5 Why?

This is a basic Six Sigma problem-solving method.
It allows us to go beyond symptomatic causes and find the root causes (on which we can then act to eliminate the problem once and for all).
The principle is to trace back to the root cause by asking the question why 5 times in a tree structure.

5 M

5 fundamental causes of process dispersion: Material, Methods, Means, Environment, Manpower.
These 5 causes are already set out in Shewhart's first book.
We generally dissociate the production process with its 5Ms and the measurement process with its 5Ms (Mesurande replaces Matière).

5 S

Lean method structured in 5 phases corresponding to 5 Japanese concepts (Seiri, Seiton, Seiso, Seiketsu, Shitsuke / Sort - Set in order - Shine - Standardize - Sustain).
Based on the hunt for waste in an area or piece of equipment, the 5S method is used to create a good working environment for value-added operations. It enables operational improvement that goes beyond order and cleanliness, through the implementation of standards and team maturity.
It can be used in both industrial and service applications.

A 3

Methodology invented by Toyota to manage problem-solving, summarizing the actions taken (project progress).

Alpha, Alpha risk, First-species risk

The Alpha risk corresponds to the risk of concluding that there is a difference between two samples, when in reality there is no difference.

FMEA

Failure Mode and Effect Analysis. This is a tool for listing risks and evaluating them according to 3 criteria: severity, ease of detection and frequency of occurrence.

Sequence analysis

Process mapping method that visualizes the sequence of elementary tasks in a process or part of a process. Each task is valued (time, distance, quantity, etc.) with the aim of improving the process.

Stakeholder analysis

Stakeholder analysis is part of the DEFINE stage. We look for all the people impacted by the project, i.e. :

• the people involved in the project
• people interested in the project
• people who can influence the project.
  The stakeholder analysis is carried out by the GB or BB. It facilitates the constitution of the project team and the creation of the communication plan. It evolves throughout the project

ANAVAR (ANOVA)

ANAlyse de la VARiance, ANalysis Of VAriance. This is a hypothesis test used to compare the mean of several samples.

ANDON

An ANDON is a warning light or panel which lights up when the operator presses an alert button. The ANDON is followed by problem solving at increasingly expert levels, depending on the time spent on solving the problem. The aim is to minimize disruption or stoppage of production.

APC

Evolution of the MSP/SPC to adapt to the case where there is a dependency between several characteristics and where precise adjustment is required with a minimum of information.

Decision tree - Regression tree

Machine learning method adapted to discrete Y responses to find a prediction model from a decision tree such that :

Decision Tree Regression

0 =: (0 <= 0.989) && (1 <= 1.7525)
0 =: (0 1.7525) && (3 <= 0.5)
0 =: (0 1.7525) && (3 > 0.5)
0 =: (0 > 0.989) && (2 -0.681)
0 =: (0 > 0.989) && (2 <= 0.5) && (1 0.5)
1 =: (0 > 0.989) && (2 > 0.5)
1 =: (0 > 0.989) && (2 <= 0.5) && (1 <= -0.681) && (3 <= 0.5)

Cause tree

It's a risk analysis tool that starts with a feared event and works backwards to identify possible causes in the form of a cause tree.

Hypothesis tree

The starting point is an objective to be reached. We will list hypotheses to reach this objective and list experiments to validate these hypotheses (the branches of the tree).

Attributes

There are two types of data. CONTINUOUS and DISCRETE data. Discrete data are qualified by attributes that can take 2 forms (Good / Bad, 0 / 1 , OK / NOK...) or more than 2 (Good, Average, Bad).

Self-control

Summarizes all the methods and tools designed to enable the operator to guarantee - independently - the control of safety, quality, deadlines and costs, directly at his workstation.

Baseline

Baseline can be translated as "reference point". The aim is to know the initial capability of the process before making any improvements, in order to compare the situation BEFORE and AFTER the Lean Six Sigma project.

Benchmarking

A method originating in marketing which consists in comparing one's way of doing things or one's performance with that of the best companies in the same sector, or with that of companies with a recognized performance in the same field of expertise.

Beta Beta risk Second-species risk

Beta risk is the risk of concluding that there is no difference between two samples, when in fact there is.

Binomial

The binomial distribution applies to discrete data. It calculates the probability of drawing k characters X from a sample of size n drawn from a batch whose probability of having X is equal to p.

Black Belt (BB)

A "Black Belt" is a person trained in Six Sigma methodology whose main role is to lead improvement projects using Six Sigma methodology and tools.

BoxPlot moustache box

A graphical representation of statistical data. The box plot (also called box diagram, Tukey box or box plot) gives an idea of the distribution of values in the form of a box (50% of the population) and whiskers (2 lines covering the whole population).

Brainstorming (unpacking ideas)

Brainstorming is a process that enables the rapid generation of a large volume of ideas, in an atmosphere free from criticism and judgment. First, the brainstormer poses the question to be solved. He or she then stimulates the generation of ideas, without acting as a judge. Once all the ideas have been presented, they are reread with the group to: check the clarity of the ideas, check whether they answer the question, eliminate duplications.

Capability

Process capability is the measure of a process's ability to meet customer expectations. It is the ratio between required and actual process performance.

Capability - Attributes

In the case of attributive criteria, capability is calculated with an equivalent Ppk which gives the same proportion of defects as in the case of measurements.

Measurement process capability

It validates the possibility of using a measurement process for a characteristic.

Control Board

A control chart can be used to visually monitor and control a process to identify drifts and ensure its stability over time.

Multivariate map T²

A multi-dimensional control chart to study whether a set of several characteristics remains under control.
An indicator T² is calculated, which is the distance between each point and the center of the point cloud. If the distribution follows a multidimensional normal distribution, we can calculate a control limit on this T² distance.

Value Stream Mapping (VSM)

Value stream mapping is a graphical representation of the flow of materials and the flow of information.
We generally make two maps:

• Current value stream mapping: this is a graphical representation of the current flow of materials and information.
• Ideal value stream mapping: a graphic representation of the flow of materials and information as they might ideally be.

Process mapping

Process mapping is a graphical representation of process steps (in chronological order) that clearly shows how process inputs are transformed into outputs for the customer.

Functional process mapping (Swimlane mapping)

Functional process mapping enables you to visualize process steps across the company's departments or businesses. (Each department has its own corridor)

Root cause

That's the origin of the problem... in our possible area of intervention.

Special cause

Source of irregular variation difficult to predict. For example, machine misalignment, wear or breakage.

Common causes

Common causes form the intrinsic variability of the process. They are due to a sum of small causes that satisfy the central limit theorem. When only common causes remain, the distribution of values follows a Gaussian distribution.

Champion

The champion is an important person in the Lean Six Sigma organization, as he or she is in charge of deploying the method throughout the company.

Project Charter

The project charter is a key element in the construction phase of a project.
It summarizes the problem to be solved, the objective to be achieved and the scope of the project. It will also define the project's roles and responsibilities, as well as the main time milestones. The Six Sigma method rightly stresses the importance of this document, and of the commitment of each project member to seeing the project through to a successful conclusion.

Correlation coefficient

Numerical value characterizing the relationship - meaning and importance - between two random variables or two statistical variables.
This value can be positive, negative or zero. Note that the value of the coefficient is zero when the two variables in question are independent.

Determination coefficient

The coefficient of determination (R², i.e. the square of the linear correlation coefficient r) is an indicator of the quality of a simple linear regression. It measures how well the model fits the observed data, or how well the regression equation describes the distribution of points.
By abuse of language, it is often called the correlation coefficient.

Collecting data

Part of the MEASURE step. The data collection plan is organized to maximize the chances of success in the ANALYZE stage.
A data collection plan is prepared by asking the following questions:

• WHO is concerned?
• WHAT to collect
• WHERE to collect?
• HOW to organize your collection?
• WHEN? When? How often?
• HOW MUCH? What sample size?

Short term

Period short enough to consider that the process is under control

Cp

Short-term capability indicator for a process, regardless of its focus. Six sigma requires a Cp > 2.

Cpc

Control process capability. It indicates whether the control process is suitable in terms of repeatability and reproducibility.
Cpc is equivalent to R&R (Cpc = 1/R&R).
In general, we consider :

• Cpc ≥ 4 (R&R% ≤25%), the process is able to
• Cpc ≥ 3 (R&R% ≤33%), the process is borderline
• Cpc 3), the process is not suitable

Cpk

Short-term capability indicator for a process that takes into account its centering.

Cpm

Short-term capability indicator based on Taguchi loss.

CTC

CTC stands for "Critical to Cost". Cost-related CTS are called CTC.

CTD

CTD stands for "Critical to Delivery".
CTS linked to the product or service lead time are called CTD.

CTQ

CTQ stands for "Critical to Quality".
CTS related to product or service Quality are called CTQ.

CTS (CTQ, CTC, CTD)

CTS stands for "Critical to Satisfaction".
CTS are the main measurable characteristics of a service or product whose specification limits must be respected to satisfy the customer. CTS are defined according to the Voice of the Customer (VOC).

Operational definition (of data)

Precise description indicating how to obtain a value for a characteristic to be measured: it describes the characteristic, and indicates how to measure it.

DFSS Design For Six Sigma

It's a best-practice methodology for designing a product, process or service that directly achieves the Six Sigma objective. DFSS follows the DMADV approach.

Vertical strip chart

Vertical bars indicate the evolution of a variable over time.

Stacked bar chart

Used to represent a matrix of two discrete variables.

Circular diagram

Used to represent the distribution of the modalities of a discrete variable.

Affinity diagram (also called KJ diagram)

It's a basic quality tool. It is used, for example, to clarify and synthesize the results of a brainstorming session. All ideas are grouped by affinity. Each grouping is given a title.
The workgroup's ability to bring ideas together enables team members to take ownership of them. It encourages creativity and the search for innovative links between ideas.
The 5 M diagram is in fact an affinity diagram for which the grouping headings are predefined.

Point diagram

A dot plot divides the sample values into small intervals and represents each value or small group of values by a point along a number line. The dot plot is suitable when the sample size is less than around 50.

Scatter plot (point cloud)

Used to visualize a possible correlation between two quantitative variables.

Process management diagram

The process management diagram documents the mapping of the process, the measures to be controlled at each essential stage of the process, and the plan for responses and actions when the process does not remain within operational limits.

Kano diagram

Diagram for classifying attributes along two axes:

• the axis of realization from the unprocessed attribute to the fully processed attribute
• the satisfaction axis from dissatisfied to very satisfied. Attributes are classified into 3 categories:

1. Obligatory, indispensable - (Must be)
2. Attractive - (attractive)
3. Proportional or linear - (One-Dimensional)

We have also added 2 new categories:

1. Indifferent - (Zone of indifference)
2. A double tranchant - (Reverse)

Time series diagram

A time series is a series of values for the same variable observed at regularly spaced points in time (day, month, year).
The vertical bar chart or control chart are examples of time series diagrams.
In a time series diagram, the X axis represents time.

5M diagram or cause/effect diagram or Ishikawa diagram or fishbone diagram

This multi-named tool is a workshop facilitator's tool for finding causes. It helps to generate a list of ideas during a creativity session by structuring thinking around the 5Ms: Manpower, Material, Methods, Machines, Environment.

Tree diagram

The tree diagram is a tool for finding ways to improve a situation, and for planning the sequence of steps involved in implementing solutions, while identifying key points to be observed for follow-up.

Bar chart

Used to represent the distribution of the modalities of a discrete variable.

Spaghetti diagram

The spaghetti diagram is a tool used to provide a clear view of the flow of parts or individuals. It plots all the paths traveled on a map. It takes its name from its resemblance to a dish of spaghetti. This visualization makes it possible to identify redundant flows and recurring crossings, and to measure the distance traveled by each product or person.

Stem & leaf diagram

A diagram that simultaneously strips the data from the distribution of a statistical variable and produces a graphical representation in the form of a histogram.
Each individual item of data is represented by its stem (first digits common to several items of data) and its leaf (last digits of the same items of data).

DMADV

Define, Measure, Analyse, Design, Validate

DMAIC

Define Measure Analyze Innovate Control (Define, Measure, Analyze, Improve, Control). Six sigma project management method.

Continuous data

Data represented by a real number, which can (theoretically) take on an infinite number of values.

Discrete data

Data that can take on a finite number of values. Example: Machine A, B and C

Ordinal discrete data

Discrete data that can be ordered. Example: Good Average Bad

DPMO

Defect Per Million Opportunities. Unit used in Lean Six Sigma to define the capability of a process. DPMO = [Number of defects X 1,000,000] / [Number of parts X Number of opportunities].

Henry's right

The Henry line is a graphical method for fitting a Gaussian distribution to that of a series of observations (of a continuous numerical variable). When fitted, the points line up on a straight line. It can be used to quickly read the mean and standard deviation of such a distribution.

Standard deviation

Standard deviation is a measure of the dispersion of data in relation to the mean. It is calculated as the square root of the mean of the deviations from the mean squared.
Standard deviation makes sense even when normality is not demonstrated.

ECC (Critical Customer Requirements)

These are the needs that must be met, or customer dissatisfaction will result. ECCs must be measurable.

Sample

It's a subgroup of a population.

Sampling

This involves taking a sample from a population.
There are several ways of sampling: random, stratified, cluster or systematic.

Random sampling

Each element in this sample has the same probability of being selected as all the other elements in the target population.

Cluster sampling

This involves subdividing a homogeneous population into clusters (subgroups), then randomly selecting clusters and considering all the elements in each cluster. Example: high school classes.
We take all the elements of a few randomly chosen classes.

Stratified sampling (quota method)

This involves subdividing a heterogeneous population into strata (subgroups). This method consists in finding the same proportions in the sample for each stratum, according to the characteristics chosen for the study in the target population.

Systematic sampling

This method consists of listing all the elements of the target population and determining the following ratio :
(Number of population elements)/(Sample size)
Example: A telephone directory contains 4,000 names. I want a sample of 200 individuals. So I calculate 4,000/200 = 20. Then, starting from the beginning of the directory, I select the 20th, the 40th, the 60th, always in steps of 20.

Systematic sampling

This method consists of listing all the elements of the target population and determining the following ratio :
(Number of population elements)/(Sample size)
Example: A telephone directory contains 4,000 names. I want a sample of 200 individuals. So I calculate 4,000/200 = 20. Then, starting from the beginning of the directory, I select the 20th, the 40th, the 60th, always in steps of 20.

Measurement error

No measurement system is perfect and therefore introduces a difference between the actual value and the measurement. This difference is the measurement error. There are several types of error (Bias, Linearity, Stability, Repeatability, Reproducibility).

Scope

It's a dispersion indicator. The range is calculated as the difference between the maximum value minus the minimum value.

Influential factors

Influential factors are those that have a direct impact and influence on the Ys in the process. They are often the cause. Hence the importance of identifying them (Analysis Stage) and dealing with them (Innovation Stage).

Flow chart

A flowchart is a type of diagram that represents a workflow or process. A flowchart can also be defined as a schematic representation of an algorithm, a step-by-step approach to solving a task.
The flowchart shows the steps in the form of boxes of various kinds, and their order by linking the boxes with arrows. Flowcharts are used to analyze, design, document or manage a process or program in various fields.

Green Belt

A Green Belt is a person trained in DMAIC methodology whose part-time role is to lead process improvement projects.
Green Belts continue to report to operational managers in departments or business lines.

GRR% - Gage R&R%

Control process capability. It indicates whether the control process is suitable in terms of Repeatability and Reproducibility.
Cpc is equivalent to R&R (Cpc = 1/R&R).
In general, we consider :

• Cpc ≥ 4 (R&R% ≤25%), the process is able to
• Cpc ≥ 3 (R&R% ≤33%), the process is borderline
• Cpc 3), the process is not suitable

GTV (Time-value graph)

Graphical representation of Value-Added (VA) and non-VA tasks on a workstation or activity.

Histogram

A histogram is a graphical representation of the distribution of a continuous variable using vertical columns.

Hypergeometric

The hypergeometric law applies to discrete data. It is used to calculate the probability of drawing k characters X from a sample of size n drawn from a batch of size N with D characters X.

Hypothesis H0

Assumption that the difference between samples is zero.

Hypothesis H1

Assumption that the difference between samples is non-zero.

Confidence interval

Confidence Interval. The confidence interval is a deviation calculated around the mean to guarantee, with a certain degree of confidence, the difference between the true value of the population and the sample.

Accuracy (Bias)

Systematic deviation between true value and displayed value.
Measuring Device Management guarantees the accuracy of the device, not the process.

KPIS

Key performance Indicators. The translation is disputed because the meaning of performance in the two languages does not correspond perfectly. Sometimes we speak of Key Process Indicators.
These key indicators summarize the achievement of project objectives. They will be identified as early as the Define phase. They will form part of the project dashboard and will be updated regularly.

Control limit

Limits to random variations in the observed statistic, which may be an individual value, a mean, a median, a range, a standard deviation, a number of non-conformities, or a proportion.
The limits are calculated to within ± 3 standard deviations of the distribution of the statistic under study. As long as the statistic is within the limits, the hypothesis of a deviation due to common causes is not rejected; the process is under control.

Lower Control Limit (LCI)

When a point is below the ICL, we reject the hypothesis of a deviation due to common causes. The process is out of control.

Upper Control Limit (UCL)

When a point is beyond the ICL, we reject the hypothesis of a deviation due to common causes. We are in the presence of a special cause; the process is out of control.

Acceptable Quality Limit AQL

This is the maximum limit of non-quality accepted in a Customer/Supplier relationship.
Above this limit, the alpha risk (of being refused a batch of quality p) is greater than 5%.

Specification limits (Tolerances)

A Specification Limit is a value that defines the acceptability of the performance of a service, product or process from the customer's perspective. There is often an Upper Specification Limit and a Lower Specification Limit.

Lower specification limit (LSL)

A lower specification limit is a value above which the performance of a service, product or process is acceptable. Specifications are defined by the customer, based on what is needed to satisfy requirements.

Upper Specification Limit (USL)

An upper specification limit is a value below which the performance of a service, product or process is acceptable.

Linearity (assumption)

This assumption is particularly true of 2k experimental designs. This assumption should be verified, for example, by carrying out a test in the center.

Linearity study

Study to check the linearity of a measurement process in its operating range.

Flow chart

A flowchart is an analysis tool that provides an ordered, sequential representation of all the tasks or events involved in a given activity. It is made up of a set of symbols linked by arrows.
Each symbol represents an event or task, and the arrow represents the relationship of precedence or succession between two consecutive tasks.

Binomial law

The binomial distribution applies to discrete data.
It calculates the probability of drawing k characters X, in a sample of size n, drawn from a batch whose probability of having X is equal to p.

Normal Law - Gauss Laplace Law

The normal distribution is a symmetrical distribution characterized by a mean and a standard deviation. When a process satisfies the central limit theorem, the distribution follows a Gaussian distribution.

Long term

Normal process production period.

Statistical Process Control SPC

Statistical Process Control. Regroups methods and tools for controlling a characteristic using control charts and capabilities.

Visual management

Visual Management is first and foremost a management tool that makes performance levels and deviations from targets clearly visible.
It provides useful information for operators and project members.
It makes deviations from the target obvious.
It enables day-to-day performance management (anticipation of results).
It allows you to manage resources: who does what and when?
It facilitates the organization of meetings around a visual communication board.

Master Black Belt (MBB)

A Master Black Belt (MBB) is an expert in Lean Six Sigma methodologies, specially trained to coach and mentor process improvement and design teams. This person is not part of any specific project team, but rather serves as an expert resource for several teams. The MBB participates in project reviews, trains Black Belts and Green Belts, and ensures the proper use of Lean Six Sigma methodology and tools.

Choice matrix - Solution selection matrix

It's a process of synthesizing and selecting solutions.
Selection items are identified and weighted.
Each solution is given a score from 1 to 10, and a weighted score is calculated for each solution.

Prioritization matrix - Eisenhower matrix

The prioritization matrix is used to rank actions according to two criteria (e.g.: Chance of success/Time to implement), which are placed on the X and Y axes respectively. We place (for example) the actions in the graph, which we divide into 4 quadrants. Actions that are quick to implement and have a high chance of success are favored over those that take a long time to implement and have little chance of success.

Pugh matrix

It's a process of synthesizing and selecting solutions.
It enables us to evaluate different solutions that will be the easiest to implement, the least costly, the most visible, the quickest to deliver the best results, the best return on investment, the least resistance to change...
Judgment is not absolute, but relative to a reference solution. A weighting is assigned to each criterion, and the matrix gives a "score" for each solution.

X/Y matrix - C&E matrix

Cause and Effect Matrix or X/Y Matrix
It enables us to relate several Xs to several weighted Ys. For example, by rating the impact of Xs on each Y from 1 to 10, we can rank the Xs in order to focus on the essential Xs.
The C&E Matrix is a bit like the Causes/Effects diagram, which is why we prefer to call it the X/Y Matrix.

Median

Value such that 50% of the population are on either side. It characterizes the central position of a population. It is less sensitive to outliers than the mean.

Taguchi method

The Taguchi method is an original approach to making Ys insensitive to noise factors, using a design of experiments approach.

MLG (GLM)

Generalized Linear Model.
This is an extension of the least-squares regression method to suit discrete Y variables. The optimization algorithm to obtain the maximum likelihood of the parameters is no longer a simple least-squares calculation, but an iterative process.

Mode

In the case of a discrete variable, this is the value with the maximum number of members.
In the case of a histogram, this is the class with the largest number of values.

Average

Sum of values divided by the number of values.
It characterizes the central position of a population.

MSA

Acronym for Measurement System Analysis. It is a benchmark standard for measurement process control.

MSP (SPC)

Statistical Process Control. Regroups methods and tools for controlling a characteristic using control charts and capabilities.

Acceptable Quality Level - AQL

This is the old name for the Acceptable Quality Limit.

Non Added Value - NVA

An activity that requires time, resources or space, but does not add value to the product itself. The activity may be necessary under current conditions, but from the customer's point of view, it adds no value to the product, but only production costs.

Opportunity

This is a characteristic measured on a unit that must comply with customer specifications. It corresponds to the "possibility" of not meeting specifications.

Pareto

A Pareto chart is a plotting tool that graphically illustrates the "Pareto Principle" or "80/20 Rule", which postulates that 80% of problems result from 20% of causes, the "major causes". Pareto graphs rank a list of causes according to an order of priority that depends on the frequency of occurrence.

Stakeholders

All persons impacted by the project, i.e. :

• the people involved in the project.
• people interested in the project.
• people who can influence the project.
  Stakeholder analysis is an important part of the DEFINE stage.

Plan Do Check Act - PDCA

Problem-solving methodology to be used when the solution is known.
Plan: define scope, obtain data, formulate hypotheses, draw up test program.
Do: implement, train, inform.
Check: verify results, find a solution quickly, identify root causes.
Act: prevent recurrence, define standards, communicate, identify new improvements.

Process driver (or process owner)

The Process Manager is the person responsible for all aspects of process execution. He/she is also responsible for continuous improvement of performance and control.

Design of experiments (DOE)

DOE, Design Of Experiment.
A design of experiments (DOE) is a set of tests planned in relation to a given objective. The aim is to carry out the right number of trials to adapt to the desired model, while maximizing the precision of the results.

Data collection plan

Part of the MEASURE step.
The data collection plan is organized to maximize the chances of success in the ANALYZE stage.
A data collection plan is prepared by asking the following questions:

• WHO is concerned?
• WHAT to collect?
• WHERE to collect?
• HOW to organize your collection?
• WHEN? When? How often?
• HOW MUCH? What sample size?

Clearing plan - Screening plan

Design of experiments to prioritize Xs. In this type of design, we're not looking for a predictive model, but simply for the Xs that have the greatest impact on the Y(s).
Taguchi's L12, L18 and L20 tables are very well adapted. So do the 12- and 20-trial designs by Plackett and Burman. In fact, these are the same tables.

Sampling plan

Reasoned selection of a representative subset of the data relating to a process. It enables us to arrive at accurate conclusions (at known alpha and beta risks for two reference points) from a relatively small sample.

Taguchi Design

Taguchi has proposed an original and simple organization for experimental designs. This makes the use of experimental designs easily accessible to non-statisticians.
The Taguchi method, on the other hand, is an original approach to making Ys insensitive to noise factors, using a design of experiments approach.

Response Surface Design of Experiments

These designs allow us to study quadratic polynomial models (with x² terms), but with fewer trials than the 3k design.
They are designed to minimize the confidence interval on the coefficients. The most famous are centered composite designs.

Factorial experimental designs

These are plans for which the factors are fixed either on modalities (discrete X criteria) or on levels (continuous criteria). All combinations (complete plans) or only part of the combinations (fractional plans) are realized.

Fish

Poisson's law applies to discrete data.
It calculates the probability of drawing k characters X, in a sample, knowing that on average, lamdba (λ) will be drawn.

Poka-Yoke

Anti-error system that eliminates the possibility of making a mistake. Example: notch on a smartphone SIM card.

Population

Statistical term representing all individuals.

Power to discriminate

Ratio between measurement dispersion and dispersion of measured products.

Pp

Long-term process capability indicator, independent of process focus

Ppk

Long-term capability indicator for a process, taking into account its centering. In general, a Ppk > 1.33 is required.

PPM

Acronym for Part Per Million. Typically used in the context of Defects Per Million Opportunity. Synonymous with DPMO.

Ppm

Long-term capability indicator based on Taguchi loss.

Problem

A problem is defined by a gap between a current situation and a desired situation.

Process

A process is a method, a technique, a way of acting.

Procedure (or Operating Mode)

A procedure is the documented sequence of steps and other instructions required to perform an activity.

Process

A process is a series of steps that transform inputs into outputs (product or service) to meet a customer's STC.
Any activity can be described in terms of a process. The ultimate aim of a process is to add value for the customer. A process generally cuts across several departments or functions.

Customer process

A set of activities that a company must carry out to transform customer demand into products or services that meet their requirements.

Management processes

A set of activities designed to establish and deploy an organization's guidelines, control and correct its activities, and analyze and improve its operations.

Support processes

Support function processes that contribute to the smooth running of customer processes.

Processes, Activities, Tasks

Process: a process is made up of a set of activities.
Activity (or operation): an activity is made up of a set of tasks.
Task: basic elements of a process.

Process owner (or Process Driver)

The Process Owner is the person responsible for all aspects of process execution. He/she is also responsible for continuous performance improvement and control.

Measurement protocol

Detailed description of the measurement of an X or Y variable, generally based on QQOQCCP questioning, enabling all aspects to be anticipated.

Power

The power of a test is the probability of declaring a difference as significant.
The power curve is plotted by varying the deviation on the X axis and placing the power on the Y axis.
The power of a test determines the sample size required.
The power of a test is the complement of the beta risk P = 1 - Beta.

Q1

This is the first quartile. 25% of data are below this value.

Q2

This is the median. This is the value where 50% of points are below and 50% of points are above.

Q3

This is the third quartile. 75% of data are below this value.

QQOQCP

QQOQCP is an empirical questioning method. Indeed, any analysis involves a preliminary phase of "systematic and exhaustive questioning", the quality of which determines the quality of the analysis itself.
So we question WHO? WHAT? WHERE? WHEN? HOW ? WHY ?
For each item, we try to be as factual as possible by answering the COMBIEN question.
This method is used to correctly pose a problem, but also to structure the presentation of the results of their analyses, or to create a data collection plan.

QRQC - Quick Response - Quality Control

QRQC is a hands-on approach to problem-solving.
Above all, QRQC requires us to base ourselves solely on facts (the real world first and foremost), and not on assumptions made far from the field, which very rarely describe reality.
Unlike "in-room" problem-solving methods, QRQC treats the problem as it arises, by directly observing anomalies, analyzing objective measurements, and involving the people involved.

R

Correlation coefficient (varies between -1 and +1)

R&R - Repeatability & Reproducibility

Control process capability. It indicates whether the control process is suitable in terms of repeatability and reproducibility.
Cpc is equivalent to R&R (Cpc = 1/R&R).
In general, we consider :

• Cpc ≥ 4 (R&R% ≤25%), the process is able to
• Cpc ≥ 3 (R&R% ≤33%), the process is borderline
• Cpc 3), the process is not suitable

R&R attributes

Control process capability by attribute. It indicates whether the control process is suitable in terms of repeatability and reproducibility.
In this test, between 20 and 30 parts are taken, previously checked by an expert (deemed true value). Each part is checked several times (repeatability) by several operators (reproducibility).

R²

Coefficient of determination (sometimes called correlation coefficient).
It varies between 0 and 1.

RACI/RASCI

RACI in management represents a matrix of responsibilities, indicating the roles and responsibilities of those involved in each process and activity. This matrix represents the organization of work by linking in a common table the Project Breakdown Structure (WBS) and the Project Organizational Structure (OBS).
The RACI matrix provides a simple, clear view of who does what in the project, helping to avoid redundant roles or diluted responsibilities. For example, approval responsibility ("A") should be assigned to a single person within an activity, while several people may be responsible ("R") for its execution. There should be at least one "R" per activity. In most cases, the person approving the activity ("A") is the superior of the person carrying it out ("R").
The French translation could therefore be :

1. R: Director
2. A: Authority or person in charge
3. C: Consulted
4. I: Informed
   RASCI is the same thing, but we add:
5. S: Support (people or authorities who can provide support. It provides R with resources).

Repeatability

Dispersion of a measurement under stable conditions for the 5 M's of the measurement process: Measurand, Means, Method, Medium, Manpower.

Reproducibility

Dispersion of a measurement when one of the 5 M's of the measurement process is modified. This is often Manpower, but it can also be Means, Method or Environment.

Resolution

This is the reading granularity of a measuring device. In concrete terms, it's the number of digits after the decimal point that can be read on the measuring instrument.
Insufficient resolution: the resolution must be equal to one tenth of the tolerance!
Example 10 ±0.02 -> tolerance 0.04 mm -> resolution 4 mm

Project review

This is a structured meeting to ensure that the project has delivered the deliverables required for each DMAIC phase (project charter, capability analysis, etc.).

Process review

Reviews enable us to monitor performance and ensure that process outputs are in line with objectives and the expectations of process customers. Process reviews enable us to identify dysfunctions and propose actions likely to improve the effectiveness and efficiency of the process. Process reviews are conducted by the process pilot or owner.

Project risk

Every project has its risks, which need to be identified as early as possible. TOHE risks can be classified as Technical, Organizational, Human and Economic.
Several tools are available to analyze these risks.
Examples include the SWOT diagram (MOFF), the FMEA, the cause tree...

RTY - Rolled throughput yield

The RTY is calculated by multiplying the yields of each process step.
Calculations can become increasingly complicated as more parallel processes are introduced.
RTY example for serial processes :
RTY = yield of process step 1 * yield of process step 2 * ... * yield of process N

RTYL

Rolled throughput yield Loss -> RTYL = 1-RTY

Chronological series

A time series is a series of values for the same variable observed at regularly spaced points in time (day, month, year).
The vertical bar chart or control chart are examples of time series diagrams.
In a time-series diagram, the X axis represents time.

Sigma

The process Sigma value (denoted z) is a process performance indicator.
This corresponds to the number of standard deviations that can be inserted between the mean and the tolerance.

SIPOC

SIPOC stands for Suppliers, Input, Process, Output, Customers.
The SIPOC is a high-level description (helicopter view or moon view) of the process.

Six Sigma

It's a 5-step DMAIC (Define, Measure, Analyze, Innovate, Control) problem-solving method for reducing process variability to improve customer satisfaction.

SMART

Characteristics of good lenses :

1. Specific (in their definition)
2. Measurable (with objectivity)
3. Ambitious / Achievable
4. Realists
5. Defined in time

SOP

A standard operating procedure (SOP) is a set of step-by-step instructions compiled by an organization to help workers perform routine operations. SOPs aim to achieve efficiency, output quality and consistency of performance, while reducing communication problems and non-compliance with industry regulations.
SOPs are the supports of Self-Control.

Output

An output is the result (product or service) of a transformation carried out by a process. Outputs are the results of processes delivered to customers.

Technical specifications

Precise indication of a set of conditions to be met by a product, material or process, including, if necessary, the methods for determining whether these conditions are met.

Standard

A standard is an operating procedure defining best practices.

Descriptive statistics

Descriptive statistics are statistical tools used to describe the behavior of a continuous or discrete variable X or Y.

Inferential statistics

Inferential statistics are statistical tools used to provide evidence of a non-random relationship between a Y and one or more Xs.

Stratification

Principle of dividing data into different subgroups. For example: by product family, by machine, by failure type...

SWOT (MOFF)

( Strengths - Weaknesses - Opportunities - Threats ) ou MOFF pour les Francophones ( Menaces - Opportunités - Forces - Faiblesses )
This is a very practical tool for the DEFINE phase. It has the advantage of summarizing the project's strengths and weaknesses in relation to the opportunities and threats generated by its environment.

Contingency table

Table used to count the distribution of two discrete variables with two or more modalities.

Task

A task is an action that is carried out, usually individually assigned, and must be completed within a given timeframe.

Test 1p - Proportion test

Compares a proportion with a theoretical proportion.

Normality test

Used to test whether the normality hypothesis can be accepted, the main tests are :

• Testing Anderson Darling
• Kolmogorov Smirnov test
• Chi2 adequacy test
• Testing Ryan Joiner

Chi-square test

There are two chi² tests

1. It can be used to compare a variance with a theoretical variance.
2. It compares the distribution of two discrete variables with 2 or more modalities (contingency table).

Test F

Compares two Variances.

Test Kappa

It's an R&R element with attributes.
The Kappa value is used to measure the correlation between several operators or between an operator and the expert.

Paired t-test

Used to compare two averages in the case of paired samples.

Equivalence tests

Statistical tools used to assess, with a certain degree of risk, whether 2 or more populations are equivalent.

Hypothesis testing

Statistical tools used to assess, with a certain degree of risk, whether there is a difference between 2 or more populations.

Two-sample t-test

Compares two averages.

One sample t-test - Test t-théorique - One sample test t

Compares an average with a theoretical value.

Central limit theorem

Theorem stated by Gauss.
Any system subject to numerous factors, independent of each other and of an order of magnitude of equivalent effect, generates a normal distribution.

TOHE

Technical Organizational Human Economic
A tool for analyzing project risks by classifying them into 4 categories:
T: All technical risks - example: machine capacity.
O: All organization-related risks - e.g. machine availability.
H: All human-related risks - example: available skills.
E: All risks related to the economic side - e.g. insufficient budget.

Transformation Box - Cox

It is a mathematical transformation (xλ) that transforms data that do not follow a normal distribution into data that do follow a normal distribution. This transformation makes it possible to use standard analysis tools such as control charts, statistical tests, etc., which assume the normality of the data.

Outlier

This is data that is abnormally distant from the others. It does not follow the normal distribution.

Variance

A variance is a measure used to characterize the dispersion of a sample or population. Variance is equal to standard deviation squared.

VOB - Voice of Business

Voice of the Business.
It's the voice of the company: what are the objectives in terms of costs, deadlines, return on investment...

VOC - Voice of the Customer

Voice of the Customers.
It's the voice of the customer. It clarifies the expectations of both customers and project stakeholders.
It is gathered using all available proactive means (interviews, surveys, polls, exchange groups, etc.) and analysis methods (analysis of historical data, complaints, etc.).

VOP - Voice of the Process

Voice of the Process.
The Voice of the Process (VOP) refers to the various pieces of information used to measure process performance.

VOS - Voix du Social

Voice of the Social.
It's the voice of the Environment, Safety and Employees.

VSM

Value Stream Mapping.
It's a visual tool that displays all the critical steps in a specific process.
It easily quantifies the time and volume required for each stage. Value stream maps show the flow of materials and information as they move through the process.

X

X is the factor that influences the output of a process.
Factors that have a major impact on the performance of a process output are referred to as "critical Xs" or "essential Xs".

Y

Y is the measure of a process output

Z

Called "process Sigma value". This is the number of standard deviations that can be inserted between the mean and the tolerance.