The real cost of the last mile (and where it hides)

The last mile is the shortest leg of a delivery’s journey and reliably the most expensive one. The reason is structural: every other leg moves parcels in bulk, and the last mile moves them one door at a time. A container amortises its cost over thousands of items; a doorstep amortises over one.

That much is well known. What is less discussed is where the cost actually hides in a normal operation, because it rarely shows up as a line called “last mile”.

The four hiding places

1. The second attempt. A failed delivery does not cost one visit; it costs the failed visit, the return handling, the re-planning and the repeat trip, plus the customer service conversation wrapped around all of it. First-attempt success is the single most leveraged number in the last mile, which is why notifications and live ETAs matter: customers who know when to expect a delivery are home for it.

2. The unmeasured stop. Drive time is visible on any map; service time is not. Minutes at the door (finding the entrance, waiting at reception, searching the van) are invisible in plans that only model driving, and they accumulate into routes that mysteriously never finish on time. Measuring real service times, then planning with them, closes the gap between the plan and the shift.

3. The empty mile. Poor sequencing does not look like waste; it looks like driving. Crossing the same area twice, backtracking for a time window that could have been planned, sending two vans where one full one would do. This is the cost route optimisation attacks directly.

4. The paper trail. Proof that lives on paper costs twice: once in the office time spent chasing and filing it, and again in the disputes that cannot be settled because the note is gone. Electronic proof of delivery converts both costs into a record that files itself.

Measuring it honestly

A useful last mile cost number needs the operation’s own data: stops per route, first-attempt rate, minutes per stop, failed-delivery reasons. Generic industry percentages make slides, not decisions. The practical path is to instrument the operation (every stop timed, every failure coded, every proof attached) and let the report answer with your numbers.

That instrumentation is a side effect of running deliveries through a system like MLogTech: the metrics fall out of the operation, instead of being a project on top of it.