4 ms·

> Deming’s methods became systematized into Six Sigma I think Deming and anyone who has seriously studied his work would take exception to that. Most of the un

> Deming’s methods became systematized into Six Sigma I think Deming and anyone who has seriously studied his work would take exception to that. Most of the underlying mental models are missing from Six Sigma, and all that remains is a watered down version of Statistical Process Control 101.

mikey_p 7y ago
Yeah, I think this is really missing the point of Deming. While statistical quality control was a big part of what he taught, he was almost more interested in the philosophy of management than process management. Everything you read about him talks about statistical quality control, but nothing ever mentions his Fourteen points, or Seven deadly sins. I read about him as a kid, but alot of it has stuck with me through multiple careers, and his Fourteen principles could easily be applied to most software and engineering teams and companies. Things that are really key to his philosophy (and deeply impressed me) were: * Cease dependence on mass inspections for quality control (Do you rely on manual testing for every release?) * Break down barriers between departments (i.e. devops) * Eliminate slogans and quotas, goals, targets etc. (Maybe try the release train process for shipping software instead?) And my personal favorite: * Eliminate barriers to pride in workmanship, for hourly workers, management and engineers (is part of your team embarrassed by bad UX or known bugs that they can't fix because they are busy with other work?) Deming's philosophy basically wrote an entire ad campaign for Ford, "Where quality is job #1."
kwhitefoot 7y ago
The way Six Sigma was presented to me by Mike Harry was that it is a tool for finding out where to apply your effort. I've not been involved in it for a long time and am sad to hear of his death. The great thing about it as he presented it was that it did not require a great deal of sophisticated tools or a lot of statistical knowledge just a spreadsheet and an appreciation of multi-way analysis of variance. When I was working with engineers using Six Sigma in the late 80s, early 90s, Six Sigma was not really to do with day to day statistical process control. It was all about finding assignable causes of variation and how variability in each part's properties contributed to the variability of the whole. The connection with process control was that it helped you decide what tolerances one could allow. In several cases the Six Sigma investigation resulted in widening the tolerances instead of narrowing them as it was found that they were pointlessly tight and resulting in more costly processes. In many others it resulted in a change to the process, adding steps, replacing parts, etc., and not just an adjustment to the process' parameters.

mikey_p 7y ago

Yeah, I think this is really missing the point of Deming. While statistical quality control was a big part of what he taught, he was almost more interested in the philosophy of management than process management. Everything you read about him talks about statistical quality control, but nothing ever mentions his Fourteen points, or Seven deadly sins. I read about him as a kid, but alot of it has stuck with me through multiple careers, and his Fourteen principles could easily be applied to most software and engineering teams and companies. Things that are really key to his philosophy (and deeply impressed me) were: * Cease dependence on mass inspections for quality control (Do you rely on manual testing for every release?) * Break down barriers between departments (i.e. devops) * Eliminate slogans and quotas, goals, targets etc. (Maybe try the release train process for shipping software instead?) And my personal favorite: * Eliminate barriers to pride in workmanship, for hourly workers, management and engineers (is part of your team embarrassed by bad UX or known bugs that they can't fix because they are busy with other work?) Deming's philosophy basically wrote an entire ad campaign for Ford, "Where quality is job #1."

kwhitefoot 7y ago

The way Six Sigma was presented to me by Mike Harry was that it is a tool for finding out where to apply your effort. I've not been involved in it for a long time and am sad to hear of his death. The great thing about it as he presented it was that it did not require a great deal of sophisticated tools or a lot of statistical knowledge just a spreadsheet and an appreciation of multi-way analysis of variance. When I was working with engineers using Six Sigma in the late 80s, early 90s, Six Sigma was not really to do with day to day statistical process control. It was all about finding assignable causes of variation and how variability in each part's properties contributed to the variability of the whole. The connection with process control was that it helped you decide what tolerances one could allow. In several cases the Six Sigma investigation resulted in widening the tolerances instead of narrowing them as it was found that they were pointlessly tight and resulting in more costly processes. In many others it resulted in a change to the process, adding steps, replacing parts, etc., and not just an adjustment to the process' parameters.