Published March 2013
When I left a great company after 25 years, I was prepared to miss three things:
Five years removed, I can now confirm that those three anticipated voids were certainly justified. And while I have no regrets, in honest retrospect I’ve noted a fourth unanticipated change that may be the most significant.
The majority of manufactured assemblies undergo a final production test to provide a final stamp of approval prior to being packaged and shipped. As such, the percentage of units which pass the test on the first try, commonly referred to as first pass yield (FPY), is the ultimate gauge as to how well an organization’s quality systems are working.
One might argue that fallout at the external customer is the ultimate quality measurement, but this has at least two severe limitations. First, feedback occurs too late for any effective problem solving. Second, not all customers bother to complain; many just quietly take their business elsewhere. No news is not necessarily good news.
When a unit fails final test, it’s generally due to one of four reasons:
Let’s look at each of these in a bit more detail.
Honest human assembly errors happen. The potential for assembly errors, however, can be dramatically reduced by creative mistake proofing during the design process. Some of the most positive feedback our design team received was when a colleague reported that a group of 7th grade girls that she was introducing to the engineering profession assembled our new hydraulic motor without instructions.
Relative to gross defects in components, again thoughtful planning during the design and process planning stage will identify the most likely mishaps and how they can be prevented, or at least detected, prior to reaching assembly. Choosing production methods and suppliers based on proven quality, and not solely on cost, also goes a long way towards minimizing this issue.
A test stand is nothing more than a custom gauge. Skilled design is needed to ensure that gauge is accurate (it measures correctly), repeatable (it measures the same item consistently), and reproducible (it provides the same answer regardless of who operates it). An ongoing verification program to ensure these qualities is paramount, not only for preventing false positives but also to ensure real defects are weeded out.
Finally, production test is certainly not the place to sort out if a particular combination of “to-spec” components results in a winner. During design, specifications for individual components should be chosen which ensure “to-spec” components result in a good assembly … every time. Manufacturing processes and vendors are then selected to ensure only “to-spec” components will be provided.
Our understanding and implementation of quality systems advanced eons over the course of my quarter century career. In the early years pushing the start button on the production test stand was akin to spinning a roulette wheel. By the time I left, a failure was news. I assumed the world of manufacturing was learning and applying at the same pace. That was a faulty assumption.
Built-in-quality, as opposed to inspected-in-quality, is a key ingredient to any Lean initiative. It’s almost impossible to make meaningful improvements in an assembly cell when efforts are constantly being diverted to reworking defects.
If your organization sounds more like the company I joined back in the early 1980s than the company I left five years ago, realize that you’re leaving a lot of money on the table every single day. Find yourself an experienced quality professional who has already led an organization through this transformation.
—
Back to the Working Great! archivesCheck out the Working Great! archives for columns on other pertinent business issues
Copyright 2013 Brimeyer LLC. All Rights Reserved.