Delivery zone and radius: the extra-kilometer myth and what the numbers actually say

Step 1: split your sales by distance ring before you touch the map

Start by pulling the last 90 days of orders and sorting them into three rings —0 to 3 km, 3 to 5 km, 5 to 7 km— with the real contribution margin of each one, because without that cut you are arguing about your radius with an opinion instead of with the cash register. The Chapinero grill I mentioned earned 31% margin in the near ring and 6% in the far one, same menu and same platform commission: the deliverable here is a three-row table showing orders, average ticket, travel minutes and margin after delivery cost per ring. You verify it by adding the orders across the three rings: if the total does not match the platform report, some address got misclassified and the whole diagnosis collapses. Radius is defined in TIME, never in distance, and you make that conversion by timing the real trip at peak hour instead of the one the map promises at three in the afternoon.

Step 2: turn kilometers into minutes, the unit your operation actually pays for

Take ten representative addresses along the edge of your coverage, run a stopwatch between 7:30 and 9:00 at night, and write down the minute the rider hands over the food and the minute the rider is available again; that second number is the one almost nobody records and the one that decides your rotation. In a dense Latin American city, 2.5 to 4 km lands between 12 and 18 minutes of travel, and the hard threshold sits at 20 minutes. This step is done when the average and the spread of those ten stopwatch readings are written down, not remembered. Cut the map at the ring where contribution margin per order still covers the rider's hourly cost divided by the deliveries that fit inside that ring, and not one meter beyond. The arithmetic fits on a napkin: eleven deliveries per shift spread the motorcycle's fixed cost across nearly twice as many tickets as six deliveries do, and that is where the margin appears that the owner spent two years hunting for in the price of a dish.

Step 3: set the radius where the ring margin still pays the rider's shift

With sector wages climbing —base hourly pay in United States restaurants rose 4% to 14.20 dollars in 2024, according to the 7shifts Restaurant Workforce Report— every dead minute on the road weighs more each year. Your deliverable: one kilometer figure, written, signed and loaded into every platform's settings. Your delivery zone is not a circle, and drawing it as one hands absurd trips to your operation: a river, an avenue with no turnaround or a hillside turns 2 kilometers of map into 22 minutes of riding. Draw the polygon block by block using the stopwatch data from step 2, exclude the sectors where real time beats the threshold even though distance looks short, and pull in the corridors where a fast avenue gives you cheap kilometers. Rappi, Uber Eats and DiDi Food all accept custom polygons; the default circle is operational laziness, not a limitation of the tool.

Step 4: trim the zone by real neighborhoods, not by a perfect circle

Verify it by ordering from three edge addresses on your own phone: if any of them quotes an estimate above twenty minutes, the polygon still carries fat. Coverage and reach are different things, and mixing them up costs orders: coverage is how far your motorcycle rides, reach is where your brand shows up when somebody opens the app hungry at 8:40 at night. Platform search engines rank by probability of conversion and that probability sinks with every added minute of estimated time, so owning three neighborhoods puts you on top in three neighborhoods, while scattering across eight leaves you invisible in all eight. Diego F. Parra works it this way inside Masterestaurant audits: the delivery polygon and the brand's local pages describe the SAME territory, with the neighborhood names written into the listing, the menu and the content. It is done when your Google listing and your in-app zone name the same sectors.

The four mistakes that ruin the trim, and how to dodge them

The costliest mistake is widening the radius when sales dip, because it looks free and it is not: demand gets spread worse, EVERY order takes longer —including the close ones, since the rider travels farther and returns later— and the store rating drags downward. The second is timing the trip off-peak and then signing off on a radius with that number. The third, watching the order count on the monthly report instead of margin by ring, an accounting mirage that lifts volume while margin slides from 31% to 6%. And the fourth is trimming the map without telling the kitchen, which keeps producing for demand that no longer exists. One discipline handles all four: no zone decision enters without its minutes figure and its margin figure sitting beside it. Shrink the radius and gross sales fall for the first two weeks: that will happen, and it has to be said beforehand so nobody reverses the decision on the worst day.

What happens if the trim drops your volume and the owner panics?

What climbs alongside it is margin per order, motorcycle rotation and average delivery time, and that last number is the one platforms reward in their ranking, so volume comes back through relevance rather than through kilometers.

If volume has not returned by day 45, the problem was never the radius but the offer: check average ticket, menu photography and listing conversion before you reopen the map. The market is on your side —cloud kitchens grow 12.6% a year between 2026 and 2033, according to Grand View Research— but that growth rewards whoever delivers fast across few neighborhoods. You are finished when five things are written and verifiable, not when the map looks pretty. One: the ring table with real margin per ring, signed. Two: the peak-hour travel average, measured by stopwatch, below 18 minutes. Three: the polygon loaded identically across all three platforms, with no default circles. Four: the same neighborhoods named in the Google listing, the digital menu and the brand's local pages.

Closing checklist: how to know the zone was built right

Five: a review booked at 45 days using the same indicators to compare against the baseline. If any of those five lives only inside somebody's head, the zone is not built, it is improvised. Open the last 90 days of reports today and build the ring table: everything else in this guide runs on top of that number. A tight radius turns every motorcycle into a rotation asset: eleven deliveries per shift against six means the courier's fixed cost spreads across nearly twice the tickets, and that is where the margin the owner blamed on plate pricing reappears. A wide radius buys gross volume and sells margin: order count climbs in the monthly report while contribution margin per order collapses from 31% to 6%, an accounting mirage you only detect once you split revenue by distance ring. In digital capture the gap runs deeper, because app search and the Google local pack reward geographic relevance: a virtual brand dominating three neighborhoods ranks high in those three, while one scattered across twelve ranks high in none.

Where the two decisions truly split?

Five-star reviews come from hot food arriving earlier than promised, and that outcome depends more on the polygon than on the recipe book;

when a customer rates lukewarm chicken with 2★ they are not judging your kitchen, they are judging your geometry. A physical restaurant tolerates a wide radius because its dining room absorbs fixed cost, whereas a dark kitchen has no such cushion, so in dark kitchen vs physical restaurant polygon discipline stops being advice and becomes a survival condition.