Seth Godin, marketer supreme, also has commented recently on quality. In Quality and Effort, Seth relates a story about creating systems to avoid bad experiences for users. In his first iterations, the result was pretty good, but not perfect. He describes how he changed the system of creating quiz questions/answers to have half the team create the questions and the other half try to come up with better answers and pay a bounty if the new answers were more correct than the original.
The result of the new system? Zero error for the next 5,000 questions.
My last example is Philip Crosby, author of Quality Is Free and Quality Without Tears. His main definition of quality was conformance to requirements – the product or service should meet or conform to whatever requirements are set for that product or service. Another part of his quality system was that processes should be done right the first time, that there should be zero defects in the process and the product/service. Many companies and industries resist this level of quality. but Crosby points out that many companies have zero defects in their payroll operations. To quote from Quality Without Tears, pp. 75-76:
Payroll doesn’t make mistakes.
Is that because they are such dedicated souls? Certainly they are, but the importance of the work does not necessarily raise performance standards. If that were so, one would think that people working in space exploration would never err. However, you can get used to anything, and bad performance standards occur eventually.
The reason that payroll does so well is that people just won’t put up with errors there. They take it very personally when something is wrong with their paycheck.
So, in this example, the expectation/standard is “zero defects” – no errors.
As a software engineer working in the aviation industry, where we have many processes and checklists, I work to create software and artifacts (requirements, tests, etc.) without defects, but most of the time, my fellow workers and I manage to make a few mistakes along the way. Maybe it is time to see if we can construct our systems to prevent such errors. It’s worth a try!