A senior quality engineer with two decades of experience signs off on a PPAP package. She is fast, she is thorough, and she catches errors that others miss. On Tuesday, she had a migraine and approved a package with an incorrect material certification. On Thursday, rushing to beat a production shutdown, she missed an out-of-spec dimension on a flow test report.
These are not hypothetical failures. They are the inevitable output of a quality system built on the assumption that the right expert, given the right training, will always make the right call. ISO 9001 requires competence. IATF 16949 requires qualified personnel. Neither standard guarantees infallibility. No human being can maintain perfect judgment across hundreds of decisions under time pressure, fatigue, and cognitive bias.
Medicine invented peer review because doctors recognised the lethal limits of individual judgment. Aviation requires dual verification for the same reason. Nuclear power plants engineer redundant safety systems into their architectures. Your quality system needs the same mechanism: structured, independent peer review at the decisions where single-point failure is most likely to occur and most costly when it does.
The Myth of the Expert Gatekeeper
Most quality assurance architectures revolve around individual human judgment. Your procedures specify exactly who can approve a deviation, release a first article, or sign a control plan. You select your most experienced people for these roles because their expertise is your best line of defence against defects escaping to the customer.
Expertise is essential, but equating expertise with infallibility is a systemic design flaw. I have audited plants where a single highly competent engineer was the only barrier between a nonconforming process and full-scale production. When that engineer missed a tolerance stack-up error, the system had no mechanism to catch it. Three months later, the plant was managing a massive field recall.
The solution is not to add more signatures to a form. A second signature that merely confirms the first decision adds bureaucracy without adding detection capability. True peer review is a structured process where a qualified second person independently evaluates a quality decision before it becomes final and actionable. The key word is independently.
Sign-off vs. Peer Review
Rubber-stamp sign-off
- Reviewer sees the original conclusion first
- Reviewer is the original preparer's direct report
- Review happens after the decision is enacted
- Criteria are generic: 'does this look acceptable?'
Independent peer review
- Reviewer evaluates raw data blind to the original outcome
- Reviewer operates outside the preparer's reporting line
- Review occurs before the decision is authorised
- Criteria are specific checklist items and risk factors
Where Peer Review Belongs in Your System
Not every quality decision needs peer review. Attempting to implement it across all routine inspections and administrative approvals will paralyse your operations. The discipline must be applied selectively at high-risk nodes where the consequences of an error are severe and the likelihood of single-point failure is elevated.
Control plan approval is a primary target. A single missing characteristic, incorrect specification, or inappropriate sampling frequency in a control plan can produce defects for months before anyone notices the discrepancy. A second quality engineer, who was not involved in drafting the document, must independently verify it against the drawing, the PFMEA, and the process flow.
First Article Inspection (FAI) reports are equally critical. First article approval is the gateway to production; if the FAI is wrong, everything that follows is built on a false foundation. FAI reports are frequently reviewed by a single person under intense time pressure to release the part. A targeted second review of the most critical characteristics is the most efficient way to prevent a false foundation.
Engineering and Supplier Qualification Reviews
Approving a new supplier is one of the highest-leverage decisions in any manufacturing organisation. A bad supplier approval, often driven by purchasing pressure to speed things up, can result in years of escapes and warranty claims. A second supplier quality professional must independently review the audit findings, the PPAP documentation, and the risk assessment for gaps in evidence.
Deviation and concession approvals require a different kind of peer review. When you authorise a departure from specification, you are betting that the deviation will not affect form, fit, or function. A second qualified person, typically from engineering rather than quality, must evaluate the deviation against design intent. They answer a different question: not 'does this meet spec?' but 'will this work in the application?'
Corrective action effectiveness verification is another notorious failure point. Too often, the same engineer who implemented the corrective action is tasked with verifying its effectiveness. This is the quality equivalent of grading your own exam. An independent reviewer must evaluate whether the objective evidence genuinely demonstrates that the root cause was addressed, guarding against confirmation bias and premature closure of the 8D.
| Quality decision | Primary risk | Required peer reviewer |
|---|---|---|
| Control plan approval | Missing characteristics, wrong specs | Quality engineer outside original drafting team |
| Deviation / concession | Functional failure in the field | Design engineer evaluating application intent |
| Supplier PPAP approval | Over-reliance on supplier claims | Independent supplier quality professional |
| 8D effectiveness verification | Confirmation bias, premature closure | Engineer uninvolved in the corrective action |
| Measurement System Analysis (MSA) | False confidence in bad data | Second MSA practitioner checking study design |
The Psychology of Productive Friction
Quality peer review functions as a cognitive intervention, not merely a procedural safeguard. The Hawthorne Effect dictates that people perform better when they know someone is observing their work. When quality engineers know their PPAP packages and FAI reports will be independently reviewed, they prepare them more carefully. The error rate drops before the review even happens.
Cognitive diversity provides the second mechanism. Two professionals with different backgrounds will see different things in the same data. A reviewer steeped in GD&T will catch every geometric tolerance error but might miss a material specification issue. A second reviewer with a materials background catches what the first missed. This is the productive friction of disagreement.
The value of peer review is not agreement; it is the productive friction of documented disagreement.
Ego activation completes the psychological framework. When professionals know their work will be scrutinised by a peer, their intrinsic motivation engages. They do not want to be the person who missed an obvious failure mode. This is not driven by fear of punitive action, but by the professional pride of craftsmanship. Peer review elevates the standard of the entire department.
Implementation Without Paralysis
The most common objection to peer review is the time investment. The concern is valid if you attempt to apply the mechanism to every routine inspection or low-risk document. The solution is rigorous risk-based triage. Map your quality system and identify the specific points where a single error could result in customer complaints, safety incidents, or regulatory nonconformities.
Once you identify these mandatory review points, define specific criteria. The reviewer must know exactly what they are checking against. Generic acceptance criteria like 'does this look acceptable?' must be replaced with detailed checklists: What exactly is the reviewer verifying? What constitutes a discrepancy? What is the reviewer explicitly not responsible for?
Structured Peer Review Workflow
- 011. Define triggerDocument the high-risk nodes that legally require peer review before processing.
- 022. Assign independent reviewerSelect a qualified peer outside the original preparer's reporting line.
- 033. Conduct blind reviewReviewer evaluates raw data against checklists without seeing the initial conclusion.
- 044. Document discrepanciesRecord all differences in evaluation; disagreements are data, not failures.
- 055. Resolve and authoriseEscalate unresolved differences to a technical authority for final arbitration.
Metrics, Failure Modes, and Detection
If you do not measure the peer review process, it will degrade into rubber stamping within months. The discrepancy detection rate is your primary indicator of process health. If your peer review process catches zero discrepancies over a sustained period, your review is superficial. Track how many of the caught discrepancies would have resulted in customer-facing defects.
Monitor your first-pass error rate over time. If the peer review process is functioning, the error rate of documents before review should gradually decrease. The preparers internalise the knowledge that their work will be scrutinised, improving their initial output. This upstream improvement is where peer review delivers its highest return on investment.
Review fatigue is the most dangerous failure mode. When the same few people are reviewing dozens of complex documents per week, the quality of each review degrades. The solution is to expand the reviewer pool, rotate assignments, and enforce strict cycle-time limits. If peer review is adding days to your engineering change process, you have a capacity problem that requires management action, not a quality problem that requires compromise.