What is FMEA?
Failure mode and effects analysis is a team-based, step-by-step review of everything that could go wrong with a product or process. It is required or expected by IATF 16949, ISO 13485 and ISO 14971 (medical devices), and is a key input to the control plan. A good FMEA is a living document: update it when you see new failures, complaints or process changes. To go deeper on a single failure, use fault tree analysis; to add quantitative criticality, use FMECA.
FMEA rating scales (1 to 10)
| Rating | Severity | Occurrence | Detection |
|---|---|---|---|
| 9–10 | Safety or regulatory effect (10 without warning) | Extremely high to almost certain | Very remote to cannot detect |
| 7–8 | Loss or degradation of primary function | High to very high | Remote to very low |
| 4–6 | Loss or degradation of secondary function, annoyance | Moderate | Low to moderately high |
| 2–3 | Minor or very minor effect | Low to very low | High to very high |
| 1 | No discernible effect | Prevented through controls | Almost certain (error-proofed) |
Adapt the descriptions to your industry and keep the scale the same across all FMEAs in your organization.
FMEA example
The example is a process FMEA of a sterile pouch sealing and packaging line. A weak seal (severity 9, occurrence 4, detection 5) has an RPN of 180 and High Action Priority: continuous temperature monitoring with an interlock is planned, which brings it to 9 × 2 × 3 = 54. A wrong component in the tray has the highest RPN (192) but Medium priority, and gets a poka-yoke tray nest. Note that an illegible label (RPN 75) ranks low by both measures, while a missing leaflet ranks low on RPN (84) but carries severity 7: severity always deserves a separate look.
Tips for a useful FMEA
- Work as a team: design, production, quality and maintenance see different failures.
- One cause per line: different causes need different controls and get different occurrence ratings.
- Prevention before detection: reducing occurrence is better than catching defects.
- Close the loop: re-rate after actions, and update occurrence from real nonconformance and complaint data.
How to use this FMEA tool
- Define the scope. Choose process or design FMEA and name the process or product and the team.
- List functions and failure modes. For each step or item, describe its function and how it could fail to perform it.
- Describe effects and causes. Record the effect on the customer or the next step, and the cause of each failure mode.
- Rate S, O and D. Rate severity, occurrence and detection from 1 to 10 against your current controls.
- Prioritize and act. Use Action Priority and the ranked chart to plan actions, owners and dates.
- Re-rate and export. Re-rate after the actions are done and export the FMEA to Excel or PDF.
Frequently asked questions
What is FMEA?
Failure mode and effects analysis (FMEA) is a structured method to identify how a product or process could fail, what the effects would be, and what causes the failure, before it happens. Each failure mode is rated for severity, occurrence and detection, and actions are planned for the highest risks.
What is the difference between a DFMEA and a PFMEA?
A design FMEA (DFMEA) analyses how a product design could fail to meet its requirements, for example a seal that leaks because of the wrong material. A process FMEA (PFMEA) analyses how a manufacturing or service process could fail, for example a seal that leaks because the sealing temperature drifts.
How is the RPN calculated?
The risk priority number is RPN = Severity × Occurrence × Detection, each rated 1 to 10, so it ranges from 1 to 1000. A high RPN suggests a higher priority, but the same RPN can hide very different risks (for example 10 × 1 × 10 and 4 × 5 × 5 both give 100).
What is Action Priority (AP)?
Action Priority, introduced in the AIAG and VDA FMEA Handbook (2019), replaces RPN thresholds. It uses a lookup table that gives most weight to severity, then occurrence, then detection, and classifies each failure mode as High, Medium or Low priority. High AP items require action or a documented justification.
What do the severity, occurrence and detection ratings mean?
Severity rates the seriousness of the effect (10 = hazardous without warning). Occurrence rates the likelihood that the cause happens, given the prevention controls (10 = almost certain). Detection rates the ability of current detection controls to find the cause or failure before it reaches the customer (1 = almost certain to detect, 10 = cannot detect).
Should I use an RPN threshold such as 100?
Fixed RPN thresholds are discouraged by the AIAG-VDA handbook because they can ignore high-severity items with low occurrence. Use Action Priority, and always look at severity 9 and 10 items separately. You can still set a threshold in the tool if your customer requires one.