TECH 50800 QUALITY and PRODUCTIVITY in INDUSTRY and TECHNOLOGY Before we begin: Turn on the sound on your computer. There is audio to accompany this presentation. Audio will accompany most of the online presentation materials through out the semester. 1 TECH 50800 QUALITY and PRODUCTIVITY in INDUSTRY and TECHNOLOGY Week 11 Lean Six Sigma Basics: Analyze 2 Lean Six Sigma Analyze Phase Introduction 3
LEAN SIX SIGMA PROCESS Champion Define Measure Lean 101 Analyze Improve Lean 201 Control Identify Opportunities Set Team Ground Rules Current State Process Map Identify Waste Examine Process and Data Finalize KPIVs Identify Constraints Determine Process Control Plan Select Project Voice of the Customer Analysis Potential KPIVs Identified Quick Hit Improvements KPIVs Verified Develop/ Evaluate Solutions Exploit Constraints Recognize/ Reward Complete Project Charter Determine KPOVs Data Collection Link KPIVs to KPOVs Future State Process Map Finalize Financial Estimates Launch Project Team Link KPOVs to CTQs Create Basic KPOV Graphs Pilot Study Implementation Plan Full Implementation Initial Financial Estimates Analyze Phase 4 THE ANALYZE PHASE Analyze Examine Process and Data KPIVs Verified Link KPIVs to KPOVs Analyze Phase Goals Use Data Driven Decision Making techniques Continue to utilize statistical analysis to understand the data. Appropriately apply advanced graphing techniques o Scatter Plots 5 THE ANALYZE PHASE Analyze Examine Process and Data KPIVs Verified Link KPIVs to KPOVs Analyze Phase Steps Examine the process/data Verify or eliminate KPIVs. Conclusively link KPIVs to KPOVs using data analysis and graphing techniques. 6
THE ANALYZE PHASE Analyze Examine Process and Data KPIVs Verified Link KPIVs to KPOVs Analyze Phase Expected Outcomes KPIV data collection completed KPIVs Analysis used to determine critical KPIVs Critical KPIVs conclusively linked to KPOVs 7 Lean Six Sigma Analyze Phase Analyze Tools 8 LEAN SIX SIGMA TOOLS Measure Lean 101 Lean 201 Champion Define Analyze Improve Control Affinity Diagram Ground Rules Worksheet Process Mapping Process Observation Worksheet KPIV Analysis Solution Matrix Theory Of Constraints Process Control Plan Project Selection Matrix SIPOC CTQ Tree Ishikawa Diagram Spaghetti Diagram Advanced Pivot Tables and Charts Impact Effort Matrix Little s Law Recognize Improvement Achieved Project Charter Voice of the Customer Analysis Create Data Collection Plan 5S Advanced Graphing Techniques Future State Process Map Variability Principle ROI Tool CTQ Tree Measurement Systems Analysis Visual Controls Pilot Implementatio n Checklist Project Management ROI Tool Basic Statistics Process Modeling and Simulation Implementatio n Checklist Basic Graphing Techniques Analyze Phase 9
Lean Six Sigma Analyze Phase KPIV Analysis 10 BECOMING a DEFECT DETECTIVE Data Analysis Using data to find patterns, trends and other clues to support or reject KPIVs. Process Analysis A detailed look at existing processes to identify waste. 11 KPIV ANALYSIS Steps 1. Examine the process/data 2. Verify or eliminate KPIVs. 3. Conclusively link KPIVs to KPOVs using data analysis and graphing techniques. 12
KPIV ANALYSIS CSI Approach you must create a robust case to convict each KPIV. We suspect that each of the KPIVs have an impact on the KPOVs. Use your data to build evidence to prove which KPIVs have an impact on the KPOVs. 13 SUMMARIZING and DISPLAYING DATA Time as a KPIV Time (Date/Time of day) should always be considered as a KPIV If evidence is found, investigate further into the time dependent variables that may be impacting the KPOV What are the KPIVs that are typically responsible for time dependencies within a process/service? 14 DISPLAYING DATA EXAMPLE 15
KPOV PARETO CHART Drive Thru Delays by Category Food Not ready Waiting For Cashier Customer Adds to Order Line At Speaker Money Not ready CC Machine down Customer Delay Other 16 KPOV RUN CHART Average Daily Drive Thru Time (minutes) CC Machine Down Cashier Absent Additional Cook Hired 17 LEAN SIX SIGMA TOOLS Additional Analyze Phase Tools: Measurement Scales Review Capability Analysis Correlation and Regression 18
Lean Six Sigma Analyze Phase Measurement Scales 19 MEASUREMENT SCALES Scale Description Example Nominal (Categorical) Ordinal (Ranking) Interval Ratio Data consists of categories only. No ordering scheme possible. Data arranged in some order but differences between values cannot be determined or are meaningless. Data arranged in order and the differences between values can be found. However ratios are meaningless. Interval scale with a zero starting point and values that are multiples. Gender, Ethnicity, Group, Department Service Arrival, Sequence by Type Satisfaction Scale Age, Length of Service, Time 20 MEASUREMENT SCALES 21
MEASUREMENT SCALES Scale Center Spread Nominal (Categorical) Ordinal (Ranking) Significance Tests Mode Information Only Chi Square Median Percentages Sign or Run Test Interval Arithmetic Mean Standard Deviation Ratio Arithmetic Mean Geometric Mean Harmonic Mean Standard Deviation Percent Variation Ftest, t test, correlation analysis Ftest, t test, correlation analysis 22 Lean Six Sigma Analyze Phase Capability 23 CAPABILITY INDICES Capability indices are a statistical measure of process capability: o Process capability The ability of a process to produce output within specification limits. The concept of process capability only holds meaning for processes that are in a state of statistical control. 24
CAPABILITY INDICES Capability indices: o Measure how much "natural variation" a process experiences relative to its specification limits. o Allows different processes to be compared with respect to how well an organization controls them. o Are important tools in process improvement efforts. 25 VISUALIZING PROCESS CAPABILITY 26 CAPABILITY INDEX C p Estimates what the process is capable of producing if the process mean is centered between the specification limits C p USL- LSL 6 ˆ 27
VISUALIZING C p 28 29 CAPABILITY INDEX C p C p USL- LSL 6 ˆ Capability indices use short term variation ˆ where the standard deviation is estimated ( ). Capability indices are based on ±3 or where the process falls 99.7% of the time. If the process mean is not centered within the specifications, C p overestimates process capability. 30
ADDITIONAL INDICES USL - ˆ ˆ - LSL C mi n, pk 3 ˆ 3 ˆ C pk estimates what the process is capable of producing when the process mean is not centered between the specification limits. C pk can also be used for single sided metrics upper or lower specification only by selecting the appropriate formula. 31 ADDITIONAL INDICES ˆ - LSL USL - ˆ C C pl 3 ˆ pu 3 ˆ C pl Estimates process capability for specifications that consist of a lower limit only. C pu Estimates process capability for specifications that consist of a upper limit only. If the specification is two sided, the off centered capability index is the smaller of C pu and C pl. 32 33
EVALUATING CAPABILITY C pk Sigma level (σ) Process yield Process fallout (PPM) 0.33 1 68.27% 317311 0.67 2 95.45% 45500 1.00 3 99.73% 2700 1.33 4 99.99% 63 1.67 5 99.9999% 1 2.00 6 99.9999998% 0.002 34 EVALUATING CAPABILITY C p represents the potential capability C pk represents the current capability If C pk is worse than C p, it can be improved by centering the process. When centering is perfect, C p = C pk. C pk can never be better than C p. 35 STABILITY and CAPABILITY A process is said to be stable when only Common Causes are present and no special cause is active. A process can be Stable, but still incapable of meeting customer specifications. Stability has nothing to do with Capability Control Limits Process Limits 36
PROCESS PERFORMANCE INDICES Process Performance Indices are an estimate of the capability of a process using measured or long term variation () and mean (). Process and capability indices are formulaically identical. However, the estimated and for capability indices have a higher level of uncertainty: C p P p C pk P pk C pu P pu C pl P pl 37 Lean Six Sigma Analyze Phase Correlation and Regression 38 REGRESSION ANALYSIS Regression Analysis includes techniques for modeling and analyzing several variables where the focus is on the relationship between a dependent variable and one or more independent variables. Regression Analysis attempts to explain how the typical value of the dependent variable changes when any one of the independent variables is varied while the other independent variables are held fixed. 39
REGRESSION ANALYSIS Regression analysis is used for trend analysis, prediction and forecasting Regression analysis is also used to understand which among the independent variables are related to the dependent variable, and to explore the forms of these relationships. Regression Analysis methods include: o Linear Regression o Non Linear Regression o Multiple Linear Regression 40 REGRESSION ANALYSIS 41 ASSUMPTIONS Sample is representative of the population. Error is a random variable with a mean of zero conditional on the explanatory variables. Independent variable is measured with no error. Predictors are linearly independent, Errors are uncorrelated. Variance of the error is constant across observations (homoscedasticity). 42
CORRELATION COEEFICIENT Linear Correlation Coefficient (r), measures the strength and the direction of a linear relationship between two variables. Value of r range is 1 < r <+1 data. 43 CORRELATION COEEFICIENT Positive correlation: o If x and y have a strong positive linear correlation, r is close to +1. o An rvalue of exactly +1 indicates a perfect positive fit. o Positive values indicate a relationship between x and y variables such that as values for x increases, values for y also increase. 44 CORRELATION COEEFICIENT Negative correlation: o If x and y have a strong negative linear correlation, r is close to 1. o An rvalue of exactly 1 indicates a perfect negative fit. o Negative values indicate a relationship between x and y such that as values for x increase, values for y decrease. 45
CORRELATION COEEFICIENT No correlation: o If there is no linear correlation or a weak linear correlation, r is close to 0. o A value near zero means that there is a random, nonlinear relationship between the two variables. o Note that ris a dimensionless quantity; that is, it does not depend on the units employed. 46 CORRELATION COEEFICIENT A perfect correlation of ±1 occurs only when the data points all lie exactly on a straight line. If r = +1, the slope of this line is positive. If r = 1, the slope of this line is negative. If r = 0, there is no correlation 47 CORRELATION COEEFICIENT A correlation greater than 0.8 is generally described as strong, whereas a correlation less than 0.5 is generally described as weak. 48
COEFFICIENT of DETERMINATION Coefficient of Determination (r 2 ) gives the proportion (percentage) of the variance of one variable that is predictable from the other variable. It is a measure that allows us to determine how certain one can be in making predictions from a certain model/graph. The Coefficient of Determination is the ratio of the explained variation to the total variation. 49 COEEFICIENT of DETERMINATION The range of r 2 is 0 <r 2 <1 The Coefficient of Determination represents the percent of the data that is the closest to the line of best fit. o For example, if r = 0.922, then r 2 = 0.850. o The value indicates that 85% of the total variation in y can be explained by the linear relationship between x and y. o The other 15% of the total variation in y remains unexplained. 50 COEEFICIENT of DETERMINATION 51
NON LINEAR REGRESSION Examples of nonlinear functions include Exponential Logarithmic, Polynomial, Power and Moving Average. e.g. Exponential The graph of y = e x is upwardsloping, and increases faster as x increases. 52 NON LINEAR REGRESSION e.g. Logarithmic growth describes a phenomenon whose size or cost can be described as a logarithm function of some input. e.g. y = C log 53 REGRESSION ANALYSIS Excel performs Regression Analysis for plotted data using the Trendline option in the Chart Tools: 54
REGRESSION ANALYSIS Excel performs both linear and non linear regression and can calculate the Coefficient of Determination 55 REGRESSION ANALYSIS Linear Correlation 56 REGRESSION ANALYSIS 5 th Power Polynomial Correlation 57
MULTIPLE LINEAR REGRESSION When there is more than one independent variable, the regression line cannot be visualized in the two dimensional space However a Multiple Linear Regression equations can be computed and has the form: Y = a + b 1 *X 1 + b 2 *X 2 +... + b p *X p 58 END OF WEEK 11 MATERIAL Resources: Paper Diffusion of Innovations Theory. Statistics Review I and II Assignment: Assignment Homework #4 59