Last-mile route planning
The last mile, from the depot to the front door, is the most expensive part of a delivery. Plan it with that cost in mind.
- Free 30-minute call
- No commitment
- You own the code
In short: A last-mile route plan covers the final leg of a delivery, from the local depot to the customer’s door. It groups parcels into routes, sets the order of the stops and keeps time windows and city access limits.
The problem
At half past two, a van is halfway through its afternoon round. Three customers weren’t home, so their parcels ride back to the depot tonight. Two stops are side by side on the map, but a one-way street forces the driver to go round the block. Tomorrow the van drives the same streets again for those three parcels.
Why the last mile costs most
From the factory to the distribution centre and on to the depot, goods travel in large loads, and each kilometre is cheap. On the last mile, one van carries single parcels street by street, often to someone who isn’t home. Here each stop costs the most, and you earn the least per parcel. A few stops in the wrong order can turn a profitable route into a loss.
Signs your last-mile plan is costing you
Vans criss-cross town because their rounds overlap. Failed deliveries come back to the depot every evening. Drivers handle loading bans and closed streets on the spot, because the plan never knew about them. Urban delivery route optimisation has to handle low-emission zones, narrow streets and streets that are closed to vans at certain hours. Most generic tools skip all three. So the route on the screen is short and the route on the road is long.
What changes
Today, you split the day’s parcels into rounds by postcode, and drivers fix the order on the road. Our model groups nearby stops into tight rounds and sets the order in each one. It knows the one-way streets, the loading bans and the windows customers chose. You check the rounds, and drivers get a sequence they can actually drive.
Last-mile route planning software: what you get
Our last-mile route planning software builds on vehicle routing optimisation and can replan during the day. The model groups the stops and sets their order in one step. Every time window is a rule the plan always keeps. The model also knows which vehicle may enter which street, and when. You judge the result on four numbers: cost per stop, how many deliveries fail, how many vehicles you need and how many deliveries arrive on time. We test it on your own orders and your worst days first, so you see how the plan handles a closed street before your drivers do.
- A model written around your own last-mile route planning rules.
- Results measured in the numbers you already track.
- It slots into how you already work, and you can check every answer it gives.
Where a delivery day really goes
Routing software loves kilometres. On a dense parcel round, the time at the door often matters more. Take a round of 120 stops at two and a half minutes each. That’s five hours at doors before the van has driven a metre. Shave thirty seconds off every stop, and you’ve won an hour.
So we take the time at each stop from your own scan data. That’s the record of when your drivers scan each parcel at the door. One average for all stops isn’t enough. A block of flats with a lift, a shop with a loading bay and a farmhouse up a track take very different times. With measured stop times, the planned end of the route is right, and your drivers finish when the plan says.
How it works
We write your depots, your customers, your fleet and your rules into one model. You get the routes that cost the least and keep every rule. You still have the last word: you check the result and send it out.
What last-mile route planning solves
Last-mile route planning gets every parcel to every door in the fewest kilometres, with the fewest failed attempts, inside the windows customers chose. Parcel delivery optimisation pursues these three goals at once: few kilometres, few failed deliveries and every time window kept. How well the model does that decides whether you earn money on a route.
Stop clustering and parcel sequencing
The first lever is stop clustering: grouping nearby deliveries into tight batches before you decide their order. Each route then covers one compact area, and vans stop criss-crossing town. You need fewer vehicles and drive fewer kilometres for the same orders.
Then the parcel sequencing algorithm sets the order within each cluster. It respects one-way streets, access points and turning bans. Two stops a hundred metres apart on the map can be a ten-minute detour in a van. We solve clustering and sequencing together, because no sequence can repair a bad split.
Keeping every time window
A delivery after the customer’s window is a failed delivery. So is one after the shop closes, or after the recipient has left for work. And a failed delivery is the most expensive outcome there is, because you drive the route twice.
The vehicle routing problem with time windows (VRPTW) keeps every stop inside its promised window while keeping distance low. We model the service time at each door, the windows customers chose and the cost of a missed window. The model accepts a few extra kilometres for a delivery that arrives in its window.
Urban access rules and failure risk
Each access rule becomes a hard limit in the model: which vehicle may enter which street, and when. At some addresses and at some times of day, a delivery fails far more often than elsewhere. A model that knows this risk can plan those stops for a better time. When disruption hits mid-route, real-time rerouting puts the remaining stops in a new order without throwing the plan away.
Common questions.
Something else? Ask us directly.
Not for a delivery round. Google Maps takes up to nine stops per route, including the final destination, and you drag them into order yourself. It knows nothing about vehicle capacity, time windows or how long each doorstep takes. For one van with a few drops it’s fine. For a fleet with a hundred stops each, you need a planner that sets the order for you.
Most often the service time is wrong. Many tools use one flat time per stop, but a flat with stairs and a shop with a loading bay are very different jobs. Multiply a small error by a hundred stops, and an hour disappears. Feed measured stop times from your scan data into the model. Then the planned finish is a time your drivers can meet.
In a dense area, many stop sequences look almost equally good if you measure in straight lines. The differences come from one-way streets, turning bans, parking and which side of the road the door is on. A model that only knows the addresses as points on a map picks a sequence that looks short but is slow to drive. Road-network distances and access rules fix most of it.
Sometimes it can, and if that’s the case, we’ll tell you. Standard software is built for an average operation, though. Yours runs on its own rules: your contracts, your exceptions, the things you know by heart. A custom model writes those rules in, and that’s where the better plan comes from.
You don’t need either to start. We build the model from the data you’ve already got, even if it lives in three spreadsheets and one person’s head. You get a system you run yourself. We train your team until they no longer need us.
You see a first result with the prototype, before you commit to anything bigger. We start with a free 30-minute call about your operation and the model that would fit it. Then we build a working prototype and run it on your numbers.
Similar pages.
Terms on this page
Every abbreviation explained in plain words, in our article “Metrics and methods”.
Tell us what you’re planning.
In a free 30-minute call with a specialist, you find out whether optimisation can help.
- Free, no commitment
- You talk to a specialist directly
- A clear go or no-go