Reproducibility between inspectors documenting Nonconformance's

[Candi1024]

I need ideas.

I have 3 qualified inspectors in final inspection. Our product is highly customized, we make box trucks (not the Chassis) and each unit is completed to customer specifications.

My issue is that each of my inspectors have different average NCRs. (nonconformances) We have a few standard forms we use for inspection, then the work order is the bible. Our NCR form is free fill form, in other words they describe each NC they find, in their own words, on a single line. They also inspect the unit while it is undergoing final finish, as I'm still fighting for dedicated inspection time. One inspector will have a long list and poor communication with the operators. On the other end, we have an inspector with great communication who will write hardly anything on the list, due to when he finds something they are right there to fix it. And he does this while they are completing the unit.
It is impossible with our current systems to make an exact checklist for each one.

So in other words, the number of NCRs I get are repeatable within one operator but not reproduceable under multiple operators.

How can I standardize my results? I've spoken to the operators, and the one operator is writing down more. But I need a more formalized way to manage this.

My ideas:
- a standard checklist, but we don't have the systems to automatically create one based on customer order
- a dedicated inspection where production hands over the unit to inspectors. I get push back on this because they want our inspectors to find it as soon as possible.

HELP!

 

[Sidney Vianna]

On the other end, we have an inspector with great communication who will write hardly anything on the list, due to when he finds something they are right there to fix it. And he does this while they are completing the unit.

Isn’t that the best practice you want to have the others emulate? I would.

Also, from what you describe, you are not identifying defects, but punch items that get resolved right away. That is typical for complex assembly jobs. Having the inspectors working hand in hand with the operators help in letting the assembly workers knowing what are the acceptable workmanship standards so they don’t repeat the same mistake next time they assemble a similar unit.

The idea of letting an assembly getting fully built and allowing a dedicated inspection time is not as effective for obvious reasons. The best (hardware) quality is the one that is built in the equipment, not a quality by inspection approach.

Good luck.

 

[Candi1024]

Isn’t that the best practice you want to have the others emulate? I would.

Also, from what you describe, you are not identifying defects, but punch items that get resolved right away. That is typical for complex assembly jobs. Having the inspectors working hand in hand with the operators help in letting the assembly workers knowing what are the acceptable workmanship standards so they don’t repeat the same mistake next time they assemble a similar unit.

The idea of letting an assembly getting fully built and allowing a dedicated inspection time is not as effective for obvious reasons. The best (hardware) quality is the one that is built in the equipment, not a quality by inspection approach.

Good luck.

Thank you for responding!

Yes, I understand you can't inspect quality into the product, but at this point they are using us to create a checklist of what needs to be finished, instead of taking the responsibility on themselves. By separating the inspection, I could call out escapes and work to prevent them.

Also we do have inspections further up the line. This isn't the only one. But it is the one that is most integrated with the process. If we don't record what the operators are missing, how can we take further actions to prevent them?

 

[Kurt Smith]

If your goal is to be able to keep general track of the kinds of defects and rates, you might consider defect codes rather than having the defects written out every time. If the problem can be solved faster than writing a sentence, anyone would understandably fix it first and document never.

You could make categories for minor issues and reserve the full explanations for issues that require more significant investments to resolve. If you keep the categories simple and quick to use, your inspectors could use tally marks or check boxes to indicate how many instances they find per inspection. That way you can push for the higher communication style inspections while also keeping some basic metrics to point you towards improvement areas. There will be a tradeoff between speed and data collection. The proper balance varies business to business.

The challenge with this kind of data collection is the culture it brings. If the workers feel they will be penalized for findings they will hide them. And the inspectors will as well to try and be kind to their fellow workers.

 

[Golfman25]

Have your one inspector train the others.

Standardize anything you can.

 

[ScottK]

Do you have standardized list of defects that the inspectors are working from? I understand that the work order is the bible but how to the inspectors know what to call out?
With that I might standardize and simplify how defects are called out so on the inspection sheet it's really more about documenting how many of what types of defects are found in process and use that info for root cause analysis and process improvement.

 

[ChrisM]

I agree with Sidney, use the knowledge of the inspectors to effectively enhance the training of the operators and raise the "quality standard" of normal work. Helping operators to see that what they may think is acceptable will actually become a nonconformance if left alone, and therefore they need to "do better" for want of a more explanatory phrase, is surely better than constantly returning assemblies for rework.

It will also help the mindset that the inspectors aren't there as "the police" to reject anything that they don't think is correct, but they are there to help make sure that things are done correctly in the first place..... what was the old addage..... "Get It Right First Time" (GIRFT)

 

[Bev D]

I’m really trying to understand what the problem is…although the specific non-conformances may differ from build to build there will be common themes. And their nature will dictate the problem solving methods to bring to bear.
It is not necesary to have detaield accurate tracking of what has happened in order to solve the problems. If one is successful in solving the problems - without blaming the operators and instituting ‘more training’ and more discipline’ you will get to a point where the data will be accurately recorded for new and recurring problems.

What kind of problems are you seeing? Are they actual defects like cracked metal (physics) or are they missing components (human error). If you give us a few specific examples we can better help you…

 

[Jtavo]

Could be a case of quantifying the qualitative? The formalisation is to get all of them to behave like the fixing communicator. Be the coach for this and measure how effective that might be (though that might come out as a different measure down the line..) Oh gawd - another Deming quote has popped up (from another reply of mine) "Cease dependence on mass inspection to improve quality. Eliminate the need for inspection on a mass basis by building quality into the product in the first place.." Get the star player amongst 'em..