Dine-In vs Takeaway vs Delivery: How to Set Up Restaurant Order Channels (2026)

August 12, 2026 at 9:44 PMTabres Team

order channels dine-in vs takeaway vs delivery restaurant online ordering takeaway vs delivery pickup orders restaurant channel mix order types

Most restaurants switch on takeaway and delivery with two toggles, copy the dine-in menu into both, and call it a strategy. Then Friday at 7 p.m. arrives and the kitchen is cooking for three rooms with one line.

Short answer: dine-in, takeaway and delivery are not three versions of the same order. They're three different products with different costs, different guests and different failure points. Set each one up on its own: its own hours, its own menu, its own prices, its own quoted prep time, and its own cap on how many orders it can send the kitchen at once. Takeaway is usually your most profitable channel and the one nobody optimises. Delivery is the most expensive and needs the tightest rules. Dine-in still pays for the building, because it's the only channel that sells a second drink.

The rest of this is the setup, channel by channel — what to change, what to leave alone, and what one order in each channel actually leaves behind.

What an order channel really is

An order channel is how the food leaves your building. That's it. Everything else follows from that one fact.

Most systems ship four:

Dine-in — guest sits, staff serve, plate comes back.
Takeaway (pickup) — guest orders ahead, collects, carries it themselves.
Delivery — someone drives it, and the clock keeps running after the food leaves the pass.
Self-service — the guest orders from their own phone, usually at a table with a QR code. It's dine-in, but the ordering happens without a waiter.

That fourth one confuses people, so here's the line: self-service is a way of taking the order. Dine-in, takeaway and delivery are ways of finishing it. A guest can scan a QR code and choose "collect at the counter" — that's a self-service takeaway order, and your kitchen needs to know both halves.

The three channels are three different businesses

Same kitchen, same recipes, completely different economics. Here's what one $40 basket really leaves behind.

Channel	Guest pays	Food + channel costs	You keep	Ties up a table?
Dine-in	$40	$12 food + $5.25 service and wash-up	$22.75	Yes, ~75 min
Takeaway	$40	$12 food + $2.25 packaging and counter time	$25.75	No
Delivery, your own driver	$44.99	$12 food + $12.30 per drop	$20.69	No
Delivery, app, no markup	$40	$12 food + $2 packaging + $10 commission	$16.00	No
Delivery, app, 25% markup	$50	$12 food + $2 packaging + $12.50 commission	$23.50	No

Assumptions: 30% food cost, a $4.99 delivery fee on your own channel, roughly $12.30 all-in cost per drop, and a 25% app commission. Your numbers will differ. Run them anyway.

Look at row two. Takeaway keeps more of a $40 order than dine-in does, and it doesn't hold a table for an hour and a quarter. That's the whole argument for pickup in one line.

Now the honest counterpoint, because that table hides something. Dine-in tickets carry drinks. A $6 beer costs you about $1.20. A second round, a dessert, a coffee — those are 80% margin items, and a takeaway bag almost never has them. So dine-in wins on the size of the basket, and takeaway wins on the cost of serving it.

You need both. You just need to stop treating them as one.

Takeaway is the channel nobody optimises

Everyone argues about delivery apps. Meanwhile pickup sits there quietly outperforming, and most kitchens have never once looked at its settings.

Think about what a takeaway order actually is. No driver. No commission. No table turned. No waiter walking eight times to one seat. The guest does your last mile for free — and turns up at your door, where they can smell the place and see the specials board.

Yet in most restaurants, takeaway is set up like this: same menu as dine-in, same prices, "about 20 minutes" for everything, and a bag left on the end of the bar with a name scribbled on it.

Five fixes that take an afternoon:

Quote a real ready time per order, not a flat number. Two burgers is 12 minutes. Two burgers on a Friday at 8 p.m. is 25. Guessing low doesn't make the food faster — it just puts an annoyed guest in your doorway.
Give pickup a name and a number. Dine-in tickets are called by table. Pickup tickets need the guest's first name and an order number, printed big. "Order 47, Maria" beats "the tall guy in the blue jacket".
Build a real collection point. A shelf, a labelled bag, a place that isn't the service station. If your collecting guests block your seated guests, you've turned a profit into a complaint.
Pack at a packing station, not on the pass. Bagging in the middle of plating is how the fries end up in someone's dine-in order and the sauce ends up nowhere.
Read the bag back. Ten seconds. Every item, out loud, against the ticket. Missing items are the number one reason a pickup guest doesn't come back, and they're free to prevent.

Pickup is also the safest place to test your off-premise menu before you ever hire a driver or sign with an app. If a dish can't survive a ten-minute car ride, it definitely can't survive a courier and a traffic light.

Setting up dine-in so the other channels don't break it

Dine-in is usually already running, so this is less "set up" and more "tighten before you add load".

Every table needs an identity. A number, a seat count, a section. If a table can't be named, an order can't be attached to it, and every other feature you turn on later inherits that mess.
Give each table its own QR code, not one shared code. A shared code strips the table number out of the order, and somebody has to put it back by hand — usually while holding a hot plate. We wrote up why in table-specific QR codes.
Turn on "add to order". Guests who can add a coffee to an open table order add a coffee. Guests who have to flag someone down usually don't.
Decide what a table status means, and use the same words all shift. Occupied, bill requested, needs cleaning — if half the team uses the screen and half uses shouting, the screen is fiction.
Watch your table turn time before you switch on delivery. If dine-in already runs slow at peak, a second channel won't expose the problem. It'll multiply it.

Self-service ordering sits on top of all this. It's the cheapest capacity you'll ever add — the guest does the typing — but only if the table identity underneath it is clean.

Setting up delivery so it doesn't eat the kitchen

Delivery is the channel with the most settings, and the one where a wrong setting costs real money every single night.

The four decisions, in order:

Direct or third-party — or both. Apps bring reach and take 15% to 30% of it. Your own channel keeps everything but the card fee, and brings nobody. Most independents run both on purpose: apps for discovery, direct for regulars. If you're weighing it up, we did the real math of third-party delivery and a step-by-step on direct delivery without the apps.
Fee, minimum order and zones. These three numbers decide whether delivery makes money, and almost nobody calculates them. There's a full breakdown in how to set delivery fees, minimums and zones. The short version: cover 40% to 60% of your cost per drop, set the minimum by formula, and draw zones by drive time — never by a circle on a map.
Its own hours. Delivery should not copy your opening hours. It should own the dead hours your dining room doesn't want, and step back during the 90 minutes your line genuinely can't cook for two rooms. That's a whole argument on its own: why delivery hours shouldn't match dining room hours.
A last-order cut-off. 30 to 45 minutes before the kitchen closes. Without it, someone orders a full rack of ribs four minutes before close and your team leaves an hour late.

One more thing that sounds small and isn't: the delivery ticket needs the address, the phone number and the courier note on the driver's copy — and none of that on the kitchen's copy. The kitchen needs food and allergens. It doesn't need a guest's address on a screen the whole room can see.

The five settings that must differ per channel

If you change nothing else after reading this, change these five. Cloning them across channels is the single most common setup mistake in hospitality.

1. Hours

The dining room, the pickup counter and the delivery zone don't share a clock. Set them per weekday, not once for the whole week. Sunday delivery and Tuesday delivery are different businesses.

2. Menu

Not every dish should exist in every channel. This is where the travel test comes in — see the next section. Anything that fails it stays a dine-in dish, and that's fine. A shorter delivery menu is faster, travels better, and gets fewer refunds.

Keep availability separate too. You can be out of sea bass for delivery and still serving it in the room, because the room eats it in 90 seconds.

3. Prices

Dine-in: your baseline.
Takeaway: same price. Don't punish the channel that costs you least.
Your own delivery: a 10% to 15% markup is normal and guests accept it. Add the delivery fee on top, visibly, on the menu page — not at the last checkout screen.
Third-party apps: mark up 20% to 30%, or the commission comes straight out of your margin. Just know that some app agreements have rules about price parity, so read yours.

Never bury the whole delivery cost in menu prices. Guests compare your two menus, and they will notice.

4. Quoted prep time

One flat "20 minutes" for everything is a promise you break all night. Prep time should move with the kitchen's actual load. When you're buried, quote 40 minutes instead of switching delivery off — a guest who's told 40 and gets it in 35 is happy. A guest told 20 who waits 40 leaves a review.

5. A capacity cap

This is the setting almost no one uses, and it's the one that saves service. Decide the most orders per channel your kitchen can take in a 15-minute window — say six dine-in tickets, four pickups, three deliveries. Past that, the channel goes to a longer quote or pauses for ten minutes. Throttle, don't collapse.

The travel test: box it, wait twelve minutes, eat it

Here's the cheapest menu decision you'll make this year.

Take a dish. Box it exactly the way you'd send it. Set a timer for 12 minutes — that's a realistic pickup drive plus a lift. Then sit down and eat it, cold coffee and all.

If you wouldn't pay for what's in that box, it doesn't belong on your off-premise menu.

Usual failures:

Anything fried and sauced together — batter turns to paste in about four minutes
Fries in a sealed bag with no vents
Poached eggs, soufflés, anything that was timed to the second
Salads dressed in the kitchen
Melted-cheese dishes that set into a solid sheet
Ice cream, unless you've solved the bag, and you haven't

Usual winners: braises, stews, curries, roasted meats, grain bowls, pizza with a vent hole, anything that was already fine at 60°C, and anything where the sauce ships in its own pot.

Two fixes solve most failures: separate the wet from the dry, and vent the hot. Sauce on the side, fries in their own vented sleeve, lid with a hole. It costs cents and saves refunds — which matter, because a refund on delivery costs you the food, the driver and the guest, all at once. If you're seeing a lot of them, missing-item refunds have their own playbook.

One kitchen, four exits: the sequencing problem

Every channel meets in the same place — the pass. That's where good setups fall apart.

The ticket has to say how the food leaves. Not in small print. Big, at the top, in a way a cook sees at a glance: dine-in, pickup, delivery, self-service. Because the same three dishes get plated three different ways depending on that word, and a cook plating a delivery burger on a hot plate has just wasted six minutes.

Then give the line one priority rule and stick to it. Most kitchens land on some version of:

Dine-in courses that are already partly served — nobody eats alone at their own table
Delivery, because a driver is standing there and the clock is against you
Pickup, because the guest is walking and has a few minutes of slack
New dine-in tickets

Argue with that order if you like. Just have one, and have it written down, or every ticket becomes a negotiation.

Route by station, not by channel. Grill items go to grill whether they're leaving in a bag or on a plate — one map, all channels, set up once. What changes at the end is the plating and the packing, not the cooking.

And use an all-day count. When you can see the kitchen owes 14 portions of fries across every open ticket, someone drops 14 at once instead of dribbling them out three at a time. This is where multi-channel kitchens win or lose their peak.

Tax: the boring part that gets restaurants in trouble

The same sandwich can carry a different tax rate depending on which channel it left through. This is not an edge case — it's the normal situation in most of Europe and in a lot of US states.

In the US: whether a delivery fee is taxable varies by state, and it can even depend on whether the charge appears separately on the receipt. Ask your state's department of revenue or your accountant. Don't guess, and don't copy what the pizza place across the street does.
In the EU and UK: dine-in is usually treated as a service and taxed at the standard rate, while cold takeaway food is often reduced or zero-rated, and hot takeaway sits somewhere in between. We broke the whole split down in VAT on dine-in, takeaway and delivery.

The practical setup point: your system should be able to apply a tax rule to one channel only. If it can't, you're either overcharging guests or underpaying tax, and both get expensive. Most decent systems let you scope a tax to dine-in, takeaway or delivery — check yours before you switch a channel on, not after your first quarter.

What to measure once all three are running

Total sales tells you almost nothing when three channels share a kitchen. Five numbers do.

What to track	Why it matters	Check it
Channel mix — % of sales by channel	Tells you where the business actually is, not where you think it is	Weekly
Contribution per order, by channel	The only fair comparison. Revenue lies; contribution doesn't	Monthly
Average ticket, by channel	Delivery should be your highest. If it isn't, your minimum order is too low	Weekly
Refund and remake rate, by channel	Delivery over 3% means a packaging or travel-test problem	Weekly
Prep time, by channel	If pickup times climb on busy nights, your cap is too high	Weekly

A rough starting benchmark for an independent with a dining room: something like 55% to 70% dine-in, 15% to 25% takeaway, 10% to 20% delivery. Wildly different mixes aren't wrong — a lunch spot in an office district might be 70% pickup — but a mix you can't explain usually means one channel is quietly running the kitchen.

One warning about the numbers themselves. Compare channels on contribution, not revenue. Delivery almost always looks like your fastest-growing channel, because bigger baskets and fees inflate the top line. Look at what's left after the drop cost or the commission and the ranking often flips. Your average ticket by channel is the number to watch move.

A sane rollout order

Don't launch three channels in a month. Do this instead:

Weeks 1–2 — fix dine-in. Table map, table QR codes, clean statuses, honest turn times. Add nothing.

Weeks 3–4 — add takeaway. Cheapest, safest, most profitable. Run the travel test, build the collection point, quote real times. Learn to pack while the stakes are low.

Weeks 5–8 — add delivery, small. One tight zone. Your own channel first, not the apps — you'll learn more from 20 direct orders than from 200 app orders you never see the guests of. Fee, minimum, separate hours, cut-off time.

Week 9+ — decide about apps. Now you know your real cost per drop, so you can tell whether a 25% commission is reach you're buying or margin you're donating.

Then review the whole thing every quarter. Wages, packaging and courier rates all moved in the last year. A channel setup from 2024 is a guess about a world that no longer exists.

Most free restaurant platforms now let you toggle dine-in, takeaway, delivery and guest ordering per branch, with separate delivery hours, fees and minimums attached to each. Tabres works this way, for example. Whatever you use, the test is the same: can you change one channel without touching the other two? If not, you don't have channels. You have one menu wearing three hats.

Mistakes that turn three channels into one mess

Cloning the dine-in menu into delivery. The fastest way to build a one-star review out of a perfectly good dish.
One prep time for everything. A promise you break by 8 p.m. every Friday.
No capacity cap. The kitchen finds its own limit, usually in the middle of service, loudly.
Switching delivery off at peak instead of throttling it. You lose the order and the guest. A longer quote loses neither.
Same price everywhere, including the apps. That's a 25% discount you're giving to a company that isn't your guest.
Punishing pickup with a service fee. You're taxing your best channel for the crime of costing you less.
A ticket that doesn't say how the food leaves. Every plating mistake in a multi-channel kitchen starts here.
Judging channels on revenue. Delivery always wins that contest and often loses the real one.

Order channels FAQ

What are restaurant order channels? The ways an order reaches the guest: dine-in, takeaway (pickup), delivery, and self-service ordering from the guest's own phone. Each one has different costs, different prep needs and often different tax treatment, so each should be set up separately.

Which is more profitable, takeaway or delivery? Takeaway, almost always. There's no driver, no commission and no table held, and the guest collects the food themselves. On a $40 basket, pickup typically keeps $5 to $9 more than the same order delivered.

Should delivery prices be higher than dine-in prices? On your own delivery channel, a 10% to 15% markup is standard and guests accept it. On third-party apps, 20% to 30% is normal, because the commission has to come from somewhere. Keep takeaway at dine-in prices.

Should I have a different menu for delivery? Yes — a shorter one. Run the 12-minute travel test on every dish and remove whatever fails. Fried, sauced, timed-to-the-second and ice-cold-required items usually don't survive the trip.

Can I turn off one channel during busy hours? You can, but a longer quoted prep time usually works better. Switching a channel off loses the order entirely. Quoting 40 minutes instead of 20 keeps most of them and sets the expectation honestly.

Do dine-in, takeaway and delivery have different tax rates? Often, yes. In most of Europe, dine-in is taxed as a service while cold takeaway food may be reduced or zero-rated. In the US, whether a delivery fee is taxable depends on the state and sometimes on how it's shown on the receipt. Check with your local tax authority or accountant — the rules change.

What's a healthy channel mix for a restaurant? For an independent with a dining room, roughly 55% to 70% dine-in, 15% to 25% takeaway and 10% to 20% delivery is a common starting point. The right mix depends on your location — a lunch spot in an office district can be mostly pickup and perfectly healthy.

Do I need a separate system for each channel? No, and you shouldn't want one. Separate systems mean separate menus to update, separate tablets to watch and orders that don't add up at the end of the day. One system with per-channel settings is the goal.

Here's the part that takes the longest to accept: adding a channel doesn't add capacity. It adds demand. The kitchen stays exactly the same size.

So the work isn't turning channels on. It's deciding, on purpose, what each one is allowed to do — its hours, its menu, its prices, its promises and its ceiling. That's an afternoon of settings and one honest conversation with your head chef.

Do it this week, before the next busy Friday. Then go box up a burger and let it sit for twelve minutes. That one test will tell you more about your delivery menu than a month of reports.