The template
| Step | Function | Failure mode | Effect | S | Cause | O | Current control | D | Priority | Action | Owner | Due | New S/O/D |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
A filled row
| Step | Function | Failure mode | Effect | S | Cause | O | Current control | D | Priority | Action |
|---|---|---|---|---|---|---|---|---|---|---|
| Seal carton | Form hermetic seal | Incomplete seal | Leak; product spoils; recall risk | 8 | Seal-bar temperature drifts after clean | 6 | Visual check at start of shift | 7 | High (RPN 336) | Add temperature check to post-clean checklist; alarm on drift |
What do the rating scales mean?
| Rating | Severity (effect) | Occurrence (cause) | Detection (control) |
|---|---|---|---|
| 1–2 | No noticeable effect | Remote; almost never | Certain to be detected |
| 3–4 | Minor annoyance; rework | Low; a few per year | Likely detected |
| 5–6 | Degraded performance; customer complaint | Moderate; monthly | May be detected |
| 7–8 | Loss of function; safety or regulatory issue with warning | High; weekly | Unlikely detected |
| 9–10 | Safety or regulatory failure without warning | Very high; daily | Not detected; no control |
Agree the scales for your site before scoring and write them on the template. Consistency matters more than precision.
How do you prioritise?
The traditional method multiplies the three ratings into a risk priority number (RPN, 1–1000) and acts on the highest. Its weakness is that a severity of 10 with low occurrence can score below a nuisance that happens often. The AIAG-VDA 2019 handbook replaced RPN with an action priority (AP) table that reads severity first: any row with severity 9–10 and more than remote occurrence is high priority regardless of detection. Whichever you use, act on severity first.
How do you run an FMEA?
- Map the process steps with the people who do them.
- For each step, list the ways it could fail. Use the history of defects, breakdowns and near misses.
- For each failure mode, list effects, causes and current controls, and rate S, O, D.
- Prioritise and define actions that reduce occurrence (prevent the cause) or improve detection (catch it sooner).
- Assign owners and dates, do the actions, re-rate.
- Revisit after any real failure or process change.
FMEA and daily management
The FMEA predicts; the tier board records. Defects, near misses and breakdowns raised at the daily meeting are evidence that an occurrence or detection rating is wrong, and the actions the FMEA generates belong on the same board as everything else, with owners and dates, where they are reviewed until done.
Step by step
- Map the process. List the steps and their functions with the people who run them.
- List failure modes. For each step, every way it could go wrong, using defect and breakdown history.
- Rate S, O, D. Effects and severity; causes and occurrence; controls and detection, on agreed 1–10 scales.
- Prioritise. Severity first, then RPN or the action priority table.
- Act and re-rate. Actions with owners and dates to prevent the cause or improve detection; re-rate when done.
- Keep it alive. Revisit after every real failure or change to the process.
A worked example
A pharmaceutical packaging line ran its first PFMEA on the blister sealing step after two deviations in a quarter. The team of two operators, the line engineer and a QA specialist listed nine failure modes; incomplete seal from temperature drift scored severity 8, occurrence 6, detection 7. The actions were a temperature check added to the post-clean checklist and a drift alarm. Six months later occurrence was re-rated at 2 and detection at 3, and no sealing deviations had been raised.