APMG LSS Green Belt cheat sheet

SIPOC maps Suppliers, Inputs, Process, Outputs, Customers at a high level to bound project scope.

Project charter elements: problem statement, goal statement, scope, business case, team, milestones, and sponsor sign-off.

SMART goal rule: Specific, Measurable, Achievable, Relevant, Time-bound; a goal should be quantified with baseline and target.

Voice of the Customer (VOC) is translated into Critical to Quality (CTQ) requirements via a CTQ tree.

Kano model classifies features as must-be, performance, or delighter based on customer satisfaction impact.

Cost of Poor Quality (COPQ) = internal failures + external failures + appraisal + prevention costs.

Pareto principle: roughly 80% of effects come from 20% of causes; used to prioritize problem focus.

Stakeholder analysis plots influence against interest to plan communication and engagement.

A problem statement describes gap, location, timing, and magnitude without stating cause or solution.

Data types: continuous (variable) data is measured on a scale; discrete (attribute) data is counted or classified.

Measurement System Analysis (MSA): Gage R&R assesses repeatability (equipment) and reproducibility (appraisers).

A rule of thumb: Gage R&R below 10% of tolerance is acceptable; above 30% is generally unacceptable.

Process capability Cp = (USL - LSL) / 6sigma; Cpk = min(USL - mean, mean - LSL) / 3sigma.

Cp compares spread to specification width; Cpk additionally accounts for centering of the process mean.

Sigma level relates to DPMO: 6 sigma is about 3.4 defects per million opportunities.

Defects per Opportunity (DPO) = defects / (units x opportunities per unit); DPMO = DPO x 1,000,000.

Baseline data collection plan defines what, how, when, and by whom data are recorded to ensure validity.

Sampling: random sampling reduces bias; sample size affects confidence interval width.

Cause-and-effect (fishbone) diagram organizes potential causes into categories such as Man, Machine, Method, Material, Measurement, Environment.

5 Whys drills from a symptom to a root cause by repeatedly asking why; stop when cause is actionable and verified.

Scatter plot examines relationship between two continuous variables; correlation does not establish causation.

Hypothesis test logic: null hypothesis assumes no difference or no effect; p-value below alpha leads to rejecting the null.

Type I error (alpha) is rejecting a true null (false alarm); Type II error (beta) is failing to reject a false null; power = 1 - beta.

ANOVA compares means of three or more groups; t-test compares one or two group means.

Regression models the relationship between a continuous response and one or more predictors; R-squared is proportion of variance explained.

FMEA uses Severity x Occurrence x Detection to compute a Risk Priority Number (RPN) for prioritization.

Multi-vari analysis studies within-piece, between-piece, and temporal variation to isolate dominant variation sources.

Brainstorming generates candidate solutions; affinity diagrams group and organize the resulting ideas.

Pugh matrix compares solution concepts against a datum using weighted criteria for selection.

Design of Experiments (DOE) manipulates factors at set levels to estimate main effects and interactions.

Full factorial with k factors at 2 levels requires 2^k runs; fractional factorials reduce runs at cost of confounding.

Kaizen is continuous incremental improvement; kaizen blitz is a short, focused rapid-improvement event.

5S: Sort, Set in order, Shine, Standardize, Sustain - a workplace organization method.

Poka-yoke is error-proofing that prevents defects from being made or passed forward.

SMED (Single-Minute Exchange of Die) reduces changeover time by separating internal and external setup tasks.

Pilot the chosen solution on a small scale before full implementation to verify effect and reduce risk.

Statistical Process Control (SPC) charts distinguish common-cause (inherent) from special-cause (assignable) variation.

Control chart rule: a point beyond the control limits signals special cause; limits are typically mean +/- 3 sigma.

X-bar and R charts are used for continuous data in subgroups; p, np, c, and u charts are for attribute data.

Process capability indices (Cp, Cpk) should be re-evaluated after improvement to confirm sustained gains.

Control plan documents what is monitored, the method, frequency, owner, and reaction plan for each key characteristic.

Standardized work captures the current best method and is the baseline for future improvement.

Mistake-proofing and checklists in the control phase help sustain gains and prevent reversion.

A response plan specifies the action taken when a control chart signals an out-of-control condition.

Project closure includes handover to process owner, documentation, and lessons learned.