November, 2018

A Product Management Fable…

Your company spent two million dollars reaching six-sigma goals by implementing the best quality management system it can afford and earning its SAE AS9100 Registration. Your CEO is so proud of this accomplishment that the company has adopted the motto “Quality Comes First.” With renewed confidence the firm has bet its future on launching a new product line projected to increase its Aerospace and Defense sector market share by 15%.

But wait, only one out of the 200 test units passed thermal testing! You urgently convene a Failure Review Board. After six weeks of investigation no one can find the root cause. Your company’s reputation is on the line and the CEO is embarrassed, especially after making a very big deal over quality standards.

All the units produced were found to be built to the engineering drawings and passed pre-environmental testing reviews. What the heck is going on? You do use parts from a government approved overseas supplier, so you worry “is this a failure in their quality control or worse, perhaps even intentional sabotage?”

Here Come the Consultants…

The CEO has no other options at this point but to call in outside help. Every step taken to find the root cause and provide documentation is reviewed. Shop orders show that part substitutions were on the approved part substitution list.

The consultants discover that your six-sigma approach to product quality eliminated tracking the as-built configurations. Since it didn’t make any difference if the substituted parts were equal to or better than those called out in the released engineering it was thought to be non-value added.

The new consultants also learn that you eliminated integrated product teams by creating functional Centers of Excellence (CoE), as recommended by the old consultants. Each CoE was set up with a center-unique change authority and operates independently. If one CoE interfaces with another it is done with a statement of work and a specification. After product level requirements are allocated to the centers by the Systems Engineering COE, the other CoEs are free to add factors of safety as they see fit.

The outside experts next spend time with each of the CoEs. Whenever parts or part substitutions come up they are referred to “good ol’ Bob” who runs the parts supply room. When auditors speak with Bob they ask if he remembers anything out of the usual in the process of kitting parts for the test units. He pauses and replies. “The Supply Chain CoE ordered enough parts for the first 199 units. After they went through receiving and inspection I sent them off to assembly. The Quality CoE decided they wanted one more build because running six-sigma metrics on 200 units made the math easier. I had to substitute the more expensive stuff we already had in stock on that one. I wrote it all out in the log.”

Findings…

In less than a week the experts present their findings. They conclude that the failure of 199 units in thermal test was due to a form of sabotage, but it was internal, not from an external supplier.

The Systems Engineering CoE did not inform the Test CoE that the requirements allocated to them already had a 150% thermal factor of safety built in. This lack of knowledge resulted in the Test CoE applying another 150% factor of safety and thermal testing at 225% of the required operational environment. The investigation further revealed that the Systems Engineering CoE had unilaterally adjusted all environmental tests by a 150% factor of safety before allocating them to the Test CoE. Had the 200 units not failed thermal testing they probably would have failed acoustic, operating angles, acceleration, EMI, vibration, or one of the other environmental tests.

Had that communication between the Systems Engineering CoE and the Test CoE taken place it is doubtful that any units would have failed. The single unit that survived was a fluke. It contained substitute parts that were less susceptible to heat. This over-testing resulted in a loss of $1M in labor costs, parts, and two production runs plus consulting fees.

By not embedding Configuration Management as an integral part of a company-wide quality management strategy your own organization assured the failures. It made the classic mistake of not being able to capture in less than five minutes what was issued to the factory floor or the configuration of the unit under test. You are not alone. Knowledge workers spend 30% of each day looking for data with a 50% success rate. Configuration-related data is a very valuable asset that needs to be as professionally and aggressively managed as any other quality control asset.

Lessons Learned…

The lack of a robust configuration management implementation turned out to be the greatest quality management issue. The company sabotaged itself by creating information silos which did not operate under a common Configuration Management Plan as an integral part of the overall Quality Management Strategy.

The CoEs agreed they can’t truly run like separate companies. Information exchange through Configuration Management was just too critical to be ignored or lost in silos. The CEO changes their charters to mandate:

  • Integrating Configuration Management at the enterprise level and flowing throughout the organizational functions.
  • Integrating test with design and systems engineering at the product level.
  • Product level change control boards with representatives from all the centers.
  • Vetting of changes across all stakeholders with wider distribution of released engineering.
  • Capture of As-Built, As-Tested, As-Delivered and As-Maintained configurations by serial number.
  • The entire organization to understand that configuration management, just like quality control, is not a function to be relegated only to the CM group or the QA department but belongs to everyone.

What Next…

While this management fable was an imaginary case study you can learn more in a real-world demonstration of how A&D industry-focused Configuration Management software tools like PDMPlus CM contribute to improved product quality by contacting CMstat at information@cmstat.com or requesting a demo HERE.