A bad part costs you three times
Scrap and rework feel like a cost of doing business, so they hide in plain sight. But a part made wrong costs you three times over: the material you put into it, the time and capacity you spent making it, and the time you now spend fixing it or making it again. And rework almost always jumps the queue, so it pushes good, paying work back behind it. The bench in the corner where bad parts pile up is losing far more than the pile suggests.
Measure first-pass yield, honestly
The one number that tells the truth is first-pass yield: of 100 parts started, how many came out right the first time with no touch-up, no rework, no adjustment. Most floors have never counted it, and most are surprised. Under 95 percent is quietly costing you twice on one in twenty parts. You cannot improve what you will not count, so count it for one real run before you change anything.
Find where the defect is born, not where it is caught
A defect is usually caught far downstream of where it was created. The instinct is to add an inspection step where it is found, which catches more bad parts but makes none fewer. Walk it backward instead: for your top one or two defects, trace to the exact operation where the part first went wrong. That is the only place a fix changes the number.
The usual causes, in order
On most floors the biggest sources are the same handful: a setup or first-article that was signed off on a bad part, so a whole run is wrong; a worn or wrong tool nobody flagged; a spec or drawing the operator is interpreting differently than the next shift; and material that was out of spec before it was ever touched. None of these are fixed by inspecting harder. They are fixed at the source: a real first-article check, tool-change triggers, one clear standard, and incoming checks on the material that actually causes problems.
Make the standard impossible to miss
Rework thrives where ‘right’ lives in someone’s head. A marked setting instead of ‘adjust until it looks right’, a go/no-go gauge instead of a tape measure and judgment, a photo of a good part at the station, a fixture that only loads the correct way. Small, cheap, shop-floor changes that make the good outcome the easy one and the bad one hard to produce.
What good looks like
A floor that measures first-pass yield, fixes defects at their source, and builds the standard into the station instead of the operator’s memory usually recovers most of its scrap and rework without new equipment. The money was never really in the bad parts. It was in the capacity they ate and the good work they pushed back.
Finding your two biggest sources of scrap on an unfamiliar floor, in a single walk, is a large part of what a walkthrough does.
Request your free walkthrough and we will find where your floor is losing the most time, no charge.
