Six Sigma, Lean and DFSS

I don’t think that many of us would know about 6s if it weren’t for the recognizing that 80% of our GDP comes from service organizations AND the similarity of a manufacturing process with a business process. Originally, the main focus of 6s was eliminating defects as defined by customers (it moved onto cost reduction and increasing value, I will get to that). So, once we have a customer’s requirement, we recognize the variation (defects) that hinders our ability to reliably deliver high-quality services. We then use analytical tools to break down and specify where this inefficiency is coming from. Then can make a decision based upon data, not just a hunch (ie. tweaking the machine).

There are a lot of statistical tools available to determine where in the process that the variation is coming from (Design of Experiments, Statistical Process Control, Regression, and to a lesser extent simulation). But don’t get caught up in the technical aspects. Just think of your process as a flow of information/commitments/paperwork/invoices/contracts/communications and see what it is that caused one project to get done in an hour, and a similar project to get done in a week. Do, however, enlist a Master Black Belt or a Black Belt to lead the project.

It has been stated that Lean is an excellent combination to 6s. In service applications, the cost related to work that adds no value in your customers’ eyes, is higher than in manufacturing, in both percentage and in absolute dollars. Queue time kills satisfaction. Lean focuses at removing the waste in the process using value stream maps and other lightly analytical tools.

The evolution of 6s began at Motorola (the focus was on defect reduction). GE championed it in the late 90s (focus on cost reduction). DuPont is the newest player in the spotlight and is focusing its improvement on value as viewed by the customer.

For a successful recipe of 6s there are a couple of things that your organization is going to need to do. This is going to have to be sponsored by the top management. There has to be a supporting infrastructure in place or at least the recognition of the need of one (that would be a formal project selection process). You are going to have to put your best people on these projects. There needs to be in house pilot projects (major ones with impact) and the results need to be tied directly to finances (why TQM is generally considered a failure). You will need to establish performance metrics to the projects and the players in the projects, and you will need to tie these evaluations to payments to the employees. After all, we measure what we value (and most frequently we only measure employee turn over rate).

The sustenance of executing a 6s project is the DMAIC process; Define, Measure, Analyze, Improve, and Control (similar to the Shewhart Cycle, or PDSA - Plan Do Study Act). This is a constant on every project. First we define the project, understand our Measuring devices (and verify that the variability is not coming from the gauges) and get the data, then we Analyze the data and understand the variability, next is to Improve the process (the only time in the entire 6s process where it is ok to modify the process), then we Control (which is to say that we make the change long term).

The applications are only as narrow as what is slated for improvement. 6s can be used to, with a very high degree of reliability, consistently take the trash out, improve your golf game, and reduce your commute home. But that is not really one of my goals so the project would be trashed (through the formal selection process of value in the eyes of the customer). Do you understand now that 6s is simply a business process improvement methodology (and not holistic unless combined with lean because lean addresses waste removal and not variation reduction)?

 

Curious as to how 6 Sigma got its name? Standard deviation is a measure of variability. In the bell curve 99.7% of the data are within 3 standard deviations of the average. If a process has 100 steps and every step is 99.7% reliable, we still have a 16% failure rate. If the variability of the process is reduced so much so that the variability is within 6 standard deviations from the average where our specification limits are we only have 3.4 parts per million defective. The Greek letter sigma is commonly used as notation for standard deviation. Therefore, the only way to increase the value of the customers is to lower their expectations, or to tighten our reliability.

One more thing for the advanced; Design for Six Sigma (DFSS). There is only so much blood that can be wrung out of a turnip; and eventually we are going to need to rethink our business process (rather than starting with what we have and improving it). Then you will have the customer in focus from the start!