Gotta Pull Defects Out by the Roots


The current Boeing fiasco – rampant quality problems with the 787 – offer some great lessons on quality control. According to the Seattle Times, “mechanics had installed nonregulation fasteners on its 777 jets in 2008.” While Boeing stopped using the wrong fasteners, the FAA got to the very heart of things in criticizing Boeing because they “did not immediately address the quality-control issues that had allowed mechanics to install the wrong fasteners.” In other words – OK, you used bad fasteners – no major issue – mistakes are made – but what was the root cause – the non-performing process – that made the wrong fasteners possible? It is not much use to simply correct the results without getting at the root of things.

It is interesting to read the comments many of the readers submitted to the article. I suspect many of them ae from Boeing insiders who have some not-so-flattering things to say about Boeing’s lean journey. Included is the statement, “it looks to me like Boeing "cherry picked" all the sexy" parts of JIT, Lean, cellular manufacturing, and all the other "World Class Manufacturing" concepts that Deming, Juran and others taught Japan and they left out most of the parts that it takes to make those things work.” Sounds like a pretty fair criticism.

My two cents is that the problem re-enforces the case for defect mapping. Central to the concept is that every reasonable defect opportunity – not everything under the sun that theoretically, hypothetically might go wrong, but everything a reasonable person thinks has a even remote chance of going wrong – should be identified; and then looked at to determine exactly what is in place to keep it from going wrong.

Human error of the sort at Boeing (using the wrong fastener) clearly qualifies as a defect opportunity, and hoping humans don’t make mistakes is the poorest quality control imaginable. It is a recipe for inevitable failure – not an ‘if we make a mistake’, but ‘when we make the mistake’ scenario. In fact, old Motorola studies showed that quality controls that rely on a smart person looking at things to catch all of the mistakes has about a 72% probability of mistakes. At the other end of the scale is poka yoke – mistake proofing – designing processes such that the mistake is impossible to make in the first place.

The lesson from Boeing’s 787 fastener debacle is that there is no substitute for identifying and addressing defect opportunities before they become actual defects. After the fact quality control is virtually no quality control; especially when it only addresses the resulting defects and not the weaknesses that underlie the process that makes the defects possible to be created, and then not caught.

Subscribe to Our Blog

Related posts

Digitizing the Invoice Approval Process to Turn AP into a Profit Center
4 Ways Technology Can Help Manufacturers Improve Efficiency and Profits
Playing to Win, or Playing to Tie
You Can't Improve If You Think You Are Perfect