Snowball to HELL!
There is only one way to describe this scenario and that is via a story. Consider the organization that is coming to the end of the project. The product is a complicated subassembly that goes into a larger system and has numerous interactions and incarnations of the design. They are late in the delivery of this key subassembly. To make the scheduled delivery date, they reduce the testing to a spot check and then launch the product upon their customers (as in let slip the dogs of war [Shakespeare, William. Julius Caesar. Act 3, Scene 1, line 273).
After some months pass, they discover they have a serious quality problem. Not life threatening, but costing a few millions of dollars every month in warranty dollars. The company develops the corrections to the problem. How do you suppose they check the adequacy of the correction? Do they use maximal effort? Did they learn from the previous fiasco?
Due to the high warranty cots, the company chooses the time honored technique of spot checking again. The corrective action goes into production after a spot check. The company’s actions to “cut corners” and hurry the release cost them a quality problem. Now for work that was not planned as part of the project, they find themselves in a position to do the same things they did before to reduce the warranty cost problem. In their efforts to expedite the closure of the warranty problem, they are in fact, subjecting themselves again to a warranty problem.
The company loses money due to the quality problem. However, it does not end there. Now they have resources that should have been dedicated to new project work still cleaning up the newly-launched product. To be able to meet both of these demands, we cut corners on the development work (dispersing our engineering talent). The company is setting themselves up for another, similar failure as they diffuse their engineering talent to handle the quality problem and the new development work. This cycle continues as we watch the cost of poor quality steadily rise.
The truth is, testing cannot guarantee the quality of the product. Sufficiently complicated products take considerable time to test. We may also suffer from environmental exposures or customer use issues that we could not imagine as part of our testing protocol. However, it is equally true, if you do not look you will not find the problems that will come back to haunt you. Abbreviated looks or spot checks do not ensure the quality of the product and increase the organization’s exposure to the cost of poor quality. Whenever we spot check, we should have a contingency budget to handle the inevitable issues with poor quality. We think we are making things better by meeting the delivery date, but the cost of poor quality should be weighed against the cost of introducing the product later.
We named this snowball to hell, because when the snowball hits the hellish heat source, we see an exposion of steam. Some wonder, how this could happen so quickly. In reality, like the financial crisis, it did not happen instantly – the end result was driven – action – by action culminating in the visible conclusion.