Somewhere right now a line cook is folding a cardboard box around someone else's dinner. The ticket says DoorDash. The food is the restaurant's. The recipe is the restaurant's. The labor is the restaurant's. The customer is not.
Two blocks away the owner is checking last week's deposit. Tuesday was the best night in three months. Sixty orders through the app on top of a full dining room. The deposit doesn't show it. Thirty percent of Tuesday left through the platform before it reached the bank account.
She knows the math. She runs it every week. She keeps running the app anyway, because turning it off feels like turning off the lights.
Twenty years ago that same restaurant would have bought a POS system for $15,000. It sat on the counter. It ran for a decade. The owner owned it. The technology company made its money on the sale and moved on.
In 2025 that restaurant pays a monthly software subscription, a payment processing fee on every transaction, a delivery commission of 15-30% on every off-premise order, and a marketing fee to appear in search results on the platform that took its customer.
The technology didn't just get better. The business model flipped. And nobody asked the restaurant if that was okay.
Three eras. Two shifts.
Three business models. Each one changed what the operator paid. Only one changed who owned the customer.
Owned a tool. Rented a service. Now pays a tax on every transaction. That cost escalation is real, and it runs through all three eras. Each step made the technology company's revenue more recurring and the operator's costs harder to escape. Toast is part of that progression. So is DoorDash.
But here's what makes the third era different from everything that came before it. The restaurant industry has been chasing convenience for decades. Drive-thrus. Phone-in pizza delivery. Frozen meals. Chili's to-go windows. Every step made food easier to get. And at every step, the customer stayed with the restaurant. The drive-thru customer was your customer. The pizza delivery customer called your number. The to-go customer picked up at your counter. Convenience evolved. Ownership didn't.
Until now. Toast replaced a $15,000 purchase with a monthly fee and a percentage of every swipe. More expensive over time, arguably. But the restaurant still owned the customer. Still owned the data. Still owned the relationship.
DoorDash didn't replace anything. It inserted itself between the restaurant and the guest, and it charges 15-30% for the position. Toast changed what the operator pays. DoorDash changed who the operator serves.
That's never happened before in the history of this industry.
The margin gap
Here is what the technology layer keeps compared to the industry it serves.
On net margin the gap is narrower than most people assume. DoorDash keeps 6.8%. Toast keeps roughly 5%. The restaurants on their platforms keep 3-5%. The distributor keeps 2.25%. The farmer loses money. Nobody is getting rich here on the bottom line.
On gross margin the gap is a different conversation: DoorDash keeps 51 cents on every dollar before operating expenses, Toast keeps 26 cents, the restaurant keeps 40-50 cents. Comparing gross margins across a software company and a business that buys perishable ingredients is like comparing a landlord's yield to a farmer's. Both are real. They measure different things.
What matters is this: the delivery platforms extract their commission before the restaurant calculates its own margin. The restaurant paying Toast still owns the customer. The restaurant paying DoorDash does not. The problem isn't the fee. It's what the fee buys. Processing buys a service. The commission buys the customer relationship, and the platform keeps it.
What a delivery order actually costs
The published commission rate is the starting point. It is not the total. Once promotions, payment processing, refunds, and marketing are included, the effective cost per order is 30-40% of revenue.
On the restaurant's own food. Cooked by the restaurant's own people. With ingredients the restaurant sourced, stored, prepped, and plated. The platform provided the customer and the driver. The restaurant provided everything else.
The line cook who packages that order knows. They trained to cook. Tonight they're folding boxes. The customer's name is on the bag but it doesn't matter. They'll order again tomorrow. From the app. Not the restaurant.
The platform owns the relationship. The customer searched by cuisine, not by name. They paid through the app. They left a rating on the app. They'll order again through the app. If the restaurant leaves the platform, the customer doesn't follow. They order from the next result in the feed.
The restaurant lost the revenue. But the customer lost something too. There is no host at the door of a delivery order. No server who remembers last Tuesday. No moment where someone says "try this, I think you'll like it." The bag on the doorstep is food without hospitality. The platform didn't just take the customer from the restaurant. It took the restaurant from the customer.
The real cost stack
Delivery commissions are the number operators know. They're not the number operators pay.
That 30-40% is the number that shows up in the weekly deposit. Not the one on the contract. Most operators don't calculate the difference until they wonder why a busy month didn't feel like a profitable one.
Why the fees are what they are
The platforms aren't charging 25-30% because they can. They're charging it because they have to.
I've sat across from operators who hear that and still feel like they're being gouged. They're not wrong to feel it. But a human in a car driving across a city for 30-45 minutes costs what it costs. Minimum wage legislation is pushing driver costs higher in every major market. The platforms spent years pricing below cost to capture market share. DoorDash wasn't profitable until 2024. The commission increases arriving now are the bill for that growth phase coming due.
And on top of the delivery economics, the platforms have built an advertising business. Uber Eats generates over $2 billion in annual advertising revenue. DoorDash sells promoted listings and sponsored search. The restaurant is now paying to be visible on the platform it's already paying 25% to be on. The advertising layer is where the real margin lives. It's growing faster than delivery itself.
What the restaurant gets
Customers it wouldn't otherwise have. That's the value proposition, and it's real. I want to be honest about that before anything else.
A restaurant on a side street with no foot traffic, no marketing budget, and no brand recognition gets placed in front of millions of people who are hungry right now. 80% of Grubhub's orders come from small and mid-size businesses. Not chains. Not brands. Restaurants that genuinely need the platform to reach delivery customers because they have no other way to find them. For those operators, the commission is the cost of access. It's not cheap. But the alternative is silence.
For a chain with brand recognition, an existing customer base, and the ability to build its own ordering channel, the same commission is a 25% tax on convenience. Those operators are paying for customers who would have found them anyway.
That's the split that matters. The lifeline and the tax are the same product at the same price. The platform doesn't distinguish between the restaurant that needs it to survive and the one that's paying out of habit. The fee is the fee. And the operators paying it can't tell which of their orders were incremental and which were cannibalized until it's too late to matter.
The growth that wasn't
Every operator who signs up believes the same thing. More customers, more revenue, the commission pays for itself. The busy months will prove it. So they watch the orders climb and wait for the deposit to match. It doesn't. Here's why.
Stanford research found that only 30 to 50 cents of every dollar spent through delivery apps represents genuinely new sales. The rest is cannibalized from dine-in and takeout, orders that would have happened anyway, now carrying a commission they didn't carry before. A typical restaurant sees delivery orders jump 40% after joining a platform, but total revenue only grows 15 to 20%. The gap is existing customers who shifted to a more expensive channel.
Think about what that means on a $2 million restaurant. Delivery might add $300,000 in gross revenue. At 25% commission, the platform takes $75,000. But half that revenue was already walking in the door at full margin. The commission on cannibalized sales alone costs $37,500 for revenue the restaurant was already earning. The net incremental profit on the genuinely new $150,000, after food cost, labor, packaging, and commission, is close to zero. In many cases it's negative.
The macro data confirms it. Between 2013 and 2019, delivery's share of total restaurant sales doubled from 4.2% to 8.4%. Inflation-adjusted industry revenue grew at roughly 1.5 to 2% annually. Delivery exploded. The industry barely moved. The platforms redistributed spending from high-margin channels to low-margin ones and called it growth. Restaurants in markets with high delivery app penetration are 23% more likely to close within two years. The platforms brought volume. They also brought a cost structure that consumed more margin than the volume replaced.
I've been on the other side of this
During my time at Sodexo, we were in deep conversations with multiple delivery partners on exactly this question. The company was looking to invest in and grow its off-premise and delivered capabilities. But in the end, it was the customer and the data that were the dealbreaker. Sodexo needed to own that. With these platforms, it simply wasn't possible.
So we built our own app. Worked with delivery companies only on last mile, not on the customer relationship. The ordering, the data, the loyalty, the brand experience: ours. The driver: theirs. That was the architecture that made sense. If a company with the scale and sophistication of Sodexo couldn't negotiate ownership of the customer relationship through a delivery platform, the 15-unit regional chain isn't going to either. The platforms aren't withholding the data out of spite. The data is the business. Giving it up would be like a landlord handing over the deed.
Something is breaking
When a CEO says publicly that his operators would make no profit on delivery orders and not a single franchisee disagrees, the market is telling you something. When the federal government calls the pricing practices deceptive, the market is telling you something louder.
The question is whether the industry responds by negotiating better rates on the same model or by building something different.
This isn't a technology story. It's an ownership story. And it's older than DoorDash.
Ownership is not an abstraction. It is the ability to greet someone by name. To remember their order. To send something to the table they didn't ask for but will love. When the platform owns the customer, the restaurant can still cook. It can package. It can fulfill. But it cannot do the thing that made most of these people open a restaurant in the first place.
The food supply chain has always concentrated margin at the intermediary layer. Distributors take their cut. Grocery chains take theirs. GPOs take theirs. DoorDash didn't invent extraction. They built a faster, stickier, harder-to-leave version of it. And they earned it. The 80% of Grubhub orders from small operators who genuinely need the platform is not a statistic you wave away.
But earning the position and earning 30% on every order forever are different things. The question is whether the people growing, making, and cooking the food can build a sustainable business when every intermediary takes theirs first. The operators who answer that question intentionally will own their future. The ones who never ask will keep renting theirs.
Who owns the customer?
Not the restaurant. Not yet.