Four in the morning

It was a few minutes past four in the morning when I rolled the rig into my third delivery of the night, and nobody at the receiving end had any idea I was on my way. That was not bad planning. That was how the system actually worked. I drove 24-metre Scania rigs for three years (2004 to 2007), and most mornings started in exactly that shape: an address on the manifest, an approximate time, and no contact person at the other end.

I turned into the industrial estate. The gates were closed, there was no sign, and there were no instructions. I called the dispatcher, who called the warehouse, who called the security guard. The guard came out after twenty minutes, looking like he had been pulled from a chair he did not want to leave. Meanwhile I stood still with the engine idling and the clock ticking toward the next delivery.

This was not a bad day. It was a normal day. The waiting times were not exceptions, they were the system.

What the loading bay actually looked like

The loading bay was occupied, or it was the wrong bay, or it was the right bay but nobody had staged the goods. I reversed in, waited, reversed out, drove around, reversed in again. Sometimes loading took forty-five minutes. Sometimes three hours. The difference rarely depended on the volume of goods. It depended on whether someone had prepared for the arrival.

There were terminals that worked. Not many, but a few. They had one thing in common: the driver knew what to expect before arriving (which bay, what time, who was receiving). It was not advanced. It was basic coordination, and it was missing at most places.

I know this sounds like a small problem at the level of an individual driver and an individual gate. It is not small at the level of a fleet. A driver who waits costs money. The truck costs money. The goods that do not arrive on time cost money downstream. The bay’s inefficiency spreads backward through the chain, so the truck that lost two hours at gate A also arrives late at gate B, which makes gate B believe it has a capacity problem when what it really has is an inherited coordination problem.

Three hours of waiting times forty drivers per day times one year. That adds up to thousands of hours that nobody sees, nobody measures and nobody reports.

The information gap nobody measures

There is a specific kind of cost hiding inside what I am describing, and it is worth naming. The problems did not show up in any system. No TMS registered the waiting time. No KPI captured the coordination gap. In the PowerPoint version we delivered “on time.” In reality we sat idling outside a closed gate, then logged the delivery as completed once the goods finally moved.

So the operational cost is real, but the information gap is the more interesting thing. An organisation that does not see its own friction cannot improve it. Worse, an organisation that has accepted that friction starts to believe its dashboards are the territory.

I stopped driving trucks in 2007. I got two master’s degrees from KTH and Stockholm University. I started a tech company that built tools to solve exactly this problem, and fifteen years later I sat at Volvo Group looking at the same pattern in a much larger system. The gap between the person doing the work and the person making the decisions is not a communication problem at the personal level. It is a structural problem. The information sits with the driver, and the decision-maker sits six organisational layers away. It takes weeks, sometimes months, before an operational insight reaches someone with the mandate to change anything.

That costs. Not in abstract terms. It costs in waiting time, in fuel, in drivers who quit, and in accident risks nobody measures.

Why 2025 is a better year for this problem than 2005

This is the part that has changed, and it is the part that should change how operations leaders think about the gate at four in the morning.

In 2005 the tools to close this gap were heavy. You needed a TMS implementation, an EDI integration with each customer, a portal that someone had to log into, and a change project to make any of it stick. The economics rarely worked for a single terminal trying to fix its own coordination problem, so most terminals did not try.

In 2025 the economics are different. A shared arrival window between dispatcher and warehouse can be a single message in a channel both sides already use. A pre-arrival check (“which bay, what time, who is receiving”) can be generated automatically from the manifest and pushed to the driver’s phone the night before. AI can read free-text dispatch notes and flag the deliveries that are missing a contact person before the truck leaves. None of this requires a transformation programme. It requires somebody with operational credibility to look at the gate at four in the morning and ask why it is closed.

The technology is not the constraint anymore. The constraint is whether the organisation has a channel back from the driver to the person who can change the standing instruction.

What an operations leader should actually do

If you run operations, logistics, or a manufacturing site that receives goods, here is what the four-in-the-morning story implies.

First, measure waiting time at the gate. Not estimated, measured. If your TMS does not capture it, instrument it manually for two weeks with a clipboard and a timestamp. You will find that the variation between sites is larger than you expect, and that the worst sites do not know they are the worst sites.

Second, give the driver a channel back. A shared inbox, a number that gets answered, a five-minute call after every shift with the dispatcher. Whatever fits the operation. The point is not the format, it is that information from the gate reaches a person with the mandate to change the standing instruction within days, not months.

Third, treat pre-arrival information as a deliverable, not a courtesy. The receiving site should know which truck is arriving, which bay it should go to, and who is receiving, before the truck leaves the previous stop. If your process cannot produce that information reliably, the gap is in your process, not in the drivers.

Fourth, when you evaluate AI or digitalisation projects in your operation, ask one question first: does this close the information gap at the gate, or does it add another layer that the driver has to wait for? Most AI pitches I see in logistics are the second kind. The useful ones are the first.

The driver knows how the system works. The question is whether there is a channel back, and whether someone at the top of the chain is willing to listen when the answer is uncomfortable.