An operator pulls a part off a machine, checks it against the print, and it is out of tolerance by a few thousandths. What happens in the next five minutes says more about a shop's quality system than almost anything else that could be observed on a floor walk. Does the part get set aside in a clearly marked location, or does it sit on the same bench as conforming parts with a mental note to deal with it later? Does the operator know whether they are authorized to scrap it themselves, or does it need to go somewhere for a disposition decision? Clause 8.7 of ISO 9001:2015 covers control of nonconforming outputs in a few paragraphs, but the gap between those paragraphs and what actually happens in the first five minutes is where most nonconformance programs quietly fail.
The problem with informal containment
Most shops have a nonconforming material process on paper: identify, segregate, disposition, document. Most shops also have an informal version that runs alongside it under time pressure, where a part gets mentally flagged as bad but stays in the flow of production a little longer than it should, because moving it to a proper hold area takes an extra minute and the operator is trying to hit a shift target. That informal gap is where mixed lots happen, where a nonconforming part ships alongside conforming ones, and where the eventual investigation into how it happened traces back not to a process failure but to a five-minute lapse in physical segregation that the written procedure assumed would never occur.
Physical segregation is a deceptively low-tech control, but it is one of the most effective ones a shop has, precisely because it does not depend on anyone remembering to update a system or fill out a form before the part moves. A clearly designated, physically distinct hold location, with rules that nothing leaves it without a documented disposition, closes the gap that policy language alone cannot close on its own.
Disposition authority: who actually gets to decide
Clause 8.7 requires that nonconforming outputs be dealt with through correction, segregation, containment, return, or suspension of provision, and that disposition authority be defined. In practice, this is where shops either over-centralize or under-define. Over-centralizing means every nonconformance, no matter how minor, waits for a quality manager who may be off-site or buried in other work, which creates a bottleneck that pressures people to route around the system rather than through it. Under-defining means disposition authority is effectively whoever happens to be standing nearby, which produces inconsistent decisions on materially similar nonconformances depending purely on who was on shift.
The workable middle ground most shops eventually land on is a tiered authority matrix: operators can disposition certain low-risk, clearly defined nonconformances themselves, following a documented rule, while anything ambiguous, customer-facing, or safety-related escalates to a defined quality role. Getting this tiering right requires being honest about which nonconformances are genuinely routine and which only look routine until someone examines them closely, and that distinction is worth revisiting periodically rather than setting once and leaving unexamined for years.
Documentation that survives being looked at later
A nonconformance record filled out quickly, in the moment, often captures the immediate disposition but loses the surrounding detail that becomes important if the same issue recurs: which machine, which operator, which lot of raw material, what the actual measured value was rather than just pass or fail. Six months later, when a pattern of similar nonconformances starts to emerge and someone needs to look for a common thread, records that only captured the disposition decision and not the underlying detail make that pattern invisible even though the data existed at the time each individual record was created.
This is one of the clearer arguments for structured nonconformance capture over free-text notes on a paper tag. When nonconformance records are tied consistently to work order, operation, and material lot data rather than recorded ad hoc, spotting a recurring pattern becomes a query instead of a manual review of a filing cabinet. A manufacturing qms that captures this structured detail at the point of disposition, rather than relying on someone transcribing a paper tag into a spreadsheet days later, tends to catch repeat patterns considerably earlier than shops still working from disconnected paper records.
Closing the loop back to prevention
Handling an individual nonconforming part correctly is necessary but not sufficient. Clause 10.2 expects the organization to evaluate whether corrective action is needed, which means every nonconformance, even a minor one dispositioned by an operator under a routine authority rule, needs some visibility at a level where patterns can be evaluated across the whole floor rather than one part at a time. A shop that handles every individual nonconformance correctly but never aggregates the data to look for recurring root causes is still missing the second half of what clause 8.7 and clause 10.2 are jointly asking for, and that second half is usually where the real cost savings live.