The Chicago Fashion Study is a very well-known 1996 multi-sponsor market study – by Professor Douglas Tigert and his associates – that provides data on key aspects of how female fashion shoppers view the Chicago Market.
The Objectives of the Exercise
The purpose of this set of exercises is to help you use the Chicago Fashion Study database to explore the use of market research data as an input to market segmentation and the identification of market opportunities.
Your Task
Assume you are doing a project for a major European clothing manufacturer that is considering marketing one or more of their products through American retail chains -- or for one of the US retailers. The company has asked you to analyze the market for four one of the following four product lines: dressy blouses; casual blouses; special occasion dresses; and jeans. Your analysis will be the basis for a report that describes and explains the segments of one of these four product classes and makes suggestions to the manufacturer or retailer about how to market to this segment. You have learned that you can purchase access to the Chicago Fashion Study – a multi-sponsor market study. The data appear useful and you have purchased the right to use them.
PRICE SEGMENTATION
Your first analysis will include a focus on segmenting the market for one of these products on the basis of price, and more specifically, the range of prices paid, from lowest to highest, for one of the categories. You must address two basic questions:
The Data File
The data file with which you will be working has been taken from a much larger one, called "The Chicago Female Fashion Study 1996". The original data file contains over 300 variables for 1740 respondents. Your data file contains only 74 variables extracted from the main file. You have a "coding sheet" for the 74 variables and selected pages from the questionnaire. You can also get access to the complete 47-page questionnaire.
The data were collected via a self-completed questionnaire placed into the homes of a sample of the greater Chicago metropolitan area. Assume that the 1740 respondents are representative of women fashion shoppers in the Chicago area.
Accessing the Data File. The data file is located at http://faculty.babson.edu/isaacson/M_E7000/gurney.mtw. We will work with the file in MiniTab.
The Nature of the File. The data file is a Minitab data file. You can Call up Minitab on GlobeNet and then get the file with the FileOpen Worksheet. command. (Note that the student version of Minitab, sold in the Bookstore, may not be capable of handling the entire data set simultaneously. So, use only Minitab 11.12, 32 bit, which is on the Babson system.)
The Assignment
You will probably want to form and use a study group for this assignment. Then select a product from this list:
Step 1) Hypothesizing About the Number of Segments: An Educated Guess
The 1740 women who participated in the survey said they paid anywhere from $8 to $902 when they last bought a special occasion dress. (See column 37.)
How do we know this? The screen below shows how to ask for a frequency distribution and cumulative percentages for column 37, which is labeled ‘P_SPECD’. To do this tally in Minitab, select Stat, Tables, Tally. In the dialogue box, "check" counts and cumulative percents.
Table 1 shows the resulting frequency distribution and cumulative percentages for ‘P_SPECD’.
You will note from this frequency distribution that only 1297 respondents indicated a price paid, and that there is a "*=443" indicating that 443 respondents did not answer this question. They may not have purchased the product or been unable to remember how much they paid.
Based on insightful analysis of these data, your next task is to determine a suitable number of meaningful segments defined by the price paid for the special occasion dress. In other words, what price ranges can be created so that each price range not only represents customers who are looking for the same type of dress, but also represents customers who are very different from customers of any other price group? Are there three, four or five meaningful segments? (Or more, or less?) What should be done with "outliers" who spent much less or much more for their dresses?
Here’s an example of how you could proceed. You might begin by hypothesizing that there are least four price groups or segments -- a low-end price group, a mid-range group, a high-end group, and a very high-end group. In our example in the next screen, we have chosen to create four approximately equal sized groups, by using the cumulative frequency distribution from the previous exhibit to determine the cut-off points, at approximately 25 percent, 50 percent, and 75 percent.
Summary Statistics for
Discrete Variables
| 8
1 0.08
10 4 0.39 15 2 0.54 16 2 0.69 17 2 0.85 18 4 1.16 19 3 1.39 20 25 3.32 22 1 3.39 23 1 3.47 24 3 3.70 25 22 5.40 26 2 5.55 27 2 5.71 28 1 5.78 29 3 6.01 30 40 9.10 31 2 9.25 32 3 9.48 34 1 9.56 35 28 11.72 36 1 11.80 38 4 12.10 39 6 12.57 40 60 17.19 42 2 17.35 43 1 17.42 44 1 17.50 45 28 19.66 46 1 19.74 48 4 20.05 49 19 21.51 50 77 27.45 52 4 27.76 54 1 27.83 |
55
14 28.91
56 3 29.14 57 2 29.30 58 2 29.45 59 16 30.69 60 73 36.31 62 2 36.47 65 25 38.40 66 1 38.47 68 4 38.78 69 15 39.94 70 39 42.95 72 3 43.18 73 2 43.33 75 38 46.26 76 1 46.34 78 1 46.41 79 8 47.03 80 67 52.20 81 1 52.27 85 20 53.82 87 2 53.97 88 2 54.12 89 16 55.36 90 34 57.98 92 1 58.06 94 1 58.13 95 15 59.29 97 2 59.44 98 4 59.75 99 12 60.68 100 76 66.54 105 4 66.85 108 4 67.15 109 9 67.85 |
110
23 69.62
112 1 69.70 115 4 70.01 118 3 70.24 119 5 70.62 120 47 74.25 122 1 74.33 125 20 75.87 126 1 75.94 129 8 76.56 130 10 77.33 135 4 77.64 138 1 77.72 139 3 77.95 140 11 78.80 145 3 79.03 150 57 83.42 153 1 83.50 154 1 83.58 158 2 83.73 160 14 84.81 164 1 84.89 165 2 85.04 170 6 85.51 172 1 85.58 175 16 86.82 179 1 86.89 180 15 88.05 185 1 88.13 189 1 88.20 190 5 88.59 193 1 88.67 197 1 88.74 199 2 88.90 200 43 92.21 |
208
2 92.37
210 1 92.44 220 3 92.68 225 7 93.22 230 2 93.37 235 1 93.45 240 3 93.68 249 1 93.75 250 18 95.14 252 1 95.22 256 1 95.30 264 1 95.37 275 4 95.68 289 1 95.76 290 1 95.84 299 1 95.91 300 19 97.38 310 1 97.46 320 1 97.53 345 1 97.61 350 6 98.07 375 1 98.15 400 5 98.54 450 2 98.69 459 1 98.77 480 1 98.84 495 1 98.92 500 5 99.31 503 1 99.38 559 1 99.46 600 1 99.54 700 4 99.85 900 1 99.92 902 1100.00 |
To do this in Minitab, you would CODE the price data by selecting: Manip,
Code,
Numeric to Numeric.
You may disagree with the price ranges, in which case you should create your own. The screen above is for explanatory purposes only. The very wide range for the fourth price range category ($126 to $902) may reflect more than one type of consumer. It’s up to each team to decide the most appropriate way to create the price segments.
Note that we have not only collapsed the original price distribution into four new categories, we have also created a new variable called ‘CODE_DRS’. The Minitab program will automatically put this new variable into the next available open column, which in this case is C75, and it’s name will be ‘CODE_DRS’. This new variable is the "column" variable for all of the cross-tabulation tables that you are now going to produce.
Step 2) Describe the Segments
If you have defined meaningful segments, each segment may have distinctive and differentiating demographics (e.g., age, income, marital status, etc.), psychographics (e.g., attitudes toward fashion, cost, etc.) and shopping profiles (i.e., the store where they last bought a special occasion dress). All of the remaining variables in your data set are candidates to be used to develop profiles of each price segment. The profiles should not only describe what is common to each segment, but they also should differentiate each segment from the others.
Here is how you do it in Minitab. Select Stat, Tables, Cross Tabulation. Click on your first variable (in our case, "AGE") and select, and then on your second variable (in our case "CODE_DRS"), and select again. An appropriate dialog box might look like:
NOTE: Minitab wants the row variable listed first. In this example this is the ‘age’ variable.
Use the TABLE command to examine your segmentation with the segment label at the top of the column and a demographic, psychographic or shopping variable in the row. Take a look at the column percents to determine whether there is a pattern.
The above screen is asking for a cross-tabulation table of ‘age’ versus ‘CODE_DRS’. Note in the screen that we have "X"ed the box labeled "Column Percents" and the box labeled "Chi-square Analysis". Those two combined commands produce the cross tabulation table shown below, which is easy to read and provides a statistical measurement that can be very useful.
1 13.80 11.87
9.80 7.99 10.97
49
38 30 25
142
2 26.48 20.94 22.55
22.36 23.18
94
67 69 70
300
3 31.83 27.81 25.49
21.73 26.89
113
89 78 68
348
4 13.24 16.56 22.55
27.16 19.63
47
53 69 85
254
5 8.45 11.25
11.76 13.42 11.13
30
36 36 42
144
6 6.20 11.56
7.84 7.35 8.19
22
37 24 23
106
ALL 100.00 100.00 100.00 100.00
100.00
355 320 306 313
1294
CHI-SQUARE = 44.545 WITH D.F.=
15
CELL CONTENTS --
% OF COL
COUNT
Note that there are 1294 respondents in the above table. The remaining respondents from our original sample of 1740 did not answer either or both of the two questions on age and price paid for a special occasion dress.
How do we interpret and draw conclusions from a cross-tabulation table? In this example what we want to know is whether respondents in the four different price segments have different age distributions. In this case, the answer is "yes" because respondents who pay the lowest price points ($8 - 50) tend to be somewhat younger. About 40 percent of them are in the youngest two age brackets compared to only 30 percent of the respondents who pay the highest prices (126 - 902). Thus, there is a statistically significant relationship between age and price paid for this dress. The Chi-square is also statistically significant.
Step 3) Confirm the Segments
You now have 70 or so other cross-tabulation variables to work with. Your job is to pick what you think might be the most relevant variables to cross-tabulate against price paid (for your product class). Try to select descriptors that will draw a clear picture of the typical customer in each segment. Look for descriptors that complement each other. For example, if there is a segment (e.g. the high-end shoppers) that shops just as often at Wal*Mart as it does at Marshall Fields for a special occasion dress, there is something wrong. You may have defined the segment too broadly to be meaningful.
Step 4) Summarize Your Analysis
After you have run all the cross-tabulation tables, you are ready to start interpreting them and drawing overall conclusions about the different price point segments. For example, where did they last buy the product? How does their lifestyle profile differ from other segment consumers? How do they differ demographically? Are their store choice criteria different? What are they looking for in a store? Where do they seek out fashion information? Are they more likely to buy on sale? And so on...
Deliverables
Each team should build a 5-10 minute presentation of its key findings.