The Real Cost of Third-Party Delivery: A Year-Long Breakdown for a Pizzeria
Anytable · August 27, 2026 · 8 min read
Most pizzeria owners know the commission on delivery apps is high. Fewer have sat down and added up the real cost of third-party delivery over a full year, including the ad fees, the tips that never reach their drivers, and the fees customers pay that the restaurant never sees. So let’s do it with a hypothetical pizzeria and round numbers, then compare it to running the same orders through your own ordering page.
Meet Tony’s (a made-up pizzeria)
Let’s call it Tony’s. It’s a neighborhood pizzeria with a dining room, a counter, and two part-time drivers. Tony’s is on one major marketplace app and does a steady 500 delivery orders a month through it. The average order, before the app’s fees to the customer, is $30. That’s 6,000 orders and $180,000 in marketplace sales a year.
Tony’s is fictional, and every number below is a round figure chosen to make the math easy to follow. Your own numbers will differ. The point is the structure of the costs, not the exact totals.
What the marketplace actually takes
The headline number is the commission. Third-party marketplaces typically charge 15–30% per order, depending on the plan and whether the app supplies the driver. Tony’s is on a plan where the app handles delivery, so let’s use 25%.
Commission: 25% of $30 = $7.50 per order. Over the year, that’s $45,000.
But commission is rarely the whole bill. Tony’s also learned, like most operators, that being on the app doesn’t mean being seen on the app. To stay near the top of the list, Tony’s runs sponsored placement and the occasional “free delivery” promotion, where the restaurant covers the cost. Call it an average of $500 a month in marketing and promotion fees, or $6,000 a year. Some months it’s zero; some months, when a competitor opens nearby, it’s more.
Then there are the costs that don’t show up on Tony’s statement at all because they never pass through Tony’s hands:
- Tips. Customers tip an average of $4 per order. That goes to the app’s driver, which is fair, since the app’s driver did the work. But it means $24,000 a year in tips flows through the transaction without touching Tony’s business or Tony’s own drivers.
- Customer-side fees. The app charges the customer a service fee and a delivery fee on top of the $30. Tony’s doesn’t see that money, but customers do, and they mentally attach it to Tony’s prices.
Tony’s own drivers, meanwhile, sit idle on marketplace orders. That’s a cost too, in a sense, because Tony’s is paying for delivery capacity and then paying the app for delivery capacity again.
What the same orders cost through direct ordering
Now imagine the same 6,000 orders came through Tony’s own branded ordering page, with Tony’s own drivers delivering. Here’s the cost structure on that side.
Software. A commission-free platform charges a flat monthly fee instead of a percentage. Anytable, for example, is $29.99 a month with 0% commission, which works out to about $360 a year. The exact figure matters less than the shape of it: flat, not proportional to sales.
Card processing. This is unavoidable on any online order. Using a common structure of about 3% plus $0.30 per transaction on a $30 order, that’s $1.20 per order, or $7,200 a year. On the marketplace, processing is buried inside the commission, so this looks like a new cost even though Tony’s was paying it before.
Delivery labor. Tony’s has to deliver these orders itself. Say a driver averages three deliveries an hour at $15 an hour plus $1 per trip in mileage reimbursement. That’s roughly $6 per order, or $36,000 a year. This is the biggest line on the direct side, and it’s the one owners most often forget when they compare the two models.
Tips. On direct orders, 100% of the $4 average tip goes to Tony’s drivers, since the money lands in Tony’s account and gets paid out. That’s $24,000 a year that now supports Tony’s own team instead of the app’s. It doesn’t reduce Tony’s costs directly, but it means drivers earn about $10 per delivery all-in, which makes the job easier to staff and lets Tony’s hold base pay steady.
Delivery fee. Customers are used to paying a delivery fee. Tony’s charges $3 per direct delivery order, which is less than most apps charge the customer. That’s $18,000 a year that goes to Tony’s rather than to the marketplace.
The year-long comparison
Here’s the full picture for 6,000 delivery orders at $30 each.
| Line item | Marketplace (25% commission) | Direct ordering (flat fee) |
|---|---|---|
| Orders per year | 6,000 | 6,000 |
| Average order | $30 | $30 |
| Gross food sales | $180,000 | $180,000 |
| Commission (25%) | −$45,000 | $0 |
| Marketing and promo fees | −$6,000 | $0 |
| Ordering software | $0 | −$360 |
| Card processing (~3% + $0.30) | Included in commission | −$7,200 |
| Delivery labor (~$6/order) | $0 (app’s drivers) | −$36,000 |
| Customer delivery fee ($3) | $0 (kept by app) | +$18,000 |
| Driver tips ($4 avg) | $24,000 to app’s drivers | $24,000 to Tony’s drivers |
| Net to Tony’s | $129,000 | $154,440 |
| Net per order | $21.50 | $25.74 |
The difference is about $25,000 a year, or a little over $4 per order, after paying Tony’s own drivers. On a pizzeria that might net 5% to 8% on total sales, $25,000 could be a large share of the year’s profit.
The pickup version
The table above assumes every order is delivered. But a big share of online pizza orders are pickup, and pickup changes the math dramatically. On a marketplace, Tony’s still pays a commission on pickup orders (often a lower rate, but not zero). On the direct side, there’s no delivery labor at all.
If half of Tony’s 6,000 orders were pickup, the direct side would save roughly $18,000 in driver costs and give up $9,000 in delivery fees, pushing the gap between the two models closer to $34,000 a year.
The realistic version
Nobody moves 100% of their orders overnight, and you probably shouldn’t try. A more honest scenario for year one: Tony’s keeps the marketplace listing, promotes its own ordering link on every box, receipt, and social profile, and shifts about half its regulars to direct within twelve months.
At 3,000 orders moved, the savings are around $12,500 in the first year, growing as more customers switch. And there’s a second-order effect that doesn’t fit in a table: direct orders come with the customer’s name, email, and order history, which Tony’s can use to bring them back. Marketplace orders come with none of that.
Month by month, what changes
Annual numbers can feel abstract, so here’s the same shift viewed as it would actually happen at Tony’s.
Months 1 to 3. Tony’s launches a branded ordering page, puts a QR code and a “order direct and save” message on every pizza box and menu, and adds the link to its Google Business Profile. Marketplace orders barely move; maybe 5% of regulars try the direct link. Savings are a few hundred dollars a month.
Months 4 to 6. The habit builds. Regulars who tried direct once keep using it because it’s saved on their phone and the total is lower without app fees. Tony’s is now at 20% to 25% of delivery volume direct, and the savings run around $1,000 a month.
Months 7 to 12. Tony’s starts emailing its direct customers a monthly special and turns on automatic review requests after each order. Direct share climbs toward half. Marketplace spend on promotions drops because Tony’s no longer depends on the app for visibility. By year end, monthly savings are around $2,000, and the customer list has grown to a few thousand names.
Platforms built for this, Anytable among them, handle the operational pieces: delivery zones and minimums, scheduled orders, real-time alerts to a staff app, and reporting on order counts and average order value so Tony’s can watch the shift happen.
The balanced conclusion
None of this means Tony’s should delete the marketplace app. Marketplaces are still useful for discovery. A new customer scrolling for pizza on a Friday night may never have heard of Tony’s, and the app puts it in front of them. Paying 25% on that first order is a customer acquisition cost, and as acquisition costs go, it’s not a terrible one.
The problem is paying 25% on the fiftieth order from the same customer. That’s not acquisition; that’s rent on a relationship you already built. The goal is not to be on one channel or the other, but to use the marketplace to meet new people and use your own ordering to keep them.
Where to start
If you want to run this math for your own restaurant, you need four numbers: monthly marketplace orders, average order value, your commission rate, and your monthly marketing and promo spend on the app. Multiply it out for a year. Then estimate what it would cost you to deliver those orders yourself, or how many are pickup and cost nothing extra. The gap is what you’re paying for convenience, and once you can see it, you can decide how much of it you’re willing to keep paying.
