Six Sigma — Practice Q&A
Q: Why does DMAIC insist on completing Measure before Analyze, rather than allowing teams to jump straight to root-cause analysis based on experience?
A: Analysis based on assumption or anecdote rather than actual data risks misidentifying the real cause of variation — teams skipping rigorous measurement often converge on an intuitively appealing but factually incorrect root cause, then implement a solution that doesn't actually address the real problem. Measure establishes an actual data baseline that Analyze then investigates, ensuring root-cause conclusions are evidence-based rather than assumption-based.
Q: What's the difference between common cause and special cause variation, and why does misclassifying them waste effort?
A: Common cause variation is the natural, inherent variation present in any stable process, arising from many small always-present factors — addressed by redesigning the process, not reacting to individual instances. Special cause variation comes from an identifiable, specific, often intermittent source — addressed by finding and eliminating that specific source. Treating normal common-cause variation as if each instance has a specific findable cause wastes effort chasing phantom causes, while treating genuine special-cause problems as unavoidable variation misses addressable, fixable issues.
Q: Why is Voice of the Customer considered essential to the Define phase specifically, rather than something addressed later in the project?
A: A Six Sigma project can technically reduce process variation and defects while still missing the point if it optimizes against the wrong requirement. Voice of the Customer grounds the entire DMAIC effort in what customers actually need from the very first phase, ensuring later phases' data-driven rigor is applied to a genuinely relevant problem rather than one based on internal assumption about what matters.
Q: Why does Design of Experiments (DOE) outperform testing one process variable at a time?
A: Testing one variable at a time while holding others constant can miss interaction effects — cases where two variables' combined effect differs from what their individual effects alone would predict. DOE's systematic, simultaneous variation of multiple inputs can detect these interaction effects and typically requires fewer total experimental runs to fully characterize a process than exhaustive one-variable-at-a-time testing, making it both more complete and more efficient for processes with multiple potentially-interacting inputs.
Q: Why does Six Sigma include a Control phase after Improve, rather than considering the project complete once a solution is implemented?
A: Without genuine Control-phase discipline — control charts, a documented Control Plan, sustained monitoring — process improvements frequently regress toward prior performance over time as initial attention fades. Control exists specifically to prevent this regression through structured, ongoing monitoring, distinguishing normal variation from a genuine special-cause signal requiring investigation, rather than treating improvement as a one-time fix followed by disengagement.
Q: How does Lean Six Sigma combine two methodologies with different core questions?
A: Lean asks "is this step adding value, or is it waste to be eliminated," focused on eliminating waste and improving flow. Six Sigma asks "how much does this process vary, and why," focused on variation reduction through statistical rigor. Lean Six Sigma combines both perspectives, reflecting the recognition that genuine process improvement benefits from addressing both waste elimination and variation reduction rather than treating either as sufficient on its own.