Download presentation
Presentation is loading. Please wait.
1
a.k.a. Voice of the Customer
The 17th Annual International Deming Research Seminar Test it in Service a.k.a. Voice of the Customer
2
What We Plan to Cover (AIM)
Why Test in Service? Why Good Sample Design? Accuracy versus Precision Some Statistical issues Some Methodological Issues Some Data Issues Some Questionnaire Issues 23 March 2011 The 17th Annual International Deming Research Seminar
3
Deming Lecture Japan 1950 23 March 2011
The 17th Annual International Deming Research Seminar
4
Evolution of Shewhart Cycle
Production Make Specification Design Inspection Sell (a. Old Way) 4. Test in Service 5. Redesign Specification 1. Design Production 3. Sell 2. Make Inspection (b. Original Shewhart 1939) (c. Deming 1950 Version)
5
Production Viewed as a System
Design and Redesign Consumer Research Suppliers of Materials and Equipment Consumer A Receipt and test of materials Consumer B Production, assembly, inspection Consumer Consumer Consumer C Consumer D Tests of Processes, machines, methods, costs Consumer 23 March 2011 The 17th Annual International Deming Research Seminar
6
Process GAP Input Conversion Output Voice of the Customer
Voice of the Process GAP In order to be sure that the conversion is operating as intended, one samples the output. Selecting which output of which sub-process to measure is the subject of another chapter. In general, it depends on what is a matter of importance to the customer. A method with the acronym of “QMS” (Quality Measuring System) helps management to determine where to measure to get the most value from the measurement. Any deviation that lessens the value of the item to the customer is subject to measurement. As Mr. William Scherkenbach pointed out, “There is powerful leverage in going upstream” (Scherkenbach, 1986). Upstream means the first input source than can be impacted. For example, if the material (or information) used is faulty, it is not likely to improve as it works its way through the network of processes. The process input is upstream while the process output is downstream. To determine the effectiveness of the conversion one needs two measures called “Voice of the Customer” and “Voice of the Process”. The first measure gives us an understanding of the customer’s needs. The second tells us how well the conversion process is meeting those needs. For the voice of the customer one uses proper market research methods. A method that obtains a great deal of information is a quadrant analysis developed by Mr. David Saunders (1992). To determine how well the conversion process meets the customers’ needs one gets the voice of the process. This text is concerned with how to get the appropriate measure from the process. One should keep in mind the customer may well be an internal one. For example, the accounting function has as its primary customer the Board of Directors and the officers of the organization. Input Conversion Output 23 March 2011 The 17th Annual International Deming Research Seminar
7
How do we get the Voice of the Customer?
Test it in Service Consumer Research In order to be sure that the conversion is operating as intended, one samples the output. Selecting which output of which sub-process to measure is the subject of another chapter. In general, it depends on what is a matter of importance to the customer. A method with the acronym of “QMS” (Quality Measuring System) helps management to determine where to measure to get the most value from the measurement. Any deviation that lessens the value of the item to the customer is subject to measurement. As Mr. William Scherkenbach pointed out, “There is powerful leverage in going upstream” (Scherkenbach, 1986). Upstream means the first input source than can be impacted. For example, if the material (or information) used is faulty, it is not likely to improve as it works its way through the network of processes. The process input is upstream while the process output is downstream. To determine the effectiveness of the conversion one needs two measures called “Voice of the Customer” and “Voice of the Process”. The first measure gives us an understanding of the customer’s needs. The second tells us how well the conversion process is meeting those needs. For the voice of the customer one uses proper market research methods. A method that obtains a great deal of information is a quadrant analysis developed by Mr. David Saunders (1992). To determine how well the conversion process meets the customers’ needs one gets the voice of the process. This text is concerned with how to get the appropriate measure from the process. One should keep in mind the customer may well be an internal one. For example, the accounting function has as its primary customer the Board of Directors and the officers of the organization. Input Conversion Output 23 March 2011 The 17th Annual International Deming Research Seminar
8
Test it in Service Thought Data Knowledge 23 March 2011
The 17th Annual International Deming Research Seminar
9
Survey Impact on Management
Wrong Design Wrong Information Wrong Decision 23 March 2011 The 17th Annual International Deming Research Seminar
10
What is the aim of the survey?
Who determines the aim? The Subject Matter Expert Why is the aim important? Without an aim there is no system 23 March 2011 The 17th Annual International Deming Research Seminar
11
Who is involved in Surveys
Subject Matter Expert Statistician Respondents 23 March 2011 The 17th Annual International Deming Research Seminar
12
The Subject Matter Expert
Commissions the survey Determines the aim Can act on the results Has budget authority 23 March 2011 The 17th Annual International Deming Research Seminar
13
Aspects of Sample Design
Statistical Sample size Variance Accuracy Methodological Sample selection Computation 23 March 2011 The 17th Annual International Deming Research Seminar
14
What is a sample Complete coverage gives answer
A subset of the interest group Draw Sample 23 March 2011 The 17th Annual International Deming Research Seminar
15
The Frame The Frame is a list of sample units
Frames are not always complete The Gap 23 March 2011 The 17th Annual International Deming Research Seminar
16
Data vs. Information Data The sample yields data Analysis
Results Data Analysis The Subject Matter Expert needs Information Information 23 March 2011 The 17th Annual International Deming Research Seminar
17
Information contains Result Measure of uncertainty 23 March 2011
The 17th Annual International Deming Research Seminar
18
Which is more important
Sample size? or Sample method? Sample size? or Sample method? Bias Total Error Sampling Error 23 March 2011 The 17th Annual International Deming Research Seminar
19
Accuracy and Precision
Sample Method Not Accurate Accurate Not Precise Sample Size Precise 23 March 2011 The 17th Annual International Deming Research Seminar
20
Sample Size (n) Expected result percentage Desired precision
Where is the actual result t the number of standard deviations under the Normal Curve Then t Area for Mean ±ts 1 68% 2 95% 3 99.73% 23 March 2011 The 17th Annual International Deming Research Seminar
21
Sample Size Example Q. Was the Conference useful? Please check one: Yes □ or No □ If one wants 95% reliability that the actual responses fall within 10% of the response of 50% then . 23 March 2011 The 17th Annual International Deming Research Seminar
22
23 March 2011 The 17th Annual International Deming Research Seminar
23
Because it is the maximum sample size
23 March 2011 The 17th Annual International Deming Research Seminar
24
Sample Method Clear Aim Random Sampling Unbiased Questions
23 March 2011 The 17th Annual International Deming Research Seminar
25
Clear Aim What does the owner wish to accomplish with the data?
How will the data be used in practice? Is the data measureable Cause Effect Is the data useable? e.g. measure engagement with an Association e.g. measure how often contacted by Association 23 March 2011 The 17th Annual International Deming Research Seminar
26
Random sample Equal probability of selection Random numbers
Random Start nth Sample Replicated design with random start per zone Deming's Replicated Design: Total sample cut into zones like sliced bread One zone Total sample size 23 March 2011 The 17th Annual International Deming Research Seminar
27
Reduce Bias in Questionnaire
If questionnaire… …is for interviewing then train interviewer to Avoid loaded questions Avoid verbal bias Avoid body language bias …is for respondent's completion Avoid question overload 23 March 2011 The 17th Annual International Deming Research Seminar
28
Open Ended Questions An open-ended question is designed to encourage a full, meaningful answer using the subject's own knowledge and/or feelings. Opposite of closed-ended question, which encourages a short or single-word answer. Open-ended questions also tend to be more objective and less leading than closed-ended questions. Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
29
Open-Ended Question Examples
Closed-Ended Questions Open-Ended Questions Do you get on well with your boss? Tell me about your relationship with your boss. Who will you vote for this election? What do you think about the two candidates in this election? Is that a photograph of your children? Tell me about the children in this photograph. Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
30
Prevent Response Bias Avoid Ambiguity: Questions not focused enough to obtain the needed information. Poor question: How often do you buy fast food? Better question: The last 10 times you’ve eaten lunch out, how often have you purchased each of the following types of food? Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
31
Prevent Response Bias Avoid questions that are too general. Poor question: Do you like orange juice? (Do you mean taste? texture? price?) Better question - more specificity: Do you like the taste of fresh-squeezed orange juice? Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
32
Prevent Response Bias Avoid loaded questions: use of language likely to bias response: Do you actually support Proposition X? Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
33
Prevent Response Bias Avoid leading questions: Questions that lead respondents toward a particular answer. Isn’t it true that women are more likely to talk on the phone while driving? Courtesy Myriam Ochart, ASQ CSSBB, CMQ/OE
34
What We Covered Why Test in Service? Why Good Sample Design? Accuracy versus Precision Some Statistical issues Some Methodological Issues Some Data Issues Some Questionnaire Issues 23 March 2011 The 17th Annual International Deming Research Seminar
35
Questions? 23 March 2011 The 17th Annual International Deming Research Seminar
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.