The tour the driver does not understand

An experienced driver should not be the permanent correction of a system that does not learn.

In The system is faster than you I used a picture: a tour planned for a new driver, a shop that opens at ten, a stop placed at half past eight. This essay takes that picture from the other side. Not the system that did not know the opening hours, but the driver who knows them, and the question of where what they know goes.

What I have and what I do not

What I have: the work in last mile coordinating drivers, tours and operational problems during delivery, today as a remote dispatcher. Drivers in the ramp-up phase who, as I wrote in High-volume days: the standard under pressure, were not yet fully used to the procedures. And the change of August 2025 to tour planning, which aimed at more realistic plans and provided for the use of information from the field.

What I do not have: a driver who told me specifically that their tour made no sense. A list of the questions drivers ask me most often about tours. How new drivers learned the tours, and whether knowing the area changed performance. Who exactly had the authority to change a tour during the day, and how it was done technically. And whether what drivers knew about an area was kept anywhere.

That is a lot of “I do not know” for one essay. But that is exactly the point. What a driver knows about their own tour is among the things an operation has least written down, even when it has a new system that asks to use it. So this essay is my analysis and what I would recommend today, not a process I have built. The scenes in it are pictures.

The planned tour and the driven tour

A system plans a tour on the data it has: the addresses, the delivery times, the distances, the average times. The driver, meanwhile, drives the tour with the conditions they have: the entrance that is behind the building, the gate that opens only with a bell, the shop that closes at lunchtime, the street that is blocked every Tuesday by the market.

Neither is wrong in itself. They are two descriptions of the same tour, from two different places. The first is fast, complete for all tours at once, and knows nothing that has not been written down for it. The second is slow, limited to the areas the driver knows, and knows things nobody has written down.

The change of August 2025 accepted this, in its own language: a more realistic plan also needs information from the field. But that does not prove there was a complete process in which every correction by a driver was recorded and improved the plan. And without that process, information from the field stays where it was born: in the driver's head.

The system plans the tour with the data it has. The driver drives it with the conditions they have. Neither has the whole of it without the other.

Neither the driver nor the system

It would be easy to write this essay as the story of the wise driver and the blind system. It would not be true.

The driver is not always right. A driver may change the order because it seems better to them, and their route may be longer, or delay a delivery that had a time slot. An obstacle they met once may no longer be there. And a driver in ramp-up, not yet used to the procedures, may call something “a tour that makes no sense” that does make sense, but that nobody has explained to them.

So the question is not who is right. The question is: when a driver discovers something the system did not foresee, does the organisation have a way to check that information? If it does, the information can become an improvement. If not, there are only two ways: either the driver ignores the system in silence, or the system ignores the driver.

The new driver and the absent driver

Imagine an area that only one driver knows well. They drive the tour every day and drive it well, because they know where to park, which door to use, which customer to leave for later. None of this is in the plan. The plan does not need it, because they are there.

Then one day they are absent. A new driver takes the tour, perhaps in ramp-up. The new driver follows the plan, because the plan is all they have. And the day stalls in the same places where the experienced driver never stalled.

This is not a case I have documented. It is the picture of what I wrote about in The system is faster than you as “full distrust”: knowledge that stays in one head. And it is the same thing I wrote about in “Watch how I do it” is not training, only here it is not the trainer who is missing. What is missing is the place where their knowledge could have been written down.

An experienced driver should not be the permanent correction of a system that does not learn. Not because they have no value. Because they have too much value for it to stay in only one place.

Four steps from the driver to the plan

I have not built a process like this, and I do not know whether one existed where I worked. But today I would focus on how information from drivers is turned into an improvement in planning, and I would recommend these four steps.

1. A simple way to report a problem that repeats on a tour. Simple means the driver does it in a minute, at the end of the tour, without a long form. If reporting takes longer than correcting in silence, silence will win.
2. The difference between a temporary obstacle and a recurring error. A road closed today is not a planning error. The same wrong entrance every week is. The first passes by itself, the second comes back.
3. Someone responsible who verifies the information before the tour is changed. Not because the driver is not trusted, but because a change to the plan affects everyone who drives it afterwards.
4. A way for the confirmed correction to be available to the other drivers too. Otherwise the correction has solved the problem for one person, for one day.

And as in the first essays: it has to be clear who can propose the change, who approves it, and how it is recorded.

What I would do differently

As a dispatcher, communicating with drivers and following up problems during delivery are part of my work. But what I would do differently today is not in the conversation. It is after it. A problem a driver tells me about and that I solve on the phone is solved for that day. If I do not write it down somewhere someone else can find it, tomorrow someone else will solve it again, on the phone.

The problem is not only when the system does not know the tour. The problem is when the organisation does not learn from the people who drive it every day.

A problem that repeats on a tour

The card below is for a problem drivers meet more than once on the same tour or in the same area: what they meet and where, whether it is a temporary obstacle or a recurring error, who verifies it, who proposes and who approves the change, and how the next driver gets the correction. Fill it in when you hear it the second time, not the fifth. Delay Analyzer shows you where the delay starts, at the ramp or on the road, so that a problem with the tour can be told apart from a loading problem before you take it to the plan.

The tour the driver does not understand is sometimes an error in the plan, sometimes something nobody has explained to them. The difference is made only when someone listens, verifies it and writes it down.