A flawless prototype proves only that you can make one good part. It does not prove the process is under control. When an auditor asks how you guarantee the next batch will meet specification after a tool change, a material swap, or a shift change, hope is not a viable defence.

Process validation is the systematic, documented evidence that a manufacturing process consistently produces output meeting predefined quality requirements. It is mandatory under IATF 16949 for automotive, AS9100 for aerospace, and FDA 21 CFR Part 820 for medical devices. In every sector where failure carries a high cost, validation bridges the gap between capability and certainty.

Without validation, you are gambling on consistency. Validation replaces guesswork with statistical proof. It demands that you prove your equipment is installed correctly, operates within limits across all conditions, and yields repeatable results under normal production loads.

IQ — Installation Qualification: Verifying the Foundation

Installation Qualification (IQ) is the first step. It answers a strict question: is the equipment installed exactly per the manufacturer specifications and your site requirements? IQ does not test whether the machine produces good parts; it tests whether the system is physically and digitally ready to begin operation.

During IQ, engineering teams systematically verify hardware placement, utility connections, software versions, and safety interlocks against the design specifications. If the machine requires 8 bar of compressed air with a specific tolerance, you verify the actual line pressure and connection diameter. If the system integrates with a SCADA or MES network, you confirm the communication protocol is live and stable.

I once validated a new crimping line for an automotive plant where the supplier's installer connected the machine to a 6 bar air supply because it was readily available. The specification strictly demanded 8 bar. The machine ran, but under increased load, it would have failed to maintain consistent crimping force.

We caught this defect on day one of IQ. The supplier rectified the fault within 48 hours. Had we skipped IQ, the pneumatic shortfall would have surfaced weeks later during Performance Qualification, causing massive delays and scrap. One day of rigorous inspection saved months of rework.

Validation begins long before the first cycle runs; it starts with verifying the exact utilities and interfaces specified in the design.
Validation begins long before the first cycle runs; it starts with verifying the exact utilities and interfaces specified in the design.

OQ — Operational Qualification: Testing the Edges

Operational Qualification (OQ) proves the equipment functions according to process specifications across all operating conditions. Here, the machine runs. But it does not merely run under ideal settings; it runs in prescribed edge-case and limit scenarios.

OQ tests what happens at the upper and lower tolerances of temperature, pressure, speed, and feed rates. It verifies that alarms trigger correctly when parameters breach limits and that safety interlocks physically halt the process when required. Reproducibility is verified by running identical cycles and confirming the outputs remain stable.

Engineers use OQ to stress the system. If a thermal chamber is specified to operate between 40°C and 80°C, OQ tests the physical response at exactly 40°C, 80°C, and during rapid ramp-up transitions. Functional tests verify that every automated axis, valve, and sensor performs exactly as the functional specification dictates.

OQ does not yield final commercial products. It acts like a dyno test for an engine. You prove the mechanical and logical tolerances hold up under stress before the process ever touches the factory floor.

The IQ/OQ/PQ Sequence

  1. 01Installation QualificationVerify utilities, hardware, software, and documentation against design specs.
  2. 02Operational QualificationTest equipment functionality at upper and lower parameter limits and alarm states.
  3. 03Performance QualificationProduce consecutive batches under normal conditions to prove long-term stability.
  4. 04Continuous Process VerificationMonitor critical parameters live using SPC and control plans post-launch.
Validation follows a strict chronological sequence; skipping phases inevitably corrupts the statistical evidence.

PQ — Performance Qualification: Proving Serial Capability

Performance Qualification (PQ) is the final hurdle before full production release. It answers the critical question: can this process repeatedly produce products meeting all requirements under normal, everyday manufacturing conditions?

PQ leaves the laboratory behind. It executes on the shop floor using actual operators, standard raw materials, and standard cycle times. Standard practice requires the successful completion of at least three consecutive production batches. All parts from these batches must fully comply with technical specifications.

PQ documents process consistency. It demands statistical proof that the process is capable, with Cpk and Ppk indices meeting or exceeding the required thresholds. It proves that different operators across morning, afternoon, and night shifts deliver identical quality, and that process performance does not drift as the cumulative part count rises.

Statistical Proof: Why Cpk Is Not Just a Number

Process validation without statistics is just a collection of opinions. Statistical tools convert raw production data into scientific evidence. They prove mathematically that your process is centred, stable, and capable of holding tolerance over time.

Cpk (Process Capability Index) measures whether your process fits within the tolerance limits and remains centred. A Cpk of 1.33 is the baseline minimum for standard parameters. For critical-to-quality characteristics or safety-relevant features, customers and standards typically demand a Cpk of 1.67 or higher.

Ppk (Process Performance Index) evaluates long-term variability. If your Ppk significantly lags behind your Cpk, the process is not stable. You use X-bar/R or I-MR control charts to monitor stability in real-time, ensuring points remain within statistical control limits throughout the production run.

Analysis of Variance (ANOVA) further validates the data by comparing variation between batches, operators, and shifts. It isolates variables to ensure differences are not statistically significant. These tools collectively replace subjective observation with mathematical certainty.

Core Capability Thresholds

1.33Cpk MinimumStandard threshold for general process parameters.
1.67Cpk CriticalRequired for safety-critical or key product characteristics.
3Consecutive LotsStandard minimum batch count required for PQ evidence.
Minimum statistical requirements for accepting a validated process into serial production.

Execution Discipline: Protocols Before Power

Before any validation testing begins, you must write a Validation Master Plan. This document defines exactly what is being validated, identifies Critical Process Parameters (CPP) and Critical Quality Attributes (CQA), sets the scope of testing, and assigns clear responsibilities.

You must write separate, detailed protocols for IQ, OQ, and PQ. A protocol is drafted entirely before the testing occurs. It dictates the test method, the required data collection, and the exact acceptance criteria. You cannot write a protocol after seeing the results.

Writing a protocol to match observed outcomes is not validation; it is the documentation of coincidence. Every test result must be recorded in real-time. Backdating data sheets or noting that things were 'probably fine' destroys audit credibility instantly.

Finally, a formal Validation Report summarises all findings. It compares actual results against the predefined acceptance criteria, lists any deviations, and provides a formal PASS or FAIL conclusion. This report requires sign-off from all responsible quality and engineering parties before production launches.

If your protocol is written after the test results are in, you are documenting coincidence, not validation.

Common Failures and Continuous Verification

The most frequent failure I audit is the belief that validation lasts forever. It does not. If you change a material supplier, adjust the tooling, update software, or even alter cycle times, you trigger a revalidation requirement. Your Change Management procedure must strictly define what constitutes a significant change.

Another critical error is testing solely under ideal laboratory conditions. If you validate at a constant 22°C but your facility operates at 35°C during summer, your validation is void. You must test across the entire operational range. The extremes are exactly where real production failures originate.

Finally, validation does not end when PQ passes. FDA guidelines and modern quality management systems demand Continuous Process Verification. You must maintain SPC charts, strictly follow control plans, and conduct periodic reviews to ensure the validated state is maintained over the product's entire lifecycle.

Processes inherently degrade, tools wear, and environmental factors shift. Without continuous monitoring, the statistical proof you established during PQ degrades into historical fiction. Ongoing verification keeps the process locked into its validated state.