The Hidden Cost of Third-Party Delivery Apps: What a Pizza Shop Really Pays

Somewhere in the last decade, a generation of pizza operators made a trade that looked rational on the surface: shut down in-house delivery, hand the last mile to DoorDash, Uber Eats, or Grubhub, and turn a messy driver operation into a clean percentage. No hiring, no scheduling, no vehicles, no reimbursement compliance. One line on the statement.

That line is the problem. The commission is the only cost the marketplace itemizes, and it is not the largest one. The full price of the trade includes fees on customers you already owned, an advertising toll to stay visible on your own orders, a menu markup that quietly suppresses demand, and, most expensively, the customer relationship itself, which the platform keeps. This article walks through the whole stack, then gives you a calculator to run your own shop's numbers against a compliant in-house program.

The visible stack: what the statement shows

Start with what you can see. Marketplace delivery commissions run 15 to 30 percent of the order depending on tier, and the lower tiers buy less: reduced delivery area, worse placement, fewer promotions. Most delivery-dependent shops end up at or near the top tier, because delivery radius is the product.

Then the pay-to-play layer. Organic placement in a crowded category is effectively zero, so shops buy sponsored listings and in-app promotions that commonly add another 5 to 15 percent of order value across the channel. This is the quiet escalation of the last few years: the commission held still while visibility became a separate purchase.

Then the friction charges: order-error adjustments and refunds that default against the merchant, tablet and integration fees, and the processing spread. Individually small, collectively another point or two.

On a $35 pizza ticket at a 28 percent effective commission plus 5 percent marketing, the visible stack alone is $11.55. That is the number to hold while we look at what the statement leaves out.

The invisible stack: what never shows up

1. You pay full commission on customers you already had

The marketplace pitch is incremental demand: new customers you would never have reached. Some orders are exactly that. But for an established pizza shop, a large share of app volume, operators who audit it typically find 30 to 50 percent, comes from existing customers who simply migrated to the app because it was installed and convenient. On those orders the platform delivered nothing incremental; it inserted itself into a relationship you built and now taxes it at 30 percent, forever. In the calculator below, this line alone is usually five figures a year.

2. The platform keeps your customer

Every first-party order gives you a name, a phone number, an address, and an order history: the raw material of reorders, win-back texts, and the regulars list every strong shop runs on. Every marketplace order gives you a ticket. The platform keeps the identity, and it uses that asset the way any owner would: to route the customer's next hungry moment wherever the platform earns most. Which leads directly to the next item.

3. The reorder screen ranks your competitors beside you

The moment your customer reopens the app to order from you again, they are standing in a marketplace where your competitors can outbid you for the top of that screen. You built the craving; the platform auctions it. In-house, the reorder path is your phone number and your website. On the marketplace, the reorder path is a search results page you have to pay to win, for a customer who was already yours.

4. The markup that fights your own demand

Most shops recover part of the commission by marking up app menu prices 10 to 20 percent. The recovery is real and the calculator credits it. But a $35 pizza at $41 changes ordering behavior: frequency drops, attach items fall off, and price-sensitive regulars drift. You are, in effect, running a permanent price increase on your most convenient channel and paying the demand cost of it, while the app's side-by-side layout shows the customer exactly who did not mark up.

5. Your name eats the delivery experience

The marketplace driver carrying four other orders, no hot bag, and no relationship with your shop delivers a 48-minute pizza, and the one-star review lands on your listing, not the platform's. Delivery quality is part of the product for pizza in a way it is not for most categories, and outsourcing it means the most reviewable part of your product is performed by someone you cannot train, schedule, or keep.

What in-house actually costs, honestly

The in-house side deserves the same unsentimental accounting, because operators who romanticize it get surprised. A compliant in-house delivery program costs, per delivery:

• Labor share: the driver's hourly cost to you, cash wage plus payroll taxes, divided by deliveries per hour. A tipped-wage driver at roughly $9 fully loaded running 2.5 deliveries an hour is $3.60 per run. Pizza's tight zones and stacked runs are exactly why this number works; it is the category's structural advantage.
• Vehicle reimbursement: the recorded round-trip miles times a documented per-mile rate. Five miles at a documented $0.48 is $2.40. This is the line the law cares about, and the one most operators historically got wrong in both directions: flat fees that underpay drivers, or the $0.76 IRS habit that overpays the real cost by $0.25 or more per mile.
• Overhead: hired and non-owned auto coverage, hot bags, delivery tech. Typically $0.30 to $0.75 per delivery at any real volume.

That stack lands near $5.40 to $6.50 per delivery for a typical shop, against $9 to $12 all-in on the marketplace for the same ticket, and the in-house delivery ends with your customer's number in your system and your driver at the door in your shirt. Tips are excluded on both sides: the customer pays them either way, though who they retain is not neutral, since tips are the reason experienced in-house drivers stay.

Run your shop's numbers

Considering the move in the other direction? The companion in-house delivery break-even calculator models wages, mileage, and HNOA insurance against commissions, and this article walks the insurance question in full.

Two things worth noticing as you move the sliders. First, the cannibalization percentage barely changes the per-delivery math but dominates the strategic picture: it is the share of your fees buying nothing. Second, the reimbursement rate matters more than it looks. The in-house advantage is built partly on paying the documented cost of the mile rather than the $0.76 IRS default; at IRS-habit rates, a meaningful slice of the annual gap hands itself back.

The caveat that makes or breaks the whole comparison

In-house delivery is cheaper only when the reimbursement layer is done right. The wave of shops that fled to the marketplaces in the first place were often fleeing exactly this: drivers on flat per-run fees, no mileage records, and a wage-law exposure nobody had priced. That exposure has only grown. Federal appellate courts have rejected the "we pay the IRS rate" defense, flat-rate reimbursement lost as a "reasonable approximation" in early 2026, and the recordkeeping burden now sits with the employer. An in-house program built on guesswork does not beat the marketplace; it trades a visible 30 percent for an invisible contingent liability with a multi-year lookback.

The switch-back playbook

None of this argues for deleting the apps tomorrow. The operators doing this well run a hybrid, deliberately:

1. Keep the marketplaces as paid discovery. Treat commission on genuinely new customers as advertising, because that is what it is, and it can be efficient advertising at that.
2. Move the reorder in-house. Box toppers, a first-party ordering discount that undercuts the app markup, a "text us next time" card with the driver. Every migrated regular is a permanent 30-point margin repair on their lifetime orders.
3. Rebuild delivery on documented rails from day one. Dispatch-recorded miles, a documented per-mile rate for each store's ZIP and vehicle class, clean tipped-wage mechanics, and the methodology file that answers an auditor. Per our complete pay guide, this is a structure, not a burden, and it is what makes the in-house number in the calculator real rather than aspirational.
4. Re-run the calculator quarterly. Commission tiers, ad costs, and your documented mile all move. The trade should be re-priced, not re-guessed.

RatesReady maintains the documented layer that makes in-house the winning side of this math: ZIP-code-level reimbursement rates across 20 vehicle classes, refreshed monthly, audit trail attached, from $49 per location per month. Request a demo and we will run your ZIPs against your current structure live.

This article and calculator are illustrative economic models for general information, not legal, tax, or financial advice. Commission structures, marketplace fees, and platform terms vary by provider, market, and contract; wage and reimbursement obligations vary by state. Verify your own agreements and consult qualified counsel about your specific obligations.