Six Sigma for Process Control
Takeaway: Introduction to six sigma methodology and how it helps improve quality and remove variation.
Control; In business and life we want control. We want control of outcomes. We need predictability and confidence in our systems and processes. This gives us hope. It gives us the ability to sleep at night knowing things are under control and will work out in our favor. We need to know that our processes are robust and will remain that way.
Six Sigma is a tool we can use to gain control. We use Six Sigma to measure variation and remove it. It operates on the theory that no two objects are exactly alike. If you can’t detect the differences, it does not mean variation is not there. It means your measurements are not sensitive enough to detect them.
In business, we often speak regarding percentages or averages. The truth is, customers don’t experience percentages/averages, they experience variation. Uncovering variation is essential for any operation.
Let’s say that every day I visit my favorite fast food restaurant for lunch. I experience differing amounts of time as I wait for my food.
Time spent waiting:
- Monday – 6 minutes
- Tuesday – 10 minutes
- Wednesday – 7 minutes
- Thursday – 20 minutes
- Friday – 13 minutes
If we average the wait times we get is 11.2 minutes, an amount of time I never actually experienced and has little meaning to me as a customer. What I do know is that it can take me between 6 to 20 minutes to get my food.
This is what Six Sigma offers us as a methodology, the ability to identify and remove variation.
The term “Six Sigma” is a registered trademark by Motorola. The name is trademarked, but the tools and methodology are not. You would be hard pressed to dive into books or articles on process improvement and not come across the six sigma methodology. It was introduced by Bill Smith in 1986 while working as an engineer for Motorola.
Six Sigma Methodology
Six sigma is the measure of the amount of variation between and within subgroupings. It is a form of statistical analysis. Six sigma is used to identify a process’s standard deviation. It recognizes the ability of a process to hold specifications. This may not show the process is capable of what we would like it to do; Meet customer requirements.
We use Six sigma when studying the variation of a single attribute such as weight, diameter, time, temperature, scrap, etc. We are taking raw data (measurements) and converting it into information for process analysis, control, and improvement.
Nuts and Bolts
Once we determine what attribute we are going to analyze, we need to collect our data. Conventional wisdom says we collect between 50 and 100 measurements. Each measurement represents a data point. As an example, if we had equipment that produced shafts for a customer. We can measure the diameter of 50 shafts. We now have 50 data points.
We then use software or tools to analyze the distribution of data points, looking for variation. We need to set our upper and lower limits as defined by the customer. We want all data points within those set limits. If all measurements are well within the set limits and our process is robust enough to maintain those tolerances, life is good. If the process can’t stay within set limits, we will have unhappy customers.
A tool called a Histogram is often used to create a visual chart. A typical histogram will use a table of set intervals for the recorded measurements to be placed in. This should reveal a bell curve with most the data points in the middle of the bell. The intervals work as baskets to put the data points in. Each data point will only fit within one of the baskets.
Example of set intervals (baskets) and data showing a distribution:
Don’t get bogged down here learning about histograms. I mention it only as a tool (option) we may use. We need to see our data points in this type of a table to see our distribution.
Six Sigma methodology:
In our measurements, we are looking at six standard deviations between the process mean and the nearest specification limit in sigma units. This should capture 99.73% of our measurements. This gives us confidence we are producing good product for our customer with only a small percentage of out of specification parts.
The goal is to work to improve processes and remove variation continuously.
In all that we do in business, we want a high level of customer satisfaction. This means creating processes that fire on the following three cylinders:
Reliable– We must develop processes that not only meet customer requirements, but that are reliable. We want robust processes. We want to avoid spending large amounts of effort and attention to maintain a set tolerance. We need to be running at the center of our specifications and always working to improve the stability of all processes. We want confidence we are maintaining process control and delivering quality on the first try. We may have equipment that is repeatable and capable, but if it breaks down often, or is prone to jam ups, etc. it is not reliable.
Repeatable– Means our process is stable and consistent in producing quality products. We know what to expect from the process. We know the process performance history and that it can maintain consistently good quality.
We may have processes that are capable of meeting quality standards and are reliable, yet they are not stable. They easily float out of the specification tolerance. The process needs constant babysitting.
Capable-Means we know our processes are capable of meeting customer requirements and demand. We are well within tolerances of set specifications and have good control of variation.
Lack of capability would mean the equipment or process was not able to meet the customer’s needed specifications. They will experience lots of variation.
The focus of this blog is on introducing six sigma and highlighting its benefits. As a customer, I hope the companies producing products I purchase continually work to remove variation and increase reliability. Six sigma provides a set of tools and framework that help to achieve that.
The six sigma approach will help take raw data and create useful information we can use to reduce waste while improving stability and quality.
Look at your processes. What can you improve?
I recommend using engineers trained in this methodology. I have seen many improvements made with this approach along a wide range of processes. It’s a popular method of improving quality for a reason…it works.