Why does the delivery date always slip in the last week?

A familiar scene: three days before shipment somebody notices the parts are still in grinding, the customer is called and a new date is given. At the end of the month the same thing repeats on another order.

The delay did not occur in those three days. It usually occurred in the first week: the material arrived two days late, an urgent job jumped the queue and took the machine, and the outsourced process came back a day late. Each one small; together, the whole delay.

Three facts in three places

In most shops the promised date lives on the order, the real status of the job at the machine, and whether the material arrived on the purchasing desk. Until those three come together the delay is invisible — and they usually come together because somebody asks, which means late.

When can it be seen?

When plan, capacity and material sit on the same schedule, the delay is visible on the day it forms. An urgent job jumping in shows as capacity overload; missing material is flagged on the plan; a late return from an outsourced process turns that operation red. None of it is a surprise.

This does not need a clever algorithm: capacity from the real shift calendar, planned durations corrected by past performance, and material requirements compared against stock are enough.

Record the reason

When a date is missed, the reason has to be recorded: late material, a breakdown, rework, or a duration that was never realistic. Three months later that record is the most valuable thing you have when quoting the same customer for the same part again.