Metrology, the scientific study of measurement, has never been faster or more precise. Yet many quality problems that land on a CMM table were decided weeks earlier, when someone measured against a drawing that was no longer the approved one.

Closing that gap isn’t a measurement problem. It’s a revision problem. The measurement was right. The drawing was wrong.

Every quality engineer has a version of this story. A batch of housings fails first article inspection on a bore diameter. The CMM report is correct. The program is correct. The part is correct, too, against Rev. B. But engineering released Rev. C eight days ago, after a bearing supplier changed, and tightened the bore tolerance. The inspection bench never received it.

Nobody made a measurement error. The information simply stopped moving at a handoff, and the cost showed up later as scrap, a nonconformance report, and an awkward call with the customer.

Others have made the same point, but from different angles, in the posts The Real Cost of Your Quality and The Hidden Cost of Poor Quality in Regulated Manufacturing. The real cost of quality sits in how information flows, not only in how fast we measure. The revision gap is one of the most concrete places where cost appears.

Where the gap opens

With most manufacturers, the drawing and the model travel from design to inspection through three handoffs. The gap opens at each one.

1. Release: A design is “released” when someone says it is. In small and midsize shops, that often means a file moved into a folder called released, a PDF saved with Rev. C in the name, or an email to the shop floor. None of these are states that can be checked later. They are conventions, and conventions break when people are busy.

2. Export: Inspection rarely works from native CAD. It works from a PDF drawing for the inspector and a STEP file for the CMM programmer. Those files are exported by hand, often at different times, sometimes from different versions of the model. The drawing says Rev. C. The STEP file was exported before the last change. Both look official.

3. Feedback: When inspection finds a problem with the design itself, such as a datum that can’t be fixtured or a tolerance the process can’t hold, the finding goes back as an email, a marked-up printout, or a hallway conversation. It rarely stays attached to the revision it concerns, so the next revision repeats it.

Why faster metrology doesn’t close the gap

Inline sensors, automated CMM programming, and AI-assisted inspection all shorten the time between making a part and knowing whether it conforms. They assume the reference is right. If the nominal geometry or the drawing behind a program is one revision behind, a faster measurement system produces confident results against the wrong target, faster.

This is the downstream side of the “shift-left” idea. Moving quality upstream means embedding inspection intent in design. It also means making sure the design that reaches inspection is the one that was approved.

What a closed revision loop looks like

Closing the gap doesn’t require a large PLM program. It requires a few disciplines that can be checked rather than remembered.

One released state, visible to everyone: Every drawing and model carries an explicit state, such as In Progress, Under Review, or Released, and a revision that changes only through approval. Quality can see the state without asking engineering.

Named sign-off on every release: Each release records who approved it, when, and why. When an auditor or customer asks who released Rev. C, the answer is on the release, not in an inbox.

Quality on the approval, not after it: Quality engineers are named on releases that affect inspection, either as approvers or as observers who see each release as it happens. Problems with inspectability surface before the revision is final.

Production files generated from the release: PDF, STEP, DXF, and DWG files are produced from the released version at the moment of release and stored against it. The drawing the inspector reads and the model the CMM programmer imports come from the same revision by construction, not by care.

Feedback attached to the model: Inspection findings are recorded on the model or drawing they concern, where the design engineer sees them in the next revision.

Small manufacturers feel it most

Large OEMs usually have dedicated document control. Small and midsize manufacturers, contract shops, and engineering consultancies often do not. Their designs live on network or cloud drives, their engineers are spread across sites, and their quality team may be one or two people who also handle incoming inspection and supplier issues. For them, the revision gap isn’t an edge case. It’s a weekly risk.

The timing is useful, too. ISO 9001:2026 has been described as evolution rather than revolution, and the requirement to control documented information, including changes, carries over. With many manufacturers reviewing their quality systems for the transition, it’s a good moment to check how a revision actually travels from design to the inspection bench, and where it can stop.

A practical starting point

Pick one recent nonconformance and trace it backward. Which revision was the inspector working from? Where did that file come from? Who approved the revision that should have been used, and how did the inspection team find out? If the answers involve folder names, email threads, or someone’s memory, the revision gap is open.