Most advice about online ordering is written for a pickup counter. Take the order, fire the ticket, call the name. When the same order has to travel four miles in a car, the ordering screen stops being a menu and starts being the first step of a delivery. Everything it fails to collect turns into a phone call, a wrong door, or a refund later that night.
Online ordering for businesses that deliver has a bigger job than a takeout form. It has to capture a deliverable address, commit to a window you can actually hit, price the trip, and hand the kitchen a ticket that knows it’s going out on a route rather than sitting under a heat lamp.
Done properly it’s also the biggest sales change available to a small food business. Digital orders run bigger than counter orders, they don’t get rushed by a queue behind them, and the ones placed on your own site cost a fraction of the ones placed on a marketplace.
The Bottom Line
- Direct beats marketplace on economics by a wide margin. Third-party commissions run 15% to 30%, and Rezku’s 2026 breakdown puts the typical all-in cost at 30% to 40% of the order once processing and marketing stack up.
- Customers mostly want to order direct anyway. Restolabs reports 67% of consumers prefer a restaurant’s own site or app, and that restaurant-owned channels carry 62% of digital orders.
- Digital orders are bigger orders. The same research puts average digital order value 23% higher than in-person transactions.
- An order isn’t deliverable until the form collects a unit number, a phone number, a window and a drop instruction. Everything you don’t ask for, a driver has to chase.
Save 80% of delivery management time
We handle everything:
- Dedicated operations manager
- Real-time tracking dashboard
- Automated customer notifications
- Urgent issue resolution
What online ordering has to do differently when you deliver the order yourself
A pickup order needs three things: what, when, and who’s collecting. A delivery order needs those plus a whole second layer that only matters because a vehicle is involved.
The extra layer is where small operations lose money:
- A precise destination, not an approximate one. “412 Main St” and “412 Main St, Apt 3C, buzzer 14” are the difference between a two-minute stop and a twelve-minute one.
- A promise you can keep. A window, not a vague “as soon as possible,” because the window is what you’ll be judged against.
- Trip pricing. Distance, zone and minimum order, all decided at checkout rather than after.
- Routing-ready data. Enough structure that the order can be grouped with the other three going the same direction.
That last one connects straight to your unit economics. Batched stops cost dramatically less per drop than scattered ones, which is why the ordering screen controls whether your delivery orders make money at all. A form that lets anyone order anything to anywhere at any time produces routes nobody can run profitably.
Direct online ordering vs third-party apps: what each really costs
Both channels have a place. Marketplaces buy you discovery; your own channel buys you margin and the customer relationship. The mistake is treating them as interchangeable when the math is nowhere near it.
| Your own online ordering | Third-party marketplace | |
|---|---|---|
| Cost per order | Payment processing, roughly 2.5–3% plus a fixed fee, and a platform subscription | Commission of 15–30% before processing; Rezku puts the typical effective cost at 30–40% |
| Who delivers | You, your drivers, or a courier partner you choose | The platform’s driver network |
| Customer data | Yours: name, address, order history, contact | Largely the platform’s |
| Discovery | You have to drive the traffic | Built-in marketplace demand |
| Menu and pricing control | Full | Constrained, and often price-inflated to absorb commission |
| Repeat purchase | You can market to the customer directly | You generally can’t |
On a $50 order at 30% all-in, the platform takes $15. Against the 3% to 5% net margin an independent food business typically runs, that single number decides whether a delivery order is profitable or simply busy.
The sensible posture for most operators: stay on one marketplace for discovery, treat it as a paid acquisition channel, and work steadily to move repeat customers onto your own ordering page.
The fields every delivery order needs before it’s deliverable
Here’s the checklist. Every field on it exists because leaving it off causes a specific, repeatable failure.
Required:
- Street address, with unit or apartment as its own required field, rather than an optional second line people tab past
- Mobile number the driver can text
- Requested delivery window, or an explicit ASAP with a realistic estimate attached
- Drop-off preference: hand it over, or leave it at the door
Strongly recommended:
- Gate, buzzer or access code, shown only for addresses in buildings that need one
- Free-text drop-spot note (“side door under the awning”)
- Contact preference for when the driver arrives
That last cluster is what makes a leave-at-door handoff work, which is why setting up no contact delivery properly starts in your order form rather than in your driver handbook. A driver who was never given a unit number cannot place an order at the right door, and no amount of policy fixes that after the fact.
One design note: validate the address as it’s typed. Autocomplete against a mapping service catches the typo, the wrong town with the same street name, and the address outside your zone, all before payment goes through. That is the only point where fixing it is free.
How to set a delivery zone and a minimum order at checkout
Zones and minimums are pricing decisions, and the ordering system is where they get enforced. Left to a human answering the phone, they get waived.
Set the zone by drive time rather than radius. A five-mile circle on a map crosses a river, a highway with no left turn, and a neighborhood where parking takes eight minutes. Fifteen or twenty minutes of actual drive time is a more honest boundary, and most ordering platforms support drawing a polygon rather than a circle.
Tier the fee if you deliver a wide area. A flat fee across a zone that spans two to twelve minutes of driving subsidizes the far customers with the near ones. Two or three bands are usually enough.
Set a minimum that covers the trip. Work back from what a stop costs you: driver time, mileage at something like the IRS business rate, which reached 76 cents per mile for the second half of 2026, packaging, and the share the payment processor takes. Then set the minimum above it.
Show the shortfall instead of a rejection. “Add $8 to qualify for delivery” converts. “Minimum order not met” abandons.
Close ordering when you can’t deliver. Auto-disable delivery when the kitchen is beyond a set ticket count or the last driver’s route is full. A rejected order annoys someone once; a late order annoys them permanently.
How online ordering increases average order value
The sales case for direct ordering isn’t only about avoiding commission. The channel sells more.
Restolabs reports that average digital order value runs 23% higher than in-person transactions, and that customers who order online visit 67% more often than those who don’t. The mechanics behind that are not mysterious:
- Nobody is waiting behind you. A customer browsing a screen adds the dessert they’d never ask for with a queue at their back.
- Modifiers are visible. Add-ons that a cashier forgets to mention are printed on the screen every time.
- The cart does the arithmetic. A free-delivery threshold or a minimum makes the upsell the customer’s own idea.
- Reordering is one tap. Saved addresses and order history turn a considered purchase into a habit.
Two specific levers to set up on day one: a threshold that unlocks free or reduced delivery just above your current average ticket, and a single well-chosen add-on prompt at checkout rather than a wall of suggestions. More prompts don’t sell more; one relevant one does.
Where online ordering saves time in the kitchen
The efficiency half of the case is quieter but it compounds.
Transcription errors disappear. Nobody mishears “no onions” on a screen. Phone-order mistakes are remakes, and a remake on a delivery order is both a wasted plate and a blown window.
Tickets route themselves. Delivery orders can print to a different station, with the delivery time rather than the order time, so the kitchen fires against the route instead of against the clock.
Order pacing becomes possible. Throttling, which caps orders per fifteen-minute slot, is the single most useful feature for a small kitchen that also delivers, and it only exists in a digital channel.
Prep offsets get automatic. If the drive averages eighteen minutes, the system can fire a 7:00 delivery at 6:30 rather than relying on someone doing the subtraction during a rush.
And the phone stops ringing. That sounds like a soft benefit until you count how many of those calls were “where is my order,” a question a tracking link answers for free.
Connecting online ordering to your POS and your route
An ordering system that doesn’t talk to anything creates a second set of books and a tablet nobody watches.
Two integrations matter. POS integration keeps one menu, one price list and one sales report, so an 86’d item disappears from the website the moment it’s marked out of stock. Without it, someone updates two menus and eventually doesn’t.
Route or dispatch integration matters as soon as you have more than a handful of deliveries a day. What you want is for a paid order to land in whatever plans your stops, carrying its address, window and drop notes, without anyone retyping it. Retyping is where unit numbers get lost.
If you’re evaluating systems, the questions that separate them for a delivery operation are narrow. Does it support delivery zones by drive time? Can it enforce different minimums per zone? Does it throttle orders? Does it pass structured address fields, including unit, to whatever you route with? Does it own the customer data, or do you? A system that nails those five is worth more than one with a longer feature list that misses two of them.
How to move customers off the marketplace and onto your own ordering page
Assume marketplace customers are borrowable, not owned. The playbook is unglamorous and it works:
- Put your own ordering URL on every package. A sticker on the bag of an order the marketplace just delivered is the cheapest acquisition you’ll ever do.
- Include an insert with a reason to switch, a discount that’s still cheaper for you than the commission you just paid.
- Keep menu prices at or below marketplace prices on your own site. Customers do check. Inflating your own channel to match a commission-padded marketplace price is self-defeating.
- Make the delivery promise clearer than theirs. A firm window and a text at drop-off beat a moving ETA.
- Ask for the email at checkout, once, and actually use it for something worth opening.
Frequently asked questions
Is a commission-free online ordering system really free? No, it moves the cost rather than removing it. You’ll typically pay a monthly subscription plus card processing of roughly 2.5% to 3% plus a fixed fee per transaction. That’s still far below a 15% to 30% marketplace commission, but budget for it as a line item.
Should I stop using third-party apps entirely? Usually not at first. Treat the marketplace as paid discovery, measure what a customer acquired there is worth once they reorder direct, and shrink your dependence deliberately rather than cutting it off before your own channel carries volume.
What’s a reasonable delivery minimum? Whatever covers driver time, mileage, packaging and processing with margin left over, which for most local food businesses lands around two to three times the cost of the stop. Calculate it from your own numbers rather than copying a competitor’s.
Do I need a separate app? Almost never. A mobile-friendly ordering page on your own domain does the same job without asking a customer to install anything. Apps make sense at a volume where you’re funding retention, not acquisition.
The order form is the front end of your delivery operation
Treat online ordering as a sales channel and you’ll optimize the menu photos. Treat it as the first step of a delivery and you’ll optimize the address fields, the zone boundary and the throttle, which is where the money is.
The sequence that works: get the required fields right so orders are deliverable, set zones and minimums so the trips pay, then turn on the threshold and add-on prompts that lift the ticket. Discovery through a marketplace is fine while you build that. Just don’t rent your best customers forever when they’d rather order from you directly.
