What Is Six Sigma?
Also known as: 6 Sigma, DMAIC
Definition
Six Sigma is a data-driven improvement methodology that reduces process variation and defects using statistical analysis, targeting a capability level of 3.4 defects per million opportunities, typically executed through the five-phase DMAIC cycle.
Six Sigma Explained
The name refers to placing six standard deviations between the process mean and the nearest specification limit. Under a static normal distribution that would imply roughly two defects per billion, but the conventional Six Sigma target of 3.4 defects per million opportunities incorporates an assumed 1.5 sigma long-term shift in the process mean. That shift assumption is an empirical convention, not a law of statistics, and understanding it prevents the common confusion when calculated and quoted defect rates disagree.
DMAIC structures the work: Define the problem and customer requirement, Measure current performance with a validated measurement system, Analyze to identify root causes with data rather than opinion, Improve by testing and implementing changes, and Control by locking in the gain with monitoring and standard work. The Measure phase includes measurement system analysis, and skipping it is the most frequent cause of failed projects because a gauge contributing 30 percent of observed variation makes every subsequent analysis unreliable.
Roles follow a belt hierarchy. Green Belts run improvement projects alongside their regular job, Black Belts work full time on projects and coach Green Belts, and Master Black Belts own methodology, training, and project portfolio selection. Champions are executives who remove barriers and approve resources. The structure works only when project selection is disciplined; programs that let belts pick their own low-impact projects generate certifications rather than financial results.
Six Sigma's practical limits should be stated honestly. It requires enough data to characterize a distribution, so it fits repetitive processes far better than low-volume aerospace or defense production where a run may be twelve parts. It addresses variation, not speed, so a perfectly capable process can still have a twelve-week lead time. And its tools assume the problem is technical; where the real cause is scheduling policy or organizational incentives, statistical analysis will describe the symptom precisely and change nothing.
Why It Matters
- Replaces opinion-driven troubleshooting with a documented, data-based chain from symptom to verified root cause.
- The Control phase is what distinguishes it from ad hoc fixes, embedding monitoring so improvements survive personnel and shift changes.
- Reduced variation raises process capability, which directly lowers scrap, rework, inspection cost, and warranty exposure.
- Customer quality systems in aerospace, automotive, and defense increasingly expect statistical evidence of capability rather than inspection records.
In Practice
A team chases a 6 percent out-of-tolerance rate on a bore diameter and prepares to buy a replacement machine. Measure-phase gauge R and R shows the bore gauge consumes 41 percent of total tolerance, well above the 30 percent threshold for a marginal system. After the gauge is replaced and operators are retrained on technique, measured defect rate falls to 1.8 percent with no change to the machine. Roughly two-thirds of what looked like process variation was measurement variation, and the capital request was cancelled.
Frequently Asked Questions
Why is Six Sigma 3.4 defects per million and not 0.002?
The pure statistical tail beyond six standard deviations is about two defects per billion. The quoted 3.4 per million adds an assumed 1.5 sigma long-term drift in the process mean, reflecting the observation that real processes shift over time due to tool wear, material lots, and environment. It is an empirical convention adopted at Motorola, not a derived result.
Do you need a Black Belt to run a Six Sigma project?
No. Most improvement projects are run by Green Belts with Black Belt coaching, and many high-value problems are solved with basic tools such as Pareto analysis, gauge R and R, and simple hypothesis tests. Full Black Belt depth matters for designed experiments, multivariate analysis, and coaching a portfolio, not for every individual project.
Related Terms
SPC (Statistical Process Control)
SPC (Statistical Process Control) is the use of control charts and statistical rules to monitor a process in real time, distinguishing normal common-cause variation from special-cause variation that signals a genuine change requiring action.
Cp and Cpk
Cp and Cpk are process capability indices. Cp compares the specification width to the process spread, assuming perfect centering, while Cpk also accounts for how far the process mean has drifted from the center of the tolerance.
FMEA (Failure Mode and Effects Analysis)
FMEA (Failure Mode and Effects Analysis) is a structured risk assessment that identifies how a product or process can fail, the effects and causes of each failure, and existing controls, then prioritizes actions to prevent the highest-risk failures.
Go Deeper
Manufacturing Downtime Cost Calculator
Combine lost contribution margin, idle labor, and absorbed overhead into a defensible monthly and annual downtime cost - with recovery effects modeled.
Manufacturing Costing Methods Compared: Standard, Actual, Average
Standard, actual, and average costing compared for manufacturers: variance analysis, ERP setup in SyteLine and LN, DFARS considerations, and AI cost intelligence.
MES-ERP Integration for Discrete Manufacturing
MES-ERP integration guide for discrete manufacturers: architecture patterns, ISA-95 mapping, SyteLine and LN connectors, costs, timelines, and AI-driven sync.
Working with Six Sigma in a live environment? Our engineers do this every day - and our AI agents automate most of it.