Six Sigma

Six Sigma is a business management strategy, initially implemented by Motorola, that today enjoys widespread application in many sectors of industry. Six Sigma seeks to improve the quality of process outputs by identifying and removing the causes of defects (errors) and variability in manufacturing and business processes (Wikipedia 2009).

The five steps in Six Sigma (DMAIC)

  1. Define

  2. The define phase identifies potential projects; evaluates projects; selects a project; defines a project (preferably using SMART framework) and sets up the project team

    • AT&T uses the Pareto Principle (Juran Principle) to choose project(s) calling it the Pareto Priority Index (PPI):
      • PPI = (savings * probabilty of success) / (cost * years to completion)
    • Projects should deal with a chronic problem; be feasible, signfiicant, measurable and serve as a learning experience
    • Project teams usually include 6-8 persons using a cross-functional approach
  3. Measure

  4. The measure phase identifies key product parameters and process characteristics and measures the current process capability.

  5. The steps are:
    • Verify the need
    • Document the process
      • Process flow diagrams shows the sequence of steps including key process output variables (KPOVs) and key process input variables (KPIVs).
      • Defects are non-fulfillment of intended use
      • Sympoms are observable from a defect
      • Theories are umproven assertion of reasons for existance of defects and symptoms
      • Causes are the proven reasons for the existance of the defect
    • Plan for data collection
      • "In God We Trust, All Others Bring Data" (Deming)
    • Formulation of Theories - Fishbone (Ishakawa)
    • Validate the measurement system
    • Measure the baseline performance
    • Measure the process capability
  6. Analyze

    • Plan for data collection - data that enables testing
    • Collect and analyze data
    • Test theories of management controllable problems (80%)
      • Flow diagrams
      • Process capability analysis
      • Product and process dissection
      • Stream to stream analysis (multiple input streams)
      • Time to time analysis - plot of data on a timescale, time between abnormalities, rate of change & cumulative data techniques
      • Test theories by collection of new data
      • Test theories involving human error
        • Inadvertent errors: unintentional, unwitting (unaware), unpredictable: remedies include reduce dependence on human attention or help workers remain attentive
        • Technique errors - unintentional, specific, consistent, unavoidable
        • Conscious errors - witting, intentional, persistent (mgt shifting priorities)
        • Communication errors
        • Theory of Constraints
  7. Improve

  8. The improve phase designs a remedy, prove its effectiveness and prepare for implementation
    • Evaluate alternative remedies
    • If necessary, design experiments
    • Design a remedy
    • Prove the effectiveness
    • Deal with resistence to change (establish the need, start small, avoid surprises, treat people with dignity, deal with resistance)
    • Transfer the remedy to operations (revise standards, processes, staffing, equipment, materials and training)
  9. Control

  10. The control phase designs activities to hold the gains of improvement


Gryna, Chua, et al (2007), Juran's Quality Planning and Analysis, 5th Edition.