The route is forty miles either way.
What costs you is getting out of the van sixty times.
OkPilot is last-mile delivery software that plans around cost per stop rather than cost per mile. It learns how long each address really takes from arrival to kerb, keeps the day's promises visible while they can still be kept, and captures proof at the door so a failed stop is a decision rather than a surprise.
Miles are cheap. Stops are not.
Shorten a route by four miles and you save a few minutes of driving. Add four stops and you add four lots of finding somewhere to park, walking in, waiting, walking back and pulling out again. That is where the day goes. The fixed cost of a stop barely moves with distance, which is why a dense urban route with sixty drops can be slower and more expensive than a rural one with twenty.
The variable nobody estimates is time at the door. Planners assume a flat two or three minutes per stop and then wonder why the afternoon collapses. A third-floor flat with no lift, a reception that signs for everything, a gated estate where the code changed, a customer who wants the box carried through to the kitchen. These are not edge cases; they are the same addresses every week, and the duration is knowable from your own history rather than from a published figure.
A failed stop is the one event worth modelling properly. You pay for the whole stop and get nothing, then you pay for it again. The retry is rarely convenient: it lands on a different day, out of sequence, usually as the one odd address on an otherwise tight run. One failure can cost more than three successful drops. No routing engine makes that cheaper. Getting in on the first attempt does.
What happens when you type "Why is the afternoon slipping?"
“The south run is forty minutes behind. Why, and what do I do about it?”
Driving is on plan. Three stops took much longer than planned: two office receptions that average nine minutes each, and 14 Harbour Court, which has taken over twelve minutes on all four previous visits. Eleven stops are left and four have windows ending at 5pm. Moving the two latest to tomorrow's first run keeps the other nine inside their windows.
It tells you while the day can still be changed
The last three stops on the north run are projected past their 6pm windows. Pull them onto the east run, which finishes 50 minutes early, or notify the three customers now?
Flat 4B at Nelson House has failed on two of the last three attempts, both before 10am. Nobody is in that early. Put it in the afternoon block instead?
Tomorrow's city centre run has 58 stops in 9 miles. On your own door times that is about 40 minutes over a shift. Drop four to the following morning?
The driver arrives knowing what the door is like
Most of what makes a stop slow is known in advance by somebody. The app puts it on the screen before the driver parks, rather than in the head of whoever did the round last month.
- Access notes attached to the address, not the order: which entrance, which buzzer, where a van actually fits, who signs.
- A realistic duration for this address from your own delivery history, so the plan is built on what the stop takes rather than a flat estimate.
- A failed stop needs a reason and a photo before it can be closed, which is also what makes the retry schedulable rather than a mystery.
- Works with no signal. Basements, lift lobbies and underground loading bays have none.
Describe your last mile
“Six vans, about 70 drops each, mixed residential and business. Business stops are 9 to 5 only and some need a booked dock slot. If a residential stop fails we retry once, then it goes back to the depot. Drivers photograph every drop.”
That is the setup. OkPilot creates:
- Per-address stop durations that update after every visit
- Business and residential as different stop types with different windows
- A retry rule: one more attempt in a different part of the day, then return to depot
- A photo required before any stop, delivered or failed, can be closed
Since 2018
In market
2,000,000+ tasks
Handled
48 hours
To full setup
Last-mile delivery questions
How do you work out what a stop actually costs us?
Take a full week: total driver hours plus vehicle cost, divided by stops completed. That is your cost per stop. Then do it again counting only successful stops. The gap between the two numbers is what failures cost you, and it is usually larger than people expect. We will not quote you an industry average, because yours depends on your own density and door times.
Does it learn how long each address takes, or do we type it in?
It learns. Every visit records arrival and departure from the device, so the duration for an address is your own measured history rather than an estimate. You can override it where you know something the data does not, such as a reception that has just changed its signing policy.
What does it do about failed deliveries?
It makes them cheaper to recover rather than pretending to prevent them. A failure is closed with a reason and a photo, which is enough to schedule the retry sensibly: a different time of day for an empty house, a safe place where the customer has agreed one, a return to depot after your retry limit. It cannot make a failed delivery succeed.
Can customers see where the van is?
Yes. Each stop gets a tracking link with a narrowing window, which is the cheapest thing you can do to reduce failures, because somebody being in is most of the problem. It also moves the "where is my parcel" calls away from your office.
OkPilot also handles what last-mile means, first, middle and last mile, dynamic route optimisation. See all delivery features.
See it costed on your own stops
Book a demo. Bring one week of real drops, including the failures, and we will show you your cost per stop, your measured door times, and which stops were in the wrong half of the day.