Don't Say "Zero Defects" Unless You Really Mean It

[leftoverture]

For context, I have been doing the Quality business for 45 years now, my time is coming to an end in the next few years. Over the years I have seen a lot of programs come and go and I have worked in a wide variety of industries, but my primary commodity has been plastics. With that said, I can also say I have been on both sides of the coin, I have worked for OEMs and for contract manufacturers both. But in recent years I have experienced much frustration because of the concept of "zero defects". Now I can agree that "zero defects" is a nice concept...but is it a practical one?

The answer, in most cases, seems to be a resounding NO!

When quality meets commercial, and the business case is analyzed (which might be too generous of a word), most of my current customers are simply unwilling to pay for the fixtures, tools, automation, camera inspection systems, or what have you, that are required to assure zero defects. And while I currently work for a contract manufacturer, this has been true on the OEM side as well. How many times in my OEM days did I work with a supplier to develop an error-proof manufacturing proposal only to have the bean counters say "no"?

Heck, in my current business, a lot of customers don't even want to pay for the mold tooling itself anymore let alone extra NRE costs. And where automation is required to provide a truly error-proof (if anything is truly error-proof) process, it can cost many tens of thousands of dollars. One could say that the contract manufacturer should cover that as a cost of doing business, and many do in very high volume or high margin operations, but in our low to medium volume range, the part price would have to be ridiculously high to cover that cost...and the customer won't pay that either.

Yet many large corporations put in place "zero defect" programs and try to push that to their suppliers. I can think of one example where the total annual volume is but a few thousand parts, but the process is complex and thus cost-prohibitive to automate. Sure we can 100% inspect the parts, but that is not a best practice and adds a labor cost that no one wants to pay for. The process is left depending on humans performing their tasks flawlessly, but we all know that is an impossibility, so "zero defects" isn't a realistic objective.

In fact, I had an auditor criticize us in a recent audit because we stated our goal was 100% on time delivery. His critique: is wasn't a realistic, achievable goal. Well, I agree with him and advised our team against that goal in the first place; and likewise, I think it is safe to say "zero defects" is not a realistic and achievable goal, either.

So when (if) you say "zero defects" be sure you really mean it...and you're willing to pay for it!

End rant.

 

[Bev D]

I’ve always put zero defects into 2 categories:
(1) necessary because of the severity of the defect and it’s resulting failure. Cost of error proofing etc. is less than the cost of the failure…
(2) an aspirational target (exhortation or slogan) like zero safety incidents with no real expectation or actions to make it happen…

 

[Golfman25]

You are right. In reality, the only way to get "zero" defects is some type of non-human 100 inspection system. We had a molder a few doors down, did mainly automotive stuff, and yes every mold he sold also came with the inspection gear to ensure zero defects. Like you, we do smaller volumes and have customers want "zero defects." Problem is the cost to inspect for that is more than the annual volume of parts. And the defects we usually end up with (maybe 1-2 pcs. per 10,000+) are caused by rough handling during shipping.

As for "goals," we had a similar conversation with our auditor. End result was that a "goal" has to be achievable, but a "target" can be aspirational. Se we set our delivery goal at 90% and our delivery target at 100%, etc.

 

[Enghabashy]

-as zero defect ; it’s objective ; the applicable zero defect in the sampling plan ;

-as example : in the stage of single Sampling plan ; if AQL 0,25 % ; normal inspection; the batch is 5000 items , the random sample withdraw 200; we will find the Ac., Re is (1,2);--- therefore it should be changed to (0,1 )

 

[Bev D]

I think we also need to remember that not all defects are ‘visual’ and some would require 100% destruct testing…this requires good robust design of the specifications and the manufacturing/supplier controls to ensure ‘quality be design’ in addition to some sample inspection/test. In many of my past organizations this was much more prevalent than visual defects.

 

[Jen Kirley]

-as zero defect ; it’s objective ; the applicable zero defect in the sampling plan ;

-as example : in the stage of single Sampling plan ; if AQL 0,25 % ; normal inspection; the batch is 5000 items , the random sample withdraw 200; we will find the Ac., Re is (1,2);--- therefore it should be changed to (0,1 )

Just for the sake of context, I believe the topic is one of philosophy and not sampling plans. A quick question for you: where did the chart originate? Is it from ANSI/ASQ Z1.4-2003 (R2018)?

 

[Enghabashy]

ISO 2859

 

[Randy]

Where here's the deal as far as I'm concerned.........."0-ZERO" means "0-ZERO" not some equation or portion, percentage or decimal level, it means NONE-ZILCH-NADA-CERO-NULL-and so on.

So I don't care if you make 10-100-1000 or 1,000,000,000,000,000 and you only have a single defect, you ain't at "0-ZERO"

 

[Jen Kirley]

ISO 2859

We cannot post copyrighted material. Please remove it. Many thanks!

 

[leftoverture]

I’ve always put zero defects into 2 categories:
(1) necessary because of the severity of the defect and it’s resulting failure. Cost of error proofing etc. is less than the cost of the failure…
(2) an aspirational goal (exhortation or slogan) like zero safety incidents with no real expectation or actions to make it happen…

I can't argue with your categorization...and I would not categorize any of the parts we make here (in my present job) as having that level of severity, but in a past company, some 20 or more years ago, I had parts where failure could equal death. That customer happily paid for the automation.