Delivery zone and radius: the matrix that actually decides (2026)

What is the ideal delivery radius for a single-production-point dark kitchen?

Between 2.5 and 3.5 kilometers along the densest residential corridor, and that number is not an aesthetic preference but cash-register arithmetic.

Best for operations running 40 to 220 daily orders on one hot line: inside that ring, logistics cost sits in the 8-11% band of the ticket, while at 4-6 kilometers it climbs to 18-24% during peak hours, because the courier charges for distance and charges for waiting. With a 30% food cost plus aggregator commission, a dish that leaves a healthy margin at kilometer 2 loses it entirely at kilometer 5, and the app's monthly report will not break it down by ring for you: you have to separate it by hand, order by order, for two weeks. Nearly 75% of US restaurant traffic is already off-premise, according to the National Restaurant Association (2025), so this geometry governs three quarters of your sales.

What is the ideal delivery radius for a single-production-point dark kitchen — in practice?

The starting mistake is drawing the polygon with the compass point on the kitchen door.

Put it where the repeat customer lives, look at the heat map of your last 600 tickets, and cut everything outside the main corridor, even when it feels like giving away market. Delivery apps do not reward coverage, they reward compliance, and that distinction is worth more than any discount campaign. Where your operation depends on Rappi or Uber Eats for over half of sales, estimated delivery time is the variable that moves your position in the listing shown to ALL nearby users, not only the distant ones. Rappi reported 35 million active users and 150 million downloads as of August 2024 (Rappi operating results), so the listing you are fighting over is crowded and the tiebreaker is operational. A concrete case: ghost kitchen in the north of the city, a 9-kilometer circle, eleven neighborhoods, 51 minutes average, rating down to 4.3 stars.

Best for operators fighting for app listing position: the short polygon as an algorithmic lever

We cut the polygon to 3 kilometers and the average dropped to 29 minutes within six weeks, same menu, same kitchen, zero new hires. That is the trade of this business, and the map settled it: you sell more by appearing on FEWER maps, because the app hands out visibility for what you deliver, not for what you promise. Declaring an inflated service area in Google Business Profile does not buy you reach, it buys you a blurred centroid that competes worse at every concrete point on the map. This suits you if your virtual brand leans on search more than on the aggregator: when someone types restaurant near me, the engine resolves it through proximity to the centroid of the area you declared, review density and freshness, and NAP consistency across directory, site and listing. An operator who declares 14 neighborhoods to sound big ends up fifth across all fourteen, when he could be first across four.

If your traffic arrives through "restaurant near me": declared service area and the blurred centroid

Diego F. Parra keeps insisting at Masterestaurant that the zone is a geometric fact and not a commercial ambition, because the centroid does not negotiate: it gets calculated. Declare the real polygon you can serve hot, verify that the NAP matches character by character across the three properties, and let the listing compete concentrated. One detail almost nobody audits: the hours declared on the listing must match your real order-acceptance hours, or search learns that you fail. Three situations make cutting to 3 kilometers a cash mistake, and you should recognize them before touching the map. First, high average ticket: selling office catering or family menus above 45 dollars, logistics cost at 4-6 kilometers eats 6-9% of the ticket instead of the 18-24% that punishes a low ticket, and stretching genuinely pays there. Second, insufficient density: unless your 3-kilometer ring holds at least 25,000 households or an active office corridor, the short polygon delivers volume that never covers the kitchen's break-even.

When NOT to pick the popular option: three cases where the short radius costs you money?

Third, food that travels well: pizza, stews, dry items and firm desserts survive 40 minutes without noticeable damage, whereas anything fried or plated with an emulsified sauce degrades past the 25-minute mark.

The second case shows up most often and destroys the most money, because the operator confuses discipline with shrinking; when demand is missing inside the ring, the problem is where the kitchen sits, not the radius. Four concrete signals disqualify a zone no matter how cheap the rent, and all four surface in one afternoon of fieldwork. The first is the physical barrier: a river, a highway without a nearby turnaround, or an overpass turns 2 linear kilometers into 14 minutes on a motorcycle, so measure in MINUTES, never in a straight line on the map. The second is the empty courier pool: if at 1:20 pm on a Tuesday the app takes more than 7 minutes to assign a rider, your delivery time is broken before anything hits the pan.

Red flags when comparing zones and polygons before signing the lease

The third is same-category saturation: fourteen fried chicken kitchens inside one polygon means you are competing for position in a listing where the user compares price, not cooking. The fourth is commission regulation: New York caps delivery at 15% and other services at 5%, according to Restaurant Business (2023), and San Francisco set the same 15% cap back in 2020 (Restaurant Dive). Where no such cap exists, the real commission can double, and that gap swallows the margin you thought you were winning with the wider radius. Paying your own riders, the right structure is not a radius but two rings under different rules, and that decision hands you back control of variable cost. Ring A, 0 to 3 kilometers: free delivery above a set ticket, a 30-minute promise, the heart of repeat business. Ring B, 3 to 5.5 kilometers: explicit delivery charge, minimum ticket 35-40% higher, a 45-minute promise.

Best for operations with an in-house fleet: the double ring with differentiated pricing

The operation holds because ring B pays its own cost instead of being subsidized by ring A's margin, which is exactly what happens under a single radius with flat shipping. Worth keeping in view: 46% of US diners prefer ordering through third-party apps, at nearly 5 orders a month, according to DoorDash cited by Restaurant Business (2024), so your own fleet coexists with the aggregator and has to be costed separately. Review contribution margin by ring every month. Should B fall under 22%, raise its minimum ticket before cutting it: marginal volume still absorbs fixed kitchen cost. Volume rises 15 to 25% for three weeks, and then you lose the neighborhood you had already won. The chain is predictable and I have watched it run all the way through: distant orders come in, average time climbs from 29 to 44 minutes, the rating drops half a point, the algorithm demotes you for every user in the area, and the customer two blocks away who ordered three times a month stops seeing you on the first screen.

What happens if you double the radius to rescue a slow month?

By the fourth month you have fewer orders than before expanding, higher logistics spend, and a rating that takes six months to recover because the average drags the history behind it.

With more than 20,000 ghost kitchen locations operating in the United States in 2023, according to Statista, the user has an alternative two blocks away and waits 44 minutes for nobody. My reading is that the short radius is the only free growth lever left to a kitchen without new capital. Open the heat map of your last 600 tickets today and delete the farthest 20%. COST PER ORDER BY RING. In the 0 to 2 kilometre ring logistics cost runs 8-11% of ticket; between four and six kilometres it climbs to 18-24% at peak, because couriers charge for distance and for waiting. The same dish at 30% food cost leaves healthy margin close in and negative margin far out, and the aggregator's monthly report will not break that out for you: you split it by hand.

The four differences that move the till

SPEED AS AN ALGORITHMIC SIGNAL. Delivery apps do not reward coverage, they reward fulfilment. A lower estimated time improves your position for every nearby user, not just the distant ones, so a short polygon lifts the virtual brand's overall visibility. That is the paradox of the trade: you sell more by appearing on fewer maps. COHERENCE WITH LOCAL SEO. The restaurant near me query resolves on proximity to your declared point, recent reviews and listing consistency. A bloated service area splits relevance across zones you will never convert, while the neighbourhood that actually pays payroll competes against chains with a tighter listing. REPEAT PURCHASE AND DENSITY. Concentrating deliveries in a few blocks builds physical recognition: same courier, same bag, same hour. Sixty-day repeat purchase in the inner ring usually doubles the outer ring's, and repeat purchase is the only metric that lowers acquisition cost without spending another dollar on ads.