Choosing delivery zone and radius: what it really costs and which radius leaves margin in 2026

What does it cost per month to set your delivery zone and radius properly?

Choosing your delivery zone and radius properly costs between 900 and 2,600 USD a month as of August 2026, and that range covers three lines almost nobody separates in their P&L:

rent for the kitchen or dispatch point, the geo-targeted ad spend that holds the neighborhood, and marketplace commission on orders coming through Rappi, Uber Eats or DiDi. The low band, 900 to 1,300 USD, fits a single site with a short radius and minimal ad spend; the high band, 1,900 to 2,600, shows up when you run two kitchens to cover two dense polygons instead of stretching one. And here is the part that stings: stretching the radius saves nothing, because every additional kilometer of average distance adds roughly 0.38 USD to delivery cost per order, and at 1,000 monthly orders that is 380 USD per kilometer you gave away. The three bands differ by real coverage, not by finish.

What each investment band includes

From 900 to 1,300 USD a month you pay for a shared kitchen or a corner of your existing location, a 3 km radius, 180 to 300 USD of geo-targeted spend by postal code, and platform commission that across the region runs between 18 % and 30 % of the ticket depending on the plan; at an average ticket of 14.20 USD and 700 orders that is already some 2,000 USD of commission sitting separately, in the sales line. Between 1,300 and 1,900 USD you get your own kitchen with two stations, a radius of 4 to 5 km, 400 to 600 USD of ad spend, and a mixed fleet — your own rider for the core, platform for the edge. From 1,900 to 2,600 USD we are no longer talking about a bigger radius but about TWO polygons: two small kitchens, each with its own 3 km radius, its own spend and its own local reputation.

What each investment band includes — in practice?

It costs more monthly and returns more margin per order. The delivery radius is a PRICING decision dressed up as a logistics decision. You are not choosing how far your bike goes:

you are choosing how many orders per hour one rider can dispatch, and that number governs your unit delivery cost the way food cost governs the cost of a dish. At 3 km a rider closes the loop in 18 or 20 minutes and moves three orders an hour. At 8 km that same rider burns 55 minutes between ride out, wait and ride back, and cost per order triples while two deliveries from the building across the street go cold on the counter. At Masterestaurant we treat it as a P&L line, with its range, its ceiling and its cut-off rule, because a cost with no owner in the report grows while nobody watches. Global online food delivery reaches 1.51 trillion USD in 2026 according to Statista, growing at a 6.24 % CAGR through 2031: there is plenty of demand inside your 3 km.

The four factors that move the price, and how much each one weighs

Four variables explain almost the whole gap between paying 900 and paying 2,600 USD. Household density per square kilometer comes first and hits hardest: in a dense zone you reach 1,000 orders within 3 km, while a sparse zone demands 8 km for the same volume, which at 0.38 USD per extra kilometer translates into 1,520 USD a month of delivery overcost. Platform commission weighs between 18 % and 30 % of the ticket and depends on your plan and your negotiated volume. Rent per square meter swings from 7 to 28 USD depending on the area, and that apparent difference is the trap I unpack in the next passage. Geo-targeted spend, cheap at 180 USD for a small polygon, explodes the moment you try to buy visibility across half a city instead of owning six blocks. A site at 7 USD/m² in a low-density area ends up more expensive than one at 20 USD/m² in a dense area, and the number that proves it sits in delivery, not in the lease.

Cheap rent on the outskirts gets paid on the other side

Take 120 m²: the outskirts save you 1,560 USD a month in rent, a figure anyone would celebrate. But that location forces an 8 km radius to gather 1,000 orders, and those four or five extra kilometers of average distance are worth between 1,520 and 1,900 USD monthly in delivery cost. The saving evaporates, and on top of it you inherit a 46-minute average delivery time, a sliding rating and a 7 % cancellation rate, which in delivery is double what the category tolerates. The paradox resolves one way only: rent is a fixed cost you see every month; delivery overcost is variable, grows with success and shows up on no line with a name of its own. Rappi, Uber Eats and DiDi Food rank the storefront on a blend of prep time, acceptance rate and recent rating, so a virtual brand with 4.7 stars and 26-minute delivery sits at the top in its own neighborhood while the one covering half the city at 4.1 stars ranks strongly nowhere.

The algorithm rewards consistency, not reach

Competition for that space is real: DiDi Food Mexico had close to 74,000 restaurants on the app in 2024, 70 % of them local small businesses, according to DiDi Food; in India, Swiggy reported 196,000 partner restaurants across 653 cities in fiscal year 2023-24. You do not win that ranking by being the biggest. You win it by being the fastest inside a small polygon, steadily, week after week, until the algorithm treats you as the safe choice on that block. Start by trimming the radius before you negotiate anything else, because it is the one adjustment that cuts cost without asking permission from anyone: drop it to 4 km for 30 days and measure average delivery time, rating and cancellations before and after. Second, negotiate commission with proven volume in hand; platforms move that percentage several points when you arrive with a record of orders rather than a promise.

How to negotiate and optimize: five concrete moves?

Third, ask for a zone exclusivity window in exchange for schedule commitment, which usually beats a straight discount. Fourth, shift spend from a wide radius to postal-code polygons and compare cost per delivered order, never cost per click.

And fifth, before signing any new kitchen, add rent plus projected delivery overcost at 1,000 orders: if the outskirts do not win that full sum, they do not win. Take the Chapinero case: 1,340 orders, average ticket of 14.20 USD, operating margin of 4.1 %, with 9 km configured in Rappi. If that owner drops to 4 km tomorrow, he loses maybe 22 % of volume up front, leaving around 1,045 orders. But average distance falls from 6.4 to 2.8 km, which at 0.38 USD per kilometer gives back roughly 1.37 USD per order, some 1,430 USD a month. Average time drops from 46 to 28 minutes, the rating climbs and the 7 % cancellation rate corrects itself, because nearly all of it came from long waits.

What would happen if you split your radius in two tomorrow?

Operating margin moves from 4.1 % to somewhere between 11 % and 13 % with fewer orders and less noise in the kitchen.

I will admit I recommended the opposite for years, wide coverage first and fine-tuning later, and that sequence burns cash and reputation at the same time. Open your platform report today, look at last month's average delivery distance and compare it with 3 km. Cheap rent on the outskirts bills you on the other side. A unit at 7 USD/m² in a low-density area forces an 8 km radius to reach 1,000 orders, and each extra kilometer of average distance adds roughly 0.38 USD to delivery cost per order. At 1,000 orders, those four extra kilometers cost 1,520 USD a month, well above whatever the rent saved. Delivery platform algorithms reward consistency, not reach.

The four differences that move real money

Rappi, Uber Eats and DiDi rank the storefront on a blend of prep time, acceptance rate and recent rating, so a virtual brand at 4.7 stars and 26 minutes climbs to the top of its own neighborhood while the one covering half the city at 4.1 stars ranks nowhere with any force. Choosing delivery zone and radius also sets your local SEO. The query "restaurant near me" resolves on proximity to the user, and Google Maps distributes visibility by your physical address rather than your declared delivery radius, which means a kitchen placed at the center of gravity of demand captures organic traffic the outskirts will never hand you. Opportunity cost is the line nobody books. When a courier burns 55 minutes on an 8 km order, the loss is not 4 USD of fuel: it is the two nearby orders that landed while he was riding and never got dispatched, and those two orders, at a 14 USD ticket and 62 % contribution, carry 17.40 USD of margin that simply evaporated.