The real cost of a product rejected for poor quality
The lost material is the only part that reaches your accounts. And it's almost always the smallest part of the bill.
When you add up what a rejected product has cost you, you nearly always count three things: material, time and direct money.
Enter your email and you will get an alert when something new is published. Unsubscribe anytime.
That's the tip of the iceberg. And it's why the cost of poor quality is underestimated in practically every company I know.
The three blocks
What you can see — material lost, hours thrown away, energy, transport. It's the only part that reaches the accounts. And that's why it's the only part discussed in the meeting.
What costs time and you don't book anywhere — stopping the line, hunting the cause, reworking, rescheduling production, handling the return, redoing paperwork, dealing with the customer, repeating the inspection. Every one of those hours is being worked by somebody who had something else to do. Somebody of yours.
What never gets accounted for at all — the order that ships late. The customer who marks you down on the supplier scorecard. The one who starts splitting volume with a competitor "just in case". The extra audit. And the wear on a team that knows perfectly well that batch went out wrong.
| Block | Examples | Does it reach the accounts? |
|---|---|---|
| Direct | material, hours, energy | yes |
| Indirect | rework, investigation, rescheduling | rarely separated out |
| Hidden | reputation, share, customer trust | never |
The rule worth remembering
The cost of a defect multiplies with every stage it advances.
Catch it at the operation and it's an adjustment. Catch it at final inspection and it's rework or scrap. Let the customer catch it at his place and it's a complaint. And if product has to be pulled from the market, the number stops resembling any of the previous ones.
Which is why investing in catching it earlier nearly always justifies itself. Because it isn't competing against the cost of the material. It's competing against the cost of the complaint.
The bridge to 2026
This is probably the soundest economic argument I know for industrial AI. And I don't need to exaggerate it.
Checking a hundred per cent of pieces inside the line wasn't viable with human inspection. It was done by sampling because there was no other way.
But sampling has an arithmetic consequence almost nobody says out loud: the defect that isn't in the sample goes to the customer. That's all there is to it.
When the system looks at every piece and acts inside the process — rejects, warns or stops the line — that defect stops advancing a stage. And note, that doesn't cut your cost of lost material. It cuts the other two blocks, which are the big ones.
At an aerospace plant we work with, the system catches oil leaks on the test benches that had gone years without being reliably detectable. The saving, obviously, isn't in the oil.