Learning to Say No to a Deadline

The Challenge

The date had been communicated. Not widely, but to enough people that moving it would be noticed and would cost something real.

The problem was a component that was arriving at acceptable quality about seven times out of ten. Sorting was possible. Sorting at the volume required, in the time available, was not.

The Process

The first move bought three weeks and was framed internally as a supplier issue, which it partly was.

The second move bought two more and could not be framed as anything. By then it was clear the acceptance rate was not improving, because the cause was a tooling wear pattern that would keep getting worse until the tool was replaced, and replacing it had a lead time longer than the remaining season.

Important Decisions

The decision to abandon rather than move a third time was made on the argument that a third move would establish the date as negotiable.

That reasoning is recorded in a single paragraph written the same week, and it is worth quoting because it was written before anyone knew whether it was right: a date that moves twice is late; a date that moves three times was never a date. We would rather be a company that ships nothing this season than one whose dates mean nothing.

The cost was real and was absorbed. Nobody has since argued it was the wrong call, though a few people argued it at the time.

Quality Standards

What came out of it was not a quality standard but a scheduling one: a date is not communicated externally until the component with the longest lead time has been received and inspected, not ordered.

This makes announcements later and less exciting. It has also meant that no date communicated since has moved.

The standard is unglamorous and it holds because it is mechanical — it does not require anybody to be disciplined in the moment, which is when discipline is least available.