Should your restaurant have its own ordering system or use the apps

The question gets asked as though it is one or the other. It almost never is. Every operator I know ends up with both, and the real decision is what share of your orders you are willing to rent instead of own.

Sixteen years on the operating side taught me to think about the apps the way I thought about a good patio in July. Volume you did not build, on terms you do not control, that goes away when the weather turns.

What am I actually paying the apps for?

Three things, and only one of them is delivery.

Discovery. Someone hungry opens the app with no restaurant in mind. That is a customer you did not have and would not have gotten. The commission on that order is a marketing cost, and it is a reasonable one.

Logistics. Drivers, routing, tracking, and the customer service call when the food arrives cold. Running that yourself is a real job that someone in your building will inherit.

Payment and support. Card processing, refunds, chargebacks, and the person who answers when the order is wrong.

Here is the part operators miss. You are also paying commission on customers who already knew your name, already ate in your dining room, and simply opened the app out of habit. Those orders cost you the same as the new ones and are worth far less, because you did not gain anything you did not already have.

The apps are priced as acquisition and used as convenience. That gap is where your money goes.

What do the three platforms actually charge?

Published rate cards, read off each company's own merchant pricing page. These change, so check yours before you make a decision on mine.

That last block is where the shopping gets misleading. Grubhub's "as low as 5%" headline is marketing commission only. Its own pricing page says delivery is separate and "starts at 10%." A Grubhub All-Access restaurant using Grubhub's drivers pays 20 plus 10, which is 30%, the same all-in number as DoorDash Premier and Uber Eats Premium.

All three converge at 30%. That is the number to hold in your head.

Now put it against the P&L. The National Restaurant Association's cost analysis puts pre-tax profit for a typical independent restaurant at roughly 5% of sales, which means one dollar of margin takes about twenty dollars of sales to produce. A 30% commission on a $30 order is $9. To replace $9 of margin you would need roughly $180 of additional sales.

A 30% commission is not a marketing line item. It is most of the P&L on that ticket.

Am I protected by my city's fee cap?

Probably less than you think, and if you are in the Bay Area, almost certainly not.

San Francisco's Police Code Article 53 says exactly what operators believe it says: "No third-party food delivery service may charge a covered restaurant a fee, commission, or charge per online order that totals more than 15% of the purchase price of the online order."

Then read the carve-out, effective 30 January 2023. The cap does not apply to a platform that "offers all covered restaurants the option to obtain 'core delivery service' for a total fee... that does not exceed 15%... without requiring the purchase of additional services."

In plain terms: as long as a platform offers a 15% option, it can sell you a 30% one. The cap became a floor option rather than a ceiling. All three majors still sell 30% tiers in San Francisco. Uber's own rate card confirms the mechanism from the other direction, naming San Francisco, Alameda County, Seattle, Los Angeles, Portland, Minneapolis, New York and others as markets where its Lite tier starts at 15%.

New York's version has also moved and the commonly repeated description of it is out of date. The city's Department of Consumer and Worker Protection published its current structure in April 2026: 15% to deliver the order, 5% for basic service, 20% for enhanced service, and 3% to process electronic payment, with the enhanced tier only available if the platform also offers the basic service at or under 5%. "New York caps delivery fees at 20% total" predates the enhanced service tier. The counter-claim that "New York allows 43%" is also wrong, because adding every optional tier together is an arithmetic ceiling, not a cap.

Read your own contract and your own statements rather than your city's press release. The cap you were told about may not be the rate you are paying.

When does building my own ordering start to make sense?

When the repeat share of your delivery volume gets large enough to fund the work. Here is the arithmetic, with numbers you can swap for your own.

Pull one month of app orders and sort them by customer name.

• Say you did 600 marketplace orders at an average order value of $32, taken from your own POS rather than a published figure. That is $19,200 a month, or $230,400 a year.
• Say your contracted rate is 25%. Total commission: about $57,600 a year.
• Now count the repeats. Say 40% of those orders came from people who had already eaten with you. That is $7,680 a month of sales from customers you already earned, carrying about $1,920 a month in commission, or $23,040 a year.

That last figure is your rental bill on your own guest list.

Now price the alternative honestly, because direct ordering is not free.

• Payment processing, the ordering platform, and either your own drivers or a contracted delivery service. Call it 10 cents on the dollar all in. Some platforms are cheaper: DoorDash's own white-label ordering product starts at zero commission plus processing.
• Moving the full repeat volume across is $92,160 of annual sales at 10 cents, or $9,216.
• Net available: about $13,800 a year.

That is your budget for the person who runs it, the packaging, and the Friday night when the tablet drops offline. If a fraction of a manager's time costs more than $13,800, you are not there yet. If it costs meaningfully less, you have been renting your own guest list for a while.

One more conversion, because it reframes the whole decision. At the Association's roughly 5% pre-tax margin, $13,800 of recovered commission is worth about $276,000 in additional sales. You will not get $276,000 of new business out of an ordering page. You might get the $13,800, and it is the same money.

The habit of pulling one honest number before making a decision like this is the same habit behind the short list of numbers an operator should see every morning. Most bad marketing decisions are made without either.

What actually breaks when you run your own ordering?

The things nobody demos.

The order prints in the kitchen during a rush and nobody sees it. The tablet loses connection. A customer calls about a driver and gets whoever answers the phone, which on a Friday at seven is the host who has four parties waiting.

Delivery itself is the hard part. Either you hire drivers, which is a payroll and insurance decision, or you contract a delivery service, which puts a fee back into the equation you were trying to escape.

And the ordering page has to be genuinely better than the app, because a customer with a card already saved will not fight your checkout to save you money. Slow, confusing, mobile hostile ordering pages lose orders quietly, which is one of the website problems that cost covers without ever showing up in a report.

None of that is a reason not to do it. It is a reason to plan it like an operations project rather than a marketing one.

Does running my own ordering help me show up on Google?

Somewhat, and not in the way people expect.

Google does not rank you higher for having a proprietary checkout. What it does reward is a complete, readable page about your restaurant, and an ordering system usually forces you to publish your full menu as real text with real prices. That is the actual ranking benefit, and you can get most of it without building anything.

Your Business Profile also has an ordering link field. Point it where you want the money to go. That single field decides a meaningful share of orders and most restaurants never touched it.

There is a phone dimension too. Maple's analysis of 1.2 million calls across more than 1,000 US restaurant and local business locations found 12% of restaurant calls are takeout or delivery orders and another 22% are hours, location or directions. Maple sells restaurant phone software, so treat that as vendor research with a disclosed sample. Those orders arrive at zero commission, and most restaurants make them harder to place than the app does.

Real ordering pages also tend to come with better photography, because you cannot sell a dish with no picture, and photos change how a listing performs more than most operators expect.

What to do this week

Pull one month of app orders and sort them by customer name. Count the repeats. That is your rental bill.

Open your contract and write down your actual contracted rate, not the tier you remember signing up for. Then run the arithmetic above with it.

Open your Google Business Profile and check the ordering link field. Confirm where it points and whether that is where you want it pointing.

Compare your own ordering flow to the app's on your phone, in your parking lot, at dinner time. Count the taps to complete an order both ways. If yours takes more, fix the flow before you promote it.

Then, if you decide to push direct, give the customer an actual reason. Better price, larger portion, free item at a threshold, anything visible. Building direct ordering and then giving nobody a reason to use it is the most common version of this project failing.

If you rank fine most of the day but go quiet exactly when orders come in, ordering is not your problem, and the reason a restaurant vanishes from Maps at dinner time is a different fix entirely.

Questions about ordering and commissions?

Email me at eric@seod.com with your monthly delivery order count, your contracted rate, and roughly what share you think are repeat customers. I will do the arithmetic with you and tell you which side of the line you are on, in one reply, with the number written out so you can check it.

I ran restaurants for sixteen years and I answer these myself. If the answer is that the apps are the cheaper option for you right now, that is what you will get back.

Or keep reading more on restaurant marketing.