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
    60
  • Poll closed .

charanjit singh

Involved In Discussions
Talking about operator (and also inspector) errors, many years ago it was estimated after a number of experiments, that 100% visual inspection is effective only between 80 - 90% on the average.

Years ago, in my lecture classes, I used try out a game to see how far this was true.I would distribute a copy of a small paragraph to each participant, give them a fixed amount of time (e.g. 30 seconds) and ask them to write down the number of a given letter (say 'a' or 'd' etc.) they could find. After the specified period I would ask them to read their answers one by one. And there would be laughter when one person said '10'. the next said '13' followed by 8, 12 etc.. Invariably i found that the average error was within 10 to 20 % in a group of around 20 participants.
 

Mike S.

Happy to be Alive
Trusted Information Resource
Wow, it has been nearly 15 years since I started this thread and it still gets sporadic hits. It is probably a timeless discussion.

Of course, charanjit, inspector accuracy will depend on a lot of factors. For example, if you are trying to separate red items from green ones, or 1 inch items from quarter inch items, the inspection accuracy will be higher than for looking for small cracks in a part, for example.
 

charanjit singh

Involved In Discussions
That's true Mike. In fact there are many psychological and physiological factors that contribute to the errors one may make at work. What goes on one's mind during performance of job may affect his/her concentration.A feeling of hurt, depression, or mental images of one's expectations during the forthcoming vacation are only a few instances one may think of.

In this context, I remember reading somewhere that in early 1960's when Martin Marietta Corporation was building modules for a space program, they used to conduct 17 consecutive visual inspections of soldered joints of the same module, based on the fact that a single inspection may be effective only around 85% in eliminating defects. And that would give a probability of defects= (0.15)^17!. Of course the inspectors did not know it was the same module they had inspected earlier.
 

LUV-d-4UM

Quite Involved in Discussions
From the Deming Conference last weekend:

A presenter offered this explanation – only when the root cause is traceable to a Company Value that was willfully broken can the employee be at fault. These ‘company values’ are developed by everyone in the company, not just management. Everyone needs to operationalize the definitions so that all agree what is “is” and live by them as a culture.

All other failures are attributable to the System.

Please give me an example of such situation. Thank you. :applause:
 
G

gutierrez

MS and MG

I like your arguments, I do agree there is no removing human error but this error should be substantially low, as in way less than 25%, like 5%. When human error is the cause like with Joe, how do you correct it? You re-train or if their overworked, decrease workload, give Joe the next 3 days off to grieve?

Now what if you were not able to select human error or only after you have eliminated all other causes? You will find that Procedure/Process will be the most relevant cause. Given this, the procedure can be updated or process modified, which is much more beneficial than re-training especially if it was an extra ordinary circumstance like Joe, there is no way to correct that even re-training is not appropriate.

I have to go with maybe not 100% but 99% should not be human error, that will provide a much better outcome. In addition, having too much human error will be pointed out by regulatory bodies as it should be a very small percent.

Good conversation
Ginny
 

japayson

Involved In Discussions
Joseph Worker mis-packed 20 cartons. Now if this were standard product, could the packages be weighed to check and stop mis-packs from shipment? Would lack of such a check, if applicable, be a system failure?

Ii have seen cases where installing such a check was called a waste of money and vetoed, on the basis of "we PAY the worker to do it right".

I have seen physical assemblies that allowed components to be installed "right", or "wrong"---- only visible under microscope--- when a simple design change to an asymmetrical dimension would make the component fit only one way. Would operator errors be only on the operator, or partly (maybe a lot) on the system if it could have been designed to prevent the mistake from even being possible?

I think it is not a simple thing to assign blame. Instead of spending much time on blame it may be better to work on preventing the possibility--- or maybe more realistically reducing the possibility of the operator error.
 

Ninja

Looking for Reality
Trusted Information Resource
...my two cents...quoting from "Black Rain"...

Fix the problem, not the blame.

For most mgmt, it matters who screwed up. For solving the problem...ignore the people, and focus on the system.
Yes, it may in fact be the person who is the problem...but fix the system anyway...
 

Kevin Mader

One of THE Original Covers!
Leader
Admin
Deming initially attributed 94% of all problems to the System leaving only 6% attributable to an individual. Over time, he got to a 98%/2% split before his death. In short, Ninja is correct: focus on the system to fix the issue.

Sadly, Theory X managers are concerned about who to blame when in reality, management owns the blame most of the time (94-98% if Deming is right). It seems irrational to pursue this route as the primary focus but it's a common practice. As for adding a scale at the end to determine the correctness of the packaged carton, keep in mind that 'weight' may not be a valuable indicator (e.g., a hex head bolt may weigh the same as another type of bolt or one of the constituent items in the package).
 

CCaantley

Involved In Discussions
We had an incident last year where our shipping clerk shipped 10,000 of the wrong part to our customer. There are shipping instructions that give instructions on part number. number of parts per box, box size, full box weight and include a picture of each part. I'm curious as to where everyone who believes the system is 100% at fault thinks the system is to blame in this case? Having worked in shipping for a time myself, I never sent the wrong parts given the tools made available.
 

AndyN

Moved On
We had an incident last year where our shipping clerk shipped 10,000 of the wrong part to our customer. There are shipping instructions that give instructions on part number. number of parts per box, box size, full box weight and include a picture of each part. I'm curious as to where everyone who believes the system is 100% at fault thinks the system is to blame in this case? Having worked in shipping for a time myself, I never sent the wrong parts given the tools made available.

Good question! Personal experience is just that, personal. The shipping clerk isn't you. Also, I'd ask if (objectively) the situation in shipping is IDENTICAL to when you worked there? Maybe it's busier. Maybe something (unique) was going on to distract the clerk. Who knows? To imply that instructions alone and following them is going to result in things the way you did them is incorrect, IMHO. It's like saying that you've never had a car accident, all drivers have to go through drivers' ed and therefore, there should be zero accidents...
 
Top Bottom