Total Cost - Quality From A Cost Perspective

What costs are driven by quality and reliability? The answer clearly depends upon both the magnitude of the problem and the point in the process at which the defect is discovered. Some years ago, the President of Ricoh™ described the cost of a single quality problem. It was something that clearly caught my attention. In that he said that for a single problem the costs were:

Point at which the defect is caught	Cost
In design	$35.00
Before buying.	$177.00
Before manufacture.	$368.00
Before the product is shipped	$17,000.00
At the customer’s location	$690,000.00

As time has passed and business models have changed, the cost of Quality is probably even greater and while companies are able to reduce their cost of the supply chain, they still have difficulties in managing the cost of quality. With the advent of Supplier managed quality and the elimination of most incoming inspection by Buyers, the impact of a quality problem is even greater as the quality problem with the product will frequently not be identified until it becomes production level fall-out that would require either re-work, or possibly scrapping the product if the cost of the re-work is prohibitive. 

What are the costs of quality and reliability? The cost attributable to quality problems is the one of the most difficult of all the total cost calculations to make and is really driven by the point at which the defect is identified. In other calculations you only have one or two variables whereas with quality problems there are multiple variables that have to be considered.

The cost of quality consists of the following elements:

• The increased cost for material, as a result of an increased safety stock necessary as a result of quality problems.
• The increased cost of inspection at incoming inspection.
• The cost of handling a failure, when the failure occurs at incoming (including administrative costs to return).
• The cost of handling a failure, when the failure occurs in process (including administrative costs to return).
• The cost of field defects.

From a safety stock standpoint a bad item is the same as a late item. As such the calculation of safety stock cost can be calculated using the same formula as safety stock for late deliveries, which is a simple cost of inventory calculation. The amount of quality related safety stock is in addition to any safety stock required because of delivery or quantity problems. In a worst case scenario the delivery could be late, when it arrives the quantity is less than ordered, and there are quality problems with the product that was delivered.

The second element in calculating the cost of quality is the increased cost of inspection required at incoming that was necessitated by the reduced level of quality. For example, if your program was to be managed without incoming inspection and the level of quality was so poor that it now required one hundred percent incoming inspection, the increased cost of quality includes those incoming inspection costs.

In addition to the added inspection cost, you also have the cost of failures at incoming. This not the same as increased inspection costs and is not double counting. The rationale behind this is while the substandard quality levels forced an inspection initially, the failure at incoming will require an additional inspection of the repaired or replaced material.

With any reduced inspection program you will have the risk of a failure of the material in process. The Administrative Costs associated with a failure in process are virtually the same as a failure at incoming. The difference rests in the cost of the failure rather than the cost of the inspection. The cost of a failure in process can be great. For example: If a subassembly was imbedded in a larger product and did not fail until the finished product was turned on, the cost of that failure would be the cost to disassemble the unit to a level at which the failed unit could be replaced with a new one. The cost could also include the direct expense which results from the failure such as lost time in a assembly line production when the line has to stop because of a failure. In addition, if the problem was not deemed to be unique, it could also result in a purge of all potential affected items with resulting costs for all the re-work that would need to be done.

Reliability costs include costs associated with reliability problems which occur in the field such as spare parts and repairs inventories, cost of service calls, cost of replacement, processing returns of failed material, etc. The cost of reliability problems is difficult to compute for several reasons. First, most reliability failures occur in the field and you must rely upon the quality of the failure data that is received from the field. Second, because of the relative low cost of some items, it may be more economical to throw away the failed material rather than send it back through a repair loop. The elimination of the repair loop eliminates failure analysis. Third, there will be failures that occur without your knowledge, as not all material will come back through your channels for repair. There are two types of basic reliability failure costs. First, are the infant mortality costs which occur either in incoming inspection, or testing of a higher level system. The second type is when the statistical average of failures which occur in the field fail to meet or exceed the agreed to mean time between failures (MTBF) of the item. Infant mortality failures are reliability failures that usually occur in the manufacturing or test process. The cost of such failures is the same as an in process failure.

THE COST OR WARRANTY AND FIELD FAILURES.

The cost of a failure in the field is the most expensive of the quality/reliability cost factors. The cost includes:

• The base cost of a service call.
• The installation cost (After the technician arrives, how long does it take to have the part removed, the new part replaced and any testing performed).
• The administrative cost to return the item for repairs. This includes the cost of verifying the failure, issuing orders, packaging etc.
• Throw away cost for items that aren’t repairable.
• The cost of repair of the failed item. Repair costs if the items that failed are usually only covered by a warranty of a duration that is far less than the agreed to MTBF.
• The cost of repair or re-work cost for the product the failed item was assembled on.
• The increased spare parts stock. Spare parts stocking levels are initially set based on projected failure rates but will be adjusted periodically to reflect the actual failure rates.
• The down time and losses that are encountered by the end user.
• The effect on service revenue profitability (since a large part of the basis of establishing the service charges is the projected reliability).
• The effect on future sales, since sales in many cases are based on total cost of ownership or amount of "up" time.
• Less the value of any recoveries provided by warranties or epidemic defects type provisions.

When you look at all these different costs, it’s easy to see how the President of Ricoh identified the cost of a defect in the field as being multiple hundreds of dollars.

The simple fact is that standard warranty provisions only cover the repair, replacement, refund or credit for the specific part. Warranties do not cover any of the other costs associated with the defect such as:

1.     The service call to identify the defect and remove the problem item from the customer site and replace it with a working field replaceable unit (FRU).

2.     The cost to return the failed product to a repair center.

3.     The cost to “re-work” the product to remove the failed part and replace it with a new one, or the cost to scrap items that are not repairable.

4.     The cost to manage warranty replacement of the failed part.

5.     The associate cost of inventories of FRU’s required because of the expected failure rates.   

If all the Supplier pays is for the cost to replace the failed part itself, the Buyer is assuming all those other costs for a problem over which they have no control.  Here’s an example. Assume a printed circuit card that the Buyer uses de-laminates after the Buyer has placed all the components on the card. Some of those components cannot be removed from the card without being damaged, so the de-lamination would cause the scrapping of components, plus all the costs associated with removing the other components and placement of them and the new components on a new card. The cost of the card itself may be minimal in comparison to the cost of the components that were damaged and had to be scrapped, or the cost of the rework. Providing you with a new printed circuit card does not make you whole for all those other costs.

As the costs of a field failure can be staggering, any time there is a problem you can expect that even if you have all the right protection built into your contracts, the Supplier will look for every way possible to avoid that liability. It is at that point when the specifications become extremely important. To avoid the cost Suppliers will look for some factor to avoid liability. Did the product’s use meet the specified parameters? Was the problem caused by the way it was handled or assembled? Was the problem caused by causes beyond their control such as electro-static discharge? Did the Buyer provide any portion of the design that could have caused the problem? Were the damages caused by improper use of the product?

Most contracts include “limitation of liability” provisions that prevent the Buyer from recovering anything other that direct damages, and the impact of this is most of the costs a Buyer encounters with a defect in the field are incidental or consequential damages, and as such would be excluded from any potential recovery unless the commitment to pay those associated costs is excluded from the limitation of liability.

When you don’t have the incidental costs associated with a warranty failure, your cost of warranty calculation becomes slightly easier. The calculation is what is the expected number of failures that will occur in the period and what is the cost of the failure (which may be the cost of the replacement or an out of warranty repair cost). The expected number of failures time the cost per failure divided by the projected volume would be the additional cost added because of the difference in warranties. The expected number of failures can usually be calculated based on the stated reliability of the product. So if a product had a Mean Time Between Failures reliability of 100,000 hours, that 100,000 is the average and statistically there should be a certain percentage of failures that should occur in each year, the average of which creates the mean. Normally the failure rates will be high in the first year (attributed to what they call infant mortality problems) and be low, rising up over time as the product begins to reach its useful life. If you can statistically predict the failures that should occur in those periods and know what your cost of an individual failure is, you should know the potential total warranty cost difference which you would then spread over the estimated purchases to determine the unit price impact on a total cost basis