Arjan Imeri Sacha Didi

Business Analytics
Table des matières
1: Introduction to business analytics and motivation..........................................................1
Business analytics definition:......................................................................................................1
Descriptive, diagnostic, predictive, prescriptive analytics............................................................1
...........................................................................................................................................1
2 : Informations supplémentaires provenant des TPs..........................................................2
Pareto Chart :..............................................................................................................................2
Histogram:..................................................................................................................................2
Run Chart:...................................................................................................................................2
A/B testing:.................................................................................................................................2
Cause-and-effect diagram (Fish bone diagram)............................................................................3
Flow Chart:.................................................................................................................................3
3: Need for business improvement.....................................................................................4
Data driven decision making:......................................................................................................4
Business improvement:...............................................................................................................4
Doing activity = business process..................................................................................................................4
Improving activity= improvement activity....................................................................................................4
Process improvement...................................................................................................................................4
Problem solving:...........................................................................................................................................4
Data:............................................................................................................................................................. 4

4: Statistical thinking..........................................................................................................5
Definition of statistical thinking:.................................................................................................5
Steps in implementing statistical thinking...................................................................................5
Average, variation, and target.....................................................................................................6
Traditional quality and quality improvement (statistical thinking)...............................................6
Traditional quality:........................................................................................................................................6
Quality improvement:...................................................................................................................................6
Good quality does not mean perfect............................................................................................................6
Consistent performance...............................................................................................................................6
Data-driven decision making:......................................................................................................7
...........................................................................................................................................7
Improvement approach of statistical thinking:.............................................................................................7

4.1: Principles of statistical thinking....................................................................................7

3 principles:................................................................................................................................7
First principle: (process thinking or system thinking)...................................................................7
SIPOC:...........................................................................................................................................................7
Blame the process, not the people...............................................................................................................8

1
Arjan Imeri Sacha Didi

Process measurement:.................................................................................................................................8
Operational definition:.................................................................................................................................8
Second principle: (Variation exists in all process)........................................................................8
Variation:......................................................................................................................................................8
No variation:.................................................................................................................................................8
Unwanted variation:.....................................................................................................................................9
Wanted variation :........................................................................................................................................9
Third principle: (understanding and reducing variation are key to success)..................................9
Consistency: Il...............................................................................................................................................9
Two types of unintended variation in business process:............................................................................10
Three key business analytics questions......................................................................................11
Writing an objective improvement statement...........................................................................11
SMART: help to write objective statements...............................................................................................11
....................................................................................................................................................................11
Deming’s red bead experiment.................................................................................................11
4.2 : Statistical thinking strategy......................................................................................12
5 steps to improve the performance of a company:...................................................................12
Synergy: Data and subject matter knowledge............................................................................12
Statistical thinking can be applied to improve business processes.............................................13
Relationship to the scientific method........................................................................................13
Definition:...................................................................................................................................................13
Two analytics’ paradigms:...........................................................................................................................14
‘Plan–Do–Check–Act’ (PDCA) cycle............................................................................................14
Definition PDCA..........................................................................................................................................14

5: Statistical engineering..................................................................................................15
Why using statistical engineering ?............................................................................................15
Definition:.................................................................................................................................15
Principle of statistical thinking..................................................................................................15
Exemple important à lire:..........................................................................................................15
6: THREE types of variation in business processes.............................................................16
1) Special cause variation: Unpredictable (unstable).............................................................16
Eliminating special causes variation to stabilize the process:.....................................................................16
2) Common cause variation: Predictable (stable)...................................................................16
Benefits of stable processes (only common cause variation):....................................................................16
Le off-target variation...............................................................................................................16
Type of action for reducing variations.......................................................................................17
3) Structural variation...........................................................................................................17
Comparaison between special, structural and common causes:.................................................17
Robustness: an underused concept (un concept sous-exploité).................................................17
Process and product robustness.................................................................................................................17
Management robustness............................................................................................................................18

7: Process stability and capability tools.............................................................................18

2
Arjan Imeri Sacha Didi

Introduction and motivation.....................................................................................................18

Stability:......................................................................................................................................................18
Capability:...................................................................................................................................................18
Capability and stability................................................................................................................................19
Statistical inference, accuracy and precision..............................................................................19
Control charts to evaluate stability............................................................................................19
Introduction to control charts.....................................................................................................................19
Run tests/rules of thumb:.........................................................................................................20
Phase 1 and 2 of control charting:.............................................................................................20
8: Control charts...............................................................................................................21
Choice of control limits.............................................................................................................21
Link between control charts and statistical hypothesis:.............................................................22
A general model for a control chart...........................................................................................22
μ and σ are known for a normal distribution..............................................................................................22
Three sigma control limits:.........................................................................................................................23
μ and σ (from the population) are unknown for a sample:........................................................................24
Types of control charts..............................................................................................................25
Control charts for discrete data (attributes)...............................................................................26
1: Defect datas............................................................................................................................................26
2 : Defective datas......................................................................................................................................27
Control charts for continuous data (variables)...........................................................................................28
Flow chart for selecting the appropriate control chart:..............................................................................31
Control chart constants and formulas:.......................................................................................................32

9: Process capability and capable processes......................................................................33

Capability:.................................................................................................................................33
Capable :...................................................................................................................................33
Process capability analysis :.......................................................................................................33
Process stability analysis:..........................................................................................................33
Specifications/targets/business requirements:..........................................................................33
“Voice of the customer”:............................................................................................................................34
“Voice of the process”:...............................................................................................................................34
“Do not confuse control with capability”....................................................................................................34
“Specification and control limits”..............................................................................................35
Process capability ratios............................................................................................................36
Calculation: Cp:.........................................................................................................................36
Cpk:..........................................................................................................................................37
Confidence intervals for process capability ratios......................................................................................38
Deming’s funnel experiment.....................................................................................................38
10: Overall business process improvement strategy..........................................................39
Overall improvement process strategy: Steps and tools:...........................................................40
Overall improvement process strategy: Steps and tools:...........................................................43
Process control charts and continuous improvement................................................................44

3
Arjan Imeri Sacha Didi

1: Introduction to business analytics and motivation

Business analytics definition:
- Business analytics refers to the methodology employed by an organisation to
enhance its business and make optimised decisions based on data, and using
‘statistical thinking’ to improve, for example, their products, services, supply chain
and operations, human resources, financial management and marketing.

Descriptive, diagnostic, predictive, prescriptive analytics

Descriptive:
- 'What happened?' or 'what is happening now?'
o Which customers are buying what products and when they are buying these
products?
 (e.g., during the day, weekday, season)?
Diagnostic:
- ‘Why did it happen?’ or ‘why is it happening?’ or ‘what is happening?’ or ‘what are
the trends?’ or ‘what patterns are there?’
Predictive:
- ‘What will happen?’
Prescriptive:
- ‘What to do?’ or ‘how to make it happen?’ or ‘what is the best that could happen?’ or
‘how to optimise what happens?’
o Which potential customers must be contacted with what product and which
media at what time point?

2 : Informations supplémentaires provenant des TPs.

Pareto Chart :
The Pareto chart helps to identify which problems
need immediate attention and which can be looked
at later. While there are many things that cause a
large problem, it is frequently found that a few
specific factors (about 20%) account for the bulk of

4
Arjan Imeri Sacha Didi

the observed occurrences of that problem (about 80%). This phenomenon was called the
‘Pareto principle’.
- On classe les frequency dans l’ordre décroissant et ensuite on fait un pareto chart.

Histogram:
- A histogram helps to identify the distribution of the data.

Run Chart:
- Il s’agit d’un graphique qui affiche les
données observées dans une séquence
temporelle.
- When data are collected from a process
over some time, it is often important to
plot the data in the order in which they
arose. This sort of graph is called a run
chart. Generally, the emphasis is on
diagnosing the degree ‘stability’ of the
process.

A/B testing:
- A/B testing (also known as ‘split testing’ or ‘bucket testing’) is a way of conducting an
experiment where you compare a control group to the performance of one or more
test groups by randomly assigning each group a specific ‘single-variable’ treatment.
A/B testing has emerged as the predominant method of online testing in the industry
today. It is often used to answer questions like, “Is my new model better than the old
one?” or “Which colour is better for this button, yellow or blue?” In an A/B testing
setup, there is a new model (or treatment) and an incumbent model (or control).

- EX: Microsoft Skype ran an experiment for the mobile segment of the Skype
application with a treatment having combined icons with corresponding labels
(right). The control (left) only showed icons. The treatment increased the ‘calls being
made’ (‘engagement’) by 2.7% (with a corresponding p-value of 0.006). As such, this
change was implemented.

Cause-and-effect diagram (Fish bone diagram)

- Ce type de diagramme affiche les facteurs qui sont censés affecter un résultat
particulier (sortie) dans un
système (processus). Les
facteurs (intrants) sont souvent
présentés sous forme de

5
Arjan Imeri Sacha Didi

groupements ou de sous-facteurs connexes qui agissent de concert pour former «

l'effet » global du groupe.
- Cause and effect illustrate the relationship between a given outcome (output) and all
the factors (inputs) that influence the outcome (output).

Flow Chart:
- A flowchart is a type of diagram depicting a workflow or a process. In an organisation,
it helps to understand complex processes — a pre-requisite for improvement. It
allows you to break any process down into individual events or activities and to
display these in shorthand form showing the logical relationships between them.
Below are some standard symbols that can be used to construct a flowchart. Use
them for the next two exercises.
- Il vous permet de décomposer n'importe quel processus en événements ou activités
individuelles et de les afficher sous forme abrégée montrant les relations logiques
entre eux.

3:

Need for business improvement

Data driven decision making:
- Managers should also understand that good decision-making processes must be
based on the analysis of data, rather than purely on gut feeling and intuition, nor on
belief.
We have 2 jobs:
- We used to have only one job: to do our work.
- We must accept a second job: to improve how we do our work to better serve our
customers to stay ahead of the competition.
Business improvement:
Doing activity = business process
- A series of activities, each with its own inputs and outputs, are done as a sequence of
steps to produce the desired output for the customer
- Tous les processus qui transforment les inputs en outputs sont des process.
6
Arjan Imeri Sacha Didi

Improving activity= improvement activity

- There are many different approaches to improvement, but we will focus on two types
of improvement
Process improvement
- A series of activities aimed at fundamentally improving the performance of the
process
o Flowcharting’ the process to understand it better
o Collecting data to assess the current performance
o Identifying areas where the process could be fundamentally improved
o Changing the process to implement improvement ideas
o Checking the impact of improvement efforts
o Making the improvements part of the standard way of doing business.
Problem solving:
- Solving specific problems that are not part of the typical behavior of the process
- These issues are often discovered in the process improvement analysis and can be
resolved without fundamentally changing the process.
- Ce sont les process qui sont des problèmes et non les personnes.
- 3 Steps for basic problems solving:
o document the scope of the problem
o Identify the ‘root causes’ (causes premières)
o Select, implement, and standardize corrections.
Data:
- Data are the connectors (or links) between
the doing and improving activities.
- Data fuel (alimente) process improvement
and problem-solving activities and increase
their effectiveness.
- Why data help us:
o Document process performance
o Identify problems
o Evaluate the impact of proposed
solution

4: Statistical thinking
Definition of statistical thinking:
- Statistical thinking evaluates a process by collecting data in addition to past
experiences and perceptions.
- Statistical thinking is an overall approach to improvement.
- Statistical thinking can be used in all parts of an organization and in all job functions.
- It helps to identify where improvement is needed.
- It provides a general approach to take and it suggests tools to use.
- Statistical thinking is not a collection of statistical methods, but is rather a way of
thinking about improvement, and how to best utilise statistical methods.

7
Arjan Imeri Sacha Didi

Steps in implementing statistical thinking

Statistical thinking is key to business analytics in order to not neglect the following four
principles that ensure successful outcomes (which are all part of its philosophy and
methodology):
- use of sequential approaches to problem solving and improvement, as studies are
rarely completed with a single data set but typically require the sequential analysis of
several data sets over time (􏰀 ‘continuous improvement’);
- having a strategy for the project and for the conduct of the data analysis; including
thought about the business objectives (􏰀 ‘strategic thinking’ );
- carefully considering data quality and assessing the ‘data pedigree’ before, during
and after the data analysis; and
- applying sound subject matter knowledge (‘domain knowledge’ or ‘business
knowledge’, i.e. knowing the business context, process and problem to which
analytics will be applied), which should be used to help define the problem, to assess
the data pedigree, to guide data analysis and to interpret the results.

Average, variation, and target

- The terms average and variation are critical to applying statistical thinking.
- Improvement activity (process improvement and process solving) is based on:
o Reducing variations through tighter (plus stricte) control of the process or

8
Arjan Imeri Sacha Didi

o Improving the overall level (average level/value, i.e., the central value around
which the process varies) by changing the process target, which may also
result in reduced variation.

Traditional quality and quality improvement (statistical thinking)

Traditional quality:
- The traditional definition of quality is based on the viewpoint that products or
services must meet the requirements of those who use them.
Quality improvement:
- Quality improvement is the reduction of variation in processes.
- If variation in the important characteristics of a product or a service decrease, the
quality of the product or the service increases.
Good quality does not mean perfect
- The proper goal should be a consistent, predictable range of quality.
o For example, so that goods can be produced throughout the production cycle
at the lowest cost that will result in quality that is acceptable to the
consumer.
- Example (Making toast, continued). It would make no sense for a family to invest
$5’000 in a toaster guaranteed to burn no more than one in a million slices.
o A failure rate of one in 500, one burned piece of toast every three months,
say, would be much less expensive to achieve but acceptable in most homes.
Consistent performance
- Customers of today and tomorrow value products or services that have consistent
performance.
o The consistency in performance can be achieved by systematically reducing
variation in business processes.
o Process variation affects not only quality but also operating cost, cycle time,
profitability, and customer satisfaction.
Data-driven decision making:
- enables businesses to turn knowledge into appropriate decisions for the good of their
customers, and ultimately for an organization’s benefit:

9
Arjan Imeri Sacha Didi

Improvement approach of statistical thinking:

- Process  VariationData

4.1: Principles of statistical thinking

3 principles:
First principle: (slide 100-130)
- All work occurs in a system of interconnected processes — a process being a chain of
activities that turns inputs into outputs
Second principle: (slide 131 à 135)
- Variation, which gives rise to uncertainty, exists in all processes
Third principle: (slide 136 à 151)
- Understanding and reducing (unintended or unwanted) variation are keys to success.

All three principles work together to create the power of statistical thinking.
First principle: (process thinking or system thinking)
-This principle provides the context for understanding the organization, the
improvement potentials and the sources of variation mentioned in the second and
third principles.
- We cannot improve a process that we do not understand.
SIPOC:
- “Suppliers–Inputs–Process steps–Outputs–
Customers” model is a common model for process
improvement efforts.
- SIPOC is a comprehensive view of process including
consumers and suppliers.
- SIPOC n’est pas un modèle qui permet de prendre
des décisions, c’est plutôt un modèle qui résume ce
qui se passe.
- Le SIPOC est une grosse image de tout ce qui se
passe du suppliers au customers. Entre les 2, il y a les
inputs-process-outputs.
o On peut par exemple développer le process
de manière plus précise.
- The SIPOC model for a business helps everyone in
the company see the business from an overall perspective by:
o Keeping a focus on consumer needs (car c’est pour lui qu’on fait ce
processus).
o Identifying inputs and outputs for each step of the process
o Helping maintain the big business picture.

10
Arjan Imeri Sacha Didi

Blame the process, not the people

- Il faut changer le procéder et pas les gens, car c’est le process qui peut créer de
l’amélioration.
- The ‘85/15 rule’, which states that 85% of the problems are in the process and the
remaining 15% are due to the people who operate the process.
o In the vast majority of the problems, the process is the origin.
Process measurement:
- Les mesures de processus (c'est-à-dire les mesures ou les paramètres) sont
essentielles à la gestion et à l'amélioration réussies des processus.
- Are enable (Permet) effective control, management, and improvement of processes.
- Are not able to recognize causes of variation.
- Lack of ‘good’ data is typically the greatest barrier to improvement
Process measurements tracking:
- Process measurements tracked over time enable us to analyze the process in the
following five ways:
o Assess current performance levls
o Determine if the process has shifted (a change) by comparing current
performance to past performance
o Determine if the process should be adjusted (minor changes)
o Determine if the process must be improved (major changes)
o Predict future performance of the
process.
Operational definition:
- Used in the data collection process are keys!
- An operational definition is a set of specific
instructions about when, where, and how to
obtain data.
- An operational definition is a definition of a
concept in terms of specific, empirical measures.
- Operational definitions can be loosely described as
descriptions that allow two people to look at the
same thing at the same time and record the same
measurement.

Second principle: (Variation exists in all process)

Variation:
- This provides the opportunities for process improvement and hence quality.
- Variation is the enemy of quality.
- Variation is everywhere. Know it, accept it, learn to deal with it.
o To understand and improve a process we must take variation into account.
- Variation creates the need for statistical thinking.
- Determine if the process has shifted by comparing current performance to past
performance
No variation:
- If there were no variation: there would be little need to use statistical thinking.
o Processes would run better

11
Arjan Imeri Sacha Didi

o Products would have the desired quality

o Services would be more consistent
o Managers would manage better
Unwanted variation:
- It is unintended (or unwanted) variation which causes all the problems, which
ultimately results in, for example, increased customer dissatisfaction, undesired
product, or service performance.
o Ce sont les variations inhabituelles (non voulues) qui posent des problèmes.
Wanted variation :
- Les wanted variations sont les variations voulues (habituelles), qui n’impactent pas le
processus et la qualité du service
o Par exemple, le fait de commander un plat qui prend plus de temps à
préparer est une variation voulue (normal, habituelle).
Third principle: (understanding and reducing variation are key to success)
- The focus is on unintended variation and how it is analyzed to improve performance.
- First, we must identify, characterize, and quantify variation to understand both the
variation and the process that produced it.

The performance of a process:

- is influenced by its average performance and the amount of variation around the
average performance.

The average performance of any process:

- is a function of various factors involved in the operation of the process.
o EX: average waiting time in minutes to be served in a restaurant.

Understand variation in process:

- When we understand the variation in the output of the process, we can determine
which factors within the process influence the average performance.
- We can then attempt to modify these factors to move the average to a more
desirable level.
- Understanding variation is important because:
o Variation is present in everything we do and must be considered.
o Variation identifies opportunities for improvement.
o Variation increases risk and decreases quality.
Consistency: Il s’agit de la régularité, qualité.
- Businesses, as well as customers, are interested in the variation of the process output
around its average value.
o Typically, consistency of a product or a service is a key customer requirement.
- Low variation is important to customers, and they will sometimes accept less
desirable average performance to obtain better consistency.
o Si y’a des petites variations le service fournit sera PLUS CONSISTENT
o Ex : Si les clients savent qu’un resto (moyen) a une faible variation (0-5min) de
service, ils vont préférer cela qu’attendre 50min pour un meilleur resto de
bonne qualité.

12
Arjan Imeri Sacha Didi

Two types of unintended variation in business process:

Special cause variation/assignable cause of variation:
- Is outside the ‘normal’ or typical variation a process exhibits.
- The result of special cause variation may be unpredictable or unexpected values
- that are too high or too low for the customer.
- Results from unexpected or unusual occurrences that are not inherent in the process.
o Ex: Waiting for service for 30 minutes instead 10 minutes or less is typical
performance for a particular restaurant.
- Because special cause variation is atypical, we can often eliminate the causes without
fundamentally changing the process.
o Using a problem-solving approach, we identify what was different in the
process when it produced the usual result.
Common cause variation/chance cause of variation/random cause of variation:
- It is important to make changes to the process.
- Results from how the process is designed to operate and is a natural part of the
process.
- Searching for the specific reason for common cause variation is a waste of time and
resources.
- Do not treat common cause variation as special cause variation (‘tampering’)!
- Common cause variation is a normal variation.
o EX : Les TPG arrivent parfois en retard de 10 minutes durant les heures de
pointes car il y a des embouteillages sur la route (cause normale de variation).

System change
Quand on a une sorte de spécial variation qui dure, par exemple une augmentation de la
température sur les fèves de cacao dû à la météo qui dure depuis plusieurs mois, on aura
donc une hausse ou une baisse du système tout entier, comme on peut le voir juste au
dessus.

Process improvement approach :

Identifier les common causes variations :
- We try to discover the input and process factors that are the largest sources of
variation.
- Knowledge and understanding of common cause variation are best obtained using a
statistical approach and ‘designed experiments
Modifier les common causes variations: Make changes to the process

13
Arjan Imeri Sacha Didi

- Once the primary sources of variation are identified, we can determine how to
modify the process to eliminate or mitigate their effects.
Three key business analytics questions
1. How am I doing?
o Understand and agree on key process metrics or ‘Key Performance Indicators’
(KPIs).
2. What drives my business?
o Once you have KPIs identified, you need to understand what drives these
KPIs.
o Which are the most ‘important’ drivers (factors) that influence a given KPI?
3. Who are my customers? What are their needs?
o Understand customers and customize their offering, messaging, marketing
channel accordingly, delight the customers, securing their future revenue.
o Drive a given KPI in the ‘right’ direction.

Writing an objective improvement statement

-
The objective improvement statement must indicate the level of improvement
expected from improvement efforts, including specific, quantifiable amounts and the
time required to complete.
o EX : L'objectif est de réduire le nombre de pizzas livrées en plus de 30 minutes
de 55% à moins de 5% d'ici le 30 juin de l'année prochaine. Cela renforcera
davantage l'image d'excellence de la pizzeria et permettra à l'entreprise
d'économiser plus de 60 000 $ par année.
SMART: help to write objective statements

Deming’s red bead experiment

-This experiment was created to illustrate several points clearly and dramatically
about poor management practices used for improvement.
- The name ‘red bead’ is a metaphor for ‘any problem’ that could be experienced..
- L'expérience montre que même si les « travailleurs volontaires » veulent faire du bon
travail, leur succès est directement lié et limité par la nature du système (c'est-à-dire
le processus) dans lequel ils travaillent.
Concept de l’expérience:
- On prend une entreprise et on fait faire un test aux employés (peluches) de
l’entreprise.
- Les employés piochent avec un paddle dans une boîte ou il y a majoritairement des
billes blanches et également des billes rouges (red beads).

14
Arjan Imeri Sacha Didi

o Les billes blanches représentent les bonnes choses que nous faisons dans nos
jobs la plupart du temps.
o Les billes rouges représentent les problèmes ou les mauvaises choses que
nous faisons et que nous voulons réduire.
- Inputs :
o raw material (boite) + equipment (paddle) + les gens qui travaillent.
- Outputs :
o Le résultat du travail (beads).
 Read beads= bad work
 White beads= good work
- Le but est de piocher uniquement des billes blanches.
- Les managers dirigent l’expérience, et on voit que chaque manager a sa technique
spécifique mais que aucune ne marche mieux que les autres car les employés
piochent toujours des billes rouges.
- Donc le problème ne vient pas des employés, ni des managers mais du système en
lui-même (car c’est impossible de tirer que des billes blanches).
o Il faut changer l’environnement (le système) et le système est décidé par le
TOP management.
Conclusion :
- Les différences dans les résultats entre les employés est uniquement due à la chance.
- Le problème fondamental du fait que les résultats sont médiocres vient du « poor
raw material », il y a trop de billes rouges dans la boite.
- Si on enlève des red beads, la variation baisse et on aura de meilleurs résultats.
o Cela est vrai mais ça dépend de quelle était la répartition de base entre les
billes rouges et blanche (50/50, 30/70 etc.)
o La répartition qui donne le maximum de variations est 50/50 (rouge/blanche).

4.2 : Statistical thinking strategy

- This is a strategy that can improve business process.
5 steps to improve the performance of a company:
1: Study the current situation:
2: Define quality performance
3: Analyze performance
4: Focus on the main problems, collect data
5: Monitor and evaluate progress

Synergy: Data and subject matter knowledge

Definition subject matter knowledge:
- Is everything we know about the process under study.
- This could be derived from experience or from an academic study.
- This guides us in planning which data from the process would be most helpful to
validate or refine our ideas (‘hypotheses’).
- It helps us to determine where biases are likely to occur and to avoid or minimize
them.
Synergy: Data and subject matter:

15
Arjan Imeri Sacha Didi

- Il faut avoir une connaissance générale du sujet/domaine qui subit l’étude pour faire
une analyse pertinente. Mais il faut également des datas pertinentes.
- Another key point is the synergy between data and subject matter knowledge
(connaissance du sujet).
o The application of statistical thinking does not begin with data.
o It generally begins with some understanding of the process of interest and
well-defined objectives.
o La connaissance du processus permet de choisir judicieusement les données à
collecter.
o This in turn leads to an increased understanding of the process, which guides
further data collection, and so on.
o Even more important is the need to analyze and interpret the data in light of
the subject matter knowledge.

Statistical thinking can be applied to improve business processes

- First: identifying, documenting, and understanding the business process itself and
subject matter knowledge.
- Once the data are obtained, statistical techniques and tools are used to analyze the
data and to account for the variation in the
data.
o Although the analysis may confirm
our ideas (‘hypotheses’), additional
questions almost always arise, or new
ideas may be generated.
o This leads to a desire to obtain
additional data to validate or further
refine the revised hypotheses, and
the whole cycle is repeated.
- A complicated issue is the fact that business
processes are not static over time but
dynamic.
o Statistical methods can be very
helpful in determining whether the process has undergone ‘significant’
change.

Relationship to the scientific method

Definition:
- Une méthode scientifique commence par une hypothèse
énoncée (« idée ») sur un phénomène, puis une expérience est
menée pour tester l'hypothèse, et l'observation des résultats
confirme ou infirme l'hypothèse.
- Les observations d'une expérience peuvent nous amener à
réviser notre hypothèse, ce qui peut conduire à une autre
expérience pour évaluer l'hypothèse révisée.
- Applications tend to be iterative.
What did scientifics:
- Scientific investigation is a sequential learning process!
16
Arjan Imeri Sacha Didi

- Statistical methods allow investigators to accumulate knowledge!

Statistical thinking uses the scientific method to develop subject matter knowledge and to
gather data to evaluate and revise hypotheses (‘ideas’).
Two analytics’ paradigms:
Deductive reasoning:
- (‘idea (hypothesis) first’): start with an idea (‘theory’), make an experiment and
analyse the collected data to confirm or evaluate the idea (using ‘explanatory and
confirmatory analytics’)
Inductive reasoning:
- ‘Data first’): analyze the existing data to generate an idea (using ‘exploratory and
predictive analytics’).
Ces 2 méthodes sont complémentaires et doivent procéder de manière itérative (répété
plusieurs fois) et côte à côte afin de permettre une prise de décision basée sur les données
et une amélioration continue appropriée.
‘Plan–Do–Check–Act’ (PDCA) cycle
Definition PDCA
- A general management approach based on the scientific method.
- Statistical thinking can be viewed as an adaptation of the PDCA cycle that focuses on
the business process and on an understanding of variation.
Why is it important?
- Multiple iterations of the PDCA cycle until the problem is solved
- Continuous improvement in itself is a process that must be managed!
Plan step
- Corresponds to thinking about the data we need
based on our subject matter knowledge.
- Set the project goals (objectives) and
expectations.
- Develop the project plan.
- Plan a change or a test, aimed at improvement.
Do step:
- Is conducting the experiment (or the survey) to
obtain data and/or assessing the existing (available) data.
- Implement the plan.
- Assure continued process performance.
- Carry out the change or the test (preferably on a small scale).
Check Step:
- Is analyzing the data.
- Check to be sure that the plan has been carried out and that
the anticipated results have been achieved.
- Assess major problems and root causes (causes premières)
- Define the scope of the project.
- Identify the people who will be affected by the project.
- Study the results. What did we learn? What went wrong?
Act step:
- Is interpreting the analysis in light of our original hypothesis.
- Fix the obvious.
- Decide how to approach the issue or problem.
17
Arjan Imeri Sacha Didi

- The results will determine whether we need to repeat the cycle.

- Adopt the change, or abandon it, or run through the cycle again .

5: Statistical engineering
Why using statistical engineering ?
- We want to think before we calculate, not the other way around (et pas
inversément).
- To avoid this pitfall (piège), we need a framework (cadre) to help us apply the
methods in a manner (manière) that is consistent with the principles of statistical
thinking.
- This is especially important when attacking complex problems, which require us to
integrate several methods into an overall problem-solving strategy.
Definition:
- Is the study of systematic integration of
statistical concepts, methods, and tools, often
with other relevant disciplines, to solve
important problems sustainably.
- Statistical engineering links together several
statistical and non-statistical tools in a logical
sequence, which may be uniquely designed for
the specific problem at hand.
- The engineering perspective provides a clear
focus on problem solving and continuous
improvement!
- Statistical engineering is a discipline that
studies how to drive greater improvement
using the existing science or theory of statistics.
It is therefore, a horizontal slice connecting
statistical teory with statistical practice.

Principle of statistical thinking

- A system or strategy to guide the use of statistical tools is needed for the tools to be
effective.
o We need to think carefully about the problem and our overall approach to
solving it prior to getting carried away with detailed analytics.
- The impact of statistical thinking and methods can be increased by integrating several
statistical tools, and even non-statistical tools, enabling practitioners to deal with
issues that cannot be addressed with any one method.
- Viewing statistical thinking and methods from an engineering perspective provide a
clear focus on problem solving.
- Statistical engineering emphasizes having a plan of attack for complex problems and
linking various tools together in an overall approach to improvement.

Exemple important à lire:

- Telephone waiting time (Slide 224 à slide 232)

18
Arjan Imeri Sacha Didi

6: THREE types of variation in business

processes
special cause There are two first types of unintended variation that we
may need to reduce: ‘’ and ‘common cause’ variation.
Do not treat common cause variation as special cause variation (􏰁 ‘tampering’)!
1) Special cause variation: Unpredictable (unstable)
- is outside the ‘normal’ or typical variation a process exhibits (also called ‘assignable
cause of variation’).
- A process is said to be ‘out of control’ if special cause variation is present.
- Non-random and unpredictable patterns on a control chart
Eliminating special causes variation to stabilize the process:
- Because special causes are ‘assignable’ to a specific event and are not a
fundamental part of the process, they can often be addressed without spending a
great deal of time, effort, or money.
- Special causes tend to have the most impact, so addressing them first typically will
result in dramatic improvement.
- Once the process is stabilized, we can begin to study and understand it, allowing us
to make fundamental improvements.
- Very often it is desirable to reduce the (unintended) variation.
2) Common cause variation: Predictable
(stable)
- ‘Normal’ or typical variation
- A process is said to be ‘in (statistical)
control’ if the only source of variation is
common cause variation.
- If the process is stable (or ‘in control’), improvement generally requires careful study
of the whole process to identify and prioritize the common causes and their
relationships.
- Generally, the objective is to reduce variation around the desirable average (‘target’)
Benefits of stable processes (only common cause variation):
- A process has an ‘identity’ and its performance is predictable; therefore there is a
‘rational basis’ for planning
- Costs and quality are predictable
- Productivity is at a maximum and costs are at a minimum for the process
The effect of changes in the process can be measured faster and more reliably.
Le off-target variation :
- La ligne trait tillé est la target des
managers et souvent on est en-
dessous, c’est ce qu’on voit.
- La plupart les points noirs sont en
dessous de la target.

19
Arjan Imeri Sacha Didi

Type of action for reducing variations

Waste of time
and ressources

You’ll miss an opportunity

for improvement

20
Arjan Imeri Sacha Didi

3) Structural variation
- There are some ‘predictably unstable’ processes, i.e. some processes are technically
unstable, but in a predictable way.
- Structural variations are inherent part of the process.
- In general, eliminating structural variation requires fundamental change to the
process, much like common cause variation.
o The causes of this behavior are sometimes referred to as ‘structural variation’.
o EX : Le nombre de clients arrivant dans une banque par heure a tendance à
être instable dans la mesure où un « rush » se produit généralement vers
l'heure du déjeuner.
 Cette précipitation est cependant très prévisible.
 Astuce concernant l’exemple: It would be more helpful to know the
average number arriving during the lunchtime rush and during off-
peak hours so staffing could be appropriately balanced.
Comparaison between special, structural and common causes:
- Special and structural causes can often be readily identified.
-
Robustness: an underused concept (un concept sous-exploité)
- Robustness is another key aspect of statistical thinking.
o Besides (outre) controlling the process by eliminating special cause variation
and improving the system by reducing common cause variation, robustness is
another way to reduce variation.
- We want to build robust processes by anticipating variation and reducing its effects
- Robustness means to reduce the effects of uncontrollable variation in:
o Process and product design
o Management practices

Process and product robustness

- Design the process or product to be insensitive to variation.
o Concevoir le processus ou le produit pour qu'il soit insensible aux variations
 Même si y’a des variations, ça ne va rien nous faire.
o Such a robust process or product is more likely to perform as expected.
- 100% inspection is not 100% accurate (pas précis) and cannot provide robustness.
- Process performance is robust if it is insensitive to uncontrollable variation in the
process resulting, for example, from the inputs, the transformations, the activities,
the steps, or external factors.
- Product performance is robust if it is insensitive to variation in conditions of, for
example, manufacture, distribution, use
or disposal.
Management robustness
- Develop strategies that are insensitive to
economic trends and cycles.
o Insensible : mm si ya des
variations, ça ne va rien nous
faire (ca vas pas nous impacter)

21
Arjan Imeri Sacha Didi

- Design a project system that is insensitive to, for example, personnel changes,
changes in project scope or variation in business conditions.
- Respond to differing employees’ needs by, for example, adopting flexible work hours,
providing ‘cafeteria’ benefits package, and so on.
- Enable personnel (permettre au personnel) to adapt to changing business needs.
- Ensure that meeting effectiveness is not dependent on facilities, equipment, or
participants.

7: Process stability and capability tools

Introduction and motivation
- Process stability and ‘capability’ are fundamental concepts in statistical thinking.
Stability:
- Stability is more important than capability!
- It is a common mistake in business to assume that observed data come from a stable
process, often leading to inappropriate analyses and erroneous conclusions.
o We need tools to tell if our processes are stable or not.
- When evaluating stability and ‘capability’, always consider stability first!
o This is because the concept of ‘capability’ does not make much sense if the
process is not stable.
o It might look ‘capable’ today but could shift tomorrow.
o It might not look ‘capable’ because of instability, but once stabilised, be
‘capable’.
Capability:
- In addition, processes are generally designed with an inherent ‘capability’.
o We may wish they would produce better results, but if the process is not
‘capable’, it will continue to disappoint until fundamental improvements are
made.
o We need tools to tell if our processes are ‘capable’ or not.
Capability and stability
- ‘Control charts’ and ‘process capability analysis’ are the primary quantitative
statistical tools used to evaluate stability and ‘capability’.

Statistical inference, accuracy and precision

Statistical inference:
- It essentially means drawing
conclusions about a population
based on sample process
measurements.
- Relationship between a population and a sample. Note that μ
and σ are (generally) unknown (statistical) parameters of the
population, and X ̄ and S are statistics calculated on a sample.
They are random variables, i.e., the values that these
estimators will take, i.e., the estimates, depend on the selected
sample.

22
Arjan Imeri Sacha Didi

Accuracy (exactitude) and precision

- Accuracy and precision in shooting arrows (‘sample estimates’) at a target (‘bull’s-
eye’).
- Processes with underlying (population) distributions (e.g., ‘Process 4’ is accurate and
precise):

Control charts to evaluate stability

Introduction to control charts
- Control charting’ is the primary statistical tool used to evaluate stability.
- Control charts arte the way the process communicates with you.
o Problem if there are special causes.
o OK if there are common causes.
- ‘Control charts’ are one of the techniques of so-called ‘Statistical Process Control’
(SPC).
o A better naming would be ‘process control aided by statistical techniques.
o This would put the emphasis where it belongs: on the intent, i.e., process
control, rather than the tool, i.e., statistical techniques.
Purpose of control chart:
- The primary purpose of control charts is to differentiate common cause variation
from special cause variation over time.
- This will determine whether a process improvement or problem-solving approach is
warranted (garantie).
- The control chart tells the manager how to appropriately respond.

Special causes and common causes in a control chart

- Special causes:
o The control chart B indicates the last point is
a special cause (outside the ‘control limits’) and worthy of investigation.
- Common causes:
o The control chart C indicates the last point is common cause (inside the
‘control limits’) and should be ignored.

Two errors are possible when managers don’t use control charts:
- They could fail to identify a performance change as special cause and thereby lose
valuable information about the process
- they could treat a change as a special cause when it is, in fact, a part of ‘typical’
process variation — this error is called ‘tampering’.

Center line, UCL and LCL

A typical control chart contains:
- A center line (represents the average value of the
process characteristic corresponding to the in-
control state)
- An ‘Upper Control Limit’ (UCL)
- A ‘Lower Control Limit’ (LCL)

23
Arjan Imeri Sacha Didi

o which are chosen so that if the process is in control, ‘nearly all’ of the sample
points will fall between them.
Process in the control:
- As long as the points plot within the control limits, the process is assumed to be ‘in
control’, and no action is necessary.
Process out of the control:
- However, a point that plots outside of the control limits is interpreted as evidence
that the process is ‘out of control’.
o Investigation and corrective action are required to find and eliminate the
assignable causes responsible for this behavior.
o Even if all the points plot inside the control limits, if they behave in a
systematic or nonrandom manner, then this is an indication that the process
is ‘out of control’.
Run tests/rules of thumb:
Process out of the control:
- The process may be defined as ‘out
of control’ if any one or more of
the following ‘rules of thumb’
(règles de base) are met run tests.
- These are intended to help
‘calibrate the human eye’ when it
searches for patterns.

Phase 1 and 2 of control charting:

Standard control charting involves two distinct phases.
Phase I (or ‘set-up’ phase): collecter données + faire un control chart + center line, UCL et
LCL.
- is a retrospective analysis of process data to construct ‘trial (essaie) control limits’
o To determine if the process has been in control over the period of time where
the data were collected.
o To see if reliable control limits can be established to monitor future
production.
- Control charts are used primarily in phase I to assist in bringing the process into a
state of statistical control by differentiating common cause variation from special
cause variation over time.
- Evaluate process stability and eliminate special causes.

Phase II: Surveiller (monitoring) les special causes.

- Begins after we have a ‘clean’ set of process data gathered under stable conditions
and representative of the in-control process performance.
- In phase II, we use control charts (with the projected phase I control limits) to
monitor (surveiller) the process for special cause variation.
- Special causes can be detected rapidly using these charts, and problems can be
diagnosed and eliminated.

24
Arjan Imeri Sacha Didi

- The control limits will not be updated unless there has been a substantial change in
the process.

8: Control charts
Definition:
- Control charts are a tool for learning about a process over time.
- Control charts help detect special cause variation.
o The control chart itself does not eliminate special causes, it ‘only’ detects
them and provides clues to help us understand the process.
- A control chart is also of little value in improving a stable process because it will only
continue to tell us that the process is stable.
o Control charts are not designed for reducing common cause variation.
o To reduce common cause variation quite different tools and ways of thinking
are required, e.g., modify internal parts of the process using ‘planned’
investigation.

Choice of control limits

- Specifying the control limits is one of the critical decisions that must be made in
designing a control chart.
Increasing control limits (UCL, LCL):
Advantage:
- By moving the control limits (UCL, LCL) farther (plus loin) from the center line, we
decrease the risk of a point falling beyond the control limits, indicating an out-of-
control condition when no assignable cause is present.
o We decrease the risk of ‘type I error’.
Inconvenient:
- However, widening (élargir) the control limits will also increase the risk of a point
falling between the control limits when the process is really out of control.
o We increase the risk of ‘type II error’.
Decreasing control limits (UCL, LCL):
- If we move the control limits closer to the center line, the opposite effect is obtained.
Link between control charts and statistical hypothesis:
- There is a close connection between control charts and statistical
hypothesis testing as the control chart tests a hypothesis
repeatedly at different points in time.
Screen: Suppose the vertical axis in the following figure is the sample
average X ̄.
Testing H0: The control chart is a test of the hypothesis (H0) that the
process is in a state of statistical control.
Don’t reject H0:
- If x ̄, i.e., the current value of X ̄, plots between the control limits, we conclude that
the process mean μ is ‘in control’; that is, it is equal to some value μ0.
- A point plotting within the control limits is equivalent to failing to reject the
hypothesis of statistical control.
Reject H0:

25
Arjan Imeri Sacha Didi

- On the other hand, if x ̄ exceeds either control limit, we conclude that the process
mean μ is ‘out of control’; that is, it is equal to some value μ1 ̸= μ0.
- A point plotting outside the control limits is equivalent to rejecting the hypothesis of
statistical control.

A general model for a control chart

Center line, UCL and LCL:
- Let W be a sample statistic that measures some process characteristic of interest.
Suppose that the mean of W is μW and the standard deviation of W is σW .
o The center line, the upper control limit (UCL) and the lower control limit (LCL)
become
- UCL = μW +LσW
- center line = μW
- LCL = μW −LσW
o where L is the ‘distance’ of the control limits from the center line, expressed
in standard deviation units.

μ and σ are known for a normal distribution

Three sigma control limits:

1) On collecte des samples et on calcule les UCL, Center line, LCL.

26
Arjan Imeri Sacha Didi

2) On analyse nos données et si certains points sont “out of control”, alors on doit
corriger ces samples qui sont “out” et les retirer de nos calculs.
3) Donc on recalcule UCL, Center Line, LCL de nos samples (sans prendre en compte les
samples “out”.
4) Si on devait ajouter des nouveaux samples par la suite, on garde nos UCL, Center
Line, LCL du point 3.

- We use the three-sigma level because this level provides a good balance between the
risk of having a false alarm and missing special causes.
- Using three-sigma control limits corresponds to using a probability of a ‘type I error’
of α = 0.0027.
o That is, an incorrect out of control signal (or ‘false alarm’) will be generated in
only 27 out of 10′000 points.
- The probability that a point taken, when the process is in control, will exceed the
control limits in one direction only is 0.0027/2 = 0.00135.
- Moreover, 1−0.0027 = 0.9973 of the population values will fall within the control
limits.

Theory of C chart in the practical 8

μ and σ (from the population) are unknown for a sample:

27
Arjan Imeri Sacha Didi

Find an estimator of mu: double X bar

Find an estimator of the standard deviation using Range Method: σ

On calcule la différence entre la plus grande et la plus petite observation pour chaque
sample.

Find an estimator of σ using standard deviation method:

28
Arjan Imeri Sacha Didi

Types of control charts

- There are many types of control charts, but they all plot data (or statistics calculated
from data) over time (or sample number) and have statistical control limits (UCL and
LCL) representing the range of common cause variation to be expected.
-
- The types of charts are often classified according to the type of process characteristic
that they are supposed to monitor.
There are two general types of control charts:
- For discrete data, i.e., ‘attributes’, such as the number of customer service calls
received
o Example. The percentage of checks produced with incorrect amounts is a
count of a discrete occurrence and is discrete data.
- For variable data, i.e., ‘variables’, measured on a continuous scale, such as cycle
time.
 Example. Waste in a chemical factory is sometimes expressed as a
percentage of total production, but it is still a continuous
measurement (typically weight).
o Some data that are very similar to defects data do not satisfy the statistical
requirement that the defects are rare for use of the Poisson distribution.
Defect:
- A defect is any item or service that exhibits a departure from specifications. A defect
does not necessarily mean that the product or service cannot be used. A defect
indicates only that the product result is not entirely as intended.
- Suppose service in a restaurant is being evaluated. If a waiter greets his table after 5
minutes, the customer can still order and enjoy a meal even though the promptness
of the greeting did not meet expectations. Therefore, this could be considered a
defect (“late greeting”) in the service.
- Un défaut est un élément ou service qui présente un écart par rapport aux
spécifications. Un défaut ne signifie pas nécessairement que le produit ou service est
inutilisable. Un défaut indique simplement que le résultat du produit n'est pas tout à
fait celui espéré.
- Supposons que le service d'un restaurant soit évalué. Même si un serveur accueille
un client après 5 minutes, ce dernier peut toujours commander et savourer son repas
même si la rapidité de l'accueil n'a pas rempli ses attentes. Par conséquent, ceci peut
être considéré comme un défaut ("accueil tardif") dans le service.
-
Defective:
- A defective is an item or service that is considered completely unacceptable for use.
Each item or service experience is either considered defective or not—there are only
two choices.
- Before final shipment, a quality inspector evaluates auto supply parts and rates each
item as “pass” or “fail” to ensure that the company does not ship any parts that will
be unusable.
- Un non-conforme est un élément ou service dont l'utilisation est considérée comme
complètement inacceptable. Chaque élément ou expérience de service est considéré
comme non conforme ou n'est pas considéré comme tel : seuls deux choix sont
possibles.

29
Arjan Imeri Sacha Didi

- Avant la livraison finale, un inspecteur en charge de la qualité évalue des pièces

automobiles et note chaque élément comme "succès" ou "échec" pour s'assurer que
la société n'envoie aucune pièce non utilisable.

o These data are better treated as if they were continuous data, even though
they technically are not continuous.
o Example. In an account payable center, we can certainly count the number of
invoices processed daily, but we cannot realistically count the number not
processed (theoretically there could be an infinite number processed).
o Invoices are being processed continuously while the office is open, however.
o We should therefore treat the number of invoices processed as continuous
data for statistical purposes, even though it is actually a count.

Control charts for discrete data (attributes)

1: Defect datas
Definition:
- First, consider discrete data, i.e., attributes, where we count the number of defects,
nonconforming items, or instances of an attribute, such as the number of customer
service calls received.
o Où l'on compte le nombre de défauts, d'articles non conformes ou d'instances
d'un attribut, comme le nombre d'appels de service client reçus.
o Example. If we are counting the number of customers arriving in a restaurant,
we can count the number of customers arriving, but we cannot realistically
count the number of people who are not coming to the restaurant.
- Such discrete data are called ‘defects data’.

The basic control charts for defects data include:

c chart or ‘count chart’:
- Plot of the number of defects (or nonconforming items).
- Assumes that defects (or nonconforming items) are rare and control limits are
computed based on the Poisson distribution (‘distribution of rare events’); see
‘Practicals 8’.
-

u chart or ‘unitised count chart’:

- Plot of the fraction (or proportion) of defects (or nonconforming items), i.e. the
number of defects (or nonconforming items) divided by the sample size.
- The u chart simply converts the data to a common scale when the opportunity for
defects varies from data point to data point.
- Assumes that defects are rare and control limits are computed based on the Poisson
distribution.
- Does not require constant number of units (unlike
the c chart). Can be used, for example, when
samples are of different sizes.

- c chart for the number of defects: c'est-à-dire de

non-conformités, observés sur 26 échantillons
30
Arjan Imeri Sacha Didi

successifs de 100 cartes de circuits imprimés, avec des limites de contrôle à trois
sigma.
- Comme les 26 échantillons contenaient 516 non-conformités totales, la ligne centrale
est 516/26 = 19,85.

2 : Defective datas
- Second, consider discrete data, i.e., attributes, that fall in only one of two categories,
such as ‘defective’ or ‘not defective’, or ‘nonconforming’ or ‘conforming’.
- Example: En comptant le nombre de cartes de crédit produites avec le nom incorrect
ou mal orthographié, nous pouvons compter à la fois le nombre de cartes incorrectes
et le nombre de cartes avec des noms corrects.
- Such discrete data are called ‘defectives data’.

The basic control charts for defectives data include:

np chart:
- Plot of the number of defectives.
- Control limits are based on the binomial distribution.
p chart:
- Plot of the fraction or proportion of defectives, i.e., the ratio of the number of
defectives to the sample size.
- Control limits are based on the binomial distribution; see ‘Practicals 8’.
These are basically the same charts, except that one (p chart) is on a proportional scale and
the other (np chart) is not.
- For defectives (binomial) data, we may have equal or
unequal subgroup sizes.
o For example, si nous inspectons les bons de
commande pour en vérifier l'exactitude, il se peut
que nous ne soyons pas en mesure d'inspecter le
même nombre à chaque fois en raison de
circonstances indépendantes de notre volonté.
- The p chart has the advantage that groups of data with
different sample sizes can still be plotted on the same chart.

Control charts for continuous data (variables)

Definition:
- Consider variable data, i.e., variables, measured on a continuous scale, such as cycle
time.
- A process characteristic is described by an average (‘central value’) and a measure of
its variation (‘deviation around the average’).
- It is necessary to monitor both the mean value of the process characteristic and its
variation!
The basic control charts for continuous data (variables) include:
Individual’s chart (plot of individual data points):
- Used when the nature of the process is such that it is difficult or impossible to group
measurements into subgroups so that an estimate of the process variation can be
determined.

31
Arjan Imeri Sacha Didi

- The sample size used for process monitoring is n = 1; that is, the sample consists in an
individual unit — a single process measurement.
- The individual’s chart is a generic chart, and it can be used in virtually any situation
when data are measured on a continuous scale.
- The individual’s chart is therefore a good default chart when you are uncertain about
the most appropriate chart to use.

Moving range control chart :

“moving range chart” ne détecte pas les changements de
variation, bien que de très longues lignes entre des points
consécutifs sur le graphique indiquent une instabilité

Joiner chart :

X bar chart: Three sigma control limit

of X double bar chart with R:

32
Arjan Imeri Sacha Didi

X double bar= Sum(Xi bar)/total

nb of samples

R chart:

S chart:
Sbar= Sum(Si)/total nb of samples

Three sigma control limit of X double bar chart with S:

Center line= X double bar

33
Arjan Imeri Sacha Didi

Flow chart for selecting the appropriate control chart:

Exemple de Individuals and Moving range:

Which type of chart would be most appropriate if we are charting the percentage waste in a
factory each month? This value is calculated by taking the total waste in tons within the
factory each month, and dividing by total output, also in tons.

Exemple de NP ou P chart :
A sample of 50 invoices are inspected each shift for errors, and classified as either ‘good’, i.e.
without error, or ‘bad’, i.e. having some problem. If we plot our ‘bad’ invoices each shift on a
control chart, which chart should we use?
- Le nombre de facture est discret, on peut compter les bonnes et les mauvaises
factures.
- On les classe comme good ou bad, donc comme occurrence ou non occurrence.

34
Arjan Imeri Sacha Didi

Control chart constants and formulas:

Constant:

Formulas:

35
Arjan Imeri Sacha Didi

9: Process capability and capable processes

Capability:
- Refers to the ability of a process to meet its business requirements.
- Fait référence à la capacité d'un processus à répondre à ses exigences commerciales

Capable :
- Means that the process variation is low enough so that if the process average is
properly set, virtually all process measurements will meet its business requirements.
- Capable signifie que la variation du processus est suffisamment faible pour que si la
moyenne du processus est correctement définie, pratiquement toutes les mesures
du processus répondront à ses exigences commerciales.
- A process could be capable, but not currently performing well relative to its
requirements because the process average is off target.
- the common cause variation fits within the specification limits.
-

Process capability analysis :

- Process capability
analysis’ is the primary
statistical tool used to
evaluate capability once
the process is stabilized!

Process stability analysis:

- For processes that are
stable, we often wish to
evaluate the degree to which the current process will satisfy business requirements,
assuming the process continues to produce stable output.
- Process capability analysis helps accomplish this.
- Therefore, it should be applied after performing process stability analysis and finding
the process to be stable.

36
Arjan Imeri Sacha Didi

Specifications/targets/business requirements:
- The business requirements, specifications, or targets are externally imposed to the
process, e.g., by technical specifications, customer expectations or requirements.
- They are chosen to reflect what the customer wants.
o Example : Si le client souhaite que la porte du réfrigérateur se ferme
correctement et hermétiquement à chaque fois, des spécifications sont
définies sur des facteurs tels que la taille de la porte, le mécanisme de
charnière et les dimensions du cadre afin de répondre à cette exigence.
- Such business requirements determine the specification limits, i.e., the ‘Upper
Specification Limit’ (USL) and the ‘Lower Specification Limit’ (LSL).
o The specification limits indicate the ‘acceptable’ range over which the process
should perform.
o Par exemple, le processus (a) délivre
une moyenne trop élevée et devrait
être réduit pour répondre aux
spécifications.

o Any process output between the specification limits is considered as ‘good’

and anything outside as ‘bad’:

“Voice of the customer”:

- Specification limits are the ‘voice of the customer’.
o They are not determined by process data!

“Voice of the process”:

- Control limits are about the performance of the process.
o They are the “voice of the process”.
o Control charts tell what state the process is in, i.e., what the process is doing
right now.
o They are calculated from process data!

“Do not confuse control with capability”

- A process in control is stable over time.
o However, there is no guarantee that a process in control produces products
(or services) of satisfactory quality, i.e., as externally imposed by
specifications.
- Capability has nothing to do with control — expect for the very important point that
if a process is not in control, it is hard to tell if it is capable or not.

37
Arjan Imeri Sacha Didi

o If a process that is in control , example: a process that is stable, does not have
adequate capability, fundamental changes in the process are needed.
 Capability relates the actual performance of a process in control, i.e.,
after special causes have been removed, to the desired performance.
- There is no mathematical or statistical relationship between specification limits and
control limits!

“Specification and control limits”

Do not draw specification limits on any control chart!
- It is a dangerous practice that can badly mislead you.

- SO, specification limits PAS EQUAL à CONTROL LIMITS.

Are all the data within the specification limits of 113 ± 0.5?
- No.
- Roughly 30% of the basket’s weights are outside the specification limits.
- In contrast, all the points on the X ̄ chart are within (or on) the specification limits.

How can one be misled if specification limits are drawn on the X ̄ chart?
- The fact that all subgroup means are within
the specification limits tells nothing about
how many (if any) of the individual
observations are within specification.

38
Arjan Imeri Sacha Didi

- Even on an individual’s control chart including specification limits is inappropriate

and can be misleading.

Process capability ratios

Capability ratio:
- Formal capability analysis, which generally means calculation of capability ratios, is
performed to quantify how well the current process meets or could meet its
business requirements, i.e., specifications.
- By plotting the data of key variables versus their specifications, we can determine
graphically whether we have a problem.
- Calculation of capability ratios supplies a number to document how well we are, or
can be, doing.

Benefit of process capability ratios:

- They make discussion of process capability more rational and less dependent on
opinion (HIPPO).
- However, always plot the data prior to using capability ratios!
o This will avoid overlooking obvious issues such as extreme outliers or special
causes.
o Capability ratios are calculated after having collected data over a ‘reasonable’
amount of time.

Calculation: Cp:

- - Large values of Cp indicate more

capable processes.
- Estimate of Cp is obtained by replacing σ with its estimate.
- When the capability ratio is larger than 1, the process meets the specifications, and
we rarely expect to see data outside of the specifications.

39
Arjan Imeri Sacha Didi

Example:
- If the capability ratio is 1.0, the process variation exactly ‘fits’ within the specification
limits, and we would rarely expect to see data outside the specifications.
- If the process characteristic has a normal distribution, then a Cp of 1.0 means that
0.27% of the data are outside the specifications, i.e., are ‘bad’ or nonconforming.
- For a Cp of 1.33 this percentage is 0.00636%.

Example:
- If the line representing the specification width is longer than the line representing
the process width, this would correspond to a capability ratio greater than 1.0,
indicating that the process can meet the specifications and that there is a very low
probability of seeing data outside the
specification limits.

Cpk:
- A similar ratio that looks at the location of
the average, in addition to the variation, is
called Cpk.
- This is a logical extension of Cp because
one could have a Cp greater than 1.0 but be experiencing data outside the
specifications because the average is off-center.
- Cpk can be described to measure the ratio of the amount of room needed to the
amount of room available to produce product within specifications. If this ratio
equals 1, then roughly speaking, there is just enough room to do the job. If the ratio
is less than 1, more room is needed, and a higher proportion of out-of-spec holes will
be produced. If Cpk is greater than 1, there is extra room, and a higher proportion of
holes will be within specifications.
o Cpk peut être décrit comme un moyen de mesurer le rapport de la quantité
d'espace nécessaire à la quantité d'espace disponible pour produire un
produit conforme aux spécifications. Si ce rapport est égal à 1, alors grosso
modo, il y a juste assez de place pour effectuer le travail. Si le rapport est
inférieur à 1, plus d'espace est nécessaire et une proportion plus élevée de
trous hors spécifications sera produite. Si Cpk est supérieur à 1, il y a pièce, et
une proportion plus élevée de trous sera conforme aux spécifications.

40
Arjan Imeri Sacha Didi

- The Cpk of an out-of-control process provides no information about what can be

expected precisely because nothing can be expected.
o Cpk is not valid in the presence of special causes.

Confidence intervals for process capability ratios

- Much of the use of process capability ratios se concentre sur le calcul et
l'interprétation de l'estimation ponctuelle de la quantité souhaitée.
o It is easy to forget that, for example, Cpk is simply a point estimate and, as
such, is subject to statistical fluctuation.
o It should become standard practice to report confidence intervals for process
capability ratios.

- When a system is stable, telling the worker about mistakes is only tampering.

Deming’s funnel experiment

- « L’expérience en entonnoir » pour décrire les effets néfastes de la « falsification »
d'un processus (stable) en y apportant des modifications sans d'abord effectuer une
étude approfondie des causes possibles de la variation de ce processus.
- Dans l'expérience, une bille est lâchée à travers un entonnoir sur une feuille de
papier qui contient une cible.
- L'objectif du processus est d'arrêter la bille le plus près possible de la cible.
- L'expérience utilise plusieurs méthodes pour tenter de manipuler l'emplacement de
l'entonnoir pour atteindre l'objectif.
- « Si quelqu'un ajuste un processus stable à un résultat indésirable ou à un résultat
très bon, le résultat qui s'ensuit sera pire que s'il avait laissé le processus seul. »
o Attempting to adjust a stable process will make things worse!

The lessons learnt (or ‘rules’) of the experiment:

- Adjusting a process when a part is out of specifications
- Making changes without the aid of control charts
- Changing company policy based on the latest customer attitude survey
- Relying on history passed down from generation to generation.

41
Arjan Imeri Sacha Didi

10: Overall business process improvement strategy

- Prior to implementing any process improvement strategy, one should define the
scope and objectives for the improvement effort.
- As the analysis of special cause variation differs from the one of common cause
variation, the emphasis should be on removing special cause variation first.
- The following overall process improvement strategy — an enhanced PDCA approach
to improvement — shows that continuous improvement occurs in iterative
sequential steps.
If a process is stable (process improvement) (no special cause variation), i.e., showing only
common causes of variation.
- Use tools of improvement to study all the data (i.e., not just the ‘good’ or the ‘bad’
points) and identify factors that cause variation.
- Determine what needs to be permanently changed to achieve a different level of
quality.
- Time would be wasted looking for something ‘broken’ in order to ‘fix’ it.

First step: Eliminate special cause variation using problem-solving strategy

If a process is not stable, i.e., showing signs of special cause.
- Doing a process improvement framework: Immediate opportunities for improvement
would be missed.
- Use tools of problem solving to identify the potential root causes.
- Try to identify exactly when, where how and why the process changed.
- If a special cause hurts the process, develop procedures to eliminate the return of
the problem.
- If a special cause is beneficial, develop procedures to make it a permanent part of the
process.
- Because special cause variation is atypical, we can often eliminate the causes without
fundamentally changing the process.
- Using a problem-solving strategy, we identify what was different in the process when
it produced the unusual result.
o Example (Printing press, continued). A printing press is running at 10% waste
this month when typical waste varies from 1% to 3%.
 They may have purchased their raw paper stock from a new supplier
that month.
 Special cause variation is also called assignable cause of variation.
Problem solving strategy to eliminate special causes:
- Even if we eliminate the special cause variation through problem solving, we will not
fundamentally change the process.
o However, we are left with common cause variation.
o To make further improvement, we need to fundamentally change the system,
i.e., we need to improve the overall process.

42
Arjan Imeri Sacha Didi

- Reducing the inherent common cause variation typically requires studying the
process because there are no unusual results to investigate. C’est une perte de
temps.
o Example (Printing press, continued). In other words, there is no single,
identifiable reason waste was high one day and low the next.
o En d'autres termes, il n'y a pas de raison unique et identifiable pour laquelle
les déchets étaient élevés un jour et faibles le lendemain.
- Using the process improvement strategy, we study the common cause variation and
try to discover the process factors that are the largest sources of variation.
The problem-solving framework begins with a ‘document the problem’ phase. The process
improvement framework does not begin this way, but rather begins with an ‘understand the
process phase’. Which of the following are valid reasons for this difference?
- In problem solving, the focus is on the specific special cause at hand, and not on the
overall process.
- To intervene in a process to effectively improve it, we need to understand the entire
process, not just a specific problem.
- When applying the problem-solving framework, we are more narrowly focused than
when applying the process improvement framework.

Overall improvement process strategy: Steps and tools:

Tools not explicitly mentioned so far of the ‘Analyze common cause variation’ step:
‘Stratification’ : définissez un « facteur de stratification » tel que le jour de la semaine, la
machine ou l'unité commerciale.
- Partition (diviser) the factor into logical categories.
o Comparez les données pour chaque catégorie pour mettre en évidence les
différences et pour que les modèles puissent être vus ;
o Example. Patterns on a control chart may suggest the need for stratification. A
sequence of points with small up and down variation relative to the control
limits may suggest that the sequence of points comes from a single strata.

43
Arjan Imeri Sacha Didi

o The opposite situation where a sequence of points that do not have values
near the center line may indicate the
combination of two strata.
‘disaggregation’: define measures for subprocesses or
individual process steps.
- Study the variation in the individual subprocesses.
- How does it contribute to the overall process
variation?
‘Regression analysis’: existing process knowledge might
suggest one or more variables (‘inputs’) that influence the
process measure (‘output’).
- A regression analysis might ‘verify’ this opinion or
indicate that these variables have negligible ‘effect’.
o (Indiquer que ces variables ont un « effet »
négligeable.)

The regression equation Y = f(X) + ε

44
Arjan Imeri Sacha Didi

Tools of the ‘Study cause-and-effect

relationships’ step not explicitly mentioned so far:
(Statistical) ‘experimental design’: strategically and systematic planned variation of input
factors for an actual process — also known as ‘Design of Experiments’ (DOE);
- The experimenter observes the ‘effect’ of these variations on important process
characteristics.
‘Interrelationship digraph’: evaluate the cause-and-effect relationships between issues to
identify which are the ‘drivers’ and which are the ‘effects.
- This provides a way to process raw ideas that have been generated in ‘brainstorming’
(which was used to rapidly generate a diverse list of ideas or potential root causes)

45
Arjan Imeri Sacha Didi

‘Model building’: construct a (statistical) model of a process that predicts (not explains!)
process performance based on input variables

‘Scatter plot’: (nuage de point) plot of a process characteristic (output) versus a potential
explanatory variable (input)

Box plot’: boxplot to depict the relationship between a discrete variable, such as the region
of a country, and the distribution of continuous variable, such as the profitability.

Overall improvement process strategy: Steps and tools:

Tool not explicitly mentioned so far of the ‘Document the problem’ step:
‘is/is not analysis’: rigorous process for carefully documenting a problem so that
‘true’ root causes can be identified.
- Specifically, it notes when, where and how an issue occurred, as well as when, where,
and how it did not occur, but could have.

Tool not explicitly mentioned so far of the ‘Identify potential root causes’ step:
‘5 Whys’: probe beneath the symptoms of a problem to get down to the root causes.
o Sonder sous les symptômes d'un problème pour aller à la racine cause.

46
Arjan Imeri Sacha Didi

- The intent is to ask ‘Why’ a symptom has occurred, then why the cause occurred and
then why the cause of the cause occurred, and so on.
- Often, we must dig down through as many as 5 causes to get to the root cause.
- It is basically applied common sense, but less rigorous than ‘is/is not analysis’.

Tools not explicitly mentioned so far of the ‘Choose best solutions’ step:
‘multivoting’: voting multiple times.
- Multivoting is used to prioritise a large group of suggestions so that teams can follow
up on those most likely to be ‘fruitful’ first (fructueuse)
‘Affinity diagram’: approach to grouping ideas by organising a collection of ideas into
‘natural’ groupings or categories, so that they can be more effectively addressed.
- It is particularly valuable after a brainstorming exercise that has produced many
ideas.
- The ‘best’ method for handling this is to write ideas, for example, on individual ‘Post-
It’ notes that can be moved easily from spot to spot.
Example. Affinity diagram to organise a list of potential
manufacturing performance indicators:

Process control charts and continuous improvement

The emphasis which must be placed on improvement has
important implications for the way in which process control charts
are applied.
- The process of continuous improvement should be charted
over time and adjustments made to the control charts in
use to reflect the improvements made (in reducing
variation):

47
Arjan Imeri Sacha Didi

48