That day in a PFMEA meeting, my severity rating was challenged and corrected by the client
Do you remember? Last month, our team was holding a PFMEA review meeting with a client, and that time really made me break out in a cold sweat. For one process step, we originally judged the severity to be 7, because it would lead to product function degradation at the customer's end, but not a complete failure. After hearing this, the client's face turned green, and they immediately interrupted, saying: "Your scoring method is completely inconsistent with the spirit of the new AIAG & VDA PFMEA; this is clearly a 9!" At that moment, my team leader and I exchanged glances, thinking to ourselves, "Oh no, did we get something wrong again?"
Where was the problem? Severity is no longer scored "by feeling."
To be frank, in the past, when we evaluated PFMEA Severity, it was actually a bit like "subjective judgment." Everyone relied on experience, thinking if it "would affect customer function," they'd give it a 7 or 8; if it "would cause a production line shutdown," they'd give it a 9. But the most core change in the new AIAG & VDA PFMEA is that it has "standardized" this severity scale. It no longer just looks at "to what extent the product is broken," but instead emphasizes "the impact of this failure mode on the end-user and regulations."
In other words, previously, you might have thought a process with a Cpk of 1.08, while not good enough, still produced usable products, so the severity might be moderate. However, if this insufficient Cpk leads to a defect where the customer's car airbag has a 5% chance of failing to deploy during an impact, then no matter how stable your process is, the severity of this "failure mode" is off the charts.
How is it actually done? Look at the "result" rather than the "degree."
The most important aspect of the new severity scale is its judgment logic. It directly asks you a few key questions, and by simply answering "yes" or "no," you can generally set the tone.
- Will it affect vehicle safety or violate regulations?
* No → Continue to the next point.
- Will it lead to primary function failure or damage?
* No → Continue to the next point.
- Will it lead to secondary function failure or significant degradation?
Let's take a practical example: Suppose the DPMO of one of our assembly processes is 6210.
- If this assembly defect leads to the vehicle suddenly stalling while driving at high speed, then even if the DPMO is not too high, the severity is directly 9 or 10.
- If this assembly defect causes the car audio to occasionally lose signal, then the severity would probably be 7.
So the key point is, you now need to consider "what 'result' this failure mode will lead to," rather than "how 'severe' this failure mode is."
The Most Common Trap: Still Scoring with Old Thinking
Honestly, one of the easiest traps for many of our engineers to fall into is still being stuck in the "gray area" of the past. They feel that as long as it doesn't directly lead to "personal safety" issues, they'll push the score towards 7 or 8. Sometimes they even think, "the customer hasn't found out yet, so let's be conservative."
But frankly, customers are very savvy now, especially car manufacturers; their demands for the new version of PFMEA are getting higher and higher. If you're still using old methods for scoring, they'll see it immediately. For example, last time our team rated a failure mode that "could lead to abnormal dashboard display" as a 7. The client directly challenged us, saying: "An abnormal dashboard display, in certain situations, can affect driver judgment. Do you think this is just minor degradation?" At that moment, our faces really turned green. To put it simply, what customers want is for you to anticipate all possible risks, not for you to minimize them.
One Thing You Can Do Today
Re-examine the PFMEA you have, paying special attention to the "Severity" rating, and ask yourself: "Will this failure mode directly trigger a red flag for the customer or regulations?"