FailModeLens

Calculating RPN in FMEA: Severity, Occurrence, Detection Ratings and What They Mean

If you have sat through an FMEA session where half the room debates whether a detection rating should be a 4 or a 5, you know that calculating RPN is the easy part. The hard part is getting the input ratings right. This guide covers both—the formula itself and the rating decisions that make or break the number’s usefulness.

How to Calculate RPN in FMEA: The Formula

Risk Priority Number is the product of three ratings, each scored 1–10:

RPN = Severity (S) × Occurrence (O) × Detection (D)

The result ranges from 1 (lowest risk) to 1,000 (highest risk). Each factor captures a different dimension of the failure chain:

  • Severity (S) — How bad is the effect if the failure reaches the customer? A score of 1 means no noticeable effect. A 9 or 10 means a safety hazard or regulatory violation.
  • Occurrence (O) — How likely is the cause to happen during the product or process lifetime? A 1 means the cause is nearly impossible given current prevention controls. A 10 means the failure is persistent.
  • Detection (D) — How likely are current controls to catch the failure before it ships? This scale is inverted: 1 means almost certain detection (automated error-proofing), and 10 means no detection method exists.

Worked Example: Calculating RPN for a Stamping Process PFMEA

Consider a metal stamping operation producing a structural bracket for an automotive seat frame. The process flow includes blanking, forming, piercing, and deburring. During the PFMEA session, the cross-functional team identifies this failure chain:

FMEA ElementEntry
Process StepForming (Station 20)
FunctionForm bracket to 90° ± 1° bend angle
Failure ModeInsufficient bend angle (under-formed)
EffectBracket does not meet dimensional specification; seat frame assembly interference at customer plant
CauseDie wear causing progressive angle loss over production run
Prevention ControlPreventive maintenance schedule (die inspection every 50,000 hits)
Detection ControlFirst-piece and last-piece inspection with go/no-go gauge; SPC check every 200 parts

Assigning the Ratings

Severity = 7. The effect is assembly interference at the customer plant. Per the AIAG severity scale, a rating of 7 corresponds to “Product operable but at a reduced level of performance. Customer dissatisfied.” This is not a safety issue (which would push to 9–10), but it disrupts the customer’s production line.

Occurrence = 4. Die wear is a known mechanism, but the PM schedule keeps it under control. Historical data shows roughly 1 occurrence per 10,000 parts at current maintenance intervals. An occurrence rating of 4 typically maps to a Cpk between 1.00 and 1.33—the process is capable but not robustly so.

Detection = 3. The go/no-go gauge catches out-of-spec parts reliably, and the SPC chart will flag a trend before parts drift far enough to fail. A detection rating of 3 means “controls will almost certainly detect the failure mode.”

The Calculation

RPN = 7 × 4 × 3 = 84

An RPN of 84 is moderate. Whether this triggers a recommended action depends on your organization’s threshold policy. Common thresholds in manufacturing:

  • RPN > 100–125 — action required (most common threshold)
  • RPN > 200 — high-priority action
  • Severity ≥ 9, any RPN — action always required for safety items

At 84, this item falls below the typical action threshold. But watch the severity: at 7, it is borderline significant. If your customer’s specific requirements classify severity 7+ as a Significant Characteristic (SC), you need enhanced controls in the control plan regardless of the RPN value.

Why RPN Thresholds Can Mislead You

RPN has a well-documented flaw: it treats all three factors as equally important. Consider two failure chains from the same PFMEA:

Failure ChainSODRPN
Under-formed bracket (customer complaint)74384
Missing weld (occupant safety hazard)101110

The missing weld has an RPN of just 10—the lowest-priority item on the sheet if you sort by RPN. But it is a severity-10 safety failure. Any rational quality engineer would prioritize it over the bend angle issue, yet RPN says the opposite.

This is exactly why the AIAG & VDA FMEA Handbook (2019) introduced Action Priority (AP) as the replacement for RPN. AP uses a logic-based lookup table that always prioritizes severity first, then occurrence, then detection. In the AP system, a severity-10 failure mode is always rated High priority regardless of how rare or detectable it is.

RPN vs. Action Priority: When to Use Which

If your organization still uses RPN, you are not alone—the majority of manufacturers outside automotive OEMs do. Here is when each method fits:

  • Use RPN when your customer or industry standard requires it (SAE J1739, IEC 60812, many medical device firms using ISO 14971-aligned FMEAs), or when your existing FMEA library is built around RPN thresholds and migration is not practical yet.
  • Use Action Priority when you are in the automotive supply chain and your OEM customers expect AIAG-VDA format, or when you want risk decisions that properly weight safety-critical failures. AP is now the standard method in SAE and AIAG-aligned programs.

You can calculate both in the same FMEA. Many teams run RPN for backward compatibility while also determining AP for the new methodology. Our RPN and Action Priority calculator handles both methods from the same severity, occurrence, and detection inputs.

Tips for Consistent Ratings Across Your Team

The biggest source of RPN variability is not the formula—it is subjective rating disagreement. Three practices that reduce it:

  1. Anchor ratings to your own data. Map your occurrence scale to actual defect rates from your quality system. If your scrap rate for a specific failure is 500 PPM, that anchors occurrence at roughly 5–6 on most scales. Do not guess when you have data.
  2. Use the detection scale as a control-type hierarchy. A detection rating of 1–2 means automated error-proofing (poka-yoke). A 3–4 means gauging or SPC. A 5–6 means manual visual inspection. A 7–8 means sampling or auditing. A 9–10 means no control or indirect detection only. This gives the team a decision tree instead of an argument.
  3. Calibrate severity once per product family. Before the FMEA session, circulate the severity scale with examples specific to your product. “Severity 8 for this product family means a customer line stop requiring a sort at their plant.” When the team shares the same reference points, rating debates shrink from 30 minutes to 5.

Next Steps

Calculating RPN is straightforward once your team agrees on the ratings. The real leverage is in what you do with the number: setting action thresholds that reflect your risk tolerance, tracking recommended actions to closure, and re-rating after actions are complete to verify risk reduction.

If you are evaluating whether to stay with RPN or transition to the AIAG-VDA Action Priority method, try running both on a single FMEA to see where the prioritization differs. Start with the RPN and Action Priority calculator to compare outputs side by side.