EXIN LSSA Lean Six Sigma Yellow Belt cheat sheet

Lean targets speed and flow by removing the eight wastes (TIMWOODS: Transport, Inventory, Motion, Waiting, Overproduction, Overprocessing, Defects, Skills unused); Six Sigma targets variation and defects via data and statistics.

Sigma level equals the number of standard deviations between the process mean and the nearest specification limit; higher sigma means fewer defects per million opportunities (DPMO).

DPMO to sigma reference points: 1 sigma = 691,462 DPMO; 3 sigma = 66,807; 4 sigma = 6,210; 5 sigma = 233; 6 sigma = 3.4.

Defect rate = defects ÷ (units × opportunities per unit); DPMO = defect rate × 1,000,000.

Rolled Throughput Yield (RTY) = product of each step's first-pass yield (Y1 × Y2 × … × Yn); it exposes hidden factory losses that final inspection hides.

First Pass Yield = units passing a step with no rework ÷ units entering the step.

Cycle time = available work time ÷ customer demand (takt time); process lead time = work in process × cycle time (Little's Law).

The classic 1.5 sigma shift is an empirical allowance for long-term drift; a 6-sigma process is quoted as 3.4 DPMO short-term capability.

Y = f(x): the output Y is a function of process inputs x; Six Sigma projects control the critical x's rather than inspecting the Y.

DMAIC sequence: Define the problem and charter; Measure baseline and measurement system; Analyze root causes; Improve with validated countermeasures; Control and standardize to hold gains.

Each phase ends with a tollgate review that checks deliverables (charter, data collection plan, verified root cause, pilot results, control plan) before the next phase starts.

DMAIC is used for existing processes with unknown root causes; DMADV (Define, Measure, Analyze, Design, Verify) is used for new designs or processes that cannot meet specs.

Phase deliverables gate the next phase: no verified root cause in Analyze means no solution selection in Improve.

PDCA (Plan-Do-Check-Act) is the underlying logic; DMAIC maps Plan to Define/Measure/Analyze, Do to Improve, Check and Act to Control.

A DMAIC project should be scoped so it can complete in roughly 3-6 months; larger scopes are split into sequential projects.

Kaizen events compress improvement into days and usually sit inside the Improve phase of a larger DMAIC effort.

Typical Yellow Belt contribution in DMAIC: collect and validate data in Measure, map the process, support root cause analysis, and help sustain the control plan.

Problem statement content: what is wrong, where, when, magnitude with data, and business impact; it must not state a cause or a solution.

SMART goal rule: Specific, Measurable, Achievable, Relevant, Time-bound; express the goal as a metric with baseline and target, e.g. reduce defect rate from 8% to 3% by Q4.

Voice of the Customer (VOC) is translated into Critical to Quality (CTQ) requirements via the VOC-to-CTQ tree: need, quality driver, measurable requirement, specification.

CTQ must be measurable and tied to a spec limit; a vague wish such as 'customers want it fast' is not a CTQ until converted to a metric with a target.

SIPOC (Suppliers, Inputs, Process, Outputs, Customers) defines project boundaries at a high level and identifies the start and stop points of the process under study.

A project charter contains business case, problem statement, goal statement, scope with in/out-of-scope items, team roles, milestones, and expected benefits.

Stakeholder analysis maps influence against interest; high-influence/high-interest parties need close engagement and are often the project sponsor or process owner.

The sponsor provides resources and removes barriers; the process owner owns the improved process after handover; the team does the analysis work.

Scope creep is controlled by the charter's explicit in-scope and out-of-scope lists, which are re-confirmed at each tollgate.

Data types: continuous (variable) data is measured on a scale (time, length, weight); discrete (attribute) data is counted (pass/fail, number of defects, defectives).

Operational definition states exactly how a value is measured, including what is included, excluded, and the measurement instrument, so collectors produce consistent data.

Data collection plan specifies the metric, operational definition, data source, sample size, sampling frequency, collector, and recording format.

Sampling rules: random sampling reduces bias; stratified sampling compares subgroups; a sample must be representative of the population and the time period.

Measurement System Analysis (MSA) checks the measurement system before trusting data; a Gage R&R study separates repeatability (same appraiser, same part) from reproducibility (different appraisers).

Bias is the difference between the average measured value and the true reference value; linearity is bias across the operating range; stability is consistency over time.

Rule of thumb: measurement variation should be small relative to the process or tolerance variation; a common guideline is gauge variation under 10% of tolerance, 10-30% conditional.

Process mapping levels: SIPOC (high level), swimlane/functional flowchart (who does what), and detailed flowchart (decision points and rework loops).

Value stream mapping records process time and lead time for each step and marks value-adding, non-value-adding, and necessary non-value-adding steps.

Baseline capability metrics: Cp compares specification width to process spread; Cpk also accounts for centering; both require stable, normal data.

The 5 Whys asks 'why' repeatedly until the systemic cause is reached; the answer to each why must be evidence-based, not opinion.

Fishbone (Ishikawa) diagram organizes candidate causes into categories: Man, Machine, Material, Method, Measurement, and Environment (6M).

Cause-and-effect analysis generates hypotheses; verification with data (stratification, correlation, hypothesis test) is required before a cause is accepted.

Pareto principle: roughly 80% of effects come from 20% of causes; a Pareto chart ranks causes by frequency or cost to focus effort.

Eight wastes (TIMWOODS): Transport, Inventory, Motion, Waiting, Overproduction, Overprocessing, Defects, unused Skills.

Value-added test: the step changes the form or function of the product, the customer cares about it, and it is done right the first time; failing any test makes it non-value-adding.

The Seven Quality Control tools: check sheet, histogram, Pareto chart, cause-and-effect diagram, scatter diagram, control chart, flowchart.

Scatter diagrams show association between two variables; correlation does not establish causation without further evidence.

Stratification splits data by shift, machine, operator, or supplier to reveal where the problem concentrates.

A root cause is confirmed when removing it makes the problem disappear and restoring it brings the problem back.

5S workplace organization: Sort, Set in order, Shine, Standardize, Sustain; it removes motion and waiting waste and exposes abnormal conditions.

Poka-yoke is error-proofing that prevents a defect from being made or passed on: guide pins, sensors, limit switches, color coding, forced sequence.

SMED (Single-Minute Exchange of Die) separates internal setup (machine stopped) from external setup (machine running) and converts internal to external to cut changeover time.

Standard work specifies the current best sequence, cycle time, and standard in-process inventory for each operator; it is the baseline for further improvement.

Kanban is a pull signal that authorizes production or replenishment; it limits work in process and prevents overproduction.

Cellular layout and one-piece flow reduce transport and waiting by co-locating sequential operations.

Kaizen is continuous small improvement by the people doing the work; a kaizen event is a focused, time-boxed improvement workshop.

Countermeasures are selected against verified root causes; pilot them on a limited scale, measure results against the baseline, then standardize.

FMEA (Failure Mode and Effects Analysis) prioritizes risks by Risk Priority Number = Severity × Occurrence × Detection; higher RPN gets action first.

Piloting an improvement with before/after data confirms the change actually moved the metric before full deployment.

A control plan lists the metric, specification or target, measurement method, sample size and frequency, responsible person, and reaction plan for each controlled characteristic.

Control charts distinguish common-cause variation (inherent, stable process) from special-cause variation (assignable events) using control limits at ±3 sigma from the process mean.

Out-of-control signals include a point beyond a control limit, a run of 7 or more points on one side of the centerline, and a trend of 7 rising or falling points.

Control limits are calculated from process data; specification limits come from the customer. Plotting spec limits on a control chart is a common error.

Standardization tools: updated work instructions, checklists, training records, visual controls, and revised procedures with document control.

Mistake-proofing and automation in Control reduce reliance on operator vigilance and prevent reversion to the old method.

An audit schedule with defined frequency and owner verifies that the new standard is being followed after the project closes.

Project closure includes handover to the process owner, documented lessons learned, and verification that the gain is sustained over several reporting periods.

Reaction plan defines who does what when a control chart signals, so response does not depend on memory.

Yellow Belt role: participate as a team member, apply basic tools, collect and validate data, map processes, and support Green and Black Belts; typically does not lead complex projects alone.

Belt hierarchy: White Belt (awareness), Yellow Belt (basic tools, team participation), Green Belt (leads projects part-time with statistics), Black Belt (leads complex projects, coaches), Master Black Belt (program and training ownership).

Team stages (Tuckman): Forming, Storming, Norming, Performing, and Adjourning at project close.

Effective team size is typically 4-7 members with the process knowledge and authority needed to change the work.

RACI matrix assigns Responsibility, Accountability, Consulted, and Informed for project tasks; exactly one person is Accountable per task.

Meeting discipline: agenda issued in advance, roles for facilitator and scribe, action items with owner and due date, minutes distributed after.

Constructive team behaviors: share data openly, disagree on evidence not opinion, and escalate unresolved conflicts to the sponsor.

Ethical data handling: report data as collected, never discard inconvenient points without a documented, valid reason, and record all changes.