Severity Occurrence Detection Ranking Scale - 1 to 10 vs. 1 to 5

[kurisuoriari]

Hello another FMEA Topic from me.

A couple of years ago we reviewed our Severity Occurence Detection Ranking.

A broad difference exsisted in using scales of 1 to 10.
As a improvement the scales were refined to 1 to 5.

Thus I have a Max RPN of 125.
This has worked fine and provides good repeatability of classification for different users and team members.

However a recent customer auditor discussed my ranking and has documented a requirement to review the ranking. His point is that the range does not allow the correct priority since a higher range would seperate items of greater importance.

However I demonstrated that because I took all RPN of the same value with the same action priority and had no RPN's above 12 on any process in our plant my process was working effectively. This could also be justified by the quality of product supplied to this customer meeting target requirements.

Yet I fear that this auditor wants me to install a 1 to 10 ranking and if this is not carried out on the corrective action report he will not close this issue and improve my supplier ranking.

Any Thoughts ?

Cheers

Chris

[Jim Wynne]

This isn't actually a FMEA question, it's a customer-requirements question. The question is: how can I deal with customer requirements that don't make sense? First, if you've done your best to diplomatically convince your customer contact that the requirement doesn't make sense, you have two choices. You can just accept the nonsensical requirement and move on, or you can go over the guy's head. There's an excellent chance, though, that his boss will have no idea what you're talking about and find it much easier just to back his own guy. Even if the boss does rule in your favor, there's an excellent chance it will come back to haunt you, and at the worst possible time.
In short, this is a decision you have to make for yourself, because you should have a much better feel for the particulars of this situation than anyone here does. We don't know the customer, or the customer's representative, or how easy or difficult it would be to make a change and move on.

Interesting situation. I wonder if there is any evidence of the reliability of a 10 point scale versus a 5 point scale. Frankly there is not a big difference between a 9 or 10 in a 10 point scale, both of which a 5 would cover on a 5 point scale.

If they've not done so, I'd ask the customer representative to create a short example of an FMEA using both 5 and 10 point scales, to demonstrate what s/he believes would occur and what the impact would be. I would then determine the degree to which this would likely occur with your product/application, and whether or not there is any evidence it has occurred.

I would bet this is just a usual "this is the way I've always seen/done it in the past" ... imposing additional costs on a supplier. Just make sure your next bid to them includes this additional cost.

[Rob Nix]

As JSW05 (aka A.E. Newman) says, if your customer requires it, then that's the way it is (unless you can convince them otherwise). That is why a good contract review process is so vital! If discussed and documented then, there would be nothing the auditor could fault you on now.

But, on another vein,

In Reply to Parent Post by kuri...

A broad difference exsisted in using scales of 1 to 10.
As a improvement the scales were refined to 1 to 5.

Therein lies the problem with FMEAs. As I've stated in other threads, the FMEA exercise is an excellent tool. But the overemphasis on ranking numbers and risk priority numbers is counterproductive. The fact is, for each line on the FMEA, you choose either to do something to improve it or not. Period.

The skills and cross-functionality of the team members are the important part. They identify the risks and decide which risks are too high to leave alone. You make a list of those things, prioritize them, and make assignments. All of the gyrations surrounding this number or that number shifts the focus away from the whole purpose of the exercise. IMHO anyway.

[Jim Wynne]

In Reply to Parent Post by Rob Nix

As JSW05 (aka A.E. Newman) says, if your customer requires it, then that's the way it is (unless you can convince them otherwise). That is why a good contract review process is so vital! If discussed and documented then, there would be nothing the auditor could fault you on now.

But, on another vein,

Therein lies the problem with FMEAs. As I've stated in other threads, the FMEA exercise is an excellent tool. But the overemphasis on ranking numbers and risk priority numbers is counterproductive. The fact is, for each line on the FMEA, you choose either to do something to improve it or not. Period.

The skills and cross-functionality of the team members are the important part. They identify the risks and decide which risks are too high to leave alone. You make a list of those things, prioritize them, and make assignments. All of the gyrations surrounding this number or that number shifts the focus away from the whole purpose of the exercise. IMHO anyway.

I agree. The numbers are an unnecessary distraction, and lead to dangerous strategies such as "trigger" RPNs--if RPN is >X, actions are required. The trigger strategy came about as a result of auditors questioning FMEA documents that show no "right-hand activity" and auditees needing to come up with an appropriate explanation for it.

[Bill Ryan - 2007]

I haven't vented in a while..... so here goes

This is yet another example of someone trying to impose his/her "ideal" onto a supplier that, apparently, has a methodology that works for them. The fact that it DOES work for you is of no consequence. It's the "Do it my way or there's the highway" syndrome. I have to go along with JSW (something new ) in that you need to weigh the pros and cons of the battles you take on (with anyone). I also agree with Dokes in that this is how the AIAG manual presents the FMEA methodology and your customer probably doesn't understand what FMEA is all about, other than through the AIAG manual (I do, however, highly doubt the customer rep would go through the excersise you cited).

End of rant (I think)

In practice - do you really do much different if something has a severity of 9 or 10? If the difference between 9 & 10 is with or without warning, it is still a "Safety" feature that must be thoroughly dissected and kept "under control". If a Detection is ranked 9 or 10 you are probably going to go through the same excersize to lower the ranking (depending.....) because you, basically, can't "catch" the nonconformance.

What's the difference between 7 and 8? Or 5 & 6 (even with the Occurrence and Detection ratings)? I realize that customers have come out with their own, grandiose, rating tables but I still refuse to sit with a team and argue whether something should be rated as this number or maybe one higher. Maybe, just maybe, the 1-5 scale would actually help stop that kind of activity as you mentioned with the "good repeatability of classification" comment.

One of the issues could be that your customer doesn't know how to handle a "bogey" number that his organization has set (but let's not rehash that pain in the a** topic).

If I were in your position (again):
All in all - as JSW mentioned, there are occasions where it is best to "bite the bullet" and cave in. If that's the route you decide on, I wouldn't change anything you do internally. I would merely change the ratings for that customer on the document he sees and if he "knows" that you have just fooled with the numbers, tell him that's what his demands necessitated because you aren't about to disrupt a system that works well for you. That doesn't sound very "ethical" but most organizations I've seen don't handle different systems for the same process very well - and it IS your company, not theirs.

(Wow, am I wound up this morning )

[Jim Wynne]

I have to go along with JSW (something new )
One of the great things about the Cove is that it presents varying informed opinions to people with questions. Disagreement is a good thing, so long as it's not just contrarianism (did I just make up a word?).

In Reply to Parent Post by Bill Ryan

That doesn't sound very "ethical" but most organizations I've seen don't handle different systems for the same process very well - and it IS your company, not theirs.

I don't think it's necessarily unethical when you give the customer what he asks for .

[Tim Folkerts]

Since you roughly cut in half the assigned scores (5 on a 1-5 scale = 10 on a 1-10 scale), how about simply doubling the numbers for this particular customer when you report the results. Since there are 3 scores, then you would multiple the final score by 8. So 125 --> 1000, or 1 --> 8. This will tend to overestimate the score slightly, but
1) it will be in the right ballpark (which is all you can really hope for)
2) it will be in terms the customer will appreciate.

I suppose the customer could argue that a 3x3x3 = 27 on your scale could be anthing from 5x5x5 = 125 up to 6x6x6 = 216 on a 10 point scale. Still, in the grand scheme of FMEA's that's not such a huge range. You are just trying to get a rough estimate of overall importance, after all.


Tim F