Priorizing Non-conformances - What is the best way?

[Stormy]

Good Day,

I am wondering if anyone has suggestions on how to priorize non-conformances? I have looked in the other forums at this site, but I haven't found anything about priorizing ncr's.
My goal is to minimize the amount of corrective actions that are dealt with. At this time we use a single form for NCR/corrective actions and we address almost every ncr with a corrective action. This method allowed us to pinpoint where our repeatable processes were, and to pinpoint any missing instructions that would deal with regular non-conformances. But we are now at the point that we feel we should streamline the system, and concentrate on the larger issues.

I hope I haven't confused everyone with this description.

So if anyone has any suggestions, they would be appreciated.

Thanks.

[Mike S.]

Stormy,

Your problem is a common one -- one that most companies have had to deal with already, or soon will have to deal with. In fact, this issue has been discussed several times here in the last year. Use the search function to look for threads with the words "corrective action" or "CAR", etc. and I know you will find tens of posts with many opinions. If memory serves (not a given!), I think the consensus was that not every NC requires a formal CA; that sometimes "human error" is an acceptable cause; and that prioritization is usually based on risk of recurrance and potential impact.

If you can't find what you're looking for in a search, write back and ask some specific questions and you'll get answers.

Stormy,
I am sure you take on the major nonconformances before the minors. My rule of thumb is this - each nonconformance is prioritized by the importance to the end product, I work backwards from there. I base alot of priority on our Lead Auditors preferences as well. I tackle the Lead Auditor's "hot buttons" first! I may be different than most, but I write a corrective action for each nonconformance. I feel it may be more cumbersome up-front, but later, it helps track the entire process - sort of "cradle to grave". I hope this aids to your inquiry.
Tim

[Stormy]

Good Day,

Ted, thank you for responding.

This is my problem: we haven’t yet defined what constitutes a major vs. a minor. You said prioritize by the importance to the end product, but what is your definition of what is important to the end product? What are the criteria that you use, and how did you choose what was the most important factors?

In a different thread in this category a guy by the name of Ken Madder wrote “They were treating every cause as a Special Cause when in fact, most were Common Cause”. I would love to know how he defines Special vs. Common.

I am beginning to prioritize some of our NCR’s, by defining non-conformances that are repeatable up to a certain quantity, and recording those on a data excel sheet. The data is then evaluated monthly. This is where I have trouble. I can use total footage & the amount of the occurrences to decide which repeatable is causing the most non-conformances, and begin a non-conformance report on the highest footage and/or highest occurrence. But in doing this am I missing the non-conformances that are important to the customer? And it isn’t clear how I can carry this type of prioritizing into the non-conformance process?

So I am still searching for clarity, I guess.

But thanks again for replying.

[Stormy]

Sorry Ted, I wrote Tim, but I meant Ted.

[Stormy]

or vice versa (I think I will stop now)!!

[M Greenaway]

Stormy

The process is quite simple (well simple to outline).

Log every complaint in some form of database, and assign a code either to the reported complaint, or to the root cause if you wish to spend time at this point investigating every root cause (I would suggest you do personally).

Take the data from a suitable period of time that gives you a sufficient chunk of information, maybe 3 months. Count the fault or cause codes, arrange the counts in descending order, then simply prioritise your action on the most common occuring reported fault or cause.

This approach is OK, but slightly flawed if different complaints or products have different costs associated with them. Arguably a better approach would be to establish the cost of the complaints, and then target actions to the most costly.