Search the Elsmar Cove!
**Search ALL of** with DuckDuckGo including content not in the forum - Search results with No ads.

Informational Nonconformances and Defects - Operator Error, System Error, or both?

In the event of a NC or defect, what/who is at fault?

  • "The system" is always at fault.

    Votes: 8 13.3%
  • "The system" is at fault ~ 90-96% of the time.

    Votes: 19 31.7%
  • "The system ~ 80%", operator ~ 20%.

    Votes: 21 35.0%
  • It's about even.

    Votes: 12 20.0%

  • Total voters
  • Poll closed .


Involved In Discussions
It's unavoidable these days, but the point was that you might want to revisit classification. You shouldn't be counting it as 10,000 bad parts; it was one bad shipment.
I brought that up to our CEO, but he said it was by part. The customer called it parts too, so I was overruled :bonk:
Paid Advertisement - Forum Supporter

Mike S.

An Early 'Cover'
How to classify, "count" or track and report the quality escape (1 bad shipment, or 10,000 bad parts or...) is kinda off topic, depends, IMO.

How does the customer feel about it, how does it affect them, and how do they count it in your quality ratings?

1 bad shipment with 1 bad part probably impacts the customer much differently than 1 bad shipment with 10,000 (all?) bad parts. Every measure must be taken and considered in context.

I've had similar things happen. Some years there are 5 escape incidents for 6 total bad parts and some years there are 3 incidents with 130 total bad parts. Blindly looking at the numbers without context could cause management to make some suppositions that aren't true and make decisions that are not in the best interest of the company. "The rest of the story" is always important.

Kevin Mader

One of THE Original Covers!
Staff member
Andy hits on a few of the contributing factors that affect personnel/people. In the example, 10,000 of the wrong part were sent to the customer. Seems it might be more obvious if all 10,000 were the same wrong part, but perhaps less obvious if a 1,000 similar 'bolts' (weight, shape, type, size, etc.) were mixed in. The clerk probably isn't doing 100% visual inspection of each 'bolt' (automation seems like a possible solution). Human factors are real and need to be considered. People are fallible. Understanding this/these help to design processes/products that are less likely to get wrong, but never 100%. Some portion of the blame will inevitably be from human contributions. The point though is that most of the low hanging fruit reside in System faults and their elimination. Investing there will in all likelihood, yield the better results.

Kevin Mader

One of THE Original Covers!
Staff member
Mike - it may come down to the metric (e.g., count of mis-shipments or % of parts ship). Either could be useful. Context is everything (as you know)!!

Jim Wynne

Super Moderator
  1. Something bad happens. It might affect the customer, or it might affect the supplier, or both.
  2. Fix it, whatever it is. Get the customer out of trouble ASAP, if the customer is in trouble.
  3. Keep track of the number of times bad things happen, and what is done to keep them from happening again.
  4. Follow up with verification of the fix, whatever it is.
  5. Go to #1.
PPM measurement is a complete and confusing waste of everyone's time, especially if you're on the supplier end and have to deal with twenty different ways of calculating it. Most of the people who have to track PPM will never produce a million of anything. It's things like this that make me glad to be retired.

Kevin Mader

One of THE Original Covers!
Staff member
I think you are mostly right Jim (re: PPM). I think it has merit for those operations that have high volume parts (e.g., automotive hardware/fasteners). Otherwise, most of the important numbers are 'unknown and unknowable'. Picking meaningful metrics must be deliberate.

Enjoy your retirement!

Jim Wynne

Super Moderator
If things go wrong at a rate where PPM is considered useful, measuring via PPM isn't going to help anything. It's like being buried in an avalanche and trying to count the snowflakes. You don't need PPM calculations to see the bloody obvious.
Top Bottom