Operations and Supply Chain Management MGMT 3306 Lecture 04 Instructor: Dr. Yan Qin

Outline – Managing Quality What is Quality Cost of Quality (COQ) International Quality Standards 7 Concepts of Total Quality Management Statistical Process Control Variations in processes Process capability Process control charts

What is Quality? David Garvin, in his book Managing Quality, summarized five principal approaches to defining quality: Transcendental view: “I can’t define it, but I know when I see it”; Product-based view: Quality is viewed as quantifiable and measurable characteristics or attributes; (Design quality) User-based view: Quality is an individual matter, and products that best satisfy their preferences are those with the highest quality;

What is Quality? Five principal approaches to defining quality (Cont.) Manufacturing-based view: “conformance to requirements” (Conformance quality) Value-based view: Quality is defined in terms of costs and prices as well as a number of other attributes.

Two Ways Quality Improves Profitability Improved response Flexible pricing Improved reputation Sales Gains via Improved Quality Increased Profits Increased productivity Lower rework and scrap costs Lower warranty costs Reduced Costs via

Quality Program Fundamental to any quality program is the determination of quality specifications, and the costs of achieving (or not achieving) those specifications.

Quality Specification Design Quality measures how well a product meets customer expectation. Specifications of Design Quality Functions/features intended to deliver Reliability/durability Serviceability Aesthetics Conformance Quality measures how well design specifications are met in production.

Cost of Quality Cost of Quality refers to all of the costs attributable to the production of quality that is not 100% perfect. It is estimated that the cost of quality is between 15% and 20% of every sales dollar. (Philip Crosby: <2.5%) Three basic assumptions: Failures are caused; Prevention is cheaper; Performance can be measured. Cost of poor quality

Four types of costs Total Cost Total Cost External Failure Quality Improvement Total Cost External Failure Internal Failure Prevention Appraisal

More Quality Definitions Fitness for use: Design quality + Conformance quality Quality Management is the totality of functions involved in the determination and achievement of quality. Quality Control focuses on the process of producing the product or service with the intent of eliminating problems that might result in defects. Quality Assurance is a complete system to assure the quality of a product.

International Quality Standards Series of standards agreed upon by the International Organization for Standardization (ISO) Adopted in 1987 More than 160 countries A prerequisite for global competition? ISO 9000 an international reference for quality, ISO 14000 is primarily concerned with environmental management

International Quality Standards ISO 9000 series Common quality standards for products sold in Europe (even if made in U.S.) 2008 update places greater emphasis on leadership and customer requirements and satisfaction ISO 14000 series (Europe/EC) Core elements include Environmental management; Auditing; Performance evaluation; Labeling; Life cycle assessment

Total Quality Management is the management of an entire organization so that it excels in all aspects of goods and services that are important to the customer.

Deming’s 14 Points Quality expert Edwards Deming used 14 points to indicate how he implemented TQM: Create consistency of purpose Lead to promote change Build quality into the product; stop depending on inspections Build long-term relationships based on performance instead of awarding business on price Continuously improve product, quality, and service

Deming’s 14 Points – Cont. Start training Emphasize leadership Drive out fear Break down barriers between departments Stop haranguing workers Support, help, and improve Remove barriers to pride in work Institute education and self-improvement Put everyone to work on the transformation

Seven Concepts of TQM The authors of the textbook developed Deming’s 14 points into seven concepts for an effective TQM program: Continuous improvement Six Sigma Employee empowerment Benchmarking Just-in-time (JIT) Taguchi concepts Knowledge of TQM tools

1. Continuous Improvement Represents continual improvement of all processes . Walter Shewhart developed a circular model know as Plan-Do- Check-Act as his version of continuous improvement. PDCA Model

2. Six Sigma Originally developed by Motorola, adopted and enhanced by Honeywell and GE. Six-sigma is a philosophy and methods used to eliminate defects by reducing the variation in the processes. Statistically, sigma is a measure of variation. In business processes, it is a measure of how many defects or failures are likely to occur per million opportunities. 4 sigma = 6,000 defects per million 6 sigma = 3.4 defects per million

Metric in 6-sigma The calculation of Defects per million opportunities (DPMO) requires three pieces of information: Unit: the item produced or being serviced; Defect: Any item or event that does not meet the customer’s specification limits; Opportunities: an chance for a defect to occur

Examples: DPMO Example 1: The customers of a mortgage bank expect to have their mortgage applications processed within 10 days of filing. Suppose all defects are counted and it is determined that there are 150 loans in the 1,000 applications processed last month that don’t meet this customer requirement. Then what is the DPMO in this case? Example 2: Now suppose the production of a product consists of two operations. Defects can occur at each operation. There were 100 defective units out of the total 2,000 units produced. What was the DPMO in this case?

Examples: DPMO – Calculations 𝐷𝑃𝑀𝑂= 150 1×1,000 ×1,000,000=150,000 Example 2: 𝐷𝑃𝑀𝑂= 100 2×2,000 ×1,000,000=25,000

6-sigma Methodology Uses many of the same statistical tools as other quality control programs. The difference is that the tools are used in a systematic project-oriented fashion through define, measure, analyze, improve, and control (DMAIC) cycle. The focus of DMAIC is understanding and achieving what the customer wants.

DMAIC Define critical outputs and identify gaps for improvement Measure the work and collect process data Analyze the data Improve the process Control the new process to make sure new performance is maintained DMAIC Approach

6-sigma Implementation Emphasize defects per million opportunities as a standard metric Provide extensive training Focus on corporate sponsor support Create qualified process improvement experts (Black Belts, Green Belts, etc.) Set stretch objectives This cannot be accomplished without a major commitment from top level management

3. Employee Empowerment Getting employees involved in product and process improvements 85% of quality problems are due to process and material Techniques include: Build communication networks that include employees Develop open, supportive supervisors Move responsibility to employees Build a high-morale organization Create formal team structures

Use internal benchmarking if you’re big enough Selecting best practices to use as a standard for performance. Steps include: Determine what to benchmark Form a benchmark team Identify benchmarking partners Collect and analyze benchmarking information Take action to match or exceed the benchmark Use internal benchmarking if you’re big enough

Best Practices for Resolving Customer Complaints Justification Make it easy for clients to complain It is free market research Respond quickly to complaints It adds customers and loyalty Resolve complaints on first contact It reduces cost Use computers to manage complaints Discover trends, share them, and align your services Recruit the best for customer service jobs It should be part of formal training and career advancement

5. Just-in-Time (JIT) JIT (Just-in-Time) is an approach of continuous and forced problem solving via a focus on throughput and reduced inventory. Throughput is the time required to move orders through the production process, from receipt to delivery. Basically it means making only what is needed at the time when it is need!

6. Taguchi Concepts Engineering and experimental design methods to improve product and process design Identify key component and process variables affecting product variation Taguchi Concepts Quality robustness: Ability to produce products uniformly in adverse manufacturing and environmental conditions Quality loss function Target-oriented quality: a Philosophy of continuous improvement to bring the product exactly on target

Target-oriented quality Quality Loss Function Shows that costs increase as the product moves away from what the customer wants Costs include customer dissatisfaction, warranty and service, internal scrap and repair, and costs to society Traditional conformance specifications are too simplistic Target-oriented quality

7. Tools of TQM Tools for Generating Ideas Check sheets Scatter diagrams Cause-and-effect diagrams Tools to Organize the Data Pareto charts Flowcharts Tools for Identifying Problems Histogram Statistical process control chart

1. Check Sheet An organized method of recording data / Hour Defect 1 2 3 4 5 6 7 8 A B C / // /// // / / / /// / // /// // //// / According to the check sheet, there were 3 occurrences of defect A, 2 occurrences of defect B, and 1 occurrence of defect C in the first hour of operation.

2. Scatter Diagram A graph of the value of one variable vs. another variable Productivity decreases as Absenteeism increases.

3. Cause-and-Effect Diagram A tool that identifies process elements (causes) that might effect an outcome Causes are classified into 4 categories in this diagram.

Example: Cause-and-Effect

4. Pareto Chart A graph to identify and plot problems or defects in descending order of frequency Bars represent frequencies and the dashed curve represents cumulative probability.

Example: Pareto Chart

5. Flowchart (Process Diagram) A chart that describes the steps in a process

Example: MRI Flow Chart Physician schedules MRI Patient taken to MRI Patient signs in Patient is prepped Technician carries out MRI Technician inspects film If unsatisfactory, repeat Patient taken back to room MRI read by radiologist MRI report transferred to physician Patient and physician discuss

6. Histogram A distribution showing the frequency of occurrences of a variable

7. Statistical Process Control Chart A chart with time on the horizontal axis to plot values of a statistic Upper control limit Target value Lower control limit Time

Statistical Process Control (SPC) Statistical Process Control is a number of statistical techniques designed to evaluate quality from a conformance view. Situations where SPC can be applied: How many paint defects are there in the finish of a car? How long does it take to execute market orders? How long do customers wait to be served from our drive- through window?

Variations around us All processes have inherent statistical variability which can be evaluated by statistical methods. Basic forms of Variations: Assignable variation: caused by factors that can be clearly identified and possibly managed, such as workers inadequately trained. Common variation: also called random variation, variation that is inherent in the production process. It is usually a result of the type of equipment used in the process.

USL and LSL It is impossible to have zero variation! Engineers therefore assign acceptable limits for variation. The limits are known as the upper and lower specification limits (USL/LSL). Also known as upper and lower tolerance limits (UTL/LTS); The range defined by USL and LSL is called the specification range.

Example: USL and LSL Consider the production of boxes of cereal. Suppose the target value of the weight of each cereal box is 12 oz. The design specifications might be 12 +/- 2 oz. This tells the production department that the production should aim for the weight of 12 oz. However, any weight between 10 and 14 oz. is also acceptable. In this case, USL = 14, LSL = 10.

Process capability Process capability is the long-term performance level of the process after it has been brought under statistical control, that is, having only common variation. There are two popular measures for determine process capability: Process Capability Ratio (Cp) Process Capability Index (Cpk) Both of them are measured based on samples taken from production or test runs.

Process Capability Ratio (Cp) The process capability ratio is calculated as: 𝑪 𝒑 = 𝑼𝒑𝒑𝒆𝒓 𝑺𝒑𝒆𝒄𝒊𝒇𝒊𝒄𝒂𝒕𝒊𝒐𝒏 𝑳𝒊𝒎𝒊𝒕 −𝑳𝒐𝒘𝒆𝒓 𝑺𝒑𝒆𝒄𝒊𝒇𝒊𝒄𝒂𝒕𝒊𝒐𝒏 𝑳𝒊𝒎𝒊𝒕 𝟔𝝈 𝝈 is the population standard deviation of the quality measure of interest. When only samples are available, we use s instead of 𝝈 in calculation. A capable process must have a Cp of at least 1.0 Six Sigma quality requires a Cp = 2.0

Example: Mean and Sample Std. Deviation Suppose a sample from the production of a product consists of 5 data points summarized as follows: What is the mean and standard deviation of the sample? Data points 9 10 8.8 10.2 5

Example: Solution Notes: In the calculation above, 𝑥 𝑖 represents the ith data point in the sample and n refers to the sample size, that is, the number of data points in the sample.

Example: Cp Suppose for an insurance claims process, Process mean 𝑥 = 210.0 minutes Process standard deviation s = 0.516 minutes Design specification range = 210 ± 3 minutes What is the process capability ratio in this case?

Upper Specification - Lower Specification Example: Solution Cp = Upper Specification - Lower Specification 6s = = 1.938 213 - 207 6(.516)

Process Capability Index (Cpk) To determine process capability, two sets of parameters required: Mean and standard deviation of the quality measure of interest in real production (or in the test run); Target (Aim) value and Design specification limits (USL and LSL)

How to calculate Capability Index ( 𝐶 𝑝𝑘 ) σ Cpk = Minimum of , 𝒙 −Lower specification 3σ Upper specification – 𝒙

How to interpret 𝐶 𝑝𝑘 ? Question 1: What does it tell you if the 𝐶 𝑝𝑘 is determined by LSL or USL? Answer: If the value is determined by LSL (USL), it means that the mean of the sample shifts to the left (right) of the center of the specification range. Question 2: How to interpret a negative/positive 𝐶 𝑝𝑘 value? Answer: A positive value means that the sample mean is within the specification range. A negative value indicates the opposite.

How to interpret 𝐶 𝑝𝑘 ? Question 3: What does it mean when 𝐶 𝑝𝑘 <1 or >1? Answer: When 𝐶 𝑝𝑘 >1, it means the specification limits are larger than the three sigma allowed in the process. All units in the sample are within design limits. Question 4: Is 𝐶 𝑝𝑘 the higher the better? Answer: Yes. “You must have a 𝐶 𝑝𝑘 of 1.33 [4 sigma] or higher to satisfy most customers.” Joe Perito

Example: 𝐶 𝑝𝑘 The quality assurance manager is assessing the capability of a process that puts pressurized grease in an aerosol can. The design specifications call for an average of 60 pounds per square inch of pressure in each can with an upper tolerance limit of 65 psi and a lower tolerance limit of 55 psi. A sample is taken from production and it is found that the cans average 61 psi with a standard deviation of 2 psi. Question: What is the process capability index of the process?

Example: Solution 𝐶 𝑝𝑘 = min 𝑥 −𝐿𝑆𝐿 3𝜎 , 𝑈𝑆𝐿− 𝑥 3𝜎 6 sigma = Cpk = 2 =0.6667

Process Control Charts Sampling by Attributes (Go or no-go information) Defectives refers to the acceptability of product across a range of characteristics. Defects refers to the number of defects per unit which may be higher than the number of defectives. Tools: p-chart (1 defect for each unit), c-chart (>1 defect each unit) Sampling by Variable (Continuous) Amount of deviation from a set standard for a single variable. Tools: X-bar chart and R chart

p-charts In p-charts, we create the control limits for the proportion of defects. We call the limits Upper Control Limit (UCL) and Lower Control Limit (LCL). Plot the sample points and see if they fall within the control limits. If a sample point falls within the control limits, it means that the sample is under statistical control. We usually use 3-sigma control limits.

Using p-charts Let z be the number of standard deviations. For 99.7% confidence z =3. For 99% confidence, z = 2.58. We usually just set z = 3. Fraction defective

Example: p-chart Hometown Bank is concerned about the number of wrong customer account numbers recorded. Each week a random sample of 2,500 deposits is taken and the number of incorrect account numbers is recorded The results for the past 12 weeks are shown in the table on the next slide.(We therefore have 12 samples in this case.) Is the booking process out of statistical control? Use three-sigma control limits.

Example: p-chart (Cont.) Sample Number Wrong Account Numbers 1 15 7 24 2 12 8 3 19 9 10 4 17 5 11 6 Total 147

Example: Solution – Step 1 Step 1: Using this sample data to compute parameters p = Total defectives Total number of observations 147 12(2,500) = = 0.0049 σp = p (1 – p)/n = 0.0049(1 – 0.0049)/2,500 = 0.0014 UCL = p + zσp = 0.0049 + 3(0.0014) = 0.0091 LCL= p – zσp = 0.0049 – 3(0.0014) = 0.0007

Solution – Step 2 & 3 Step 2: Calculate the sample proportion defective. Step 3: Plot each sample proportion defective on the chart, Sample proportion defective = # 𝒐𝒇 𝒅𝒆𝒇𝒆𝒄𝒕𝒊𝒗𝒆𝒔 𝒊𝒏 𝒂 𝒔𝒂𝒎𝒑𝒍𝒆 𝒔𝒂𝒎𝒑𝒍𝒆 𝒔𝒊𝒛𝒆 For sample 1, the proportion of defectives is 15/2,500 = 0.0060.

Solution – Step 3 Fraction Defective Sample Mean UCL LCL .0091 .0049 .0007 | | | | | | | | | | | | 1 2 3 4 5 6 7 8 9 10 11 12 X The p-Chart Showing sample defective 7 is Out of Control

Using c-chart With p charts, each item can only have one defect. With a c chart, each item can have multiple defects.

Example: Lumber yard Lumber yard expects 4 knotholes per eight-foot board. Then,

𝑥 and R charts 𝑥 and R charts are a type of control charts that monitors a variable’s data using multiple samples collected. 𝑥 chart is a plot of the means of the samples. R chart is a plot of the range within each sample, while range is the difference between the highest and the lowest values in each sample. Preferably using 25 or so samples of size 4 or 5.

How to construct 𝑥 and R charts For x-bar chart, Upper Control Limit = 𝑋 + 𝐴 2 𝑅 Lower Control Limit = 𝑋 − 𝐴 2 𝑅 where, 𝑋 is the mean of a sample, 𝑋 = 𝑖=1 𝑛 𝑋 𝑖 𝑛 𝑋 is the mean of all the sample means, 𝑋 = 𝑗=1 𝑚 𝑋 𝑗 𝑚 n is the sample size and m is the number of samples in total 𝑅 𝑗 is the range of sample j. 𝑅 is the mean of all ranges. 𝑅 = 𝑗=1 𝑚 𝑅 𝑗 𝑚

How to construct 𝑋 and R charts For R chart, Upper Control Limit = 𝐷 4 𝑅 Lower Control Limit = 𝐷 3 𝑅 where, 𝑅 is the mean of all ranges of the samples. The values of A2, D3, and D4 are determined by sample size n.

Control chart factors Sample Size Mean Factor Upper Range Lower Range n A2 D4 D3 2 1.880 3.268 0 3 1.023 2.574 0 4 .729 2.282 0 5 .577 2.115 0 6 .483 2.004 0 7 .419 1.924 0.076 8 .373 1.864 0.136 9 .337 1.816 0.184 10 .308 1.777 0.223 12 .266 1.716 0.284

Example: X-bar and R charts The Watson Electric Company produces incandescent light bulbs. The following data on the number of lumens for 40-watt light bulbs were collected when the process was in control. a. Calculate control limits for an R-chart and an X-chart. b. Since these data were collected, some new employees were hired. A new sample obtained the following readings: 570, 603, 623, and 583. Is the process still in control? Observation Sample 1 2 3 4 604 612 588 600 597 601 607 603 581 570 585 592 620 605 595 5 590 614 608

Example: Solution Sample 𝑿 R 1 601 24 2 602 10 3 582 22 4 32 5 604 Total 2,991 112 Average 𝑿 =𝟓𝟗𝟖.𝟐 𝑹 =𝟐𝟐.𝟒

Example: Solution (Cont.) The R-chart UCL= 𝐷 4 𝑅 = LCL= 𝐷 3 𝑅 = 2.282*22.4 = 51.12 0*22.4 = 0 The 𝑿 -chart UCL= 𝑋 + 𝐴 2 𝑅 = LCL= 𝑋 − 𝐴 2 𝑅 = 598.2 + 0.729*22.4 = 614.53 598.2 – 0.729*22.4 = 581.87 b. The range is 53 (or 623 – 570), which is outside the UCL for the R-chart. A search for assignable causes inducing excessive variability must be conducted.

Sample Control charts Normal – No action UCL Variations Nominal LCL Sample number Normal – No action

Sample control chart Run – Take action UCL Variations Nominal LCL Sample number Run – Take action

Sample control chart Sudden change – Monitor UCL Variations Nominal LCL Variations Sample number Sudden change – Monitor

Sample control charts Exceeds control limits – Take action UCL Nominal UCL LCL Variations Sample number Exceeds control limits – Take action