Has anyone gone through the process where sigma six is bolted on to an existing production facility.
If so , what were you observations , pitfalls and any other storys regarding this.
Declan
Damn, I thought this about the new G.I. Joe cartoon.
You mean Six Sigma?
Yes , my bad
Its that glazed eyed look that comes from reading stuff bout it on the web.
That screwed me up
Declan
Well, it’s not something that you can bolt onto anything.
The process involves a few steps:
Study the process;
Identify problems;
Attempt to fix problems;
Evaluate the fix;
Iterate…
Sigma Six is a quasi-proprietary mix of ideas used by the Quality Assurance/Quality Control industry. It is heavy on sampling, probability and statistics.
Yes , thats part of what I was reading in the online literature for it , the bolting on comes from the fact that there are two forks that six sigma can take.
One is from design from the ground up and the other is to attack an already existing process , as you described above.
What I was asking for is any real world experience with doing it, and any problems that were found , and how and if they were overcome.
Declan
I’m a Six Sigma green belt, and our company uses Six Sigma extensively, although we’re starting to switch to ‘lean’ Six Sigma because we’ve found the process to be just a bit too intensive for the stuff we’re doing.
Really, all Six Sigma is, is a collection of various tools and methods and best practices, all geared towards making various processs more controllable, repeatable, and to discover inherent inefficiencies and flaws.
There are generally two flavors of Six Sigma - one applied to design (DFSS - Design for Six Sigma), and one for existing processes (DMAIC - Define, Measure, Analyze, Impove, Control).
Again, all this is a way to apply some rigor around analysis, to use statistics and defined roadmaps to analyze engineering problems and come up with good solutions.
Six Sigma can be applied effectively, or it can be over-used. Early on, some genius in our company decided that Six Sigma was so good that every engineer had to personally complete two Six Sigma projects per year. Given that there are almost 100 of us just in one office, it became pretty tough to come up with meaningful projects worth our time. So a lot of it was just makework, and created so much ‘noise’ that no one could tell the good from the bad. Then we tried applying it formally to software design, and that was pretty much a disaster. But along the way we discovered some parts that made sense, and we streamlined others to strip out some of the paperwork and use just the best parts, and overall it has improved our development practices. It’s all about being reasonable and smart, keeping what works and abandoning what doesn’t.
But overall, I think learning Six Sigma is very valuable, for the same reason that learning calculus is valuable. Even if you never formally build a House of Quality or do a full FMEA again in your life, the knowledge of these tools just helps you think better and more logically.
One factoid I picked up from a 6 sigma master black is that a company simply does not go from zero sigma to 6 sigma. There are iterations. You go from fucked up, to less fucked up, to unfucked, to pretty good, to 6 sigma. I mighta got some of the terminology in there wrong. Seriously, anyone that thinks they can just implement 6 sigma in a few months and reap all the benefits is a few short of a 6 pack
Well sure. But most companies aren’t at 0 Sigma. The average is more like 3-4 sigma. But the first step in the six sigma process is to make sure that the processes you are measuring are ‘in control’, which means they behave predictably and repeatedly. If your process is all over the place, then the first order of business is to get it behaving in a reliable fashion. Then you can start the process of analyzing it and figuring out how to improve it.
My company uses Six Sigma (although they call it something else). My own rather cynical view of it goes something like this;
-
Quantify something. If you can’t really quantify it, don’t worry, just find a number associated with it.
-
Make that number go down. Or maybe up.
-
Repeat
You have to find a burning platform. (E.g., we’re hemorrhaging money because the flapadoodle is taking too long.) Then you deconstruct the flapadoodle process to find some things to quantify. Then you take data and let the data speak to you.
If you’re required to do a Six Sigma project, make sure to pretend you know nothing about how to do your job. Also, pretend that you don’t know the solution. In fact, you probably should pretend that you’re the dumbest (yet most earnest) person on the planet, genuinely eager to learn and apply Six Sigma methodology to every decision you ever make for the rest of your life.
I’ve found it’s best to pretend that you did have a vague notion that you needed to do X, but after looking at the data, you realize that you actually have to do Y. Then you’ve demonstrated that you’re making decisions based on quantifiable criteria! You are so in. And the next org chart that gets printed will have a sigma printed next to your name. After that, you’re qualified to get a tattoo on your forehead to remind you of Six Sigma’s importance. The tattoo should probably be “666”…
*::twitch::
::eye tick::
::twitch::*
Curse you for making me think of such things.
My last company went through the flavor of the month improvements for a while. Then with new owners things calmed down after a year to two years of working on the same plans and controls it started working pretty good. It took a lot of buy in from everybody. From the top boss to the guy sweeping the floor. It can’t be dictated from the top. If everybody sticks with it you can see some great results. I was damn impressed when some numbers came out showing how much more efficent some of the parts were made by looking at machine time and total man hours used.
Now I’m a consultant at another factory and see them doing what my old company did four years ago. There is a lot of scepticism, and it’s well deserved. It’s being done half assed and I’m shaking my head as an outsider seeing what’s going off the rails.
I’ve quietly said so to a few people, but I’m just a nobody doing an unrelated project.
My company practically created six sigma. It works good for manufacturing or anything requiring machinery. It’s best if you can avoid the human element all together, e.g. (best one I can think of) creativity. It works best for designed processes that are measurable and have quantifiable metrics. That is tbe best serious answer I can come up with that doesn’t involve a lot of cynasism and swearing. While my company is really good at the manufacturing part (or was, before we outsourced practically everything), we went a little crazy trying to green belt and black belt everyone (I think I have a black belt, then again, I cheated).
Our bread and butter now is in IT and in particular, software. This is where Six Sigma is near useless. Since all of our stuff is custom, it’s hard to make a case for best practices and best methods. At best, you doom your company/development team to follow the same bs-track that got you down the wrong path in the first place. Secondly, everyone is too damn busy/too damn lazy to document, especially in software (and I don’t blame them). My engineers came up with a guide of trying to apply six sigma to internet dating. It didn’t make much sense to me, but they got a kick out of it.
Overall, I agree with what Sam Stone said. Even if you don’t use it, it’s good for knowing how to think logically.
When I joined my current company, we had this big thing for 5S. While it may work well for a Boing plant, it’s not so hot for a laboratory. Then we got bought by a bigger fish (probably the biggest) and it’s **6 Sigma ** time. I’m waiting for 7 something to pop up in a few years.
Yeah we are concurently starting to get more serious about the 5S , along with adopting the sigma stuff and soon the TQA program.
The good news , is that from the sounds of what a previous poster said , we are doing stuff that can be measured and metric’d.
So with that being said , what are the signs , omens and portents of a company thats going to do it wrong ?
And if we are doing it right , how soon before the y=x/f stuff starts to translate into something that someone can understand and umm, whats the end result.
Its all nice to say culture and what not , but what tangible benifits have been shown, like say a major decrease in the amount of non-compliant parts , scrappage , waste etc.
Declan
How would you go about guess what your company is on the sigma scale, before implementing the program.
Declan
How would you go about guessing what your company is on the sigma scale, before implementing the program.
Declan
Lots of ways, depending on what your company does. If you make products on an assembly line, you can look at things like number of products that failed quality control, or the amount of time the assembly line is down. It really depends on what you are measuring. Six Sigma just means a process that is accurate within six standard deviations. For example, an assembly line that has six sigma quality in defects will have only 3.4 defects per million operations. In comparison, a three sigma quality line would have 2700 defects per million.