FMEA RPN vs Action Priority

Three realistic scenarios where RPN and AP disagree. The first row is the canonical safety case that forced AIAG to retire RPN as the primary metric.

RPN = Severity × Occurrence × Detection. Teams used a threshold (often > 100) to decide which failure modes required action. The simplicity of one number made it easy to sort and report — and easy to game.

• It treated Severity, Occurrence and Detection as equally weighted, which they are not.
• An S=10 / O=2 / D=2 safety failure scored 40 — below typical action thresholds.
• An S=3 / O=5 / D=4 cosmetic failure scored 60 — above the threshold.
• Teams chased the number and ignored the actual risk.
• Inconsistent thresholds between sites and customers made benchmarking meaningless.

AP weighs Severity first. High Severity always yields elevated priority regardless of how low Occurrence and Detection appear. AP is a rule table, not a calculation, so it cannot be gamed by inflating Detection confidence. See the full logic in the Action Priority Guide.

1. Keep the existing S/O/D ratings — they are still required.
2. Replace the RPN column with an Action Priority column.
3. Recompute AP using the AIAG-VDA 2019 matrix.
4. Document Low AP decisions; document High AP justifications.
5. Update Control Plans where Detection assumptions change.

Some customers and internal legacy templates still display RPN alongside AP. That is acceptable for continuity, but AP is the decision metric. Avoid using RPN thresholds for any new program. Read the common FMEA mistakes page for related anti-patterns.

Primary keyword: FMEA Risk Priority Number. The page also covers the migration cluster — searches from teams figuring out whether to keep RPN, how to convert, and what customers actually expect in 2026.

The auditor objection to RPN is not academic. It is operational: RPN can rank a safety failure below a cosmetic one, and customers have repeatedly seen this fail in the field. AIAG-VDA retired RPN because the anti-patterns below were too common to ignore.

• Threshold arbitrariness. One site uses RPN > 100 as the action trigger; another uses 125. Same failure, different decisions, same OEM customer.
• Severity dilution. Multiplying S × O × D allows a high-Severity failure with low Occurrence and low Detection to score below the threshold. Action Priority closes this loophole.
• Detection gaming. Teams pushed Detection down to drop RPN below threshold without changing the actual control. AP's rule table makes this much harder.
• Cross-program inconsistency. RPN scores between two programs are not comparable because the rating tables drift. AP categories (H / M / L) are stable across programs.

See the Action Priority guide for the replacement logic and the audit checklist for the evidence reviewers expect during the migration.

Most automotive suppliers in 2026 still maintain legacy programs with RPN-based PFMEAs running in parallel to new AIAG-VDA 2019 programs. The migration below preserves history while converting the active decision metric.

1. Freeze a baseline. Snapshot the current PFMEA with its existing S/O/D and RPN values. This becomes the audit-of-record for the pre-migration state.
2. Re-rate Severity only if customer impact changed. Severity values usually carry over; auditors flag unexplained Severity changes during migration.
3. Recompute AP from the AIAG-VDA matrix. Do not derive AP from RPN buckets. Use the rule table — see the AP decision table.
4. Re-examine all rows that flip category. Rows that were "below RPN threshold" but become High AP need an action plan and justification.
5. Update the Control Plan in lockstep. Any AP change that affects a detection control propagates to the Control Plan.
6. Document the migration in the FMEA revision history. Auditors expect to see "Migrated from RPN to AIAG-VDA 2019 AP on YYYY-MM-DD by [owner]".
7. Keep RPN as a reference column if customers require it. Do not use it as the decision metric.

For tooling that runs both RPN and AP in parallel without divergence, see AIAG-VDA FMEA software and the FMEA generator.

Some OEMs and Tier-1s still require RPN to appear in submissions for continuity with historical data, even while accepting AP as the decision metric. This is acceptable as long as the team treats RPN as a reporting column and AP as the decision column. The risk is internal teams reverting to RPN-style threshold thinking once both columns appear in the template.

For the engineering anti-patterns this transition creates see common FMEA mistakes; for the workshop discipline to keep both columns honest see FMEA facilitator guide.

• Action Priority Guide
• FMEA Ranking Tables
• AIAG-VDA 7-Step Process
• Common FMEA Mistakes
• AIAG-VDA FMEA Software