Got something to say to this podcast? Send a voicemail, and you might hear yourself on the next episode!

Six Sigma sets a goal of 3.4 defects per million opportunities. Engineer Bill Smith introduced it at Motorola in 1986 as American manufacturers lost ground to Japanese competitors, and Motorola trademarked the name and credited it with more than $17 billion in savings by 2005. Jack Welch made it General Electric’s mandated strategy in 1995; GE reported $350 million in savings by 1998 and later more than a billion, and by the late 1990s about two-thirds of the Fortune 500 had launched programs. The name refers to fitting six standard deviations between a process’s mean and its specification limits, shifting quality control from inspecting products to controlling the process.

Projects follow DMAIC (define, measure, analyze, improve, control) to fix processes or DMADV to design new ones, run by champions and a martial arts hierarchy of master black belts, black belts and green belts. The 3.4 figure depends on Mikel Harry’s 1.5 sigma shift, an allowance for long-term drift that lets a 4.5 sigma process qualify as Six Sigma. Statistician Donald Wheeler called it goofy, and without it the rate would be about two defects per billion. Results were mixed: a 2010 Wall Street Journal report said over 60 percent of projects fail, mostly from weak leadership commitment, thin training and poor project choice.

• MD Anderson Cancer Center cut patient preparation time for certain exams from 45 minutes to five without buying new equipment.
• Harry based the shift on tolerance stack-up experiments with stacked discs, which he called benderizing.
• Philip Crosby argued that 3.4 defects per million would make every complex microchip defective and that only zero defects is acceptable.
• Design News editor John Dodge was among critics who said the method stifles serendipitous research.
• After CEO James McNerney brought Six Sigma to 3M, the company had to remove it from its research functions.