FMEA Detection Rating Guide

The full AIAG-VDA Detection scale calibrated by control type. Use this as the team reference during rating workshops.

Detection rates the existing control system, not the control system you wish you had. Detection looks only at what is in place today, in production, reliably, every shift.

Prevention controls reduce Occurrence. Detection controls reduce Detection. They are not the same and must not be mixed in the same column.

• Tool-life counter → prevention (reduces Occurrence).
• In-line vision system → detection (reduces Detection).
• Poka-yoke that makes the defect impossible → prevention, rating Detection = 1.

• Rating Detection optimistically to bring AP down — auditors look for exactly this pattern.
• Counting visual operator checks as low Detection. Manual visual is rarely better than 7.
• Crediting a control that exists in the Control Plan but is not actually performed.
• Mixing prevention and detection in the same rating.

With high Severity, Detection can still keep AP High even when Occurrence is low. See the full logic in the Action Priority Guide and the complete ranking tables.

Vision systems are the most over-rated control in modern PFMEAs. Teams default to Detection 2 because the system is "automated", then auditors find lighting drift, missing reference images, or alarms that nobody actions. A vision system earns its rating from evidence, not from its existence.

The same logic applies to torque transducers with data logging, laser micrometers, and barcode-verified assembly. Closed-loop alarm response is the dividing line between Detection 4 and Detection 2.

SPC is primarily a Prevention control. A capable, in-control process reduces Occurrence. SPC only earns Detection credit when the chart is reviewed in time to stop product before escape — typically Detection 5 with frequent sampling and timely reaction rules.

Gauges earn Detection based on coverage (sample vs 100%) and feedback (open loop vs auto- reject). A go/no-go gauge used 100% with documented reaction earns 4; the same gauge used as a sample check earns 6.

Layered process audits (LPAs) verify that controls exist and are followed. They do not themselves detect defective product, so they do not appear in the Detection column. They protect the Detection ratings you have already claimed — auditors expect to see LPA evidence supporting any Detection below 5.

Mistake-proofing (poka-yoke) splits into two AIAG-VDA categories:

• Prevention poka-yoke — the defect physically cannot be produced (fixture only accepts the correct orientation). Reduce Occurrence; Detection collapses to 1.
• Detection poka-yoke — the defect can occur but is detected and the part stopped before the next operation (presence sensor + interlock). Detection 2.

Mis-classifying a detection poka-yoke as prevention is the single most common reason an auditor rejects an FMEA's Detection ratings.

Detection is where teams lower AP without changing the process. IATF auditors and customer SQE teams know this and target it specifically. Expect challenges when:

• Detection 3 or lower is claimed without GR&R and alarm-response evidence.
• The Control Plan reaction plan is missing, generic, or names a role nobody owns.
• The control in the FMEA is not the control actually listed on the Control Plan.
• Layered audit records do not confirm the control is performed every shift.
• Detection ratings stayed flat after a known escape — the FMEA was not updated.

Treat any Detection below 5 as a claim that must be defended with evidence. See the FMEA audit checklist for the exact evidence customers ask for.

Inspection feels safe, so teams over-credit it. Three patterns to watch for:

• Inspection theatre. A check exists on paper but adds no real detection capability (visual check of a feature the operator cannot actually see).
• Detection inflation after an escape. Adding a second inspection rarely justifies more than one rating-point improvement; the root cause is usually upstream.
• Substituting Detection for Prevention. Teams add a final check instead of fixing Occurrence. AP often stays High; the audit finding lands on the Prevention column.

1. Walk the line before the workshop. Confirm the control exists and is performed every shift.
2. For every Detection ≤ 4, demand evidence: GR&R, alarm log, reaction plan, audit record.
3. Default operator visual inspection to 7 unless the team can prove otherwise.
4. Force a re-rating any time a control is removed or replaced.
5. Re-rate Detection only after the action is verified in production, not when the work order is closed.

• Severity Rating Guide
• Occurrence Rating Guide
• Action Priority Guide
• Control Plan vs FMEA
• Common FMEA Mistakes
• FMEA Audit Checklist