What In-House Delivery Actually Costs: Insurance, Reimbursement, and the Break-Even Math
Spend an hour in any operator forum and you will find the same sentence, in nearly the same words: "I want to bring delivery in-house, but the insurance is going to kill me." It is a fair fear. Commercial auto exposure is real, the coverage has a name most operators have never had to learn, and the marketplaces have spent a decade telling you that handing them 28 percent of every order is the safe choice. This article takes the fear seriously, names the coverage, prices the comparison honestly, and shows you the break-even math, because the answer to "is insurance expensive" is the wrong question. The right question is: at what order volume does in-house beat the apps? And that question has a number, which our break-even calculator computes for your shop.
Name the coverage: hired and non-owned auto
The policy behind the fear is hired and non-owned auto (HNOA) coverage. In plain operator language: when your employee drives their own car to deliver your pizza and injures someone, the injured party's lawyer sues everyone, and your business's liability in that claim is what HNOA responds to. It matters because personal auto policies routinely exclude delivery use: your driver's own insurer can deny the claim the moment delivery driving is established, which leaves the business standing behind the accident with its general assets. That is why carriers writing restaurants expect HNOA the moment in-house delivery appears on an application, alongside workers' compensation rated for driving exposure and verification that each driver carries their own valid coverage.
What does it cost? Honestly: it varies too much for any number here to be a quote. Premiums move with your market, delivery volume, driver count, and claims history, and small-operation planning ranges run from a few hundred to a few thousand dollars per year per location, illustrative only. Get a broker quote; it takes a week. The structural point does not vary: HNOA is a fixed, quotable monthly line item. Which brings us to the reframe.
The reframe: a line item vs a tax on growth
Marketplace commissions are not a line item. They are 25 to 30 percent of every delivery order, forever, scaling with every dollar of growth you generate; the full anatomy, commissions, required visibility spend, error charges, and the customer data you never receive, is in our third-party app cost breakdown. HNOA, by contrast, costs the same in a $40,000 month as in a $25,000 month. One cost is flat; the other slopes upward with your success. Any time a business can trade a percentage-of-revenue cost for a fixed cost, there is a crossover volume where the trade wins, and everything above that volume is margin you were previously donating. The only work is finding your crossover.
The break-even math, walked once
Here is the model, with every figure illustrative and modeled. Take a shop doing 35 delivery orders a day, 30 days a month (1,050 orders), at a $35 average ticket, currently on a marketplace at a 28 percent effective commission. The apps' monthly take: 1,050 x $35 x 28% = $10,290.
Now the in-house side, built honestly. Drivers at a $12 wage plus 12 percent payroll burden running 2.5 deliveries per driver-hour: 420 driver-hours costing $5,645. Mileage at 5 round-trip miles per delivery reimbursed at the $0.76 IRS rate (the July 1, 2026 figure, and hold that thought): $3,990. An illustrative $300 monthly HNOA premium. Total: $9,935. In-house already wins by about $355 a month, per order $9.46 against $9.80, and the break-even, the volume where in-house first beats the apps with insurance included, lands around 481 orders a month, roughly 16 a day.
Notice what did and did not matter. The scary fixed cost, insurance, was cleared by a mid-volume shop with room to spare. The line quietly dominating the in-house side was mileage at the IRS rate: $3,990 of the $9,935, reimbursing a national tax ceiling rather than your market's actual cost. Rerun the same shop at an illustrative documented ZIP-level rate of $0.48: mileage falls to $2,520, monthly savings jump to $1,825, and break-even collapses to 149 orders a month, about 5 a day. The reimbursement rate, not the insurance premium, is the swing vote in this entire comparison. Run your own numbers in the calculator; it models exactly this math with your inputs.
The piece operators skip: reimbursement is what gets litigated
If the math convinces you to hire drivers, understand which part of the operation carries the legal weight, because it is not the insurance. It is the per-mile reimbursement, and the last two years rewrote its rules. Parker v. Battle Creek Pizza (6th Cir. 2024) rejected the "we pay the IRS rate" defense; it is binding in Michigan, Ohio, Kentucky, and Tennessee, and persuasive everywhere else. West v. BAM! Pizza Management (D.N.M. January 2026) rejected flat per-delivery reimbursement defended as a "reasonable approximation" and treated missing mileage records as the employer's problem. California Labor Code 2802 requires reimbursement of actual costs. The precise framing, and it matters: no statute mandates a particular calculation method; what has happened is that the flat-rate defenses are collapsing under case law, leaving documented, market-level rates paid on recorded miles as the structure that holds. The full landscape is in our 50-state operator guide.
Where insurance and reimbursement turn out to be the same subject
Here is the connection the forums never make. The HNOA fear is, underneath, a fear of underinsured drivers: the personal policy that excludes delivery use, the coverage that lapsed, the claim that lands on the business. And a defensible per-mile rate, built from real filed insurance data for your ZIP codes, includes what it actually costs a driver to properly insure a vehicle in your market. Reimburse below that reality and you are structurally encouraging the underinsurance you bought HNOA to protect against; reimburse the documented rate and drivers can afford the coverage that keeps claims off your policy. Carriers see the same logic from their side: underwriters writing delivery risk increasingly ask about driver screening, ongoing motor vehicle record monitoring, and pay practices, and operations that document all three generally present as better risks. Insurance and reimbursement are not two line items; they are one system, and the documented rate is its foundation.
The day-one setup: HNOA bound before the first run, workers' comp rated for driving, driver coverage verified at hire, dispatch-recorded miles, and a documented per-mile rate for each store's ZIP and vehicle class with the methodology retained. That last piece is RatesReady: rates built from real filed insurance data across 20 vehicle classes, refreshed monthly, audit trail attached, at $49 per location per month (Starter) or $79 (Compliance). Request a demo and we will price your ZIPs against the break-even math live.
This article is an illustrative economic and legal-landscape overview for general information, not legal, tax, insurance, or financial advice. All dollar figures are illustrative and modeled; insurance premiums vary widely and only a licensed broker's quote for your operation is real. Case law application depends on jurisdiction and facts; consult qualified employment counsel and a licensed insurance professional about your specific obligations.