Six sigma

The name six sigma comes from statistics: sigma denotes standard deviation, a measure of variation. In quality management, the term represents an exceptionally capable process. The traditional account describes 99.999998 per cent conforming output and no more than 3.4 defects per million opportunities, although those figures rely on different statistical conventions and should not be treated as interchangeable without stating the assumed process shift. The managerial method seeks to approach this level by understanding, reducing and controlling variation. Better quality and process performance are means to an end; the intended outcome is measurable operational and financial improvement.

When to use it

Use six sigma when an organisation needs to identify the causes of poor operational performance, improve process reliability and translate fewer defects into customer or financial value. A project should have a defined business case, but the claim that management should reject any initiative without at least $175 000 in savings is a legacy rule of thumb, not a universal threshold. Scale the expected benefit to the process, organisation and risk involved.

Six sigma is normally sponsored from the top and delivered through a formal project structure. Leaders communicate the purpose, select priorities and review results, while trained employees conduct the analysis. Common roles include:

• Executive management champions – the CEO or other senior leaders who maintain oversight of the project portfolio and remove organisational barriers.
• Master black belts – expert practitioners who train black belts, advise teams and support consistent application.
• Black belts – project leaders accountable for the complete improvement effort.
• Green belts – practitioners who lead part-time or bounded initiatives and implement improvement work.
• Project teams – employees who contribute process knowledge and apply the relevant techniques under green-belt leadership.

Each organisation adapts the infrastructure, but effective implementation usually requires:

• Sound understanding of the statistical methods being used.
• Enough time and evidence to define the problem correctly.
• Adequate resources for implementation, not only analysis.
• Visible leadership commitment and active governance.
• Attention to the cultural change that new controls may require.
• A communication plan that explains purpose, progress and consequences.
• Practical training for the people doing the improvement work.
• Black belts who can facilitate across functions as well as analyse data.

Origins

Motorola established six sigma as a corporate quality approach during the mid-1980s. Engineer Bill Smith is widely associated with formulating it, while chief executive Bob Galvin sponsored the organisational commitment and Mikel Harry helped develop the technical method and training infrastructure. The work linked process capability, defect measurement and improvement projects within a managed deployment system rather than treating statistics as an isolated quality exercise. AlliedSignal and General Electric later helped popularise the approach beyond Motorola and reinforced its association with trained project roles and financially accountable improvement.

What it is

Motorola intensified its quality effort in 1987 while facing strong Japanese competition. Its engineers concluded that measuring defects per 1000 units was too coarse for the reliability they sought and shifted attention to defects per million opportunities. AlliedSignal and General Electric later extended and popularised the method, reporting substantial savings alongside improved customer outcomes. Six sigma subsequently moved beyond manufacturing into service processes.

The core premise is that process variation drives defects and unpredictable results. Statistical tools help a team describe current fluctuation, distinguish signal from noise and estimate likely outcomes. When performance is inadequate, further analysis identifies the inputs and causes that most influence the result, allowing the team to improve the process and establish controls that sustain the gain.

How to use it

A standard improvement project follows five DMAIC stages: define, measure, analyse, improve and control.

1. Define. Select the process and problem, clarify customers and requirements, establish scope and agree SMART goals—specific, measurable, acceptable, realistic and time-specific.
2. Measure. Establish reliable definitions and collect data that describe current performance and create a baseline for comparison.
3. Analyse. Determine the gap between present and desired performance, then test which causes account for the variation or defects.
4. Improve. Design, test and implement changes that address the verified causes.
5. Control. Formalise the improved process, assign ownership and monitor the measures needed to prevent regression.

Final analysis

Six sigma combines technical and behavioural disciplines. Its hard elements include a structured problem-solving sequence, statistical process-control tools, DMAIC and formal project management. Its soft elements include leadership, creativity, communication, motivation and the ability to change established working practices.

Benchmarking can support a project by comparing important characteristics of the product, customer experience, internal process or production system with relevant alternatives. Process-level comparison helps financially oriented managers see where six sigma analysis may create value, but external differences must be interpreted carefully rather than copied mechanically.

Vision and enthusiasm are useful, but they cannot replace infrastructure. Sustainable results depend on capable project selection, training, expert support, coordination and clear accountability for maintaining the improved process.

Top practical tip

Define the financial or customer value before launching the project, then connect every technical measure to that outcome.

Top pitfall

Do not mistake the belt structure or statistical vocabulary for improvement. Results come from a well-scoped problem, valid data, verified causes and controls that people can sustain.

Further reading

Breyfogle III, F.W. (2003) Implementing Six Sigma: Smarter Solutions using Statistical Methods. Hoboken, New Jersey: John Wiley & Sons.