[RD] Six Sigma

[BvBPL]

Six Sigma is a project management model developed at 3M and subsequently widely adopted across many industries. It is goal is the reduction of customer-facing defects in products.

Have you worked with Six Sigma?
What is your opinion of it?
Are you certified in it?
How difficult is it to become certified?
What are the weaknesses of Six Sigma?
Where does it work well? Where is it less applicable?
What is your opinion of other models?


If you have an ongoing, iterative process, as opposed to a finite goal, is Six Sigma the most applicable model to adapt? If not then what would be superior?

 

[Mongoloid Cow]

My dad used to work in management before retirement and had to implement six sigma for a time before he was able to get his company to not use it further. He reckoned that it was embarrassingly ridiculous (karate belts! hai!), illogical, and ultimately bad practice. To really get the most out of it you need the ability of time travel. If you lack access to the company Tardis then there are other methods that work much better. I don't really understand what it is so I can't really explain his reasons.

 

[Samez]

I am a cerrtified green belt and was offered the black belt training but refused.
Six sigma is imho a set of classical quality management tools combined in a systematic approach.
It works really well in some situations but wont work in others. The heavy focus on the define phase is usefull as it is always good for projects to know what is your current state and what is your target - and when you are drifting off the planned course.
Six sigma also has some serious downsides: it does not work well with new processes/ for development, low statistical numbers / small series production is an issue as many tools won't work and it will also not work well for problems with a lot of equally important influence factory.
The last two issues will be relevant for all statistical methods and are not counted exclusively relevant for six sigma. In my workplace there are usually lot sizes between 1 and 50. So if there is an issue and until it is recognised there are no statistical relevant numbers left to gather data reliable- it is possible to preemptively gather it but there are limits.