Six Sigma

Six Sigma treats quality as a measurable property of a process rather than as a final inspection result. The approach gained prominence at Motorola in the 1980s, where engineers sought to reduce variation in manufacturing and design; it subsequently spread through companies such as General Electric and into service organizations.¹ In statistical terms, sigma denotes standard deviation, a measure of dispersion. A process operating at a nominal six-sigma capability has very small overlap between its output distribution and specification limits, although the frequently quoted figure of 3.4 defects per million opportunities assumes a conventional long-term 1.5-sigma shift.² The method therefore combines statistical process control with organizational routines, financial prioritization, and formal training.

DMAIC is the principal Six Sigma route for improving an existing process, while DMADV or DFSS is used when a product or process must be designed or substantially redesigned. In Define, teams establish the customer requirement, project boundaries, and business case; Measure checks the measurement system and establishes a baseline; Analyze tests likely sources of variation; Improve changes and validates the process; and Control maintains performance with monitoring and documented responses.¹ Typical tools include process maps, cause-and-effect diagrams, control charts, regression, design of experiments, and capability analysis. Organizations often assign executive sponsors, project champions, and practitioners called Green Belts or Black Belts, with Master Black Belts providing advanced coaching.

Six Sigma is applied wherever a repeatable process produces measurable outputs, including manufacturing, logistics, software, banking, healthcare, and public administration. Its strongest projects connect a critical customer or safety requirement to a defined defect, a reliable measurement system, and an intervention whose effect can be tested. Statistical process control distinguishes ordinary background variation from special causes requiring investigation, while capability indices compare process variation with specification limits.² The approach is not a universal substitute for engineering judgment: poorly defined requirements, unstable processes, rare outcomes, and ambiguous defect opportunities can make the numerical target misleading. Reviews of the management literature also identify risks of treating Six Sigma as a certification program or short-term cost campaign rather than as a system for learning and sustained process governance.³

Six Sigma’s lesser-known dimension is its dependence on measurement validity and experimental reasoning, not merely on a named sequence of steps. A team can obtain an impressive capability estimate from a biased gauge, inconsistent sampling plan, or shifting definition of a defect; measurement-system analysis is therefore an early practical safeguard. The “six” in the name is a performance aspiration, not a requirement that every process literally reach six standard deviations. Long-term process behavior, autocorrelation, non-normal data, and multiple failure modes may require methods beyond a simple normal-distribution calculation.² Six Sigma is also frequently combined with Lean: Lean emphasizes waste and flow, whereas Six Sigma emphasizes variation and causal evidence. The resulting Lean Six Sigma repertoire is common, but the two traditions retain different historical and analytical emphases.

DMAIC

A structured improvement cycle meaning Define, Measure, Analyze, Improve, and Control.

DFSS

Design for Six Sigma, a family of methods for designing products or processes to meet requirements from the outset.

Defect

A failure to meet a stated customer, engineering, regulatory, or process requirement.

Process capability

The relationship between a process’s observed variation and its specification limits.

Control chart

A statistical graph used to distinguish common-cause variation from special-cause variation over time.

Green Belt

A trained participant who applies Six Sigma methods, often part-time or within a project team.