Best feedback loop practices in defect communication between Quality and Manufacturing

Bev D

Heretical Statistician
Super Moderator
Maybe maybe not. I’ve run several COPQ programs in 3 different organizations. In two I had it about 95% automated. While it helped with some leadership it fell on deaf ears in others. Think of Boeing. No amount of COPQ could have swayed that organization. Why? Because the leadership made oodles and oodles of money before the door blew off and 2 planes flew into terrain. They got theirs… and that’s all that mattered.

I’m not saying the OP shouldn’t calculate the COPQ, they probably should if the defect rate itself doesn’t move the needle. It is a lot of work. Later I’ll post what I’ve done and what to avoid - as the OP asked.

Bev D

Heretical Statistician
Super Moderator
Here are few important cautions about COPQ:

Don’t try to estimate the cost of things that can’t be accounted for. These include goodwill, Customer loss or defection, regulatory or legal fines, and morale/employee turnover. While these things matter, they are impossible to determine - or be believed - until after they occur. The naysayers will simply try to argue that the events probably won’t happen.

Deming famously said that some things matter can’t be counted and some things that can be counted don’t matter. Understand the difference.

Don’t worry about the small costs. They do add up but are probably the last things you should go after. These include ‘low cost scrap’ that just get tossed and one-offs that have negligible severity.

You will need to also track the severity of the failure if it is an external failure (and some internal failures if safety is involved). I don’t know what your product is nor do I know what your Customer is (another manufacturer or the end user?) but don’t discount low cost items that fail frequently. Defect rate matters as it affects the Customer in ways that are difficult to dollarize. Example: I worked for a recreational automotive parts company that made after market winches. We had several motor failures a year - very high cost to replace. We also had hundreds of relay switch failures a year. Very low cost to replace, in fact the dealers often just kept them in stock. Both caused the winch to be inoperable and the user only found out when they were trying to use the winch. The user didn’t care about the cost; they were p!$$ed that they were right and truly stuck in the middle of nowhere. That cost the user time and money or embarrassment if their buddy had to winch them out.

Be absolutely transparent about the rules you use to determine the cost of poor quality. ‘Production number only’ and lazy or incompetent people will tend to spend more time trying to tear your cost figures apart than it would take to just solve the problem. These people like to ‘yeah but’ stuff to death.

Be ready for this one: we don’t have time to get it right the first time; we need to get to market fast and be the first. The profit from the early launch will exceed the failure costs and we can always fix it later…again see Boeing. I have lived this in more than one market. It goes hand in hand with executives making oodles of money and not experiencing any consequences for poor quality. No matter what, too many Boards will excessively compensate their executives including the golden parachute…

Time can be more important than money. Another reason to track defect rates or Yields in addition to cost of failure. Example: I worked for a third tier automotive company once and we were coming up on Job 1 (which is not quality it is the first production run of a new model). We had a 50% yield on a critical part. The cost of failure was total scrap and rebuild of the part, but it wasn’t horribly expensive. The real problematic effect was that at 50% yield loss we were incapable of meeting build quantities even if we worked 24-7. That would have meant the End of our contract. Now there were a lot of opinions about the cause and the ‘fix’. Some were plenty expensive and some were cheap. None of them worked. Finally we went in and actually determined the causal mechanism; it only took 2 hours of actual time and 2 days elapsed time (heat treat was involved) The cost per scrapped part was $120. (Straight out cost of the part per finance) and the solution cost us $25 more every 18 parts. In my experience the cost for the solution when you actually know the true causal mechanism is far less than the cost of the failure. I’ve seen a lot of very expensive solutions that never worked because they didn’t address the cause.

DON’T include ‘prevention’ costs. This is the old Cost of Quality (COQ) metric. It is an old abandoned by most metric because the cost of prevention is the cost of doing business right the first time. It has been used for decades to justify never solving anything and cutting corners. (Again see Boeing). Almost everything you do is a prevention cost. From documentation to buying materials that cost a lot of money. Remember you are determining the cost of poor quality, not good quality.

DO include a reasonable finance person…this helps with credibility and can be an eye opener to them.

These are simple obvious things. There are other more subtle things to COPQ but that is probably worthy of a couple of chapters in a well written book…

Bev D

Heretical Statistician
Super Moderator
And here is a list of ideas/examples of things to quantify for COPQ

DIRECT COSTS: all materials, components and special equipment used to correct defective material
  • Rework: manpower, materials, components and re-inspection and test costs.
  • Scrap: the cost of all material, components and manpower used to create the product to the point of detection. If scrapped the inspection and test costs are included as the replacement product will also under inspection and test.
  • Packaging and shipping costs of any returns.
  • Containment inspection and test including inventory testing and inspection to determine the extent of escaped product.
  • Remediation/replacement cost of correcting or recalling defective product from the field Materials, shipping, packaging, travel, etc.
  • Warranty and concession costs.

INDIRECT COSTS: all manpower performing work to correct and prevent future occurence of defective material.
  • Containment manpower: developing containment actions and performing containment actions
  • Engineering and Technician cost of investigating, developing and implementing corrective actions
  • Manpower cost of Customer Service
  • The administrative cost of handling NonConforming Part Reports (NCRs).
*Replacement material and component costs are cost (purchase price of the replacements) not price (charged the Customer)
*Manpower cost is typically the average wage of all personnel who perform rework, Customer service, warranty investigation of returns and warranty repairs/rework.
*NCR cost is the administrative cost of processing a NCR; it is typically an average cost or estimation. It doesn’t include disposition assessment, scrap cost, rework cost, or investigation & correction costs. These are covered separately as they are unique.

It is very helpful to create process maps on internal and external failures that include steps for once a defect is detected to when it is corrected. This will help identify what costs are incurred and if you have an ERP system where to collect the data…

‘SOFT’ COSTS - These are very difficult to assign a dollar value; they do cost ‘time’
  • Increased cycle time
  • Delayed/Late shipments
  • Higher inventory (storage)
  • Manpower needed to expedite and track orders delayed by defect rates, disposition defective material, manage inventory (cycle counting and movement), reorder material, explain variances…
  • Reduced confidence in quality and delivery can slow sales
  • Loss of Customers
  • Regulatory and other legal fines
Top Bottom