Six Sigma

Six Sigma is a data-driven quality improvement methodology that aims to reduce process variation to achieve fewer than 3.4 defects per million opportunities. It uses structured problem-solving (DMAIC or DMADV) and statistical analysis to identify and eliminate the root causes of variation.

Why companies use it

·Provides a rigorous, statistical approach to reducing defects in high-volume or high-cost processes
·The certification belt system (Green Belt, Black Belt) creates a structured internal improvement capability
·GE, Motorola, and later the semiconductor and medical device industries adopted it to drive measurable financial returns
·DMAIC gives a clear framework for scoping improvement projects and measuring results rigorously

What hiring managers look for

·Green Belt or Black Belt certification signals that a candidate has led or contributed to a structured improvement project
·Statistical literacy (hypothesis testing, regression, DOE) is required to do Six Sigma properly and is a differentiating skill
·Candidates who can scope a Six Sigma project, define a Y metric, and identify X variables show systems thinking
·Combined lean + Six Sigma (Lean Six Sigma) capability is the most requested profile in manufacturing quality

Typical interview questions

What does "Six Sigma" mean statistically? How many defects per million opportunities is that?

Walk me through the DMAIC process. What is the goal of each phase?

What statistical tools would you use in the Analyse phase of a DMAIC project?

What is the difference between DMAIC and DMADV, and when would you use each?

Describe a Six Sigma project you have led or contributed to. What was the problem, your role, and the result?

Common mistakes

·Starting the Improve phase before validating root causes — implementing solutions for the wrong causes wastes time and money
·Defining the project Y metric too vaguely — if you cannot measure it precisely, you cannot prove improvement
·Skipping the Control phase — improvements without a control plan revert within months
·Using Six Sigma for problems that do not require statistical rigour — a simple kaizen or 5 Whys often suffices
·Confusing sigma level with process capability index: a 6-sigma process has a Cpk of 2.0 (with 1.5-sigma shift assumed, Cpk ≈ 1.5)

Real engineering example

A PCB assembly line was experiencing 4,200 defects per million solder joints (approximately 3.6 sigma). A Green Belt DMAIC project identified that 68% of defects were cold solder joints on the same 12 component types, all QFP packages. DOE on reflow profile parameters showed that increasing pre-heat ramp rate by 0.8°C/second eliminated the temperature gradient causing poor wetting. Post-improvement DPMO fell to 380 (5.3 sigma), saving €1.2M annually in rework.

Topics covered

Six SigmaDMAICGreen BeltBlack BeltDPMOstatisticsquality