What to Charge for a Delivery Fee: The Math

Ask in any operator group what to charge for delivery and two answers come back, both of them shrugs: "whatever the shop down the street charges" and "$3.99, because that is the number customers see on the apps." Neither answer is stupid. Delivery fees are visible, customers compare them line by line at checkout, and a fee that reads high next to the shop across the road genuinely does cost you orders. But a competitor's fee is not information about your costs. It is information about their nerve. This piece names the four inputs inside a delivery run, walks the math once with every figure illustrative and modeled, shows why your fee never has to cover the whole run, and separates the fee you charge the customer from the reimbursement you owe the driver, because those are two different obligations with two different sets of rules.

Start from cost to serve, not from the shop down the street

A delivery fee is not a price. It is a partial recovery of a cost you have already committed to the moment a driver pulls out of the lot. That cost is knowable to the dollar for your shop, and it is made of four things: the miles, the minutes, the insurance and vehicle program, and the packaging. Everything else in the pricing conversation, what the neighbors charge, what the marketplaces display, what your regulars will tolerate before they switch to pickup, is a constraint on the answer. None of it is the answer.

There is a practical reason to start with your own number. A good share of the shops setting fees by comparison are copying a competitor who is underwater on delivery and does not know it yet, because the loss shows up as thin overall margin rather than as a line called delivery. Copy the fee and you inherit the loss. Compute cost to serve first, then decide, deliberately, how much of it the fee should carry and how much your food margin should absorb.

Name the four inputs: miles, minutes, insurance, and packaging

• Miles, priced at a rate you can defend. The IRS standard mileage rate is $0.76 per mile, effective July 1 2026, and it is the simplest per-mile number to stand behind because you did not invent it. Pull the round-trip average from dispatch data, not from the radius drawn on your marketing map, and count the drive back: the return leg is half your miles and it earns nothing. Background in our guide to mileage reimbursement for pizza delivery and the July 2026 rate increase.
• Minutes, fully loaded. Road time, out the door to back behind the counter, multiplied by the loaded hourly cost of the driver: wage plus payroll taxes plus whatever else rides on the hour. In-shop time is not a delivery cost; that driver is folding boxes. Stacking two orders on one run is the single biggest lever on this input, and it is free.
• Insurance and the vehicle program. The coverage is hired and non-owned auto (HNOA): when your employee drives their own car on your errand and injures someone, that is the policy that responds, because personal auto policies routinely exclude delivery use. Add motor vehicle record checks and mileage tracking, then allocate the monthly total across monthly deliveries. See what HNOA actually costs and MVR checks for delivery drivers.
• Packaging and last-mile consumables. Corrugated, bags, sleeves, sauce cups, and the heat bags you replace twice a year. Pennies per order, real money per month, and the only input on this list that scales cleanly with order count.

The math, walked once: five miles, twelve minutes

Here is the model, with every figure illustrative and modeled. Take a shop averaging 5 miles round trip per delivery and 12 minutes of road time, running about 1,000 deliveries a month. Miles: 5 at $0.76 is $3.80. Minutes: 12 at a loaded $17 an hour is $3.40. Insurance, tracking and MVR checks allocated across those 1,000 runs: $0.35. Packaging: $0.20. Cost to serve, all figures illustrative and modeled: $7.75 per delivery.

Notice what did and did not matter. Insurance, the cost operators fear most before they price anything, is $0.35 of a $7.75 run, under 5 percent of the total. The two inputs that decide your number are miles and minutes, and both are operational rather than financial. A radius trimmed from 6 miles to 4, or a stacking rate that moves from 1.1 orders per run to 1.5, moves the fee conversation further than any carrier or any menu price ever will. Before you raise a fee, look at whether you can shorten the run.

The fee does not have to cover the whole run

Set a fee of $4.95, illustrative, against that $7.75 modeled run and it covers about 64 percent. The uncovered part is $2.80, modeled. That sounds like a leak until you remember the order also carries food margin: on a $34 average ticket, illustrative, $2.80 is roughly 8 percent of the order, and a delivery ticket typically runs larger than a walk-in one. The fee is not supposed to make delivery free. It is supposed to keep the uncovered share small enough that your food margin swallows it without noticing.

Hold that number up against the alternative. A marketplace taking 28 percent of the same $34 ticket costs $9.52 on that single order, illustrative, and it scales with every dollar of growth you generate; the full anatomy is in our third-party app cost breakdown, and the crossover volume is what our break-even calculator computes. A $2.80 gap you control beats a $9.52 commission you do not.

So state the rule as a share, never as a dollar. If the uncovered gap stays under roughly 10 percent of your average ticket, leave the fee alone; it is doing its job. Between 10 and 15 percent, the first moves are an order minimum and a tighter radius, not a higher fee. Past 15 percent, the problem is the run itself: too far, too slow, or too few orders per trip. Our delivery cost index tracks how those inputs have moved.

Three ways to price it, and when each one fits

1. One flat fee across the whole radius

Simplest to explain, simplest to POS, and it cross-subsidizes: the two-mile orders pay for the six-mile ones. It fits a tight, dense radius where the miles do not vary much. It fails the moment your longest run is triple your shortest, because you are quietly paying customers to live far away.

2. Banded by distance

Two or three bands, priced from actual round-trip miles at $0.76 per mile plus the minutes each band really takes, all figures illustrative. Customers accept distance bands more readily than operators expect, because the logic is obvious. The cost is menu complexity and a POS that has to know which band an address falls in.

3. Folded into menu prices, with no fee at all

Some shops raise delivery-menu prices a few points and advertise free delivery. It converts well and it hides the cost from the checkout comparison. It also charges your pickup customers nothing extra only if you run separate price books, and it makes your cost-to-serve number invisible to you, which is exactly how shops end up underwater. If you go this way, keep computing the run cost internally every quarter even though no customer ever sees it.

Sandwich shops carry the harder ratio: a $14 ticket cannot absorb a $7 run

Cost to serve barely moves with ticket size. The same 5 miles, the same 12 minutes, the same insurance allocation apply whether the bag holds two large pizzas or one turkey club. That is why the pizza math above does not transfer to a sandwich shop, a bakery, or a salad concept: on a $14 average ticket, illustrative, a $7.75 modeled run consumes essentially the entire gross margin on the order, and no fee a lunch customer will pay closes that by itself.

The fixes are structural rather than pricing. An order minimum that pulls the average delivery ticket up toward $25, illustrative, changes the ratio more than any fee change. A radius half the size of your pizza neighbors' is correct, not timid. Batching the lunch wave so a driver leaves with three or four bags is the difference between a viable channel and a charity. And group and office orders, where one stop carries eight sandwiches, are the only delivery your economics genuinely love: price them to win.

The fee is not the reimbursement: keep the two lines apart

Charging the customer $4.95, illustrative, does not discharge anything you owe your driver. The fee is revenue to the business. Driver reimbursement is an expense of the business, and under federal wage law an under-reimbursed vehicle expense can be treated as a kickback that pulls effective pay below the minimum wage, which is the theory behind most delivery wage suits; our explainer on the FLSA kickback theory walks it. In Parker v. Battle Creek Pizza, Inc. (6th Cir. 2024), the court rejected both bright lines, the IRS rate as an automatic floor and a loose reasonable approximation as an automatic safe harbor, and pointed employers back toward actual vehicle costs. Parker is binding only in the Sixth Circuit, meaning Michigan, Ohio, Kentucky and Tennessee, and persuasive everywhere else. The same is true of Bradford. West v. BAM! Pizza Management (D.N.M., January 2026) shows the same fight continuing outside that circuit.

State law can raise the floor on top of that. California Labor Code section 2802 requires indemnification of necessary business expenses and you cannot contract around it, whatever your delivery fee says; see our California guide and the state-by-state rates. Two practical rules follow. First, never let the customer-facing fee become your reimbursement policy; set the driver rate from mileage records, as covered in how to pay delivery drivers legally. Second, say plainly on the menu and at checkout that the delivery fee is not a gratuity and is not paid to the driver, because customers assume otherwise and that assumption creates its own problems. Then run a periodic driver reimbursement compliance audit so the per-run records exist before anyone asks for them.

RatesReady logs each delivery run, computes per-driver reimbursement at the rate you set, and keeps the per-run mileage records that make both your cost-to-serve number and your reimbursement defensible, so the fee you charge and the rate you pay stop being guesses. It is $49 per location per month. If you want to see it against your own dispatch data, request a demo.

This article is general information for restaurant operators and is not legal advice or tax advice. Wage, expense and pricing rules vary by state and by situation; consult a qualified attorney or accountant about your own shops.