Lean

Six Sigma

A quality-management methodology that reduces process variation through statistical control. It uses the DMAIC approach for improvement.

Origins

Six Sigma was developed by Motorola engineer Bill Smith in 1986. The goal is a defect level of no more than 3.4 per million operations (corresponding to 6σ of a normal distribution). The methodology spread widely after Jack Welch rolled it out at General Electric in 1995.

DMAIC

The basic process-improvement cycle in Six Sigma is DMAIC:

  • Define — define the problem and the project goals
  • Measure — measure the current state of the process
  • Analyze — analyze the causes of the problems
  • Improve — implement improvements
  • Control — lock in the results through control procedures

For entirely new processes the related DMADV cycle is used (Define, Measure, Analyze, Design, Verify).

Tools

Six Sigma makes active use of statistical tools:

  • Shewhart control charts
  • Pareto analysis
  • The “fishbone” (Ishikawa) diagram
  • FMEA — failure mode and effects analysis
  • Regression analysis
  • Design of experiments (DoE)

The belt hierarchy

Six Sigma is known for its colored hierarchy of specialists:

  • Yellow Belt — a basic awareness level
  • Green Belt — leads small improvement projects
  • Black Belt — leads complex projects full time
  • Master Black Belt — an enterprise-level coach and strategist

Strengths

  • Strict measurability of results
  • Proven economic effectiveness
  • A well-structured training methodology

Limitations

  • Bureaucratic — requires significant documentation
  • Focused on reducing variation rather than radical transformation
  • Less effective for service and creative processes