Rider management and delivery times: the cost BEFORE and the cost AFTER

What does managing riders and delivery times actually cost today?

As of September 2026, rider management gets paid for in three ways, each with its own range: the platform's fleet costs you between 18% and 30% commission on order value, with no payroll involved;

an in-house fleet starts at USD 1,450 per rider per month in a large Latin American city, once you add salary, benefits, insurance and the motorcycle; and the hybrid model, which covers peak hours with your own people and valleys with the platform, runs between USD 900 and USD 1,900 monthly depending on how many time bands you decide to cover. The number that matters isn't the total but the cost per delivery. With the platform, a USD 12 ticket at 24% commission leaves you USD 2.88 of logistics cost per order. With an in-house fleet doing three deliveries an hour, that same order costs you USD 1.60.

What each price range includes?

The 18% to 22% band applies to marketplaces where you deliver with your own people and the platform only supplies the storefront, the payment and the customer dispute;

it's the cheapest tier and also the one that hands you every operational risk. The 24% to 30% band includes the platform's rider, trip insurance, diner support and automatic refunds when food arrives cold or late, which is precisely what Rappi, Uber Eats and DoorDash charge extra for. Above 30% you're buying packages with advertising built in, paid category placement and co-funded discount campaigns. On the in-house side, the USD 1,450 base range covers one full-time courier with their own motorcycle plus a mileage allowance; if you supply the bike, add between USD 180 and USD 260 monthly for maintenance, mandatory insurance and real depreciation.

The 900-order threshold, and why density beats volume

Below 900 monthly orders, delegate: the arithmetic won't carry a payroll. But that number, on its own, misleads. A thousand orders spread across fourteen operating hours will not sustain an in-house fleet, while seven hundred concentrated between 12:00 and 14:30 and between 19:00 and 21:30 absolutely will, because within a short radius a rider completes three deliveries an hour instead of one. When I sit down with a dark kitchen client, the first thing I ask for is not the monthly consolidated figure: it's the order histogram by thirty-minute band. That's where you see whether there's a real operation or scattered noise. The second filter is geographic: once your delivery radius passes 3.5 kilometers, dead return time eats the cost advantage and you're better off back with the platform.

Twelve extra minutes cost USD 2,900 a month

Delivery time isn't a service metric: it's a SALES metric, and that distinction changes where you put your money. A ghost kitchen in Chapinero closed July at a 38-minute average against the 26 minutes of a competitor four blocks away, with 9.4% of orders ending in partial refunds for cold food; the gap cost USD 2,900 that month, with the ticket up 11% and profit down four points. Marketplaces rank results using estimated time as a weighted variable, so dropping from 38 to 26 minutes lifts your position without a single dollar spent on ads. That move is worth more than any two-for-one promotion and, unlike the promotion, it doesn't break your margin. A restaurant delivering in 25 minutes sells between 20% and 35% more than one delivering in 40, with the same menu.

Five factors that move the price of your logistics

Five variables explain almost the entire gap between a USD 1.60 cost per delivery and a USD 4.20 one. Delivery radius weighs most: going from 3.5 to 6 kilometers raises cost between 35% and 50%, because the empty return trip never gets billed. Hourly concentration moves another 30%: without clear peaks, you're paying for idle rider hours. Kitchen dispatch time —the minutes between the ticket landing and the bag leaving— adds or removes up to 20% of logistics cost, and almost nobody measures it. Food category decides the rest: a dish that survives twenty minutes in packaging doesn't play the same game as fish. And courier turnover, with the average DoorDash rider earning USD 12.23 per hour during 2024 according to Gridwise, makes every replacement cost you between USD 120 and USD 200 in lost training.

How to negotiate commission and cut cost without switching models?

Negotiate with documented volume and partial exclusivity, the only two levers an account executive can approve without escalating. Bring your monthly GTV, your cancellation rate and your rating index to the table:

above 1,200 orders a month with a rating over 4.7, a two-to-four-point commission discount is an ordinary conversation, and on a USD 14,400 monthly operation that's USD 430 flowing back to your pocket. Second move: ask to be shifted from the rider-included tier to the self-delivery tier during the bands where you already staff couriers. Third, and this one pays best, attack kitchen time before street time. Shaving four minutes off dispatch costs you no commission, requires nobody's permission, and moves the same ranking needle a faster rider moves.

What no platform puts in writing?

There's a list of costs that never appears in the contract and typically runs between 3 and 6 margin points. Cold-food refunds fall on you under most agreements, even when the delay belongs to the rider;

serious thermal packaging costs between USD 0.35 and USD 0.80 per order and becomes non-negotiable once your radius passes four kilometers; co-funded promotions get deducted from your settlement, not the marketplace's; and dispute withholding freezes cash for seven to fifteen days. Diego F. Parra insists at Masterestaurant on costing delivery as a separate channel, with its own P&L and its own break-even, because folding it into the dining room hides exactly these leaks. The useful question isn't how much the platform charges you: it's how much you keep per order after commission, packaging, refunds and co-funded promotion.

What happens if you build a fleet and volume drops

Build four in-house riders on the expectation of 1,100 monthly orders and watch volume slide to 700, and your cost per delivery jumps from USD 1.60 to USD 2.90, leaving you worse off than with the platform, except now payroll doesn't switch off the following month. That's the real risk of an in-house fleet, and my recommendation is firm: between 900 and 1,400 orders, go hybrid; above 1,400 with a radius under 3.5 kilometers, go in-house; below 900, delegate without debate. The global delivery market moved close to USD 1.4 trillion in 2025 according to Statista, and that size guarantees the platform will still be there when you need it back. Start this week by measuring one thing: the minutes between ticket and bag, band by band.

Four differences that decide the model

DENSITY beats volume. A thousand orders spread across fourteen hours will not sustain an in-house fleet; seven hundred packed between noon and 2:30 pm and between 7:00 and 9:30 pm will, because a rider completes three short-radius drops per hour instead of one. When I audit a delivery operation, the first thing I ask for is not the monthly total — it is the order histogram in thirty-minute bands. That chart shows whether there is an operation or just noise. Delivery time is not a service metric. It is a SALES metric. Marketplaces rank results with estimated time as a weighted variable, so a kitchen dropping from 38 to 26 minutes climbs positions without spending a cent on ads.

Four differences that decide the model — in practice

That move is worth more than any 2-for-1 promotion and it leaves your margin intact. Commission is visible; the cost of delay is not. Nine percent of orders partially refunded for cold food, across 1,200 orders at a 12 USD ticket, equals 1,296 USD that disappear with no dedicated line in your income statement. Nobody books them as logistics cost, yet that is precisely what they are. Both models coexist better than people assume. The operation with the healthiest margin refuses to choose: it covers the short radius with its own fleet, leaves the long radius and off-peak hours to the platform, and uses the marketplace as an acquisition channel while migrating repeat customers to direct ordering. More coordination work, and it pays.