RPN is ordinal, not calibrated probability
A change from 240 to 120 means the product of three ratings halved; it does not prove that physical risk or expected loss halved. Keep the underlying ratings visible.
Reliability
Compare current and residual FMEA risk-priority numbers under explicit controls.
FMEA PRIORITIZATION
For FMEA teams documenting action priority without converting ordinal scores into false probabilities or bypassing high-severity governance.
CURRENT DECISION RECORD
Every row is regenerated from the active inputs and carried into Copy, TXT, and the page-specific PDF payload.

| Factor | Current | Control | Residual |
|---|
CURRENT CALCULATION PROCESS
RPN=SxOxD; RPNres=Sxmax(1,O(1-rO))xmax(1,D(1-rD))
Waiting for valid inputs.
HOW TO USE
FMEA FUNDAMENTALS
DEEP RISK ANALYSIS
A change from 240 to 120 means the product of three ratings halved; it does not prove that physical risk or expected loss halved. Keep the underlying ratings visible.
S=10, O=3, D=4 and S=5, O=8, D=3 both produce 120, yet the first retains a catastrophic severity that may require action regardless of rank.
Occurrence reduction needs prevention evidence, while detection improvement needs inspection capability evidence. Treating one control as both without proof double-counts its benefit.
WORKED DECISION CASES
A connector scored S=8, O=5, D=6 has RPN 240. A keyed assembly and torque verification reduce modeled occurrence by 50% and detection difficulty by 40%, producing residual RPN 72; the team still preserves the severity-8 escalation review.
A safety-related sensor with S=10, O=2, D=2 starts at RPN 40, below many numeric thresholds. The low product must not close the issue because a single undetected failure can retain an unacceptable effect.
EVIDENCE RECORD
Retain the FMEA revision, item function, failure chain, scoring rubric, named participants, evidence for each rating, control owner, due date, implementation record, and post-action verification. The residual score is auditable only when reviewers can trace every changed rating to a completed control and its evidence.
MODEL LIMITS
FMEA GLOSSARY
FREQUENTLY ASKED QUESTIONS
Only when both analyses use the same scoring definitions, evidence windows, and escalation rules. Cross-product ranking without a common rubric creates false precision.
Usually no. Prevention changes occurrence; severity changes only when the design changes the consequence experienced after the failure occurs.
The 1-10 FMEA scale has no zero category. A control may make failure remote or highly detectable, but the rating framework still records a minimum class of one.
No. Organizations define their own action priorities, and many require mandatory review for high severity even when RPN is below a numeric threshold.
Possibly, but each effect needs separate evidence. Do not apply the same claimed reduction twice unless the control genuinely prevents causes and independently detects residual defects.
No. It records a prioritization judgment under one rubric. Safety requirements, regulatory duties, single-point failures, and verification evidence remain separate acceptance gates.
RELIABLE SOURCES
IMPORTANT FMEA LIMIT
Use this page to prioritize investigation and actions. Do not use a low current or residual RPN to waive hazard analysis, statutory requirements, mandatory severity escalation, or verification of safety-critical controls.