Random, Low Frequency Defects that Escape to the Customer are the Worst!

In my 45 years in Manufacturing at increasing levels of responsibility, I’ve seen my share of quality problems.  Many can be solved with troubleshooting.  More difficult problems require some analysis and use of cause analysis tools.  But the trickiest, most frustrating and hardest to solve seem to be very low occurrence, seemingly random issues, that escape out to the customer.

With the right approach and patience, these problems are solvable and can be addressed with permanent changes in the process – often with very little cost.  Examples we’ve successfully tackled last year to help our clients include surface scratches in pre-paint steel and torn rubber seals.

These problems get escalated to the top

When these problems slip by inspection or testing, they become a thorn in the backside of the customer causing a 2×4 to fall upon the plant management team.  It might be a single-digit ppm problem, but to the customer those 2 or 3 incidents per month put a big red mark on your customer’s scorecard.  The corresponding costs negatively impact operation performance.

To make matters worse, because they are very low occurrence issues, your OEM customer might not be checking or inspecting for the defect.  As a result, the end-users are affected (due to a fit or function failure) and the OEM takes a hit on customer dissatisfaction and high warranty cost, especially relative to the low cost of the defective part itself.

Typically, when this type of defect is found, manufacturers react quickly with containment by adding “less than effective” visual inspection and in some cases, adding very costly 3rd-party inspection.  Sometimes operations, engineering and quality personnel walk the process back, find the first process element that is the suspected root cause and they add further controls to that specific process element.

If it was just a guess, they end up adding more complexity that in reality is waste.  In most cases, this type of response does not eliminate chronic/random defects and the defect does reoccur.  Customers lose faith.  Manufacturing costs spiral upward as these containment iterations happen over and over again with each reoccurrence of a customer complaint.

Getting to the short roots

These low frequency, high severity manufacturing defects are chronic and random in their occurrence on the shop floor.  Since there is not an identifiable ‘special cause’, classical ‘Is/Is-Not analysis is not effective.  The nature of these problems is stack-up or convergence of typically 3 to 5 common cause variables which are at the extreme of their limits or which have no control limits.  If one or two are off, no problem.  If three or more are off at the same time by chance, these very low frequency problems occur.  These random interactions are difficult and nearly impossible to find.

I have been very successful with an approach which engages a cross-functional team to seek out all the potential causes of variation that could lead to the specific defect.  The process tool used is a 4M or Fishbone Diagram generating more detail than you’ve probably ever seen.  The participants are heard, ideas accelerate and buy-in to the solution is huge.  Each source of variation is ruled-out or confirmed by the team through “go see” walks on the manufacturing floor, interacting with operations personnel.  Finally, the corresponding process variation by element can be eliminated or put to “zero” with the implementation of discrete actions, usually identified and implemented by the cross-functional team.

Lasting solutions to difficult problems

Another key aspect of our approach is non-administrative solutions.  The identified fixes are not training, or updating work instructions; but rather simple process changes implemented quickly by engineering or operations.  With “zero” variation established for each process element, the random interactions that cause failures also goes to “zero”.  Hence, the random and chronic defect is eliminated.

One final note, nothing lasts forever.  Even robust error proofing and process changes can deteriorate, wear or unintentionally be modified.  To prevent reoccurrence, we have our clients verify that the solution is held in place; for example, checking for completion of specific preventative maintenance tasks, tool change frequencies, or usage of a job aid.  These quick verifications are easily implemented though a Layered Process Audit system where all levels of plant personal invest a little time to hold standards and best-practices in place.

Key take-away

While high-severity / low-occurrence problems are initially frustrating, they don’t need to be costly for very long.  The challenge is to seek out the subtle chain of events which stack-up to create these random defects.  If you would like to understand how this problem solving approach could work for you, we can help with a case study.



Submit a Comment

Your email address will not be published. Required fields are marked *