Why Your Delivery Hours Shouldn't Match Your Dining Room Hours: Restaurant Delivery Hours Guide (2026)

August 10, 2026 at 12:02 PMTabres Team

restaurant delivery hours delivery cut off time online ordering hours restaurant dayparting delivery prep time

Most restaurants set their delivery hours once, on the day they switched delivery on, by copying their opening hours. Then they never look at that screen again. It's one of the most expensive five-minute decisions in the building.

Short answer: your delivery hours shouldn't match your dining room hours — but that doesn't mean "shorter". It means different. Delivery should own the hours your dining room doesn't want: the 2 p.m. to 5 p.m. dead zone, the hour after last seating, the day you're closed. And it should step back during the 90 minutes your line genuinely cannot cook for two rooms at once. Add a last-order cut-off 30 to 45 minutes before the kitchen closes, raise your quoted prep time at peak instead of shutting the channel off, and set the schedule per weekday, not once for the whole week.

The dining room and the delivery channel share one kitchen. They do not share one clock. Once you see it that way, the settings screen looks completely different.

Your kitchen doesn't double when delivery opens

Here's the thing nobody tells you when you turn on online ordering. You've added a second dining room, with its own guests and its own tickets — and staffed it with exactly the same people.

A delivery order isn't a lighter version of a dine-in order. It's heavier. Same cooking, plus:

3 to 5 minutes of packing, bagging, sauce cups, labels, and seals
A handoff to a driver who may or may not be standing there
Zero chance to fix anything after it leaves

So when your line is doing 40 covers an hour and a delivery ticket lands, you didn't add one ticket. You added one ticket and a five-minute job nobody is scheduled to do.

That's the whole argument in one paragraph. Delivery hours are not a marketing setting. They're a capacity setting.

The 20 extra minutes nobody schedules

Dine-in food travels 30 feet. Delivery food travels 3 miles. Your ticket time is only half the story.

Look at a real Friday order:

7:07 p.m. — guest taps order
7:08 p.m. — you accept it
7:22 p.m. — food up (14 minutes, busy line)
7:26 p.m. — packed and sealed
7:31 p.m. — driver arrives, waits, leaves
7:48 p.m. — doorbell
7:51 p.m. — guest actually eats

That's 44 minutes from tap to bite. The kitchen sees 14 minutes and thinks it did a great job. The guest sees 44 minutes and leaves three stars.

On a Tuesday at 3 p.m., the exact same order is a 26-minute experience. Same food. Same kitchen. Same price. Completely different review.

Now ask yourself which of those two hours you'd rather sell.

The four switches most owners never separate

"Delivery hours" isn't one setting. It's four, and mixing them up is why so many kitchens feel out of control at 7:30 p.m.

Delivery hours — the schedule the channel runs on. Set these per weekday. Monday and Saturday are not the same business.
Last order cut-off — the moment the button stops working. This is not the same as closing time, and treating it as the same is mistake number one.
Quoted prep time — the promise you make on the checkout screen. Honest at 3 p.m., longer at 7 p.m.
Pause — the emergency brake for when the rail is buried and the schedule was wrong.

Two more you can flex by time of day, and almost nobody does:

Delivery radius. Tight at peak, wider when the road is empty. A 4-mile drop is fine at 3 p.m. and a disaster at 7:30 p.m.
Minimum order. Raise it during peak. If a ticket is going to cost you a dine-in table's timing, it should at least be a big ticket.

Stop taking delivery orders when the kitchen closes

Kitchen closes at 10 p.m. Delivery cut-off set to 10 p.m. Here's what actually happens.

An order lands at 9:58. Food's up at 10:12. Packed at 10:16. Driver's late, gone by 10:24. Your closer is still standing there with the lights on, on the clock, at 10:35 — and if it's your own driver, they're back at 10:50.

You just bought 40 minutes of paid labour and a delayed close for one $28 order.

Set the delivery cut-off 30 to 45 minutes before the kitchen closes. Use 45 if you run your own drivers, 30 if you use a courier service. It costs you a handful of late orders and it buys you a clean close every single night.

While you're in there: if you trade past midnight, check that your reports use a proper business-day boundary, or your Friday delivery sales will land on Saturday's numbers. We wrote about that specific headache in why late-night sales show up on the wrong day.

Don't close delivery at peak — cap it

Here's where most advice gets lazy. "Turn delivery off during the dinner rush." That's easy to say and usually wrong.

Your delivery peak is 6:30 to 8:30 p.m. So is your dine-in peak. If you close the channel every night at exactly the hour people want to order, you don't have a delivery business — you have a hobby.

Try this order instead, from gentlest to harshest:

Raise the quoted prep time. Go from 20 minutes to 35. Guests see an honest number and decide for themselves. Some order anyway. Some order at 9. Both are fine — nobody is angry, because nobody was surprised.
Throttle the flow. If your platform lets you cap orders per 15-minute slot, use it. Four orders a slot beats eleven orders in one slot and a dead half hour after.
Raise the minimum order and shrink the radius for those two hours.
Pause when the rail actually breaks. Ten minutes off is better than an hour of cold food.
Close the channel — last resort, and only for the one or two shifts where the line genuinely can't do both. For a 24-seat room on a Saturday night, that's a completely reasonable call.

One note on the marketplaces, since people worry about it: a scheduled closure is not the thing that hurts you. What hurts is accepting orders and then cancelling them, and quoting 20 minutes when you take 45. Apps track that. Honest hours and honest prep times are the safe play.

The window that actually pays: 2 p.m. to 5 p.m.

Everybody fights over the dinner rush. Meanwhile the best delivery hour in most kitchens is sitting there empty.

Between lunch and dinner your dining room is dead, your rent is still running, and your cooks are prepping. Delivery demand in that window is real — offices, people working from home, late lunches, early dinners for families with small kids. And your line has nothing else to do.

Look at what changes in that window:

Ticket time drops, because nothing is competing
Food arrives hot, so reviews go up
No dine-in guest is hurt by the order
The labour is already paid for

That's the closest thing to free money on the schedule. If your delivery hours have a gap from 2 to 5 because you copied your "kitchen closes between services" habit into the ordering platform, you're turning off your most profitable delivery hour to protect nothing.

Same logic after close. If your dining room stops at 10 p.m. and your neighbourhood orders at 11, a short late-night menu — 8 to 10 items that travel well — can run with one cook and a courier.

Does it pay? Do the math before you commit, not after:

Average late delivery ticket: $32, plus a $5 delivery fee the guest pays
Food cost at 30%: –$9.60
Packaging: –$2.00
Courier, flat: –$8.00
Card processing: –$1.10
Contribution per order: about $16

One cook for two hours at $18/hour is $36, plus roughly $10 of utilities. So you need about three orders per two-hour block to break even. If you can't do two orders an hour after close, don't open. If you can do five, that block is quietly one of the best on your week.

How to find your real delivery windows in 14 days

No guessing. Your own data already knows the answer.

Pull hourly sales split by order type for the last four weeks. Dine-in in one column, delivery in another. Most reporting tools do this — here's how to read a restaurant sales report if yours looks like noise.
Mark every hour your kitchen blew past its ticket-time target. Ask the head chef; they'll name the hours from memory and they'll be right.
List every refund, remake, and "it was cold" complaint from the last month with its timestamp. Then check what to do when a delivery order is missing items — those complaints cluster in specific hours, not randomly.
Any hour that shows up on both lists — slow tickets and complaints — is a capped hour, not an open one.
Any hour with near-zero dine-in and any delivery demand at all is an open hour. Extend into it.
Any hour with under one delivery order on average is an off hour. You're paying attention costs for nothing.

Run it for two weeks after you change the schedule and compare. That's the whole method.

A schedule that works for most independents

Here's a realistic shape for a kitchen that closes at 10 p.m. Adapt the times, keep the logic.

Time block	Dining room	Delivery setting
9:00–11:00	Prep / closed	Off, unless you sell breakfast
11:00–1:30	Lunch service	On, normal prep time
1:30–5:00	Dead zone	On — your best window
5:00–6:30	Building	On, normal prep time
6:30–8:30	Peak	Capped: prep time +15 min, higher minimum, tighter radius
8:30–9:15	Winding down	On, normal prep time
9:15–10:00	Last tables	Off — cut-off has passed
10:00–12:00	Closed	Short late menu, if the math works

Weekends get their own version. Saturday 7 to 8:30 p.m. is often the one block worth closing outright. Monday, if your dining room is dark, can run delivery-only from 5 to 9 p.m. with a skeleton crew — some of the best margins on the week, because there's no front of house at all.

If you're building delivery from scratch rather than fixing it, the wider setup — radius, fees, minimums, drivers — is covered in how to set up direct delivery without third-party apps.

Delivery hours are a staffing decision

This is the part that gets skipped, and then the schedule fails and everyone blames the schedule.

Every hour delivery is on, somebody owns packing. Not "whoever's free" — a named person on the rota. On a quiet afternoon that's the same cook. At 7:30 p.m. it can't be, because that cook is on the pass.

So when you extend delivery into a new window, ask one question: who is bagging? If the answer is "we'll figure it out", you've just scheduled a bad night. Rota it properly — the weekly scheduling headache gets easier when each channel has an owner per shift.

Same for a small line. If your kitchen is already tight on space, adding a packing station at peak physically slows the pass down. Fix the flow first — small kitchen, slow service covers the layout side.

Weather, holidays, and the days to just switch it off

Delivery demand isn't flat, and neither should your hours be.

Rain and cold spike delivery and kill walk-ins. Those are days to extend the window, not hold it.
Heatwaves and perfect summer evenings do the opposite. The terrace fills and delivery goes quiet.
Big match nights and holidays are delivery gold — if, and only if, you've staffed a packer. If you haven't, close the channel and protect the room.
Your first two weeks after a menu change, keep delivery tight. New dishes get slow before they get fast.

Ten seconds on the settings screen on a Monday morning is a genuinely underrated management habit.

Tell the guest, clearly

Half of delivery complaints are expectation problems, not food problems.

Show the delivery hours on the menu page, not buried in a footer. If delivery closes at 9:15, say 9:15.
Never let someone build a cart at 9:58 and then reject it. That's the single most annoying thing an ordering page can do.
Quote the honest time. A 45-minute quote that lands in 40 beats a 25-minute quote that lands in 38, every time.
If you close a window on purpose, say why in one line. "Delivery pauses 6:30–8:30 p.m. so your food arrives hot." Guests are surprisingly fine with it.

Numbers to check every Monday

Five minutes, once a week. That's all this takes.

Delivery orders per hour, by hour. Find the hours below one and cut them.
Ticket time by channel, by hour. When delivery ticket time and dine-in ticket time both spike in the same hour, that's your cap hour.
Complaint and refund rate by hour. Cold food has a timestamp. Read it.
Contribution per delivery order by window. Peak orders often earn less than dead-zone orders once you count the damage.
Cancel rate. Anything you accepted and couldn't deliver means your hours are lying.

If you want the wider weekly set, we listed them in the KPIs worth checking every Monday morning. And if the delivery channel itself still isn't paying after all this, the honest breakdown lives in the real math of third-party delivery.

Restaurant delivery hours FAQ

Should delivery hours be the same as opening hours? No. Delivery and dine-in share a kitchen but not a demand curve. Delivery should run longer in the dead hours and after close, and pull back during the dine-in peak.

What time should a restaurant stop taking delivery orders? Thirty to 45 minutes before the kitchen closes. Use 45 if your own staff drive, 30 if you use a courier. Otherwise you're paying someone to stand around after close.

Should I turn delivery off during the dinner rush? Cap first, close last. Raise your quoted prep time, tighten the radius, and raise the minimum order. Only shut it off completely if your line genuinely can't cook for both rooms.

Is late-night delivery worth it? Often, yes. On roughly $16 contribution per order, two hours of one cook plus utilities needs about three orders to break even. Under two orders an hour, don't open.

Do delivery apps punish you for closing? Scheduled hours are fine. What hurts is cancelling orders you already accepted and quoting prep times you can't hit. Be honest in the settings and you're safe.

Can I set different delivery hours for each location? You should. A city-centre branch and a suburban one have completely different curves. Any decent menu or ordering platform lets you set delivery hours per branch, per weekday — free ones like Tabres included. If yours only does one blanket schedule, that's a real limitation.

How long should my quoted prep time be? Whatever you actually hit 80% of the time, plus a little. Then add 10 to 15 minutes for your peak block. Under-promising is a strategy; under-delivering is a review.

The dining room and the delivery channel look like one business from the outside. Inside, they're two rooms fighting over one stove. Giving them the same hours means one of them always loses — and it's usually the guest sitting in front of you, waiting on a plate that got bumped for a bag going 3 miles away.

Open the settings screen this week. Split the schedule by weekday. Pull the cut-off forward 30 minutes. Extend into your dead zone. Cap your peak instead of killing it. None of that costs anything, and by next Monday the ticket times will have already told you whether you got it right.