Total Quality
Management (TQM) and
Statistical Quality
Control
1
Objectives/Agenda
• Introduction to Quality
• Cost of Quality
• Introduction to Total Quality Management (TQM)
• Evolution of TQM
• System Quality Issues (ISO 9000)
• Six Sigma
• Benchmarking
• Seven Basic Tools of Quality
• Statistical Quality Control 2
Quality
• Define Quality
• Engineer: conformance to specifications
• Customer: fitness for use
• Marketing: degree of excellence
• Customer Service: a smaller number of customer complaints
• Dictionary: “a distinctive attribute or characteristic possessed by someone or
something”
3
Quality
• Definition by Pioneers in Quality Management
• Deming: Quality is defined only in terms of the agent
• Juran: Quality is fitness for use
• Crossby: Quality is conformance for requirements
• Figenbaum: Quality means best for certain customer conditions. These conditions
are the actual use and the selling price of the product
• Boeing: Quality is providing our customers with products and services that
consistently meet their needs and expectations
4
Quality
• ISO defines quality as
“The totality of features and characteristics of a product or service that
bear on its ability to satisfy stated or implied needs”
• Alternatively, it is defined as
“compliance with some standards or requirements as mentioned by the clients”
The ability of a product or service “to consistently meet or exceed customer
expectations”
5
Quality in Different Applications
• Airlines: On-time, comfortable, low-cost service
• Health Care: Correct diagnosis, less waiting time, lower cost
• Food Services: Fast delivery, good ambience, good food
• Consumer Products: Durability, defect free, look and feel
• Automotive: Mileage, Safety Features, Comfort
• Utilities like Electricity: Less shut-downs, Proper voltage, safety
6
Dimensions of Quality (Product)
• David Garvin mentioned 8 dimensions of product quality
• Performance
• Aesthetics
• Features
• Conformance
• Reliability
• Durability
• Perceived Quality
• Serviceability
7
Source: https://hbr.org/1987/11/competing-on-the-eight-dimensions-of-quality
Strategic Quality Management (SQM)
“The process of establishing long-run quality goals and defining the
approach to meeting the goals” --- Juran and Gryna
8
Consequences of Poor Quality
9
Consequences of Poor Quality
“NBC Connecticut- 3rd December 2014” “The Full Signal- 09th Nov 2009” “Firstpost- 30th Dec 2016”
“Honda said Wednesday it will expand a recall of autos “Leading mobile handset and accessory “Samsung urged its users to return the
equipped with defective driver's side air bags made by manufacturer Nokia has issued a press device as soon as possible, and
Japan's Takata Corp. to all 50 states, The Associated release this morning warning that announced an over the air update that
Press reported. The announcement comes as Takata, certain Nokia-branded chargers may would limit the charging to sixty percent
automakers and the nation's auto safety watchdog pose an electric shock hazard. of the capacity, to reduce incidents of
appear before a House subcommittee to testify about Customers in possession of these the smartphone catching fire. Samsung
the faulty air bags. Under certain conditions, the air charger models are encouraged to shares started to plummet. The US
bags can expand with too much force, spewing return them to Nokia in exchange for officially recalled the Samsung Galaxy
shrapnel at passengers. They have been linked to at repaired units.” Note 7. The replacement devices were
least five deaths, yet Takata has so far declined to to have a green battery icon to indicate
broaden the recall nationwide despite a demand from that they were replaced and safe to use.
the National Highway Traffic Safety Administration.” In the United States, it was the largest
http://www.nbcconnecticut.com/news/national-international/NATL- http://www.thefullsignal.com/products/2951/nokia-
Takata-Tosses-Recall-Ball-Into-Automakers-Court-284602571.html issues-recall-14-million-defective-cell-phone-chargers fire and burn hazard related recall ever.”
https://www.firstpost.com/tech/news-analysis/the-samsung-galaxy-note-7-fiasco-a-blow-by-
blow-account-3695113.html#:~:text=On%20August%2024%2C%20a%20Samsung,it%20would
%20recall%20the%20smartphones. 10
Consequences of Poor Quality
• Loss of Business
• Liability
• Productivity
• Costs
11
3 Aspects of Quality
• Quality of Design: Inherent value of the Quality of Design
product in the marketplace
• Quality of Conformance: Degree to which the
product or service design specifications are met
Quality of Conformance
• Quality of Performance: Degree to which the
product or Service satisfies the customer from
the perspective of both quality of design and
the quality of conformance
Quality of Performance
12
Costs of Quality
• Prevention Costs: Costs of all the activities directed towards preventing defects
from happening
• Appraisal Costs: Costs of all activities concerned with the appraisal of
products/services by measuring, evaluating or auditing
• Failure Costs
• Internal Failure Costs: Costs related to the failure of a product prior to its dispatch to the
consumer
• External Failure Costs: Costs incurred by a company after the product has been delivered to
the customer and product failure has taken place
13
Total Quality Management
• ‘Total’--- Made up of the whole
• ‘Quality’--- Degree of excellence a product or service provides
• ‘Management’--- Act, art or manner of planning, controlling, directing etc.
• Thus, TQM is the “art of managing the whole to achieve excellence”
• In other words, it refers to “a quest for quality in an organization”
14
Total Quality Management
• There are three key philosophies in this approach
• Continuous Improvement
• Involvement of everyone
• Customer Satisfaction, i.e., meeting or exceeding customer expectations
15
Goal of TQM
“Do the right things right, the first time and every time”
16
Introduction to Six Sigma
17
Six Sigma
• A process improvement tool
• First introduced in Motorola in 1980
• Six Sigma allows managers to readily describe process performance using
a common metric: Defects Per Million Opportunities (DPMO)
18
Six Sigma Example
• Consider a situation when out of 300000 ball bearings produced, 200 are
defective. Calculate the DPMO.
• Sol: DPMO= (200*1000000)/(300000*1)=667
• Consider a situation when out of 200000 problem tickets raised by the
client, an IT company fails to fix 150 problem tickets correctly. Calculate
the DPMO.
• Sol: DPMO= (150*1000000)/(200000*1)=750
19
Six Sigma
“The objective of six sigma process is to ensure that there are no
more than 3.4 defects for every million opportunities to produce
some feature of a part”
20
Why the terminology “Six Sigma”
• What is sigma?
• How does 3.4 defects for every million opportunities linked to sigma?
21
DMAIC Cycle
22
Six Sigma Quality: DMAIC Cycle
• Define, Measure, Analyze, Improve, and Control (DMAIC)
• Developed by General Electric as a means of focusing effort on quality
using a methodological approach
• Overall focus of the methodology is to understand and achieve what the
customer wants
• A 6-sigma program seeks to reduce the variation in the processes that lead
to these defects
• DMAIC consists of five steps
23
DMAIC
• Define: Customers and their priorities
• Measure: Process and its performance
• Analyse: Causes of defects
• Improve: Remove causes of defects
• Control: Maintain Quality
24
DMAIC
• Problem: There is a formula milk manufacturing company. In every box,
the weight is supposed to be 500 grams. The government has specified the
allowable shortfall of no more than 5 grams per box. It is however,
reported that on an average there is 496.5 grams of formula milk present
in every box.
25
DMAIC-- Define
• Define: What is the critical-to-quality (CTQ) characteristic?
• The weight of formula milk in a box
26
DMAIC -- Measure
• Measure: How we can measure the extent of the problem?
• Acceptable limit is 495 grams
• LTL= 495
• Similarly, we define UTL= 505
• We go the market and take 100 such boxes and found that the average
weight is 496.5 grams with a standard deviation of 5.5 grams.
• What percentage of boxes are outside the tolerance limits?
27
DMAIC -- Measure
• Our issue is rather to find what proportion of boxes have a weight lower
than 495 grams?
• z=(495-496.5)/5.5=-0.2728
• Norm.Dist(-0.2728, 0,1,TRUE)=0.384
• 38.4% of the boxes fall outside the LTL
28
DMAIC -- Analyze
• How can we improve the capability of our processes that fills the formula
milk into the boxes?
• The goal is to Decrease Variation in the weight of the boxes
29
DMAIC -- Improve
• How good is good enough?
• Motorola’s “Six Sigma”
• 6σ minimum from process center to nearest spec
30
DMAIC -- Control
• Statistical Process Control (SPC)
• Use data from the actual process
• Estimate distributions
• Look at capability - is good quality possible
• Statistically monitor the process over time
31
PDCA Cycle
32
Plan-Do-Check-Act (PDCA) Cycle
• It is an iterative method used in businesses for the controlling of continuous
improvement of processes and products
• Also known as Deming cycle
• Plan- Establish objectives and processes required to deliver the desired results.
• Do- Execution of the process
• Check- Gather data and check the conformance
• Act-Adjust the processes if found greater deviations
33
Benchmarking
34
Benchmarking
“A process of comparison of company’s own processes and
performance metrics with some standards or with the processes
and performance metrics with other companies”
“As is” “To be”
35
Classification of Benchmarking
• Benchmarking can be classified • American Express (for billing and collection)
• Cummins Engines and Ford (for factory floor layout)
(on the basis of unit) as: • Florida Power and Light (for quality improvement)
• Honda (for supplier development)
• Internal Benchmarking • Toyota (for quality management),
• Hewlett-Packard (for research and product
• Competitive Benchmarking development)
• Functional Benchmarking • Saturn (a division of General Motors) and Fuji
Xerox (for manufacturing operations)
• Generic Benchmarking • DuPont (for manufacturing safety)
36
Benchmarking Wheel
• The purpose of benchmarking process models is to describe the steps that
should be performed when conducting a benchmarking study
37
Benchmarking Wheel
Plan
Improve • Determine the process to benchmark based on
• Plan the implementation of the organization’s critical success factors
improvements • Understand and document own process
• Implement improvements • Measure performance of own process
and monitor the
implementation progress
Analyze
• Identify gaps in Find
performance and the • Identify benchmarking
root causes for the gaps partners
Collect
• Understand and
document the
benchmarking
partners’ performance
and practice
38
Seven basic tools of quality
39
Seven basic tools of quality
Tool Pictorial Presentation Purpose
Type
Check Sheet Range Data Collection
Makes Process Steps
Flow Chart
Visible
Histogram Frequency Distribution
Pareto Diagram Trivial Many/Vital Few
Cause & Effect Root Cause Analysis
Scatter Diagram Correlation
Control Charts Controllability 40
Data Collection Forms/Checksheets
Helps to standardize the data
collection
The same data can be used
to create histograms
41
Flow Chart
The given flowchart depicts
the process steps as a part
of a SIPOC (Supplier,
Inputs, Processes, Outputs
and Customers) analysis
It is used in the Define
stage of the analysis
42
Pareto Chart
Break down the problem into
relative contributions of its
components
The idea is that large percentage
of the problems are due to a
small percentage of causes
43
Fishbone Diagram
Shows hypothesized
relationships between
potential causes and
the problem under
study.
44
Opportunity Flow Diagram
Used to separate
the value-added
from the non-value
added steps in the
process
45
Control Chart
Time sequenced charts
depicting the plotted value of
a statistic including a
centreline average and one or
more control limits
46
Run Chart
Depicts trends in data over
time, and help us in
understanding the magnitude
of the problem in the define
stage itself
Typically, they plot the
median of the process
47
Statistical Quality Control
48
Statistical Quality Control
• Quantitative aspects of Quality Management
• Processes usually exhibit variations in their output
• Variations that are caused by factors clearly identifiable and possibly managed
are called assignable variations
• A poorly trained employee that creates variation in finished product output
• Variation that is inherent in the process is called common variation
• A molding process that always leaves “burrs” or flaws on a molded item
49
Taguchi’s View of Variation
• Traditional view is that quality within the Lower Specs and Upper Specs
is good and that the cost of quality outside this range is constant
• Taguchi views costs as increasing as variability increases, so seek to
achieve zero defects and that will truly minimize quality costs
50
Taguchi’s View of Variation
High
Incremental
Cost of
Variability
Zero
Lower Target Upper
Spec Spec Spec
51
Process Capability Index
• As a production process produces items small shifts in equipment or
systems can cause differences in production performance from differing
samples
• Shows how well parts being produced fit into design limit specifications
• It is measured as follows (depending upon whether the process means are
centered between LSL and USL
CP = (USL – LSL) / 6
52
Process Capability Index
• Problem: There is a formula milk manufacturing company. In every box,
the weight is supposed to be 500 grams. The government has specified the
allowable shortfall of no more than 5 grams per box. It is however,
reported that on an average there is 496.5 grams of formula milk present
in every box.
• Calculate the Process Capability Index
53
Control Charts
54
Control Charts
• It is a tool to determine whether any manufacturing process is in a state of
control
• It mainly consists of:
• Different points representing a statistic (like, mean, range, proportion) of
measurements of any quality characteristic in samples drawn from the
manufacturing process at different times
• A center line
• UCLs and LCLs
55
Control Charts
• Mean Chart (Bar chart) measures central
tendency of a process
• Mean Control Chart or chart
• If the process standard deviation is known,
then chart is defined as
• and
where n is the sample size of each sample
56
Mean Chart
• A quality inspector took 5 samples 1 2 3 4 5
each with 5 observations, measuring A 2.25 2.24 2.30 2.28 2.34
the diameter of ball bearings. The B 2.38 2.23 2.34 2.23 2.21
same has been shown in the table
C 2.21 2.31 2.26 2.38 2.29
below , where each observation is
D 2.36 2.20 2.29 2.35 2.30
taken in millimetres (mm). Find 3
E 2.31 2.33 2.35 2.30 2.36
control limits for diameters. It is
known that the standard deviation
for the process is 0.1 mm.
57
Solution
Sample
Means 2.302 2.262 2.308 2.308 2.3
UCL= 2.296+ 3*0.1/√5 = 2.4302
LCL= 2.296- 3*0.1/√5 = 2.1618
58
Sample Range for Process Variability
59
60
Mean Chart
• In the previous example, determine control limits
• Sol: Average of Range = 0.138
• A2 for a sample of size 5 = 0.577
• UCL= 2.376, LCL = 2.216
61
Range Chart
• For monitoring process dispersion
62
Range Chart
• In the previous example, determine control limits for sample range
• D3=0, D4= 2.11 for n=5
• Average R = 0.138
• UCL= 0.2912
• LCL= 0
63
Control Charts for attributes: p-Chart
• p-charts (prepared for fraction of defectives)
• where is the fraction of defectives
• and or 0
64
p-Chart
Construct a control chart Sample n Defectives
that will describe 99.5% of 1 100 4
the process variation. 2 100 2
3 100 5
4 100 3
5 100 6
65
p-Chart
• Here
4
=0.0196
66
Control Charts: c-Chart
• c-charts used to control the number of occurrences
• and or 0
• In the previous example,
67
Control Charts
Acceptance Sampling
69
Statistical Quality Control: Lot Acceptance
Sampling
• Popularized by Dodge and Romig
• It is a sampling scheme and a set of rules for making decisions
• Acceptance, Rejection, or indecisiveness (or repeat sampling)
are based on the count of defectives obtained in any sample(s)
70
Lot Acceptance Sampling
• Lot Acceptance Sampling or Acceptance Sampling is used under any of the
three following cases:
Testing is destructive
The cost of full inspection is very high
Time taken for full inspection is too long
• Single Sampling Plan
• Double Sampling Plan
71
Single Sampling Plan
• A sample of size ‘n’ is selected
• Once the number of defects present in the sample exceeds a fixed pre-
chosen no. ‘c’, the entire batch is rejected
• Else the entire batch is selected
72
Double Sampling Plan
• A sample of size ‘n1’ is selected first
• If the no. of defects exceeds ‘c1’, the batch is rejected
• If the no. of defects is less than ‘c2’, the batch is accepted, where c1>c2
• If the no. of defects fall within (c2, c1), then another sample of size ‘n2’ is
drawn and the no. of defects in the new sample are checked
• If the sum of defects in both the samples exceed c3, batch is rejected else
batch is accepted
73
Acceptance Sampling
• Acceptable Quality Level (AQL)
• Lot Tolerance Percent Defective (LTPD)
• Risks in Acceptance Sampling are of two types:
• Type I Error or Producer’s Risk (α)
• Type II Error or Consumer’s Risk (β)
74
Example
• Producer of electric bulb claims of having 5% defective bulbs in a lot of
1000. A consumer devices a single sampling plan where he draws a
sample of size 50 (with replacement). If he finds more than three defective
bulbs, he rejects the entire lot, else, he accepts the lot. Calculate the
probability of accepting the batch.
• Sol: In Excel use, BINOM.DIST(3,50,0.05,TRUE)
• 0.7844
75
Example
• Producer of electric bulb claims of having 4% defective bulbs in a lot of
1000. A consumer devices a single sampling plan where he draws a
sample of size 50 (without replacement). If he finds more than three
defective bulbs, he rejects the entire lot, else, he accepts the lot. Calculate
the probability of accepting the batch.
• Sol: In Excel use, HYPGEOM.DIST(3,50,40,1000,TRUE)
• 0.8658
76
Operating Characteristics Curve
• It helps a sampling plan to discriminate between good lots and bad lots
• It is a plot between % defectives in a lot or batch vs. the probability of
acceptance of the lot
• The degree to which a sampling plan can discriminate between good lots
and bad lots is a function of the steepness of the graph’s OC curve
77
Operating Characteristics Curve
P (of P P (of P
defectives) (acceptance) defectives) (acceptance) OC Curve
1.000
0.000 1.000 0.090 0.021 0.900
0.800
0.010 0.953 0.100 0.011 n=80, c=2
0.700
0.020 0.784 0.110 0.005 0.600
0.500
0.030 0.568 0.120 0.003
0.400
0.040 0.375 0.130 0.001 0.300
0.200
0.050 0.231 0.140 0.001 0.100
0.000
0.060 0.134 0.150 0.000
0.070 0.075 0.160 0.000
0.080 0.040 78
P (acceptance)
Operating Characteristics Curve: Sampling Plan
c LTPD/AQL n. AQL
Consider a situation where the vendor has an 0 44.890 0.052
AQL of 2% defectives and is okay with a 1 10.946 0.355
producer’s risk of 5%. Further, the buyer 2 6.509 0.818
considers 8% or more defectives as unacceptable 3 4.890 1.366
and wants to ensure that it has a consumer’s risk 4 4.057 1.970
ok no more than 10%. Using the given 5 3.549 2.613
information calculate ‘n’ and ‘c’ 6 3.206 3.286
AQL=0.02, LTPD=0.08, α=0.05, 7 2.957 3.981
8 2.768 4.695
β=0.1 9 2.618 5.426 79
