Results
Radius is straight-line distance from your store, the way you would draw it on a map.
Your zones and tiered delivery fees
| Zone | Cost to serve | Your fee |
|---|---|---|
| Zone 1 | $ | |
| Zone 2 | $ | |
| Zone 3 | $ |
“Your cost to serve” is driver time plus mileage for a round trip to the middle of the zone. It is a guide for setting fees, not a recommended price. Edit the fees to see how much of the cost each one covers.
How this planner works
Two different things limit a delivery area. The first is food quality: past a certain number of minutes in a bag, fries go limp and the customer blames you. The second is driver capacity: every extra mile of radius makes each run longer, so the same drivers complete fewer runs per hour. The planner works out the radius each limit allows and recommends the smaller one.
runs: orders per run = 1, or your batch size when batching is on
earlier-stops delay = (orders per run − 1) × (hand-off + drive between stops)
food-quality radius = (max transit − hand-off − earlier-stops delay) × speed ÷ 60 ÷ detour factor
average customer distance = ⅔ × radius (exact for customers spread evenly over a circle)
run time = pick-up + 2 × (⅔ × radius × detour ÷ speed × 60) + orders per run × hand-off + (orders per run − 1) × drive between stops
fleet capacity per hour = drivers × orders per run × 60 ÷ run time
capacity radius = the radius at which peak orders per hour = fleet capacity × your “busiest” %
recommended radius = the smaller of the food-quality radius and the capacity radius
driver load: a = (peak orders ÷ orders per run) × run time ÷ 60 · ρ = a ÷ drivers
chance an order must wait = Erlang C(drivers, a)
average wait once food could leave = chance × run time ÷ (drivers × (1 − ρ)) × (1 + v) ÷ 2, where v = (spread of run times ÷ run time)²
wait added to the ETA = average wait × e−prep ÷ m, with m = average wait ÷ chance (a driver is requested when the order arrives, so part of the wait overlaps prep)
wait for the batch to fill = (orders per run − 1) ÷ 2 × 60 ÷ peak orders per hour
typical ETA = prep + waits + pick-up + drive to ⅔ radius + half the earlier-stops delay + hand-off
edge ETA = prep + waits + pick-up + drive to full radius + full earlier-stops delay + hand-off
quoted window = typical ETA rounded to the nearest 5 minutes, to edge ETA rounded up to the next 5
The driver-wait lines are the Erlang C formula, the same one call centres use to decide how many agents to staff, with the Allen–Cunneen adjustment because delivery runs vary less in length than the textbook formula assumes. It captures the thing owners feel on a Friday: as drivers approach being busy all of the time, waits do not grow steadily, they shoot up. That is why the planner sizes your radius against a “busiest you want drivers” percentage rather than against 100%.
Zones and fees
The recommended radius is split into three rings of equal width. Because area grows with the square of the radius, the outer ring holds five-ninths of the territory and the inner ring only one-ninth, so most of your long, expensive runs come from the ring where a flat fee covers the least. The table shows your cost to serve the middle of each ring and what share of it your fee recovers.
Limitations
- Real delivery areas are not circles. Rivers, highways, rail lines and one bad left turn bend them. Use the radius as a budget of minutes, then trim the map by hand.
- Customers are assumed to be spread evenly. If most orders come from one apartment cluster, your true average trip is shorter or longer than ⅔ of the radius.
- Speed is a single average. Friday 6 pm and Tuesday 2 pm are different cities; run the planner for your worst hour.
- The queue formula is an approximation and assumes orders arrive at random. A rush of orders in ten minutes will beat it.
- The ETA assumes a driver is requested the moment the order arrives, so waiting overlaps prep. If you only call for a driver once the food is bagged, quotes should be longer.
- The fee guide ignores tips, order size and what competitors charge.
Sources
- IRS standard mileage rate, business use: 76 cents per mile for July 1 – December 31, 2026; 72.5 cents for January 1 – June 30, 2026. Source: IRS, Standard mileage ratesAs of September 2026 (IRS page last updated July 28, 2026).
- Waiting-time maths: the Erlang C formula (A. K. Erlang, 1917) with the Allen–Cunneen approximation for service-time variability. These are standard queueing-theory formulas, used here as arithmetic, not as a data source.
Speeds, hand-off times, the detour factor and the example fees are planning assumptions that you are expected to overwrite. They are not statistics.
Putting it to work
Quote the window, not the best case. A customer told “35–50 minutes” who gets food in 40 is happy; one told “30 minutes” who gets it in 40 calls the store. If the planner says the fleet cannot sustain your peak, you have three levers, in order of cost: turn on batching, pull the radius in during the rush, or add a driver.
Then check the plan against reality. Compare the quoted window with actual door times for a week. If the far ring regularly misses, it is a minutes problem, not a driver problem, and the ring should shrink or carry a longer quote.
Running your own drivers?
KwickSpot is the GPS driver tracking built into KwickPOS: assign orders from the POS, see drivers on a live map and send customers an ETA. The calculators on this page are free either way.