Explain Six Sigma to me (in this context)

I understand the premise of Six Sigma - if you engineer something such that the rate defects in your resulting product falls within six standard deviations of some established norm (which works out to 99.8% error-free, or something like that), you are “Six Sigma compliant”. This makes perfect sense to me in a manufacturing environment (I believe Toyota pioneered the approach to manage the quality of their product line).

My company (which does not manufacture anything) is now touting Six Sigma in every way they can. For example, our project teams which are implementing new processes and functionality within one of our core systems are said to be Six Sigma compliant. We now have Six Sigma project management methodologies. We have Six Sigma processes and Six Sigma compliant “bolt-ons” on our system.

None of these initiatives involved measuring any rate of error, either to establish a baseline or to evaluate the final product. We have our JR VPs telling executive management that everything they’ve done is Six Sigma compliant, and executive management is annoucing this to the organization. As far as I could tell, they only thing the did was hire a Six Sigma “blackbelt” (whatever that is) consultant to make project management suggestions.

Does any of this make sense? Do I not understand Six Sigma, or do we just have ppl blowing smoke up other people’s asses?

Six Sigma is a mangement cult wherein foolish executives are conned by idiotic consultants into paying them shitloads of money to use a lot of buzzwords. In other words, it’s business as usual.

[minor hijack]
Please let me be the first one to tell you that hearing about Six Sigma “blackbelts” really chaps my ass. I understand that they are trying to using the term to denote a standard of excellence, but it’s hard to explain to folks that have not spent years earning a Black Belt in a martial arts discipline how really wrong that is.

Although it did afford me a measure of low-grade humor once. We had a meeting at a company I worked for a couple of years ago and they had one of these Six Sigma blackbelts. He started off by asking if any other blackbelts in the group would stand up. Of course, I did. He then started asking me a bunch of wierd questions and was not happy with my answers. When he asked how I got my blackbelt, I answered along the lines (I was a 4th Degree at the time) of “Eleven years of hard training and then a testing in front of the International Testing Committee where I presented an 83 move form, sparred several matches and broke boards with a front kick, continous motion side kick and a hammerfist strike. What did you do to get yours?”

He was not amused. But I was. :smiley:
[end hijack]

Agree, as a Project Manager for the last fifteen years, Ive seen them come and go. Six Sigma is the rage now, RUP (Rational Unified Process) was the buzz a few years ago. Before those, ISO 900X, TQM, blah blah blah were the “new” ways to seperate corporations from their money to learn new ways to preach quality.

Quality is a concept that is easy to applu to a manufacturing operation, but not so to a call center, or mortgage operation, etc. But… as long as your competition continues to implement the latest and greatest methodologies, regardless of their effectiveness, you can bet that Six Sigma will not be the last corporate religion that will be shoved down your throat.

nitpick: Motorola

In any case, you’re very correct. It seems to be used in so many contexts when, in reality, it refers to a very specific idea. It’s like how anything that’s long is referred to as “marathon”. Like: The NHL and the player’s union sat through a marathon meeting. No, a marathon is a foot race that’s 26.2 miles.

Motorola, incidently, has dropped the whole six-sigma thing.

AND if you do the math, you find that the six-sigma cult specified failure rate works out to only around 4.5 Sigma…but that doesn’t rhyme.

AND, if want to know how I really feel about such nonsense, that would require a pit thread.

Been there, done that.

It has been my experience that the engineers and managers who latch on to the latest “quality system” in a cult-like manner do so because they’ve failed at everything else. These idiots quickly figure out that subscribing to the latest quality trend and preaching “quality” can cover up the fact that they’re lousy managers and engineers.

That’s because of the 1.5-sigma long-term variation of the mean. A stable process has Zst>=6.0, but Zlt>=4.5 is okay too.

The basic ideas and tools of statistical process control have been around for a damn long time, and Six Sigma is just a package of them. But they’re still just tools, useful in specific applications. They don’t replace the human brain, but they can cover its absence, as the above testimonials indicate.

–ElvisL1ves, Six Sigma Green Belt at former employer, ACE Co-Pilot at current employer, recognizer of bullshit (but player of games when required) everywhere

They now have a qualification beyond black belt called “super black belt.” It’s like fucking Scientology. More classes = more money.


Disclaimer: My company is using Six-Sigma processes as one quality means.

Okay, so six sigma isn’t just about attaining a six sigma grade on your manufacturing processes. It’s a complete system. I can see that in some companies it could reach a cult-like status, but that’s not really the intention. Instead it’s a guided, focused philosophy towards maintaining quality goals. It doesn’t work by solving problems for you, and it won’t cover up for bad engineers. It requires good engineers to take a deliberate approach in solving their own problems.

Consider define, measure, analyze, improve, control – the fundamentals of six sigma execution. These aren’t foreign concepts to any engineer. But just like ISO qualification ensures that you “do what you say, say what you do,” six sigma methodology makes sure you stay on the right track.

Couple of links: 1 and 2.

So, given all of that, it’s true that there are managers that are not engineers that can be hoodwinked by a consultant for a large amount of money to implement six sigma programs just for the sheer buzzword. But without proper support, it really only achieves “flavor of the month” status and becomes worthy of poking fun at.

SS Green Belt here…

Six Sigma is really a combination of 50% consultant hogwash, and 50% useful tools.

I’ve seen it fail miserably enough times in the companies I’ve worked for, but not really due to any inherent faults with the system but because of the “Genuine bonafide monorail!” approach with which it is sold by the swaggering black belt consultants. It then fails because of poor implementation and understanding by the participating company who use it as a band-aid problem solver instead of a systemic philosophy.

Sadly, the SS training process itself (for green and black belts) reinforces the carnival shyster rah-rah approach. Open ears and objectivity are not to be found.

That said, there are some fantastically useful analysis tools that can be used to analyze your business process and optimize, streamline, improve quality, yada yada. However, it takes dedicated follow through and support from company leadership, which I have yet to ever personally see happen before the next “Quality Flavor of the Month” distracts them.

We have six sigma processes running in our organization. Most of the people who have suffered through the team process of applying these tools state that they were successful, but then they also state that they could accomplish the same results with fewer people using less time applying some basic logic.

My opinion - it has wasted too much of my department’s valuable time as we spent a whole month proving to a “blackbelt” who drank way to much of the consultant’s kool-aid that one of our primary processes wasn’t broke.

Elvis this is the same statement that our Six Sigma Instructor made to us during my “Blue Belt” training. I contacted a college instructor friend of mine for a comment on it. Here is my friend’s reply:

"I only teach statistics to MBA’s, so I was interested in this half-assed “4.5 - 6.0 z is close enough” answer. I mean this 6 sigma process is supposed to be totally precise and absolutely anal about measuring critical factors. I’d fail a student who gave me the excuse that 1.5 s.d. is something that could just be ignored.

If there is a 1.5 s.d. variation in the mean over the long-run, how does someone differentiate between long-term and short-term variance when they are collecting data IN THE PRESENT TIME?

The answer… you can’t. If you observe a process that produces 1 failure in 1,000,000 draws, it is a 4.5 sigma process according to EVERY statistical table of normal probability ever printed. If you don’t understand that you deserve to fail not only in a classroom but in business as well.

Now, if you study a process over time you could determine a long-term “variation of mean,” but what I wonder is how did Motorola, back before all this analysis began, recognize the validity of that 1.5 sigma rule of thumb PRIOR to any process being observed long enough to varify the 1.5 sigma fudge factor? I mean they didn’t call it the “4.5 Sigma” process right out of the block, did they?

In short, Six Sigma sounds better and makes for a better looking logo than Four-point-Five Sigma. Six Sigma is just a load of management BS to make them feel like they have control over some aspect of their company and obsolve themselves of the guilt of failing their feduciary responsibilities"

You don’t “ignore” the drift of the mean, it does matter. You have to be aware that it can be present, and it’s very much worth your while to track it and pin it down if you can. The 1.5s value is the default assumption if you don’t have any data to the contrary yet, nothing more. You differentiate between Zst and Zlt in snapshot data by adding in the mean variation - the actual value if you have it, 1.5s if you don’t.

Motorola didn’t invent statistical process control either, the “father” of it was W. Edwards Deming. The data behind it comes from his work, which includes not only the value of a typical mean variation but the fundamental fact that *any * controlled process will produce results in the familiar Gauss bell curve.

The guy at Motorola (whose name escapes me) credited with inventing Six Sigma is a friend of a friend. One time he got drunk with my friend and started sobbing about what a mess the whole thing had become. “It wasn’t supposed to be taken literally”, he cried. “That’s just stupid.” But he didn’t know how to stop the bandwagon.

Life is full of compromises whether we pretend it is or not, and they all require some wisdom. I like the idea of making things better than they might seem to require at first glance, because many of the costs of things not working well are hard to anticipate but still onerous. But the postings I see above sound pretty reasonable to me.

If you mean Mikel Harry, he appears to be coping with his disappointment.

My instructor friend flew into a small rage after reading this. I think that we may have spawned another subscription to SDMB.

My poor comprehension of statistical analysis prevents me from relaying his statements but I sensed his main argument was that your argument says that something isn’t important but then says it is.

I’m sorry that I’m arguing a point I really don’t understand but after seeing six sigma in action I don’t have a very high opinion of it. I think the process is just the latest flavor of TQM and management consutling BS hiding behind mumbo-jumbo statistical statements that don’t really ring true.

I have my own personal theory on management consulting.

Without Consulting

Company makes widgets

  1. People who run the machines and make the widgets see innefficiencies and have ideas on how to improve the process
  2. Midlevel managers think that the people who make the widgets complain too much so they ignore them
  3. Upper level management doesn’t have a clue as to how the widgets are actually made.

With Consulting

  1. People who run the machines and make the widgets see innefficiencies and have ideas on how to improve the process
  2. Midlevel managers think that the people who make the widgets complain too much so they -
    ignore them
  3. Upper level management doesn’t have a clue as to how the widgets are actually made but they want to make more money and keep their jobs.
  4. Upper level management

Without Consulting

Damn hamsters!

Company makes widgets

  1. People who run the machines and make the widgets see innefficiencies and have ideas on how to improve the process
  2. Midlevel managers think that the people who make the widgets complain too much so they ignore them
  3. Upper level management doesn’t have a clue as to how the widgets are actually made.

With Consulting

  1. People who run the machines and make the widgets see innefficiencies and have ideas on how to improve the process
  2. Midlevel managers think that the people who make the widgets complain too much so they -
    a.ignore them
    b. pass their suggestion to upper level management who replies “too costly”
  3. Upper level management doesn’t have a clue as to how the widgets are actually made but they want to make more money and keep their jobs so they-
    a. hire consultants who present the latest mumbo jumbo system
    b. consultants ask people who run the machines how to fix the process
    c. Consultants package the replies from the people who run the machines to make it look like their mumbo jumbo supplied the answer.
    d. process improves, fat checks for consultants and everybody in management

Six Sigma is really the wrong name for this process. After all, meeting 6-Σ really is completely abritrary. Who says we can’t make money at 4-Σ? Or that maybe our true requirement is 10-Σ?

The six sigma method to improving quality isn’t as simple as saying, “Make it better.” It’s about studying variation, the causes and types of causes of that variation, and dealing with those causes, and continuing surveillance. There are other ways to investigate variation and treat problems, too. But how do you define a standard process to do so? Start from scratch? Do it trial and error? Use a different system? It’s important to realize that it’s just a process – nothing more. It’s not and engineering philosophy, it’s not a cure all, and it requires dedication to make it work.

No, I have no special relationship with Six Sigma, and I’ve seen half-assed attempts at its implementation (yearly requirement for everyone!) as well as truly, hugely successful uses of it. When it’s used honestly and sincerely by people who fundamentally know what it is, it works supurbly.

Now, what really sucks is that stupid thing we do at my company where we sacrifice a goat every morning. I mean, really. If one goat is fine, wouldn’t two goats be better? Stupid managers…

Does that mean when it doesn’t work fundamentaly knowledgable people aren’t honestly and sincerely using it?

That’s the implication that the consultants state. If you gain from my process then you did so because of my process. If you fail, you didn’t understand my process.

Its just another way for the consultant to take credit for success while denying blame for failure.

I’m reminded of the Dilbert cartoon where Dogbert is posing as a six sigma consultant. He states that all companies that use his process increase their profits.

…except the ones experiencing industry downturns
…or flat growth industries
…or industries that upturnned a little bit

I’m a newb, and this ain’t the pit, but I bustin’ here.

About 6-7 years back, Larry Bossidy, CEO of fortune 5 compay AlliedSignal (Now known as Honeywell) in a video taped address to all employees announced that Six Sigma would be adopted company wide. He stated ONE reason for this.

Increased quality? nope.
Increased effiency? nope.
Increased profits? nope.
A good idea? nope.
The reason: He felt that it would impress Wall Street investors, and raise the price of the company’s fairly stagnant stock.