In one line: A rigorous, data-driven method for improving critical processes: measure how far they vary from what the customer needs, find the root causes, and fix the process so it stays within limits.
Core idea
Motorola created Six Sigma in the early 1980s, and it became famous when General Electric adopted it under Jack Welch [1]. The name comes from statistics. A process whose nearest specification limit is six standard deviations (σ) from the mean, allowing for some drift over time, produces about 3.4 defects per million opportunities. In practice, the term now covers a methodology, a set of tools and a structured improvement organization.
How it works
DMAIC, the improvement cycle for existing processes:
| Step | Question | Typical tools |
|---|---|---|
| Define | What is the problem, for which customer, and why does it matter? | project charter, SIPOC, voice of the customer |
| Measure | How does the process perform today? | data collection, measurement-system analysis, process capability |
| Analyze | What are the root causes of variation and defects? | Pareto charts, fishbone diagrams, hypothesis tests, regression |
| Improve | Which changes fix the root causes? | Design of Experiments, piloting, mistake-proofing |
| Control | How do we hold the gains? | control charts, standard work, response plans |
For new designs, a variant called DMADV (Define, Measure, Analyze, Design, Verify) underpins Design for Six Sigma.
Six Sigma programs often assign trained roles (Green Belts, Black Belts, Master Black Belts) and require projects to show measurable financial or customer benefit.
In practice
Some engineering companies require every new hire to complete a Six Sigma project. A typical early-career example: complex laboratory test benches that only one or two experts know how to assemble. Each new setup depends on finding those people, and results vary between benches. Defining the problem, measuring setup time and errors, analyzing where the variation comes from, documenting a standard build procedure, and controlling it with checklists turns tribal knowledge into a repeatable process. The lesson generalizes: wherever outcomes depend on a hero, there is an unmeasured, uncontrolled process waiting to be improved.
Questions to ask yourself
- Which process on my team depends on one person’s knowledge?
- Do I know the capability of the process I rely on, or only its average?
- Which DMAIC step do we usually skip, Measure or Control?
Connections
- Partner method: Lean. Foundation: Statistics in Engineering. Generic loop: PDCA Cycle.
- Key improvement tool: Design of Experiments. Lineage: Total Quality Management.
- Capabilities exercised: Realizing the Vision, Decision Making, Inquiry.