Action Priority vs RPN: the AIAG-VDA change explained

For thirty years, FMEA teams prioritised their work with a single number: the Risk Priority Number, or RPN. The 2019 AIAG-VDA FMEA handbook retired it in favour of an Action Priority (AP) table. If you learned FMEA before 2019, this is the change that most affects how you actually triage — so it's worth understanding properly.

The old way: the Risk Priority Number

RPN was arithmetic. You rated each failure mode on three 1–10 scales — Severity (S), Occurrence (O) and Detection (D) — and multiplied them:

RPN = S × O × D  (a number from 1 to 1000)

Then you sorted by RPN and worked the biggest numbers first, often against a threshold ("action required above 100"). Simple, intuitive — and quietly misleading.

Why RPN broke down

Multiplication treats all three inputs as equal and interchangeable, which they are not. Consider two failure modes:

• A cosmetic blemish that happens often and is hard to spot: S2 × O9 × D9 = 162.
• A safety-relevant failure that is rare and reasonably detectable: S10 × O2 × D3 = 60.

By RPN, the cosmetic nuisance (162) outranks the safety risk (60) by almost three to one. No competent engineer believes that — but the number says it, and teams that managed by the number ended up "chasing RPN" instead of managing risk. Threshold-based RPN made it worse: people nudged a Detection score from 5 to 4 to drop just under the action line, with no real change to the process.

The AIAG-VDA Action Priority table

Action Priority replaces the single number with a decision table that returns one of three levels — High, Medium or Low — for every combination of S, O and D. The table is built on a clear hierarchy of intent:

1. Severity leads. How bad the effect is matters most. Severity 9–10 (safety or regulatory) drives High almost everywhere.
2. Occurrence is next. For a given severity, a more frequent cause raises the priority.
3. Detection is the tie-breaker. Better detection can lower the priority, but it can never rescue a high-severity, high-occurrence risk down to Low.

So "High" no longer means "big number" — it means "the team should act, and document what they decided." Medium means "should act, customer may expect rationale," and Low means "action is at the team's discretion."

The same two failures, under AP

The ranking flips to match reality: the safety failure is High, the nuisance is Low. That is the entire point of the change.

What it means for your workflow

• Stop sorting by a number. There is no "AP score" to rank; you group by High → Medium → Low and work the High items first.
• Severity is non-negotiable. You cannot detection-control your way out of a High on a safety characteristic. The honest response is to reduce occurrence (better prevention) or, ideally, design the severity out.
• Document decisions, not just scores. AP is about prompting action and recording the rationale — what you'll do, who owns it, by when.

RPN isn't forbidden — some customers still ask for it — but AIAG-VDA AP is the expected method, and it's the one an auditor familiar with the 2019 handbook will look for. In SolidLaunchpad the Action Priority is computed live from the official table as you score, so you always see the current, method-correct priority. For the full mechanics of producing those scores, see How to run a PFMEA, step by step.