Rework consumes capacity twice and is almost never measured. Start counting it and the priorities become obvious.
Ask a small manufacturer for its rework rate and you will usually get an estimate. Ask for the record and there often is not one. Rework is absorbed by the people who spot it, corrected quietly and never counted β which is why it remains the largest unclaimed productivity gain in most SMEs.
A job done twice occupies the machine, the operator and the floor space twice, while the customer pays once. At a fifteen per cent rework rate, roughly one working day in seven is spent producing nothing new. That is capacity you already own and are giving away.
Before improving anything, measure. A tally sheet at each workstation recording defect type and stage takes seconds per event and produces, within a fortnight, a ranked list of causes. Almost always a small number of causes account for the majority of events, and the top two are usually addressable within the month.
Where a defect arises from ambiguity, clarify the procedure. Where it arises from variation, tighten the control or the tooling. Where it arises from a step being performed out of sequence, change the sequence so the wrong order is impossible. Design the error out before you train around it, and train before you inspect.
Sustained improvement comes from making error physically or procedurally difficult: a jig that only accepts the part one way, a form that cannot be submitted with a field blank, a kit prepared so the wrong component is not to hand. Reliance on vigilance fails on a busy day.
None of this requires capital. It requires two weeks of honest counting and the discipline to act on the top of the list rather than the loudest complaint.
If this describes a problem you are living with, an enquiry is the fastest way to find out whether it is worth engaging help.
Send an Enquiry