IQ, OQ, PQ explained simply

The first time someone hands you a validation protocol for a new line, IQ, OQ, and PQ can read like three different ways of saying the same thing. They’re not. Each one answers a different question, and answering them out of order is the single most common reason a line fails validation on the first attempt.

IQ — Installation Qualification: is it built the way the drawing says?

IQ checks that the equipment, as physically installed, matches what was specified and purchased — correct components, correct wiring, correct software version, correct documentation on file. This sounds basic, and it is, which is exactly why it gets rushed. I’ve seen OQ testing start on a station where a sensor was swapped for a “compatible” substitute during installation, with nobody updating the documentation. The OQ results were meaningless until IQ caught the discrepancy — backwards, and expensive.

OQ — Operational Qualification: does it do what it’s supposed to do, across its range?

OQ tests the equipment’s function across the full range of its intended operating parameters — not just the nominal setpoint, but the edges. If a vision system is supposed to detect a defect at a given lighting level, OQ tests it at the low and high end of the acceptable lighting range, not just under ideal studio conditions. This is where a lot of automation system architecture decisions get tested for real: did you actually design margin into the station, or just enough to pass under perfect conditions?

PQ — Performance Qualification: does it work, repeatably, with real product, over time?

PQ is the qualification that uses actual product (or a validated surrogate) running through the actual process, repeated enough times to demonstrate statistical confidence — not three good runs, but a sample size that means something. This is where you find problems that only show up under sustained operation: fixture wear, thermal drift, accumulation effects that a single OQ run would never reveal.

Why the order matters

Each qualification stage assumes the previous one passed cleanly. Running OQ before IQ is fully closed out means you’re testing function on equipment that might not even match its own specification — any failure you find could be a real functional problem or just an installation error, and you won’t know which until you go back and redo IQ anyway. The same logic applies to PQ: performance data collected before OQ is closed out doesn’t tell you anything trustworthy about repeatability, because you haven’t yet confirmed the operating envelope you’re claiming to repeat within.

For a regulated line, the qualification sequence isn’t bureaucracy on top of the engineering — it is the engineering, written down in a way someone other than you can audit. Designing the station with that audit trail in mind from day one is a lot cheaper than retrofitting documentation onto a line that’s already running.