EXIN LSSA Black Belt Practical key terms

A measurable characteristic of a product or service that directly reflects what a customer values, translated from vague customer wishes into a specific, testable requirement a project can target and verify.

The collection of stated and unstated needs, expectations, and complaints gathered from customers through interviews, surveys, and complaint data, which serves as the starting input for defining improvement projects and CTQs.

A classification of customer requirements into must-be, performance, and delighter categories, used to decide which features to fix first, which to improve proportionally, and which to add for differentiation.

The core improvement logic stating that the output result Y is a function of input and process variables x; projects identify which x variables most influence Y and control them rather than reacting to output variation.

The five-phase improvement roadmap: Define the problem and scope, Measure baseline performance, Analyze causes, Improve by implementing and verifying solutions, and Control by standardizing and monitoring the gains.

A design-for-Six-Sigma roadmap: Define design goals, Measure customer and process requirements, Analyze design options, Design the chosen solution, and Verify performance through pilots before full launch.

A formal one-page agreement stating the business case, problem statement, goal, scope boundaries, team, milestones, and expected savings, signed by the sponsor before work begins.

The senior manager who authorizes a Black Belt project, provides resources and removes organizational barriers, and is accountable for sustaining the results after the project team disbands.

A high-level map listing Suppliers, Inputs, Process steps, Outputs, and Customers for a process, used at project start to agree scope and identify where measurements must be taken.

A step-by-step diagram of how work actually flows, including decision points, handoffs, and rework loops, used to expose complexity, delays, and non-value-adding steps before measurement begins.

A lean diagram showing material and information flow with cycle times, wait times, inventory, and lead time, used to separate value-adding from non-value-adding activity and locate the largest delays.

A structured investigation using tools such as the 5 Whys, fishbone diagram, or fault tree to move past symptoms and identify the underlying process conditions that produce a defect.

A repeated questioning technique that asks why a problem occurred and then why each answer occurred, continuing until a controllable process cause is reached rather than a person or an excuse.

A fishbone-style chart organizing potential causes of a problem into branches such as people, machine, method, material, measurement, and environment, used to structure brainstorming before data collection.

A study quantifying how much observed variation comes from the measurement system itself, covering repeatability, reproducibility, bias, linearity, and stability, so that data can be trusted for decisions.

A specific MSA study that separates repeatability (same appraiser, same part) from reproducibility (different appraisers), usually expressed as a percentage of total or tolerance variation.

An MSA method for pass/fail or categorical judgments, measuring within-appraiser consistency and agreement between appraisers and against a known standard, reported with kappa statistics.

Cp compares specification width to process spread ignoring centering; Cpk also accounts for how far the process mean sits from the nearest specification limit, so Cpk is always the lower, more realistic figure.

Indices calculated with the overall standard deviation of all data rather than within-subgroup variation, describing how the process actually performed over the period observed rather than its inherent short-term capability.

A shorthand converting process capability into the number of standard deviations between the mean and the nearest specification limit, commonly quoted with a 1.5-sigma long-term shift assumption.

A normalized defect rate calculated as defects divided by total defect opportunities, multiplied by one million, allowing fair comparison between processes of different complexity and volume.

The probability that a unit passes every process step without any defect, obtained by multiplying the first-pass yields of all steps; it exposes hidden rework that final inspection yield conceals.

A time-ordered plot of a process measure with statistically derived control limits, used to distinguish common-cause variation from special-cause signals and to judge whether a process is stable.

Routine, inherent variation produced by the design of the process itself, appearing as random scatter within control limits; it can only be reduced by changing the process, not by adjusting individual points.

Irregular variation from an identifiable source outside the normal process, signalled by points beyond control limits, runs, trends, or other pattern rules; it must be investigated and removed before capability is assessed.

The deliberate grouping of consecutive measurements taken under the same conditions so that subgroups capture only common-cause noise, making the between-subgroup variation meaningful for control limits.

The state in which a process shows only common-cause variation on a control chart, a prerequisite for calculating meaningful capability indices, since unstable data make Cp, Cpk, and similar figures invalid.

A statistical procedure that compares observed sample data against a null hypothesis, returning a p-value used to decide whether an observed difference is larger than plausible random variation.

The probability of observing data at least as extreme as the sample result if the null hypothesis were true; a small value suggests the effect is unlikely to be mere chance.

Rejecting a true null hypothesis, concluding a difference exists when it does not; its probability is the significance level alpha, typically set at 0.05 before data are collected.

Failing to reject a false null hypothesis, missing a real effect; its probability is beta, and one minus beta is the statistical power of the test, improved by larger samples or smaller noise.

The number of observations needed to detect a stated effect with chosen alpha and power, driven by the size of the difference worth finding and the amount of variation in the process.

A planned set of runs in which factors are varied together according to a statistical layout, allowing estimation of main effects and interactions with far fewer trials than one-factor-at-a-time testing.

An experimental layout in which every combination of factor levels is run, either fully or fractionally, enabling efficient estimation of main effects and interactions while controlling confounding.

A situation where the effect of one factor on the response depends on the level of another factor, meaning factors cannot be optimized independently and must be studied together.

A set of designs and models, such as central composite or Box-Behnken layouts, used to fit curved relationships and locate factor settings that optimize a response rather than merely screen factors.

A mistake-proofing device or design feature that prevents an error from being made or makes it immediately obvious, removing reliance on operator attention and inspection.

A document specifying, for each key process step, what is monitored, the measurement method, sample frequency, control limits, and the corrective action taken when a signal occurs.

A documented, agreed sequence of tasks with defined cycle time, work-in-process, and quality checks that becomes the current best method and the baseline against which further improvement is measured.

The time required to complete one unit of work from start to finish of a given process step, used to calculate capacity and to balance workload against takt time.

The available production time divided by customer demand, defining the pace at which a process must complete one unit to meet demand without overproducing or building backlog.