I get a lot of calls asking if I “do” Six Sigma. While I use some elements of the Six Sigma tool kit, no, I don’t do Six Sigma.
I am not a black belt or even a green belt. Now, the reasons are not what you might be thinking. I am familiar with Six Sigma. And I have seen it work very well in many organizations. When I worked in GE, you could not avoid it. GE was an early adopter of Six Sigma. Many GE alumni went on to spread the gospel in other companies.
What is Six Sigma?
In a nutshell, Six Sigma is a structured approach to reducing the number of defects that result from business processes. In service terms, you can use the term error rather than defect. Sounds like the perfect thing for me to be doing for my clients.
So why do I steer clear of Six Sigma? Well, it’s not because the majority of Six Sigma initiatives fail, although they do. Between 60% and 74% depending on whose surveys you read.
And it is not because I believe Six Sigma projects suffer from all the usual defects of big corporate initiatives, like poor project management, scope creep, and lack of senior executive support. All the usual suspects.
Mostly, I avoid it because it does not work in the kind of areas I specialize in. time. There are simply some places that do not benefit from the full Six Sigma treatment. By that, I mean the general area of knowledge work. AKA, back office, white collar, electronic processes. So right there, that’s most of the work we do in North America.
Now, there are bits of Six Sigma that I happily and shamelessly steal. For starters, I find that the DMAIC (Define, Measure, Analyze, Improve, Control) stuff works well. It is a good discipline. But one to adapt, not follow slavishly.
Six Sigma Tool Kit
The 5 Whys tool is great for getting to the cause of a process issue. It does not rely on statistical methods and bunches of data. Two items that are often in short supply in the real world of office processes.
Likewise, I make use of fishbone diagrams. People can understand cause and effect diagrams without lengthy explanations. They are a great way to structure problems solving. But in fairness, fishbone diagrams were in use decades before Six Sigma.
The part that I am really skeptical about is Measure. The M in DMAIC. Trying to measure messy transaction flows in an office processes is really, really hard. It’s just not the same as, say, measuring the deviation of the diameter of extruded plastic in parts per million.
Six Sigma works best when there is very little variation in a given process. It really needs tons of data to shine.
Six Sigma Weakness
And going through the whole method, it is pretty ponderous. Meaning slow. What you need is something that produces results quickly. Not something that soaks up resources and training for months before it delivers. Not to mention that is expensive to train or to hire in consultants with the right certification.
The other weakness of Six Sigma in an office setting is that it focuses on a single process at a time. Nobody even thinks about that. But people working in offices work on multiple concurrent processes. Unlike assembly line workers, they are rarely dedicated to one single process.
Their problem is usually not with the steps in any particular process. Instead, what screws them up is switching between processes. That is where office people’s efficiency suffers. Every time they switch from one thing to another, there are tiny built in delays. And those add up to something that is utterly invisible to Six Sigma. It does not even consider them.
So no, I don’t do Six Sigma. I don’t recommend it for the typical office environment. Unless you have highly repetitive processes that people are more or less dedicated to, it’s not the right approach. That’s not to say you can’t borrow from the Six Sigma tool kit. Just don’t drink the Kool Aid.
So when it is a single process, you don’t need anything elaborate to produce a result.
For 90% of the organizations I come across, what works best is something simple. You can map a process on a whiteboard with a group of people in an afternoon.
People who are using the process don’t need data to tell you what is going wrong. They have been doing it forever and they are dying to tell you. I just give them that opportunity. Sometimes, I feel like a lightening conductor.
Looking for something? Try searching for it here: