1、Copyright 2002 Dell Computer Corporation Revision 071802,52,BPI Level I Module 3 - Measure,Define,Measure,Improve,Analyze,Control,Report,BPI Model,Copyright 2002 Dell Computer Corporation Revision 071802,53,Module Objectives: Measure,Describe measurements Family of measures Data types Attribute / Di
2、screte Variable / Continuous Leading and lagging Data and information control Measurement System Effectiveness (MSE),Copyright 2002 Dell Computer Corporation Revision 071802,54,Measure,A description and quantification of the dimension, quantity, capacity, performance or other characteristic of a pop
3、ulation or process Measurements during DMAICR,Copyright 2002 Dell Computer Corporation Revision 071802,55,Why Do We Measure?,Establish baselines and targets Provide a common language Focus on correct causes Demonstrate process improvement and results Determine if we are meeting customer requirements
4、,Copyright 2002 Dell Computer Corporation Revision 071802,56,Why Do We Measure? Some Principles,If you cant measure it you cant manage it You get what you measure Remember the Heisenberg Uncertainty Principle Dr. Deming and Operational Definitions What is it? How is it used? How is it measured?,Copy
5、right 2002 Dell Computer Corporation Revision 071802,57,Family of Measures,Copyright 2002 Dell Computer Corporation Revision 071802,58,Exercise: Output Measures,We must ensure a strong link between Critical Success Factors and Output Measures.,High? Medium? Low? None?,Copyright 2002 Dell Computer Co
6、rporation Revision 071802,59,Types of Data: Attribute,Attribute data can be counted and plotted as DISCRETE events Can be numeric or character dataExamples: Number of defects Pass/Fail Percent defective Defects Per Hundred Units (DPHU) Missing, Wrong, and Damaged (MWD) Initial Field Incident Rate (I
7、FIR),Copyright 2002 Dell Computer Corporation Revision 071802,60,Types of Data: Variable,Variable data can be measured and plotted on a CONTINUOUS scale Numerical data onlyExamples: Distance Length Time Weight On Time First Time Fix (OTFTF), if measured in time,Copyright 2002 Dell Computer Corporati
8、on Revision 071802,61,Characterizing Measures,Create the following table on the butcher paper Describe each metric fully Balance your set of metrics using Quality, Cost, Cycle Time and Input, Process, OutputLeave room for other columns Desirability (increase, decrease, maintain) Measure of central t
9、endency? Measure of variability? Have you considered countermeasures?,Copyright 2002 Dell Computer Corporation Revision 071802,62,PTT: Measure Tab,Unit of Measure Fully describe the metric (description, variable, unit of measure) Increment Use the drop down to indicate sampling frequency Effective D
10、ate Use the drop down to select the date Performance Before Use your baseline or historical data Cost Before Use your baseline or historical data to determine the financial cost estimate Projected % Improvement The goal from your desired-state problem statement ValueChain Enabled If selected, the su
11、pplier can see the project through the value chain portal,Copyright 2002 Dell Computer Corporation Revision 071802,63,PTT: Measure Tab Example,Copyright 2002 Dell Computer Corporation Revision 071802,64,Do You Understand? - DATA,Data control plans ask a series of questions about the characterization
12、, generation, processing, and storage of metrics Is there a documented description of the metric? Descriptor/metric/unit-of-measure Quality, Cost, Cycle Time Is there a documented formula available that explains how to calculate the data? How often is the data sampled? How is the data stored? How is
13、 the data put into storage? Is there a document that describes how to access the data?,Copyright 2002 Dell Computer Corporation Revision 071802,65,Do You Understand? - INFORMATION,Information control plans describe how the metric is transformed into a format that is capable of being used for effecti
14、ve communication and decision making How often is a report generated? Who is responsible for preparing the report? In what state is the information presented? What does a single data point on the report or chart represent? Is the goal or target known? Why is this information important? What happens
15、to us if this information is good or bad? Are these consequences documented? Who is responsible for the internal reaction plan? Is there a document that contains instructions or descriptions about how the process owners react to the information?,Copyright 2002 Dell Computer Corporation Revision 0718
16、02,66,Measurement Systems,Total Process Variability = Product Variability + Measurement System Variability,Measurement System Variability = f(Repeatability, Reproducibility),Copyright 2002 Dell Computer Corporation Revision 071802,67,Properties of Any Measuring System,Accuracy Ability to produce an
17、average, measured value which agrees with the true value or standard being used Precision Ability to repeatedly measure the same product or service and obtain the same results Stability Ability to repeatedly measure the same product or service over time and obtain the same average, measured value; u
18、sually evaluated using a control chart,Copyright 2002 Dell Computer Corporation Revision 071802,68,Accuracy and Precision,Copyright 2002 Dell Computer Corporation Revision 071802,69,Attribute MSE,Purpose is to assess the effectiveness of an attribute measurement system (e.g. test, inspection, verifi
19、cation)Key Principle: The overall effectiveness (E) of an attribute measurement system can be determined as follows:E = No. of Parts Identified Correctly (Good or Bad)Total No. of Parts* An attribute measurement system is generally consideredcapable if E is 0.90,Copyright 2002 Dell Computer Corporat
20、ion Revision 071802,70,Attribute MSE,Additionally, it is possible to determine a bias towards either false acceptance or false rejection.P(FA) = No. of bad parts accepted as good * Consumers RiskNo. of bad partsP(FR) = No. of good parts rejected as bad * Producers RiskNo. of good parts,Copyright 200
21、2 Dell Computer Corporation Revision 071802,71,Attribute MSE Exercise,Part No. Good Bad 1 2 3 4 5 6 7 8 9 10,E= P(FA)= P(FR)=,*For this exercise, long lines are good, short lines are bad.,Copyright 2002 Dell Computer Corporation Revision 071802,72,Class Exercise: Design a Dell MSE,Design a Dell MSE
22、for your project or functional area List the type, title, number of operators, number of observations/parts, and outline the MSE procedure,Copyright 2002 Dell Computer Corporation Revision 071802,73,Project Exercise,Use Post-its to identify where MSE may be useful to your project,Copyright 2002 Dell
23、 Computer Corporation Revision 071802,74,End of Measure - Moving to Analyze,As you move from Measure to Analyze, what are the things you should have completed? Identified a Family of Measures balanced across Quality, Cost and Cycle Time and through Inputs, Process, Outputs. Developed Operational Definitions for each measure Examined existing data for suitability and begun collecting new data as required Tested all measurement systems for Accuracy, Precision and Stability Established baseline performance for the As-Is Process Updated the Project Tracking Tool,
