Why the Closest Delivery Stop Isn't Always the Fastest Route
The closest stop on a map is not always the next stop on the fastest delivery route, because straight-line distance ignores the road network. We tested this with MiliRoute, a delivery route optimization and multi-stop route planning app built in Australia. On a real 24-stop route in Castle Hill and Kellyville, Sydney, two stops sat just 177 metres apart in a straight line, but about 3.1 km and roughly 6 minutes apart by road.
Key finding
A stop that looks closest on the map can be the slow one to reach. In this test, the route order that looked wrong to the eye was the fastest order we found, and every "obvious" change made the route slower.
- Stop 13 to stop 19: 177 m in a straight line, about 3.1 km of driving distance, and about 6 minutes (399 seconds) of estimated driving time.
- MiliRoute's original order: 150.9 minutes and 106.3 km.
- Moving stop 19 to directly after stop 13: about 3 minutes slower.
- Reordering stops 16, 17 and 18: about 5 minutes slower.
- Ordering the stops by straight-line distance only: about 20 minutes slower.
The route that looked wrong
We looked at a real saved delivery route with 24 stops around Castle Hill and Kellyville. On the map, stop 19 sat only about 177 metres from stop 13. Yet the route did not visit it next. It drove a long loop north first and reached stop 19 six positions later.
That looks like a mistake, and it is a fair thing to question. When two pins are almost touching, going to the second one next seems obvious. So we tested it instead of assuming.
Straight-line distance vs driving distance vs driving time
These are three different measurements, and delivery route planning depends on keeping them apart.
- Straight-line distance is the shortest path between two points on a map, as a bird would fly. It ignores roads completely. For stops 13 and 19 it was 177 m.
- Driving distance is the length of the road route between two points. For the same two stops it was about 3.1 km.
- Estimated driving time is how long road-routing data estimates that drive will take. Here it was about 6 minutes. These are calculated estimates from routing data, not times measured by physically driving each leg.
The three do not move together. Two stops can be very close on the first measure and far apart on the other two.
How we tested the route
- We took the real saved route: 24 stops.
- We got the estimated driving time between every pair of stops from Google's road-routing data, once, and saved it. Every comparison below uses exactly the same numbers.
- We calculated the total estimated driving time of MiliRoute's original order.
- We calculated the total for the changes someone might try by eye: visiting stop 19 straight after stop 13, and reordering stops 16, 17 and 18.
- We calculated the total for an order built on straight-line distance only.
- We ran the optimizer again with 10 times its normal search budget, to see whether more search time would change the answer. It returned the same order.
Customer names and street addresses are not shown in this article.
The results
| Route strategy | Total drive time | Difference |
|---|---|---|
| MiliRoute's original order | 150.9 min (106.3 km) | n/a |
| Stop 19 moved directly after stop 13 | about 3 min slower | +~3 min |
| Stops 16, 17 and 18 reordered | about 5 min slower | +~5 min |
| Ordered by straight-line distance only | about 20 min slower | +~20 min |
All rows use the same estimated driving times. The re-run with 10 times the search budget produced the same order as the original.
Why the close stops were not close by road
Because a vehicle can only travel along roads, not across the gap between two pins. The routing data put the road route from stop 13 to stop 19 at about 3.1 km, roughly 6 minutes (399 seconds), despite the two stops being 177 m apart in a straight line.
The routing data shows that the detour exists, but it does not tell us which feature of the road network causes it. Common reasons in general include limited crossings, one-way streets, divided roads and estates with a single entrance.
Some individual legs make the picture clearer:
- Stop 13 to the next stop in the original order: 244 seconds.
- Stop 18 to stop 19: 208 seconds, along a main road.
- Stop 13 to stop 19 directly: 399 seconds.
Those are single legs between two stops. They are not the same as the 32 and 35 minutes below. That comparison covers a longer sequence of several legs. Over the same stretch of the route, the original order added up to about 32 minutes (1,938 seconds), while the alternative that visited stop 19 straight after stop 13 added up to about 35 minutes (2,097 seconds). That is about 3 minutes more, which matches the first alternative in the results table.
Why straight-line optimization can produce a slower delivery route
Because straight-line distance is only a stand-in for driving time, and the two can disagree. We built an order that was best on straight-line distance, then priced it with the same estimated driving times. It came out about 20 minutes slower than MiliRoute's order.
That does not make straight-line distance useless. It is a reasonable rough guide where roads run directly between stops. It becomes a poor guide where the road network forces detours, as it did between stops 13 and 19.
What this means for delivery drivers
- Check the road route before overriding the order. If two stops look close but the route separates them, open the road route between them. The detour usually explains it.
- Judge by driving time, not by how close the pins look. Pins on a screen do not show one-way streets or missing connections.
- Reorder for reasons the data cannot know. A customer who is only home after 2pm is a good reason. Set an arrival window on that stop instead of changing the whole order.
If you are starting from a long list of stops, see how to organize a big delivery route and how to scan delivery addresses with your phone camera.
How MiliRoute optimizes multi-stop routes
MiliRoute orders stops using estimated driving times between them, not by how close the pins look. For each route it gets the estimated driving time between your stops from Google's road-routing data, then searches for the order with the lowest total estimated driving time.
That is why the route in this article came out the way it did. It is also why the route planner for Australia treats a short line on the map and a short drive as different things. Courier and small-fleet teams in Sydney can see how this applies to their own runs on the courier route optimization page, and teams planning many vehicles can look at the fleet route planner for small business.
For the bigger picture of how route optimization works and what to look for in a tool, see our route optimization guide for delivery fleets.
Limitations of this case study
This is one real route, calculated once. The times are estimates from road-routing data, not times recorded by driving every leg, and real driving times vary with traffic, time of day and roadworks. The case study shows that, for this route and this data, the visually obvious changes were slower. It is not a benchmark, and it does not show that MiliRoute will produce a faster route than every other method on every route.
FAQ
Why isn't the closest delivery stop always next?
Because straight-line distance ignores the road network. Two stops can be very close on a map yet far apart by road, and the next stop on a fast route is chosen by driving time, not by how near the pins look. In our test, two stops 177 m apart were about 3.1 km and roughly 6 minutes apart by road.
Does MiliRoute use straight-line distance?
No. MiliRoute orders stops using estimated driving times between them, taken from Google's road-routing data.
Is the shortest driving route always the fastest?
Not always. The shortest driving distance and the shortest driving time can differ, for example when a shorter road is slower. MiliRoute looks for the lowest total estimated driving time.
Why does my delivery route zigzag?
Often because the road network forces it. A route can look like it doubles back on the map when the roads leave no direct connection between nearby stops. Check the road route between the stops before assuming the order is wrong.
Should I manually reorder delivery stops?
Only if you know something the routing data does not, such as a customer who is available only after a certain time. Otherwise, check the road route first. In our test, the manual reorderings we tried were about 3 and 5 minutes slower.
How does multi-stop delivery route optimization work?
It searches for the stop order with the lowest total cost, usually driving time, using the estimated travel time between every pair of stops. The more stops there are, the more possible orders there are, so the search has to be done by software rather than by eye.
What is the difference between straight-line distance and driving distance?
Straight-line distance is the shortest path between two points on a map, ignoring roads. Driving distance is the length of the road route between them. For stops 13 and 19 in this test, they were 177 m and about 3.1 km.
Next step
Try it on your own run with the multi-stop route planner, or read how drivers share a route with another driver and capture proof of delivery.
Written by Mike Gale, AI Engineer at MiliRoute, from an analysis of a real delivery route.