There are two ways to run first-party delivery. The driver supplies the vehicle, or the operator supplies the vehicle. Most operators never frame it as a choice. They default to driver-supplied because that is how every delivery job has always run, and the cost is invisible to the operator’s P&L. It only shows up in the driver’s wallet — and from there in turnover, recruiting cost, and service quality — but those are downstream, and the operator rarely traces them back to the vehicle.

Driver-owned: the driver carries the entire auto cost stack. The operator carries wage and tip pass-through. The operator’s profit line on delivery looks strong. The driver’s take-home, after the stack, is weak. Turnover is high. Recruiting is constant. Service quality is uneven because the seat is rarely held by the same cast member for long. The operator is paying for the vehicle indirectly — through churn — and not seeing it on a line item.

Operator-owned: the operator buys, leases, or finances the vehicle. The operator carries fuel, maintenance, depreciation, and the commercial policy. The driver carries none of it. The driver’s wage is the take-home. The math the driver runs at the kitchen table now works. Turnover drops. Recruiting cost drops. The seat holds. Service quality stabilizes because the same cast member is in it.

The operator’s profit line on delivery looks worse on paper. It is not worse in the operation. The cost did not appear from nothing. It moved from the driver’s invisible column to the operator’s visible column. The total cost of running the last mile is roughly the same. The split changed. And the split is the whole point — because the split changed who can afford to take the job.

The operator who runs first-party delivery on driver-owned vehicles is running a model that the driver’s auto cost stack will eventually break. The operator who runs it on operator-owned vehicles is running a model where the math holds for both sides of the table.

Pick on purpose. Do not default.