The cost of quality escapes in manufacturing is not an inspection problem, and the proof is sitting in your last ten findings. Cost of poor quality runs 5 to 25 percent of sales depending on product complexity, according to Manufacturo’s April 2026 analysis — and in high-complexity environments like aerospace and defense, TBM Consulting has observed cost of quality approaching 10 percent of revenue. Those numbers get discussed in quality reviews as if they were the price of imperfect people. They are mostly the price of imperfect documents.

What is the 1-10-100 rule in manufacturing quality?

The 1-10-100 rule says a defect costs one dollar to prevent, ten dollars to catch and correct internally, and one hundred dollars once it reaches the customer. Every stage a defect survives multiplies the bill: material, labor, machine time, disposition cycles, engineering investigation, expedited replacement, and — at the customer — warranty, chargebacks, and trust.

Most plants know the rule. Watch where they spend anyway. The quality budget conversation ends in more inspectors, tighter final test, and faster containment — all ten-dollar money. The one-dollar column, prevention, gets a training line item and a poster.

Manufacturo’s own phrasing is the honest version: many manufacturing systems are “built to find problems, not prevent them.”

Where are your findings actually created?

Here is the exercise that changes the conversation. Pull your last ten internal findings and walk each one upstream, past the workstation where it was caught, to where it was born.

A typical trail: the instruction specifies 4.2 bar because it was written at 4.2 bar in 2021. The material changed in 2023. The person who worked out the new setting taught it by voice. First shift knows it. The weekend crew does not. The part fails final test on Monday. The finding says “operator error.” The corrective action retrains people to the document that was wrong to begin with — which is how the same finding comes back next quarter wearing a different part number.

Run that trail on ten findings and most plants discover eight of them were created at the same address: a work instruction describing a different job than the one being run. The person standing nearest the defect gets the finding. The document that caused it gets a review date.

Why doesn’t more inspection fix it?

Because inspection prices the detection stage while the defect factory keeps running upstream. Every uncaptured method change is a machine that produces future findings on schedule. You can staff final test heavily enough to catch almost everything — aerospace does — and the cost shows up as that near-10-percent-of-revenue quality burden, growing a little every year, while the creation rate never moves.

Inspection is necessary. In AS9100 environments it is mandated, layered, and non-negotiable, and it should be. The argument here is narrower and more uncomfortable: a meaningful share of what your appraisal budget buys is compensation for documentation that stopped matching the floor. That share is measurable, and almost nobody measures it.

Catching is the most expensive form of knowing.

Run the caught-vs-created audit this week

Sixty minutes, one sheet of paper, ten findings.

1. Pull your last ten internal quality findings. Any mix — final test failures, in-process rejects, receiving escapes that got caught downstream.
2. Column one — where was it caught? Name the station.
3. Column two — where was it created? Walk it upstream and classify: document wrong (the instruction describes an outdated method), method uncaptured (the good technique lives in someone’s head or on one shift), training gap (the document is right, the person never got it), or true slip (the document is right, the person knew it, the hands missed).
4. Compute the ratio. Findings created upstream of where they were caught, out of ten.
5. Price it. Take your annual appraisal spend — inspection hours, test capacity, containment — and multiply by that ratio. That figure is what documentation drift costs you at the ten-dollar stage, before a single escape reaches a customer.

If the ratio is eight of ten, your inspection budget is subsidizing your documentation gap. The subsidy renews automatically every year.

Stop buying one-dollar problems at ten dollars

The findings are real. The inspection hours are real. The wrong document is still on the floor. Unless the method that actually makes the good part gets captured and validated with the people who run it, column two never changes — and next year’s quality budget meeting reaches the same conclusion at a higher number.

Diego Echenique is the CEO and co-founder of SenseiLab. He has spent 20+ years in manufacturing around the world — launching plants, leading operations, and running Lean and Six Sigma transformations across automotive, mining, and heavy industry.