By the time goes by, the continuous initiative is growing in the industry and we still do not know what is six sigma and at this point we are too afraid to ask.
Don’t worry, this article will elaborate more on what is Six Sigma from methodology and statistical point of view.
From methodology’s point of view, Six Sigma is the vast library set of tools and knowledge used in DMAIC (Define-Measure-Analyze-Improve-Control) approach to improve a process to satisfy the customers and achieve the six-sigma level.
In this definition, we can break it into four main elements:
Set of Tools
Process improvement to satisfy customers
Six sigma level
Too many things?
Don;t worry, we will elaborate these elements in detail to understand better of the Six Sigma methodology.
Set of Tools and DMAIC Approach
In conventional problem solving, you might familiar with the Demming Cycle or PDCA cycle. The Six Sigma methodology is using the DMAIC approach instead in project management. We have discussed in detail the differences between PDCA and DMAIC in previous article.
In every phase of DMAIC, it is consisting of the quality tools to achieve the objectives in each phase. Below is the objectives and generic quality tools use in DMAIC phases.
We can see the tools of Six Sigma is consisting of almost 80% of statistical tools and 20% of conceptual tools. Thus, the measurable input and output is must for Six Sigma methodology and the project will be executed in data driven solutions, not by assumption.
In Six Sigma, the customer is the king. The decision to determine either the process is good or not is based on customer’s perspective. The defect definition is defined by the customers as well.
Therefore, the end goal of Six Sigma initiative is to close the gap between the voice of customer (what they really want) and voice of process (what do we really produce).
Six Sigma Level
At the most basic definition, six sigma level is a statistical representation for what many experts call a “perfect” process. Technically, in a Six Sigma process, there are only 3.4 defects per million opportunities. In percentages, that means 99.99966 percent of the products from a Six Sigma process are without defect. At just one sigma level below— five sigma, or 99.97 percent accuracy –processes experience 233 errors per million opportunities. In simpler terms, there are going to be many more unsatisfied customers. The table below showed why 99% is not good enough in certain cases.
The sigma in here is this lower sigma (σ) or known as standard deviation, the capital sigma (Σ) is represent the sum in mathematical language. But what really means by five or six sigma?
The six sigma is defined by there are six standard deviation can fit from centre of the bell curve to the specification limit given by the customer or in other word it is more precise and accurate. At very low levels of sigma, any process is unlikely to be profitable. The higher the sigma level, the better the bottom line is likely to be.
Let us take the real-world example.
Consider an example based on Amazon shipments. On Cyber Monday in 2013, Amazon processed a whopping 36.8 million orders. Let’s assume that each order error costs the company an average of $35 (a very conservative number, considering that costs might include return shipping, labor to answer customer phone calls or emails, and labor and shipping to right a wrong order).
So we can see in this situation, even 1 sigma shift can cost the company is around $284,000/year. In high volume production or transaction, 1 sigma can create a butterfly effect to the company.