Six Sigma is a management system that strives to achieve near-perfection in terms of quality. In statistical terms, a process that produces 3.4 defects per million opportunities or less (i.e. near perfect outcome in quality) is called a Six Sigma process. Or in other words, 99.99966% of the products produced by a Six Sigma process are expected to be defect-free.
A Brief History
The use of statistics in industrial problem-solving can be traced back to inter-war period in the US and post second world-war Japan by luminaries such as Walter Shewhart, W. Edwards Deming, Joseph Juran, Phillip Crosby, Kaoru Ishikawa and Genichi Taguchi among others. The concepts of Total Quality Management and Zero-defect propounded by them were precursors to Six Sigma. The term Six Sigma was coined by Bill Smith of Motorola in 1986 and later widely implemented at General Electric and Honeywell in mid-to-late 1990s. It has since been adopted across various organisations in vastly diverse industries (manufacturing, pharmaceuticals, banking & insurance, shipping, healthcare etc.) the world over.
A Basic Overview
The sigma in “Six Sigma” stands for standard deviation, which is the most common measure of variation. Variation is inherent in any process, but must be reduced to produce consistent, stable and predictable results. Six Sigma approach strives to identify and eliminate or control the causes of variation.
Customer’s requirements and specifications form the focal point of Six Sigma approach. Always meeting the customer’s specifications, then, becomes the goal. Let us try to understand the underlying statistics behind Six Sigma thinking through an example –
Suppose that a customer sets the specification for the length of a widget to be 10 + or – 0.3 cm. This is to say, the any widget with a length between 9.7 cm and 10.3 cm would be deemed acceptable by the customer, and any widget outside this range would be deemed a defective unit by the customer.
Now, the manufacturer of the widget produces parts with an average length (mean) of 10 cm and a standard deviation of 0.05 cm (i.e. variation from the mean). Now, six standard deviations to either side (6*0.05) from the mean for this process is 9.7 cm to 10.3 cm. This would indicate that the manufacturer’s process is a Six Sigma process and is very capable of meeting the customer’s requirements. The probability of producing a faulty or defective unit in this case is 3.4 out of 1,000,000.
Ultimately, reducing variation in the process leads to greater customer satisfaction as a result of meeting customer’s quality expectations. This in turn results in significant cost savings for the company.
Six Sigma Approach and Tools
We’ve seen that we must reduce variation in the output. How then, to identify the causes of variation and eliminate or control them?
Six Sigma methodologies espouse a structured way of problem solving. Foremost among them is the DMAIC approach (explained below):
Define – Define the problem; set target, identify the scope of the project, resources required, potential constraints etc.
Measure – Check effectiveness of the measurement system, measure “current-state” conditions, and collect data.
Analyse – Analyse the data to identify potential root causes of variation.
Improve – Validate and implement solutions to the root causes of variation and
Control – Implement control measure to prevent the problem from reoccurring.
This is the approach expected to be followed by a Six Sigma project team (usually a cross-functional team) working on process improvements. A typical Six Sigma project is expected to last 2-3 months and shows significant improvement in the process after this period.
There are various statistical and other problem-solving tools to aid the project team at each of these stages. Common among them are –
Process Mapping, Gage R&R, 7 Basic QC tools (Check Sheets, cause & effect diagrams, scatter plots, Pareto charts, histograms, graphs/ charts and control charts), hypothesis testing, regression & ANOVA, design of experiments (DOE), process capability, Taguchi methods etc.
Six Sigma approach requires individuals from various levels of an organisation to be engaged in process improvement with clearly defined roles and responsibilities. These are described below (courtesy of www.isixsigma.com) –
Benefits to an Organisation
Six Sigma provides an organisation a customer-focused & structured problem solving approach and an ethos of continuous improvement. It is predominantly geared towards quality improvement, which in turn improves customer satisfaction. In addition to customer satisfaction and retention, significant cost savings are usually achieved by reducing scrap, defects, improving cycle times and numerous other process improvements. Furthermore, the Six Sigma process can be applied to product and process development efforts as well with an eye towards in-built quality (Design for Six Sigma or DFSS). Motorola estimates that it has saved $17 billion to date through Six Sigma. Jack Welch, former CEO of General Electric remarked that “Six Sigma is the most important initiative GE has ever undertaken”.
A word of caution – Six Sigma tools, ultimately, must be used for a small and select bunch of projects only by those professionals trained in Six Sigma and statistical problem solving, and must not be used for every process improvement project. As a general rule of thumb, it is thought that 90% of problems in a process could be solved using common sense and KAIZEN™ tools & methodologies, leaving only 10% of the intractable issues to be resolved using Six Sigma tools (i.e. use these tools only when all other avenues are exhausted).