver, and no window in which the food can go cold in someone else’s car. Delivery wins on something else: it reaches people who were never going to walk through your door, and it fills hours your dining room cannot.
Most restaurants end up running both. The question worth answering is not which one is better in the abstract but which one you should be actively pushing with your menu, your pricing and your staff time. That depends on your segment, your margin and what your kitchen can absorb.
This comparison covers restaurant delivery vs take-out on the things that move money: margin per order, labour, packaging, risk and the customers each channel actually wins. If the live question for you is whether to run delivery with your own people or hand it off, that is covered separately in our guide to outsourcing your restaurant deliveries, and the commission maths sits in our breakdown of restaurant delivery apps.
The Bottom Line
- Take-out is the bigger channel in both segments. In 2025 take-out accounted for 25.2% of full-service customer traffic against 9.6% for delivery, and 43.6% of limited-service traffic against 12.4% for delivery (Technomic Inc., via National Restaurant Association, State of the Restaurant Industry 2026, retrieved 2026-11-17).
- Operators disagree about which matters most, and the split falls along segment lines. Among limited-service operators for whom off-premises is the bigger business, 54% named delivery as their most important off-premises channel against 25% for takeout. Full-service operators split almost evenly, 50% takeout to 48% delivery (National Restaurant Association, State of the Restaurant Industry 2026).
- Customers are more comfortable with what take-out costs. 73% say they were satisfied with the cost of their most recent takeout order, against 67% for delivery including fees, taxes and tips (National Restaurant Association, Off-Premises Restaurant Trends 2025, retrieved 2026-11-17).
- Take-out is also the more frequent habit. 47% of adults pick up takeout at least once a week; 37% order delivery that often (National Restaurant Association, State of the Restaurant Industry 2026).
- Delivery costs real money per order even when you do not employ a driver. Merchant fees on New York City restaurant delivery apps averaged 19.5% of subtotals, or $5.95 per delivery, in the second quarter of 2026 (NYC Department of Consumer and Worker Protection, retrieved 2026-11-17).
Earn $1,200+/week delivering for local businesses
Driver benefits:
- Steady local business routes
- Weekly pay + tips
- Dedicated support team
- Flexible scheduling
Restaurant delivery vs take-out, side by side
| Delivery | Take-out | |
|---|---|---|
| Who does the last mile | You, a courier or an app’s driver | The customer, free |
| Typical cost per order | 15-30% commission, or wages plus vehicle if in-house | Packaging and a few minutes of counter labour |
| Share of full-service traffic (2025) | 9.6% | 25.2% |
| Share of limited-service traffic (2025) | 12.4% | 43.6% |
| Who owns the customer | The app, unless you take the order directly | You |
| Control over food on arrival | Low once it leaves the pass | High until handover |
| Who gets the tip | The driver | Your staff |
| Main failure mode | Late, cold, wrong address | Customer arrives before the food is ready |
| Reach | Anyone inside the delivery radius | People already nearby |
| Peak pressure | Spikes with weather and events | Predictable, tied to commute patterns |
Read the traffic shares carefully, because they are the most misread numbers in off-premises. Delivery gets the attention and takeout gets the volume. At full-service restaurants takeout is more than two and a half times the size of delivery; at limited-service it is more than three times.
Metrobi is transforming catering deliveries
Specialized solutions for catering businesses:
- Catering-trained drivers
- Proper handling equipment
- Peak day delivery support
- 23% average cost reduction
Why take-out earns more per order
The arithmetic is not subtle. On a $40 order:
Take-out. You pay for packaging, perhaps $1.50 to $3.00 depending on what you are selling, and the counter minutes to hand it over. If the customer ordered through your own site you pay card processing. Your margin is close to your dine-in food margin, minus the service you did not have to provide.
Delivery through an app. You pay commission on the subtotal. At 20% that is $8.00 before processing, before any promotion you agreed to, and before any refund charged back to you. On most restaurant margin structures that is the entire profit on the order.
Delivery with your own driver. You pay a wage for the time the run takes, plus vehicle costs or mileage, plus insurance, plus the idle minutes between orders. That can be cheaper than commission at volume and much more expensive at low volume, because you are paying for the gaps.
This is why pickup commission on the apps, at roughly 6-10% against 15-30% for delivery, is the most underused lever most restaurants have. The same marketplace demand costs you a fraction when the customer collects.
Where delivery wins
Per-order margin is not the whole picture, and the case for delivery is about volume you cannot otherwise get.
It reaches people who will not come to you. A fast casual franchisee in Georgia put it to the National Restaurant Association this way: “Those customers are not going out to eat, so either my food gets to them or someone else’s food gets to them.” Delivery customers are not usually take-out customers who got lazy. They are a different occasion.
It is not capped by your room. An independent casual dining operator in Colorado made the point crisply: on-premises sales are limited by occupancy and space, off-premises only by operating hours. If your kitchen has capacity your dining room cannot use, delivery converts it.
Baskets are usually bigger. Delivery orders tend to be household orders rather than single meals, and fees make small orders uneconomic for the customer, which pushes basket size up.
It is how younger customers eat. 60% of Gen Z adults and 50% of millennials order delivery at least once a week, against 14% of baby boomers (National Restaurant Association, State of the Restaurant Industry 2026). If that is your trade, the channel is not optional.
Most operators think it is worth it. Among restaurants that offer delivery, 69% say it creates more opportunities than challenges. That drops to 63% at full-service restaurants, where margins are thinner and the dish is more fragile.
What each channel actually costs you
Margin per order is only part of the cost picture. Four other lines differ between the two channels.
Labour. Take-out loads the counter: someone has to bag, check and hand over, and at peak that person is not doing anything else. Delivery loads the pass and the phone instead, and if you run your own drivers it loads your payroll and your rota. Neither is free; they just land on different people.
Packaging. Delivery packaging has to survive a vehicle, a stairwell and a wait. That means sturdier containers, vented lids on anything fried, separated hot and cold, and a seal. Take-out packaging only has to survive a walk to a car. The cost difference per order is real and usually underestimated.
Risk. On take-out, the handover is the end of your exposure. On delivery, your exposure continues for twenty minutes you cannot see. 43% of delivery customers say the thing they most want restaurants to improve is maintaining proper food temperature, and the person in control of that is the driver, which is why driver quality is a margin issue and not a service nicety. We cover the standard to hold them to in how to find the best restaurant delivery driver.
Tips. On take-out, tips go to your staff. On app delivery, they go to the driver. An independent fine dining operator told the National Restaurant Association plainly: “Third-party delivery drivers get the tips instead of my staff.” That affects what you have to pay to keep people, and the numbers behind it are in our guide to the restaurant delivery driver tip.
Which customers pick delivery and which pick take-out
The channels are not competing for the same decision as often as it looks.
Take-out tends to win on price sensitivity and proximity. The customer is already out, or close by, and does not want to pay a fee plus a tip on a $15 lunch. Cost satisfaction runs higher for takeout than delivery for exactly this reason.
Delivery tends to win on friction avoidance. Weather, kids, no car, a group that cannot agree, a late shift. The fee is the price of not leaving, and for that occasion it is cheap. Convenience expectations have also been reset by retail, where fast shipping became the baseline rather than a perk — the same dynamic we cover in the pros and cons of two-day shipping. Diners now bring those expectations to restaurant orders.
Demographics split the same way. Takeout is the broader weekly habit at 47% of adults; delivery is concentrated in under-45s, where it is close to half the population ordering weekly.
Segment matters too. Full-service operators split almost evenly on which channel is more important, while limited-service operators lean to delivery by better than two to one. If you are a sit-down restaurant and you are being told delivery is the future, that advice was written for someone else’s business model.
The problem both channels share
Customers want the same two things fixed in both, and they are the same two things.
For delivery, the top improvement requests are maintaining proper food temperature (43%) and reducing the wait (39%). For takeout, they are maintaining proper temperature (40%), reducing the wait (39%) and improving food quality (38%).
That is a kitchen and packaging problem, not a channel problem. A restaurant that fixes hold times, packaging and the order-ready signal improves both channels at once, and the fix is usually cheaper than anything you can do with pricing.
Two practical moves do most of the work. Build an off-premises menu that excludes dishes that cannot survive fifteen minutes in a box. Our primer on what a restaurant menu is and what it has to do is a useful starting point if you are building a second menu for the first time. And change the sequencing so that off-premises orders are timed to the handover rather than fired with the dine-in tickets.
How to decide which channel to push
Work through these in order.
- Start with your dish. If it degrades fast or plates as part of the experience, take-out is the channel you can actually deliver well on, and delivery should be a limited menu at most.
- Check whether delivery orders are incremental. If your delivery volume is mostly regulars who used to collect, you are paying a commission for orders you already had. Push pickup instead.
- Look at your off-peak. Delivery earns its keep when it fills hours your staff are already paid for. If it forces an extra shift, the maths gets much worse.
- Compare the all-in cost to your gross margin. Commission plus processing plus promotions plus a refund allowance. If that exceeds gross margin on a typical delivery basket, delivery is losing money however busy it looks.
- Make pickup easy before you discount it. A clear collection point, an accurate ready time and a short queue convert more delivery customers to take-out than a 10% off code does.
- Price the channels separately. Delivery menu prices should reflect delivery costs. Take-out prices should not have to subsidise them.
The short answer
Take-out is the more profitable order and the bigger channel in both restaurant segments. Delivery is the growth channel, the one that reaches customers who are not coming to you, and the one that costs you most of the margin on every order it brings.
Push take-out as your default off-premises channel, and treat delivery as paid customer acquisition with food attached. Measure it that way and the decision usually makes itself: delivery is worth running where the orders are new, the dish travels, and the hours were already staffed.