I have seen last mile from three places. As a parcel courier, as a dispatcher with more than 70 drivers on a shift, and then as Team Leader and Area Manager, with four Shift Leaders, eight dispatchers and around 250 drivers under my responsibility indirectly. The same operation looks completely different from each of them.
As a courier, the day is very concrete. You take the route, load the van and go out on the road. There you do not care about the theory of the process. You care whether the parcels are in the right order, whether the route is well organised, whether you have enough time, and what you do when traffic, an address, a customer or a problem with the load breaks your plan. From that place I understood something that is not always visible from the office: a small problem in the warehouse can cost a lot of time to the person on the road.
As a dispatcher I no longer looked only at my own route. I had to check how the routes were split, how the drivers set off, the delays, the live problems and the communication throughout the day. In that position you start to understand that an individual problem can in fact be a system problem.
As Area Manager I could not solve every problem myself. My job was to see the patterns: why the same problem keeps repeating, at which point in the process it starts, and what has to change so that it does not happen again tomorrow.
The driver sees the problem. The dispatcher sees its effect on the whole shift. The manager has to understand why the problem exists.
The main KPIs I worked with were Loading Time, Delays, Incomplete and Damage, together with on-time performance. Each has its own moment in the day. Before departure, one of the most important was loading time. During the day you watched the delays and the incidents building up live. After the operation, they all had to be read together, not separately, because a loading problem in the morning could show up later as a delay, and a mistake in preparation could show up as an incomplete order at the customer.
A concrete example. The target for loading time was around 20 minutes. For a period we had an average of around 24 minutes. In other words, we were creating delay right at the start of the day.
It would have been very easy to say: the drivers are loading slowly. But 24 minutes do not tell you by themselves why you have a problem. You have to go to the ramp and see what is physically happening. There you found missing bags, trolleys that did not match the route or the floor as they should, problems with the order of the stops, loads that took longer than normal, drivers waiting. You looked at the scanners, at how dry, chilled and frozen were organised, and at whether the driver was taking everything before leaving.
The work was to go to the process, not to blame the driver straight away. We worked on very concrete things: better checks of the trolleys and the routes, scanners closer to the ramps, pre-scanning, final checks for completeness, a clearer organisation of dry, chilled and frozen, and ramp audits. In other periods loading time came down to the range of 16 to 19 minutes.
For me the lesson was not "we brought a KPI down". The lesson was that a KPI improves when you change the process that produces it.
The same goes for incomplete orders. For me, Incomplete was not just a percentage. It meant that a customer had not received their full order. You had to understand whether the product had been missing in the warehouse already, had been put in the wrong trolley, had been left on the ramp, or whether something had happened during delivery. With damage too: the number tells you how many cases you have. It does not automatically tell you why you have them.
That is why the daily KPI review, with the operational structure and the responsible leads, could not be just reporting. We looked at what had happened with damage, incomplete orders, delays, loading time and the main problems of the operation. The question was not only what the KPI was. The question was why.
If you had 58 delays, it was not enough to say 2.8% delays. You had to understand which routes they came from, which causes kept repeating, and which part of the problem was in our hands. We used Pareto to see which reasons created the largest part of the problem, and 5 Why so as not to stop at the first answer. At the end there had to be a concrete action: who takes the problem, what we change and what we check the next day. Otherwise the meeting was just reporting.
Pressure in last mile comes from many directions at once. You have the departure time. You have the customers waiting. You have the targets. You have the drivers. You have the warehouse. You have missing bags. You have one route running late while another has just created a problem. And the operation does not stop so that you can analyse the problem calmly.
My lesson was that pressure must not be passed down as panic. If a manager sees a red KPI and the first reaction is to shout, very soon you have 20 people trying to look busy, but nobody is solving the cause. I tried to turn the pressure into concrete questions. Where is the problem? How many cases do we have? Which reason keeps repeating? What can we do now? What do we need to change for tomorrow?
For me, accountability does not mean looking for someone to blame every time a number turns red. It means finding who owns the problem and following it until it is closed.
If you ask me which KPI misleads you most when you look at it only as a number, I would say Delays. The number gives you the impression that you are measuring only the driver's delay. But the delay may have started much earlier. At loading. With a missing bag. With a mixed-up trolley. With the order of the route. With an operational problem before the driver had even left the ramp. Then on the dashboard you see only one driver who is "late".
If you manage only the number, you can punish the wrong person. If you manage the process, you start to understand where the delay was really created.
A KPI tells you where to look. It does not tell you by itself what happened.