You are on page 1of 60

CM 723 Advanced Communication Research

ASSIGNMENT 4: Understanding and predicting shopping behaviors

Boston University

Christopher Wilcox Meredith Knight Xue Teng (Portia) Yajie Li (Jessie) May 1, 2012

Table of Contents
I. Introduction 1. Client problem 2. Research questions 3. Logic of analysis II. Sample description Demographics III. Data Analysis 1. Review of variables 2. Selection and testing of multiple item measures Correlations Factor Analysis Reliability 3. Building a Regression Model 4. Cluster Analysis IV. V. VI. Cluster Description Final Story and Key Takeaways Limitations

Appendices

I.

Introduction Electronic Shopping Incorporated (ESI) is developing technology to enable individuals to

shop interactively from home using their television. ESI faces two key business challenges in their initiative to transform home shopping through this new technology. The first challenge is to identify the individuals most likely to be early adopters of this new technology. The second challenge is to explore how American consumers may adopt this new technology. We used an existing dataset as a resource to answer the following research questions: 1. What are the profiles of different groups of potential customers based on their current attitudes, beliefs, and behaviors? 2. What factors cause individuals within those groups to be more likely to adopt this technology? The existing dataset was reviewed to identify variables directly related to shopping at home using the television. Three related items were combined into one composite dependent variable. All other variables were reviewed to identify potential predictors, and each predictor was evaluated to determine whether it was suitable for a regression model by determining how well those variables correlated with the dependent variable. Finally, the dataset was segmented to identify major groups within the sample who would be likely to adopt the technology.

II.

Overall sample description

Race, ethnicity, and gender Ninety three percent of the respondents report that they are not Hispanic (93.2%, SE:1.1, 95% CI). Eighty six percent report that they are white (86.3% SE:1.5, 95% CI), 6.7% are black (SE:1.1, 95% CI), 1.3% are Asian (SE:0.4, 95% CI), 4.4% are other (SE:0.9, 95% CI), and 1.4% refused to respond (SE:1.1, 95% CI). The overall gender ratio for the sample was 48.2% male and 51.8% female (both SE:2.3, 95% CI). Age The data for age was gathered in categories rather than absolute age. The highest percentage of respondents was between the ages of 30-39, representing 23.9% of the total sample (SE:2.0, 95% CI). The second highest percentage of respondents were between the ages of 4049, representing 21.8% of the total sample (SE:1.9, 95% CI), and the third highest percentage was for the 20-29 age group, representing 18.3% (SE:1.8, 95% CI). Income Over sixty percent of the sample reported income levels of between $15,000 and $75,000 (63.7% SE:2.1, 95% CI). Only five percent reported an income over $100,000 (SE:0.9, 95% CI), and 18.8% refused to answer the question (SE:1.7, 95% CI). Marital status Over half (56.3%) of the sample is married (SE:2.1, 95% CI), and slightly over a quarter are single (26%, SE:2.6, 95% CI). Nine percent (9.5%SE:1.3, 95% CI)are divorced, 6.2% are

widowed (SE:1.1, 95% CI) and 1.7 are separated (SE:0.6, 95% CI). 0.3 refused to answer the question (SE:0.8, 95% CI). Home ownership Seventy two percent of the respondents own the home in which they live (72.4%, SE:1.9, 95% CI), while 27.6% rent (SE:1.9, 95% CI). Seventy two percent of the respondents live in single-family homes (72.8%, SE:1.9, 95% CI), 15.7% live in an apartment (SE:1.6, 95% CI), 6.4% live in a multifamily house (SE:1.1, 95% CI), 2.6% are in a townhouse (SE:0.7, 95% CI) and 2.0% are in a condo apartment (SE:0.6, 95% CI). Household demographics The mean number of people in the household is 2.85, with a standard deviation of 1.5, and a median of 2. Sixteen percent of households have only one person (16.6%, SE:1.6, 95% CI), 33.6% of households have 2 people (SE:2.0, 95% CI), 19.5% have 3 people (SE:1.7, 95% CI), 17.3% have 4 (SE:1.6, 95% CI), 17% have 1 (SE:1.6, 95% CI). Eighty four percent (84.6%) of households do not have children under 5 (84.6% (SE:1.6, 95% CI). Only 10.6% of households have one person under 5 (SE:1.3, 95% CI), and 3.9% have two people under 5 (SE:2.1, 95% CI). Shopping patterns Most people shop locally. Ninety two percent go to local shopping stores (92.3%SE:1.2, 95% CI) and 91.0% go to the mall (SE:1.2, 95% CI). Forty two percent (42.4% SE:2.1, 95% CI) use the phone to shop from a catalog, and 42.5% shop through the mail (SE:2.1, 95% CI). Fewer than fifteen percent of people shop by calling 800 numbers (14.5% SE:1.5, 95% CI).

Very few people bought merchandise from infomercials (6.1% SE:1.0, 95% CI), home shopping networks (8.8% SE:1.2, 95% CI), or through online computer services (4.5%, SE:0.9, 95% CI). Seventy one percent (71.4%) of the respondents look at the Yellow Pages for information about where to purchase items (SE:1.9, 95% CI). Thirty six percent of the sample watch infomercials (36.5% SE:2.2, 95% CI), and 30.1% watch home shopping channels. Personal preferences Over half of the respondents describe themselves as outgoing (56.0%, SE:2.2, 95% CI), and almost three quarters describe themselves as cautious (73.5% SE:2.2, 95% CI). Fifty eight percent of the people say they enjoy staying at home in their free time (58%, SE:2.1, 95% CI), 31.6% prefer going out (SE:2.0, 95% CI), and 10.4% said neither in particular (SE:1.3, 95% CI). Eighty percent of the sample (80.5%, SE:1.7, 95% CI) wait before buying new products and 14.4% are among the first to buy new products (SE:1.5, 95% CI). Education Most people are either high school graduates (29.8% SE:2.0, 95% CI), have attended some college (24.5%, SE:1.9, 95% CI), or a college degree (21.0%, SE:1.8, 95% CI). Only 15.3% of the respondents are currently in school (SE:2.2, 95% CI), 84.7% are not currently in school. The level of school mentioned often is college at 48%, high school at 28% and graduate school at 16.2% (SE:21.6, 95% CI). Employment Sixty seven percent are employed (67.5%, SE:2.0, 95% CI), and 32.5% (SE:2.0, 95% CI are not). Of those who are employed, 81.9% (SE:1.6, 95% CI) are employed full time and

18.1% are part time. Seventy seven percent of those who are working work for someone else (77.2% SE:1.8, 95% CI), 14.1% are self-employed (SE:1.5, 95% CI), and 8.6% responded that they are both self-employed and work for someone else (SE:1.2, 95% CI). Fifty eight percent (58.9%) of people work someplace other than home (SE:2.7, 95% CI), while 41.1% work at home (SE:2.7, 95% CI). Respondents were also asked whether they occasionally worked from home. Slightly under half replied that they never work from home (47.4%, SE:2.8, 95% CI) 20.2% occasionally work from home (SE:2.2, 95% CI), and 14.5% rarely worked from home (SE:2.0, 95% CI). The top three occupations cited among respondents were technical, sales, or administrative support (32.7%, SE:2.5, 95% CI), management (30.9%, SE:2.4, 95% CI) or service (15.9%, SE:1.8, 95% CI). Seventy percent of respondents reported that there is another member in the household who is also employed (SE:2.4, 95% CI), 29.9% do not (SE:2.4, 95% CI). Of the respondents who are not currently working, 47.4% are retired (SE:3.8, 95% CI), 21.1% are homemakers (SE:3.1, 95% CI), 10.9% are unemployed (SE:2.4, 95% CI), and 6.0% are looking for work (SE:1.8, 95% CI). Over sixty five percent of people who have someone working from home do not also work from home themselves (65.4%, SE:2.6, 95% CI), and 34.6% have two people working from home (SE:2.6, 95% CI).

III.

Data Analysis

1. Identifying the Dependent Variable The ideal dependent variable should predict the participants likelihood of using the new TV technology to shop at home. In the surveys, four items met those requirements. The four measures are Likelihood of looking for consumer information about something you want to buy, Likelihood of looking up the store that sells an item you need to buy, Likelihood of looking through catalogs and ordering items using your TV remote and Likelihood of seeing a video shopping mall and have a sales person demonstrate the products for you. These four measures each have a distinct meaning but also share an overlap in meaning, so the items were combined into a new construct called Likelihood of using new TV technology to shop at home. After first checking the frequency distributions for each of the four measures for outliers and skew, a three-step process was used to confirm our assumption about the multiple item measures. First, the multiple item measures were validated through an inter-item correlation test, then a factor analysis was run, and finally, the reliability of the measure was tested using Cronbachs alpha. According to Table 1, inspection of the correlation matrix revealed the presence of the coefficients of .50 to.70, which show a relatively large overlap in meaning of the four measures. According to the output of factor analysis (see Table 2), Principal component analysis revealed that these four measures explained 71.4% of the total variance, with all four factor loadings around .80-.90. Based on the robust results of inter-correlation test and factor analysis, the reliability of these four measures was tested. The construct Likelihood of using new TV technology to shop at home has good internal consistency, with the Cronbachs alpha coefficient reported of .87. There is no significant improvement if any item deleted. Therefore,

the assumption that the four measures belong to the construct Likelihood of using new TV technology to shop at home was confirmed. Two new variables, home shopping sum and home shopping average, were created by computing the true scores of this construct in two ways for different analysis purposes. Table 1: Inter-item Correlations for Proposed Dependent Variable
FUT-Q.2 Look for consumer information about something you want to buy FUT-Q.2 Look through catalogs and order items using your remote FUT-Q.2 See a "video" shopping mall and have a salesperson demonstrate the products for you

FUT-Q.2 Look up the store that sells an item you need to buy

FUT-Q.2 Look for consumer information about something you want to buy FUT-Q.2 Look up the store that sells an item you need to buy

FUT-Q.2 Look through catalogs and order items using your remote FUT-Q.2 See a "video" shopping mall and have a salesperson demonstrate the products for you a. Listwise N=2005

Pearson Correlation Sig. (2tailed) Pearson Correlation Sig. (2tailed) Pearson Correlation Sig. (2tailed) Pearson Correlation Sig. (2tailed)

.658** .000

.524** .000 .672** .000

.529** .000 .620** .000 .702** .000

.658** .000 .524** .000 .529** .000

.672** .000 .620** .000

.702** .000

Table 2: Component Matrixa Component 1 FUT-Q.2 Look for consumer .795 information about something you want to buy FUT-Q.2 Look up the store that sells .876 an item you need to buy FUT-Q.2 Look through catalogs and .861 order items using your remote FUT-Q.2 See a "video" shopping mall .845 and have a salesperson demonstrate the products for you

2. Identify and Verify the Multiple Item Measures When reviewing the telephone and mail questionnaires, possible multiple item measures were identified in three different sections: general beliefs and attitudes, technology, and shopping. When items could potentially have an overlap in meaning, those items were tested by seeing whether respondents interpreted (and thus answered) the questions in a similar way. First, the possible multiple item measures were grouped conceptually. To confirm and revise the initial grouping a quantitative approach was used, running a correlation between the items, a principal components factor analysis with varimax rotation, and checking the reliability of the measure using Cronbachs alpha. If the proposed measures met the requirements for each of these steps, the individual items were combined into a new variable. The proposed constructs and corresponding measures of the general beliefs and attitudes are organized in Table 3.

10

Table 3: Proposed Constructs and Corresponding Measures of General Beliefs and Attitudes Proposed Constructs Measures Lack of spare time yl5: I Feel Like I Never Have Any Spare Time yl28: On Most Days I Can't Get Everything Done I Need to Do yl57: I Never Have Time to do the Things I Really Want to Do yl9: Free Time is Something I Make Certain I Allow Myself (Reversed) Attitude toward an yl38: I Always Finish What I Start organized life yl43: I'm the Most Organized Person I Know Seeking sense of yl8: I Find My Job Extremely Fulfilling achievement from yl11:What I Do is More Important Than What I Earn work yl7: I Think of My Job as a Means to a Paycheck, Not as Part of a Career (Reversed) yl19: If I Were Rich, I Wouldn't Work (Reversed) Satisfaction with yl10: I Like My Life Just the Way it is Now current life yl39: I Have More Self-Confidence Than Most People My Age yl40: I Consider Myself a Happy Person yl46: I Live Life to the Fullest Positive attitude tv9b: TV is My Link to the Outside World toward watching TV tv9c: I Would Love To Be Able To Watch TV 24 Hours a Day tv9h: It's the Best Free Entertainment I Can Get tv9i: I'm a Real "Couch Potato" Likelihood to be an sh4: I like to try new products when they first come out early adopter sh15: I am very interested in any new products and services yl32: I Consider Myself a Trend Setter yl58: I am Usually the First to Try New Things ct1v: I love cutting edge, high tech things Attitude toward time sh16: Saving time is more important than saving a few dollars saving ct1u: Anything that saves me time is important yl54: I Hate Waiting in Line yl56: Anything That Saves Me Time is Important Concern of taking yl3: It Bothers Me When Something Unexpected Interrupts My Daily risks Routine yl4: I Feel Uncomfortable Trying New Things yl12: Everything is Changing Too Much These Days yl14: I'm a Person Who Does Not Take Risks yl44: I Get Easily Annoyed When Things Don't Go as I Planned Attitude toward q48: Interest in Learning More About the Information Superhighway information q49: Importance of Information Highway to Your Future superhighway

11

The inspection of the correlation matrix revealed the presence of the coefficients of different ranges for each of the proposed group of measures, most of which have comparable coefficients that can be accepted. The details of the correlation matrix can be found in Appendix 1. A few measures shown relatively low coefficients compared to other measures within their proposed groups, so we raised red flags for those measures. The problematic measures with their Pearson coefficients are shown below. According to Table 4, for the proposed construct Lack of Spare Time, the correlation coefficients drop dramatically for the measure Free time is something I make certain I allow myself. Therefore, a red flag was raised to this measure at this stage. The same situation was also discovered for the construct Seeking sense of achievement from work with the measure If I were rich, I wouldn't work and the construct Concern of taking risks with the measure I get easily annoyed when things don't go as I planned. (See Table 5 and Table 6) Therefore, red flags were also raised for those two measures. Table 4: Inter-item Correlation of Lack of Spare Time YL-Q.1 Psychographics-I Feel Like I Never Have Any Spare Time 1 .359 .000 YL-Q.1 Psychographics-On Most Days I Can't Get Everything Done I Need to Do .359 .000 YL-Q.1 Psychographics-I Never Have Time to do the Things I Really Want to Do .570 .356 .000 .000 RE: YL-Q.1 Psychographics-Free Time is Something I Make Certain I Allow Myself .370 .132 .239 1 .570 .000 .356 .000 1

12

.000 .000

.000

Table 5: Inter-item Correlation of Seeking Sense of Achievement from Work RE: YL-Q.1 Psychographics-I Think of My Job as a Means to a Paycheck, Not as Part of a Career 1 .404 .000 YL-Q.1 Psychographics-I Find My Job Extremely Fulfilling .404 .000 YL-Q.1 Psychographics-What I Do is More Important Than What I Earn .241 .274 .000 .000 RE: YL-Q.1 Psychographics-If I Were Rich, I Wouldn't Work .239 .150 .177 1 .241 .000 .274 .000 1

Table 6: Inter-item Correlation of Concern of Taking Risks YL-Q.1 Psychographics-It Bothers Me When Something Unexpected Interrupts My Daily Routine 1 .364 .308 .000 .000 YL-Q.1 Psychographics-I Feel Uncomfortable Trying New Things .364 .000 YL-Q.1 Psychographics-Everything is Changing Too Much These Days .308 .000 YL-Q.1 Psychographics-I'm a Person Who Does Not Take Risks .230 .000 YL-Q.1 Psychographics-I Get Easily Annoyed When Things Don't Go as I Planned .408 .272 .000 .314 .230 .000 .000 .170 .216 .089 1 .272 .000 1 .230 .000 .314 .000 .230 .000 1

13

Besides the inspection within each proposed group, the coefficient across the groups was reviewed to identify whether each measure was assigned to the appropriate group. Since no other measure has a coefficient greater than .60, the assignment of the measures is acceptable. A factor analysis was then conducted for these measures. Principal component analysis revealed that these measures explained 54.2% of the total variance, which is acceptable (greater than 50%). Generally, the factor loading distribution (see Table 7) shows a consistency within each proposed group, which means that the measures should belong to the same construct. The details of the factor loading distribution can be found in Appendix 2. However, there are also some problematic measures found in this stage. They are discussed below. According to Table 7, the researchers found that the three red flags raised in the interitem correlation test were confirmed by the low factor loadings (Bold in the Table 7) of these measures when compared to other measures in their groups. Therefore, the researchers were more confident about the decision to remove those items from the groups. Table 7: Factor loadings for the measures Component 1 2 3 4 5 6 YL-Q.1 Psychographics-I Feel Like I .017 .156 .792 .102 Never Have Any Spare Time .082 .031 YL-Q.1 Psychographics-On Most Days I Can't Get Everything Done I .073 .155 .110 .589 .056 .040 Need to Do YL-Q.1 Psychographics-I Never Have Time to do the Things I Really Want to Do RE: YL-Q.1 Psychographics-Free Time is Something I Make Certain I Allow Myself YL-Q.1 Psychographics-I'm the Most Organized Person I Know .047 .070 .083 .048 .377 .092 .235 .162 .184 .731 .178 .078 .117 .008 .053 .051

7 .049 .055

8 .027 .011 .105 .032 .070

9 .045 .357 .032 .097 .730

.083 .028 .035

.588 .058

14

YL-Q.1 Psychographics-I Always Finish What I Start RE: YL-Q.1 Psychographics-I Think of My Job as a Means to a Paycheck, Not as Part of a Career YL-Q.1 Psychographics-I Find My Job Extremely Fulfilling YL-Q.1 Psychographics-What I Do is More Important Than What I Earn RE: YL-Q.1 Psychographics-If I Were Rich, I Wouldn't Work YL-Q.1 Psychographics-I Like My Life Just the Way it is Now YL-Q.1 Psychographics-I Have More Self-Confidence Than Most People My Age YL-Q.1 Psychographics-I Consider Myself a Happy Person YL-Q.1 Psychographics-I Live Life to the Fullest TV-Q.9 TV is My Link to the Outside World TV-Q.9 I Would Love To Be Able To Watch TV 24 Hours a Day TV-Q.9 It's the Best Free Entertainment I Can Get TV-Q.9 I'm a Real "Couch Potato" SHP-Q.1 Iike to try new products when they first come out SHP-Q.1 I am very interested in any new products and services YL-Q.1 Psychographics-I Consider Myself a Trend Setter YL-Q.1 Psychographics-I am Usually the First to Try New Things CT-Q.1 I love cutting edge, high tech things

.002 .056 .031 .011 .138 .069 .183 .005 .281 .096 .139 .002 .029 .679 .562 .613 .758 .330

.224 .079

.022 .171 .047 .128 .145 .046 .045 .180 .002 .149 .119 .001 .041 .110 .046 .262 .014 .189

.008 .057 .003 .093 .009 .226 .016 .009 .018 .104 .044 .007 .107 .011 .003 .064 .129 .002

.003 .009

.045 .021 .017 .018 .040 .166 .114 .085 .061 .057 .155 .191 .010 .133 .037 .217 .101 .407

.023 .003

.011 .712

.720 .067

.246 .236 .203 .611 .535 .740 .664 .042 .032 .027 .171 .012 .010 .212 .157 .025

.037 .008 .011 .027 .100 .085 .005 .048 .028 .129 .056 .184 .466 .085 .106 .509

.109 .023 .231 .044 .049 .083 .049 .604 .646 .661 .679 .190 .136 .045 .018 .059

.678 .591 .580 .318 .116 .095 .031 .100 .041 .069 .059 .021 .006 .003 .043 .004

.145 .128 .028 .046 .373 .106 .097 .116 .025 .023 .149 .056 .017 .142 .128 .062

15

SHP-Q.1 Saving time is more important than saving a few dollars CT-Q.1 Anything that saves me time is important YL-Q.1 Psychographics-I Hate Waiting in Line YL-Q.1 Psychographics-Anything That Saves Me Time is Important YL-Q.1 Psychographics-It Bothers Me When Something Unexpected Interrupts My Daily Routine YL-Q.1 Psychographics-I Feel Uncomfortable Trying New Things YL-Q.1 Psychographics-Everything is Changing Too Much These Days YL-Q.1 Psychographics-I'm a Person Who Does Not Take Risks YL-Q.1 Psychographics-I Get Easily Annoyed When Things Don't Go as I Planned Q.48 Interest in Learning More About the Information Superhighway Q.49 Importance of Information Highway to Your Future

.124 .026 .046 .039 .037 .258 .008 .426 .115

.004 .078 .035 .203 .095 .112 .100 .035 .114 .060 .000

.002 .022 .324 .221

.028 .167 .035 .355

.578 .714 .471 .532

.034 .154 .173 .019 .032 .206 .234 .329 .062

.001 .039 .034 .075

.105 .029 .149 .034 .022 .061 .018 .025 .089 .027 .017

.022 .088 .051 .071

.715

.141

.110 .070 .183 .043

.139

.053

.504 .590 .284

.137 .141 .041

.112 .103 .155

.100 .147 .097

.653 .057 .089

.022 .010 .001

.200

.036

.129

.063 .033

.160 .165

.832 .825

.019 .037

.025 .007

Then, the reliability for each of the constructs was tested. The Cronbachs alpha of the proposed construct Lack of spare time is .652. The Cronbachs alpha goes up to .680 when the reversed measure Free time is something I make certain I allow myself is deleted. According to this result and the red flags for this measure in the previous stage, the researchers took this measure out and computed the true scores for the construct Lack of spare time with the other three measures, which are I feel like I never have any spare time, On most days I can't get everything done I need to do, and I never have time to do the things I really want to do.

16

The Cronbachs alpha of the proposed construct Attitude toward an organized life is .54, which is acceptable. Since this construct only have two measures, its not possible to verify the consistency of the measures by conducting Cronbachs alpha if item is deleted. Therefore, the true score of the construct was computed Attitude toward an organized life with the two measures, which are I Always Finish What I Start and I'm the most organized person I know. The Cronbachs alpha of the proposed construct Seeking sense of achievement from work is .53. The Cronbachs alpha goes up to .54 when the reversed measure If I were rich, I wouldn't work is deleted. Although this is not a significant improvement, according to the red flags for this measure in the previous stage, the researchers took this measure out and computed the true scores for this construct with the other three measures, which are I find my job extremely fulfilling, What I do is more important than what I earn and I think of my job as a means to a paycheck, not as part of a career (Reversed). The Cronbachs alpha of the proposed construct Satisfaction with current life is .66, showing a good internal consistency of the measures. There is no significant improvement if any of the items were deleted. Therefore, the researchers computed the true scores for this construct with the four measures, which are I like my life just the way it is now, I have more selfconfidence than most people my age, I consider myself a happy person and I live life to the fullest. The Cronbachs alpha of the proposed construct Positive attitude toward watching TV is .586, showing a mediocre internal consistency of the measures. There is no significant improvement if any item was deleted. Therefore, the researchers computed the true scores for

17

this construct with the four measures, which are TV is my link to the outside world, I would love to be able to watch TV 24 hours a day, It's the best free entertainment I can get and I'm a real couch potato. The Cronbachs alpha of the proposed construct Likelihood to be an early adopter is .694, showing a good internal consistency of the measures. There is no significant improvement if any items are deleted. Therefore, the researchers computed the true scores for this construct with the five measures, which are I like to try new products when they first come out, I am very interested in any new products and services, I consider myself a trend setter, I am usually the first to try new things and I love cutting edge, high tech things. The Cronbachs alpha of the proposed construct Attitude toward time saving is .545, showing a mediocre internal consistency of the measures. There is no significant improvement if any item is deleted. Therefore, the researchers computed the true scores for this construct with the four measures, which are Saving time is more important than saving a few dollars, Anything that saves me time is important, I hate waiting in line and Anything that saves me time is important. The Cronbachs alpha of the proposed construct Concern of taking risks is .649, showing a good internal consistency of the measures. There is no significant improvement if any item is deleted. However, the measure I get easily annoyed when things don't go as I planned. gained two red flags at the previous stages. Therefore, the researchers decided to take it out and computed the true scores for this construct with the four measures, which are It bothers me when something unexpected interrupts my daily routine, I feel uncomfortable trying new things, Everything is changing too much these days, I'm a person who does not take risks.

18

The Cronbachs alpha of the proposed construct Attitude toward Information Superhighway is .804, showing an excellent internal consistency of the measures. Since this construct only has two measures, its not possible to verify the consistency of the measures by conducting Cronbachs alpha if item is deleted. Therefore, the researchers computed the true score of this construct with the two measures, which are Interest in learning more about the Information Superhighway and Importance of Information Highway to your future. For all the constructs mentioned above, one construct of sum and one construct for average were created respectively for different analysis purposes in the following stages. Here are the final groupings and their factor loading for the shopping part. With these groupings, 65.3% of the total variance was explained. Table 8: Final factor loading (shopping section)
Component 3 0.102 -0.047 0.833 0.832 0.032 0.051 0.123 0.187 -0.037

1 SHP-Q.1 I worry more about spending now because of the state of the economy SHP-Q.1 I sometimes worry about not having enough money to pay my bills SHP-Q.1 I tend to do a lot of research and shop around before I make big purchases SHP-Q.1 I usually know where to get a good price SHP-Q.1 I usually stick to the brands I know SHP-Q.1 If a company has been around for a long time, I feel it has better products or services SHP-Q.1 If I've found something that's good, then I don't like to change SHP-Q.1 I am more likely to buy something sold in a store than sold in a catalog RE: SHP-Q.1 I like to shop by mail so that I don't have to deal with salespeople 0.085 -0.076 0.036 0.034 0.021 0.019 0.03 0.701 0.828

2 0.13 -0.025 0.069 0.113 0.806 0.789 0.635 0.206 -0.049

4 0.761 0.82 0.05 0.005 -0.024 0.01 0.138 0.175 -0.096

5 0.017 0.014 -0.086 0.035 -0.014 0.02 -0.058 0.066 -0.162

19

RE: SHP-Q.1 I'm too busy to shop in stores, so buying through the mail is a real convenience

0.858

-0.038

-0.026

-0.039

0.032

The last group was deleted because of its low reliability. Four new variables of the multiple-item measures were created.

Worry about money (The Cronbachs alpha coefficient is .47.) Likely to buy from store (The Cronbachs alpha coefficient is .60.) Money concern (The Cronbachs alpha coefficient is .64.) Brand loyalty (The Cronbachs alpha coefficient is .72)
Table 9: New multiple item measures

Items SHP-Q.1 I worry more about spending now because of the state of the economy SHP-Q.1 I sometimes worry about not having enough money to pay my bills SHP-Q.1 I tend to do a lot of research and shop around before I make big purchases SHP-Q.1 I usually know where to get a good price SHP-Q.1 I usually stick to the brands I know SHP-Q.1 If a company has been around for a long time, I feel it has better products or services SHP-Q.1 If I've found something that's good, then I don't like to change SHP-Q.1 I am more likely to buy something sold in a store than sold in a catalog RE: SHP-Q.1 I like to shop by mail so that I don't have to deal with salespeople RE: SHP-Q.1 I'm too busy to shop in stores, so buying through the mail is a real convenience

Construct Worry about money

Money concern

Brand loyalty

Likely to buy from store

The Cronbachs alpha coefficient of Worry about money is .47 with two items. The Cronbachs alpha coefficient of Likely to buy from store is .60 with two items. The Cronbachs alpha coefficient of Money concern is .64 with two items. The Cronbachs alpha coefficient of Brand Loyalty: is .72 with two items. Since all these constructs only have two measures, its not possible to verify the consistency of the measures by conducting Cronbachs alpha if item is deleted. The true score was computer for all the constructs.

20

Here are the final groupings and their factor loadings for the technology attitude part. In this case, 69.4% of the total variance was explained. Table 10: Final factor loading (technology section)
Component 1 CT-Q.1 My answering machine is an essential part of my daily living -- I couldn't do without it CT-Q.1 I'm comfortable retrieving messages from my answering machine when I'm away from home CT-Q1 I resent it when I call and get an answering machine CT-Q.1 The telephone is my main contact with the outside world CT-Q.1 The telephone is an essential part of my daily living -- I couldn't do without I prefer to deal directly with a teller for my banking needs CT-Q.1 I prefer to use an ATM machine for all my banking needs CT-Q.1 I am very comfortable using an ATM to make deposits or pay bills Ct1m Computers and technology will ruin the quality of my life Computers and technology will eventually be used to limit personal freedom Computers and technology control too much of our lives already PC-Q.1 My personal computer is an essential part of my daily living -- I couldn't do without it -0.03 2 0.212 3 0.141 4 0.711 5 0.31 6 0.141

0.063 0.318 0.042 0.024

0.07 0.044 0.093

0.171 0.014 0.034

0.744 0.737 0.01

0.109 0.186 0.83

0.251 0.062 0.114 0.206

0.053 0.108 0.147

0.002

0.117

0.792

0.325 0.007 0.005

0.728 0.836

0.068 0.113

0.101 0.107

0.171 0.153

0.066

0.808

0.13

0.027

0.331

0.742

0.065 0.024

0.039 0.104

0.023

0.049 0.009 0.033

0.298

0.795

0.045

0.085 0.025

0.817

0.11

0.138

0.071

0.154

0.753

0.08

0.122

0.042

0.292

21

CT-Q.1 I often spend hours playing games on my personal computer PC-Q.1 I spend more time with my personal computer than I do with people

0.021 0.011

0.76

0.125

0.004 0.081

0.017

0.077

0.866

0.076

0.014

0.063

CT-Q.1 Computers and technology will improve the quality of my life CT-Q.1 I was nervous about using a personal computer at first, but now that I'm used to it, I love it

0.27

0.173

0.331

0.134

0.163

0.664

0.011

0.282

0.069

0.165

0.013

0.736

The true scores were computed and 6 new variables were created of the following multiple item measures. Role of computers in life (The Cronbachs alpha coefficient is .63.) Bad attitude about computers (The Cronbachs alpha coefficient is .54.) Time spent on computer (The Cronbachs alpha coefficient is .61.) Attitude towards using ATM (The Cronbachs alpha coefficient is .72) Attitude towards using telephone (The Cronbachs alpha coefficient is .74.) Attitude towards using answering machine (The Cronbachs alpha coefficient is .59.)

22

Table 11: Multiple item measures for technology Items CT-Q.1 My answering machine is an essential part of my daily living -- I couldn't do without it CT-Q.1 I'm comfortable retrieving messages from my answering machine when I'm away from home Ct1b I resent it when I call and get an answering machine CT-Q.1 The telephone is my main contact with the outside world CT-Q.1 The telephone is an essential part of my daily living -- I couldn't do without I prefer to deal directly with a teller for my banking needs CT-Q.1 I prefer to use an ATM machine for all my banking needs CT-Q.1 I am very comfortable using an ATM to make deposits or pay bills Ct1m Computers and technology will ruin the quality of my life Computers and technology will eventually be used to limit personal freedom Computers and technology control too much of our lives already PC-Q.1 My personal computer is an essential part of my daily living -- I couldn't do without it CT-Q.1 I often spend hours playing games on my personal computer PC-Q.1 I spend more time with my personal computer than I do with people CT-Q.1 Computers and technology will improve the quality of my life CT-Q.1 I was nervous about using a personal computer at first, but now that I'm used to it, I love it Construct Attitude towards using answering machine

Attitude towards using telephone

Attitude towards using ATM

Bad attitude about computer

Time spent on computer

Role of computer in life

The Cronbachs alpha coefficient of Role of computer in life is .63.The Cronbachs alpha coefficient of Bad attitude about computer is .54.The Cronbachs alpha coefficient of Time spent on computer is .61.The Cronbachs alpha coefficient of Attitude towards 23

using ATM is .72. The Cronbachs alpha coefficient of Attitude towards using telephone is .74.The Cronbachs alpha coefficient of Attitude towards using answering machine is .59. There is no increase of Cronbachs alpha value if each of the items in the same constructs was deleted. In this way, the true score for all the constructs was computed. 3. Building a Regression Model In order to answer the research question, it is essential to determine the variables that influence the behavior of home shopping, and how much each of those variables influence home shopping. To this end, the multiple regression was used to evaluate the variables for their contribution to the overall explanation of likelihood of home shopping behavior. The phone and mail survey items were again reviewed for variables that could have a direct effect on attitude and use of the new home shopping technology. A list of 43 variables was compiled from the mail and phone survey. Each of these 43 variables were evaluated for its contribution to the explanation, represented by the beta value, and to see whether the variable made a statistically significant and unique contribution to the explanation, represented by the significance level. The results from the regression showed that 11 of these variables were significant at the 95 percent confidence level. Table 12 lists all of the predictors with the highlighted variables as significant. After removing the variables that did not have significance at the 95 percent confidence level, a second regression was run to select the relevant variables. This process of removing nonsignificant variables and re-running the model was iterated until there were nine predictors that were significant at the 95 percent confidence level. These nine predictors explained a 22.5 percent variation in answers from respondents in the surveys.

24

The next step was to examine the standardized beta coefficients to determine the impact each of the variables had on the 22.5 percent of variation explained. After carefully removing only the variables that effected the variation less than one percent, five variables were selected. These five variables explained nearly 21 percent (20.9%) of the variation in answers from respondents in the surveys. The final list of variables is in Table 13. Table 12: List of predictors

25

Table 13: Final list of predictors

4. Cluster Analysis 5. The final regression model identified five explanatory variables of interest relevant to this investigation: having choices, being able to work from home, attitude towards the information superhighway, preference of buying from a store, and the likelihood to be an early adopter. Those five variables were used to assign all individuals to specific groups using cluster analysis. Cluster analysis is a method of looking for similarities among groups of individuals within the sample to enable the analysis of the characteristics of those different groups. Grouping the sample into clusters provides a different lens to explore and interpret the available data for strategic planning to reach these different groups. A k-means cluster analysis was run on 1,818 cases in the sample. In k-means cluster analysis, the software forms the most distinct groups possible based on the proposed number of groupings. The researchers evaluated three, four and five possible groupings, considering both the number of individuals who were in the proposed group and also whether the groupings made qualitative sense. After evaluating the different potential arrangements, the researchers decided on four clusters. The four clusters are detailed in Table 14.

26

Table 14: Cluster Analysis results Cluster 1 YL-Q.1 Psychographics-I Like Having Choices Available to Me YL-Q.1 Psychographics-If I Could Earn a Living Without Leaving My Home, I Would Prefer buy from store Don't Buy from mail or with phone avg (sh12 sh19 sh23) Attitude toward the super highway avg (q48, q49) Likelihood to be an early adopter avg (sh4, sh15, yl58, yl32, ct1v) New Dependent variable avg - use of shopping (fu2a5DV fu2a6DV fu2a7DV fu2a4) slightly high high 2 moderate 3 moderate 4 slightly high

low

high

low

slightly low slightly high high

high

high

moderate

low low

slightly low slightly low

high high

high

low

low

slightly high tech friendly open consumers

Cluster names

optimistic trendsetters

traditional routiners

skeptical latecomers

Number of Cases in each Cluster Cluster 1 2 3 4 Valid Missing 593 331 538 356 1818 318

27

1) Demographics Age Each cluster had a majority of respondents within a different age category. Twenty nine percent of the optimistic trendsetters fell within the 30-39 age category (29.3%, SE:2.1, 95% CI). Tech friendly open consumers were almost exactly split between 20-29 (21.1%, SE:1.9, 95% CI) and 40-49 (21.6%, SE:1.9, 95% CI). Skeptical latecomers had a majority concentrated within the 40-49 age category 23.2%, and traditional routiners had the oldest majority, falling within the 60-69 age group (22.1%,SE:1.9, 95% CI). A chi square test of significance confirmed that optimistic trendsetters are significantly more likely to be within the 30-39 group than the 60-69 age group when compared with tech friendly open consumers, 2(1, N=949) = 15.44, p<.001. There was no significant difference in age grouping between 30-39 and 40-49 when comparing optimistic trendsetters to skeptical latecomers, 2(1, N=1131) = 2.36, not sig., nor was there a significant difference in age grouping between 30-39 and 40-49 when comparing optimistic trendsetters to tech friendly open consumers 2(1, N=949) = 3.38, not sig. In sum, the optimistic trendsetters are mostly in their thirties, the tech friendly open consumers are either slightly younger (20s) or slightly higher (40s), the skeptical latecomers are in their 40s, and the traditional routiners are in their 60s. Gender All four clusters were almost equally balanced in terms of gender. The overall gender ratio for the sample was 48.2% male and 51.8% female (both SE:2.3, 95% CI). The optimistic trendsetters most closely matched this percentage with 47.9% male and 52.1% female (both SE:2.3, 95% CI).. The tech friendly open consumers had 50.6% male and 49.4% female (both

28

SE:2.3, 95% CI), and the traditional routiners were 51.4% male and 48.6% female (both SE:2.3, 95% CI). There was no statistically significant difference in gender between any of the groups. Ethnicity Respondents were asked to report whether they were Hispanic. A majority of all four groups were not of Hispanic ethnicity: 95.8% of traditional routiners (SE:0.9, 95% CI), 94.1% of skeptical latecomers (SE:1,1, 95% CI), 93.4% of optimistic trendsetters (SE:1.1, 95% CI), and 91.0% of tech friendly open consumers (SE: 1.3, 95% CI). Tech friendly open consumers and optimistic trendsetters were more likely to have Hispanic people within the group than the other two groups. Race A majority of respondents in all four groups reported they were white when asked about race: 92.1% of traditional routiners (SE:1.2, 95% CI), 88.3% of skeptical latecomers (SE:1.5, 95% CI), 86.5% of tech friendly open consumers (SE:1.6, 95% CI) and 83.1% of optimistic trendsetters (SE:1.7, 95% CI). The optimistic trendsetters were the most diverse group, with 8.4% black respondents (SE:2.2, 95% CI), 1.9% Asian (SE:1.1, 95% CI), 4.4% other (SE:1.7, 95% CI), and they had the highest percentage of individuals that refused to answer the question. Marital status The majority of respondents in all four groups are married. Sixty eight percent (68.0%, SE:2.1, 95% CI) of skeptical latecomers, 56.5% of traditional routiners (SE:2.3, 95% CI),

29

56.2% of optimistic trendsetters (SE:2.3, 95% CI), and 43.0% of tech friendly open consumers reported that they were married (SE:2.3, 95% CI). The second most popular status was single, never married. Tech friendly open consumers were the highest percentage of single individuals with 40.2% (SE:2.3, 95% CI), optimistic trendsetters were second at 29.2% (SE:2.1, 95% CI), traditional routiners were 22.7% (SE:1.9 95% CI), and 16.5% of skeptical latecomers were single (SE:1.7, 95% CI). Since married and single and never married were the two top choices, a chi square test of independence was used to determine whether there was a significant difference between optimistic trendsetters and tech friendly open consumers, our two top target groups. The results of the test were significant 2(1, N= 802) =15.6, p<.001, suggesting that optimistic trendsetters are much more likely to be married than tech friendly open consumers. Income To assist with interpretation and analysis, the income data available in the data set were recoded into three categories: 1. Under $35,000, 2. $35,000-$75,000 and 3. $75,000 and up. Most of the respondents in all three categories have incomes of less than $35,000, forty percent (40.0%, SE:2.3, 95% CI) of the skeptical latecomers, 38.8% of the optimistic trendsetters (SE:2.2, 95% CI), 38.7% of the traditional routiners (SE:2.2, 95% CI), and 33.7% of the tech friendly open consumers (SE:2.2, 95% CI). The tech friendly open consumers had the highest percentage of respondents who reported income of $75,000 and up. Slightly over 30% of each of the four groups had income in the $35,000-$75,000 range, with 35.9% of optimistic trendsetters (SE:2.2, 95% CI), 35.6% of traditional routiners (SE:2.2, 95% CI), 34.4% of skeptical latecomers (SE:2.2, 95% CI), and 32.6% of tech friendly open consumers (SE:2.2, 95% CI). Thirty three percent (33.7%, SE:2.2, 95% CI) reported an income of $75,000 or higher, compared to 25.7% of both traditional routiners and skeptical latecomers (SE:2.0, 95% CI), and 30

25.3% of optimistic trendsetters (SE:2.0, 95% CI).Chi square tests of independence showed no significant difference among any of the four groups in the three income groups (below $35,000, between $35,000-$75,000, and over $75,000). Education There was a wide range of education levels among the four groups, with responses in categories ranging from grade school or less to post graduate education. After reviewing the frequencies, the responses were recoded into three categories: did not graduate from high school, high school graduate, and college graduate. Over sixty percent of the respondents reported graduating from high school in three of the four clusters: sixty one percent of optimistic trendsetters (61.2%, SE:2.2, 95% CI), sixty percent of traditional routiners (60.6%, SE:2.2, 95% CI) and sixty one percent of skeptical latecomers (61.3%, SE:2.2, 95% CI). Only 49.6% of tech friendly open consumers reported they had graduated from high school (SE:2.3, 95% CI), however that cluster had the highest number reporting graduating from college or additional education (36.8%, SE:2.2, 95% CI)). The other three clusters had a lower percentage of college graduates: traditional routiners had 30.9% (SE:2.1, 95% CI), optimistic trendsetters had 28.6% (SE:2.1, 95% CI), and skeptical latecomers had 26.6% (SE:2.0, 95% CI). A series of chi square tests of independence demonstrated that none of the groups had a significance difference in high school graduates versus percentage of college graduates. A similar series of tests showed no significant difference in the percentage of college graduates versus percentage of post-graduates in each of the groups. Own or rent primary residence

31

When asked whether they own or rent their primary residence, 78.2% of the traditional routiners reported owning their home (SE:1.9, 95% CI), compared to 75.8% of skeptical latecomers (SE:2.0, 95% CI), 68.8% of tech friendly open consumers (SE:2.1, 95% CI), and 67.6% of optimistic trendsetters (SE:2.2, 95% CI). A chi square test of independence showed no significant difference between owning and renting among optimistic trendsetters and tech friendly open consumers. Number of people in household The phone survey also included questions about other individuals in the household, including the number of people who share the residence. A one-way between-groups analysis of variance was conducted to determine whether there was a difference between the four clusters. The result showed a statistically significant difference at the p<.001 level in number of persons per household for the three groups: F (3, 1817) =4.6, p=.01. The mean difference between each group was compared using Tukeys HSD test. The average number of children per household for optimistic trendsetters (M=3.04, SD = 1.49) was significantly different from the mean score for traditional routiners (M=2.63, SD = 1.73), and also the average number of children for tech friendly open consumers (M=2.67, SD = 1.40). The skeptical late comers (M= 2.95, SD=1.40) was not significantly different from any of the other groups. Type of residence Single family homes were the most commonly reported types of residence in all four clusters: traditional routiners had the highest percentage of members living in single family homes at 77.6% (SE:1.9, 95% CI), skeptical latecomers reported 76.2% (SE:2.0, 95% CI), optimistic trendsetters reported 72.7% (SE:2.0, 95% CI), and tech friendly open consumers

32

reported 62.9% (SE:2.2, 95% CI). Twenty-one (21.3%, SE:1.9, 95% CI) of the tech friendly open consumers and 16.9% (SE:1.7, 95% CI) of optimistic trendsetters reported living in an apartment, while 13.6% of traditional routiners (SE:1.6, 95% CI), and 11.9% of skeptical latecomers reported living in an apartment (SE:1.2, 95% CI). A chi square test of independence was used to determine whether dwelling type was significantly different between optimistic trendsetters and tech friendly open consumers; the results were significant 2(1, N = 831) =4.85, p<.05. Employment Respondents were asked whether they were currently employed. Of the four groups, 74.4% of the tech friendly open consumers (SE:2.4, 95% CI), 72.5% of the optimistic trendsetters (SE:2.5, 95% CI), 67.1% of the skeptical latecomers (SE:2.5, 95% CI), and 63.7% of the traditional routiners reported being employed (SE:2.6, 95% CI). The difference between optimistic trendsetters and skeptical latecomers was marginally significant, 2(1, N = 1131) = 3.93, p <.05. Respondents who responded that they were employed were also asked if they were full or part time. Tech friendly open consumers had the lowest percentage of full time workers with 77.4% reporting full time work (SE:2.3, 95% CI), optimistic trendsetters had 82.3% (SE:2.1, 95% CI), skeptical latecomers had 83.7% (SE:2.0, 95% CI), and traditional routiners reported 83.9% (SE:2.0, 95% CI). Respondents were also asked whether they worked at home or elsewhere. Of the 288 people who responded, 50% of the skeptical latecomers work from home (SE:5.8, 95% CI), 43.3% of the optimistic trendsetters (SE:5.7, 95% CI), 32.7% of the skeptical latecomers (SE:5.4, 95% CI), and 29.4% of the tech friendly open consumers (SE:5.3, 95% CI).

33

Free Time All respondents were asked whether they like to stay at home in their spare time or whether they enjoy staying home in their spare time. Of the four groups, the skeptical latecomers (64.7%, SE:2.2, 95% CI) had the highest percentage who would choose to stay at home in their free time, the optimistic trendsetters had the second highest likelihood of staying at home (60.7%, SE:2.2, 95% CI), the traditional routiners were third (55.6%, SE:2.3, 95% CI) and the tech friendly open consumers were the least likely to stay at home (45.2%, SE:2.3, 95% CI). There is a statistically significant difference between the optimistic trendsetters and the tech friendly open consumers, the two groups most likely to use the new technology, 2(1, N = 848) = 25.70, p <.001. 2) Beliefs Attitude towards TV According to F test result for the relevant variables, a one-way between subjects ANOVA was conducted to test mean difference on Positive attitude toward watching TV among the four clusters. The means of this variable differed significantly across the four clusters, F (3, 1437) = 8.209, p = .000. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 3.69, 95% CI [3.53, 3.84]) show a significantly more positive attitude toward watching TV than the traditional routiners (M = 3.07, 95% CI [2.87, 3.25]), p = .000 and the tech friendly open consumers (M = 3.34, 95% CI [3.16, 3.52]), p = .024. Comparisons between the optimistic trendsetters and the skeptical latecomers were not statistically significant at p > .05. Role of computers in life

34

A one-way between subjects ANOVA was conducted to test mean difference on the Role of computer in life among the individuals in the four clusters. The results of this test showed a significant difference across the four clusters, F (3, 1763) = 22.77, p = .000. Tukey post-hoc comparisons of the four clusters indicate that computers play a more important role in the lives of the optimistic trendsetters (M = 2.52, 95% CI [2.36, 2.68]) than the skeptical latecomers (M = 1.83, 95% CI [1.69, 1.97]), p = .000 and the traditional routiners (M = 1.69, 95% CI [1.55, 1.83]), p = .000. The optimistic trendsetters and tech friendly open consumers have a very similar and positive attitude towards computers, as the comparison between those two groups was not statistically significant (p > .05). A one-way between-subjects ANOVA was conducted to test mean difference on Attitude toward computers among the four clusters. The means of this variable differed significantly across the four clusters, F (3, 1754) = 53.18, p = .000. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 5.89, 95% CI [5.69, 6.09]) show significantly more positive attitudes towards computers than the traditional routiners (M = 4.50, 95% CI [4.23, 4.78]), p = .000 and the skeptical latecomers (M = 4.43, 95% CI [4.22, 4.65]), p = .000. Comparisons between the optimistic trendsetters and the tech friendly open consumers were not statistically significant (p > .05). Concern of money A one-way between subjects ANOVA was conducted to test the difference in the means regarding Concern of money among the individuals in the four clusters. The there were significant differences in the mean for concern of money across the four clusters, F (3, 1794) = 9.31, p = .000. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 7.56, 95% CI [7.41, 7.72]) show significantly more concern of money than the 35

traditional routiners (M = 6.84, 95% CI [6.61, 7.08]), p = .000. Comparisons between the optimistic trendsetters and the skeptical latecomers, the tech friendly open consumers were not statistically significant at p > .05. Price consciousness Respondents were also asked to comment on their ideas about price when shopping. A one-way between-groups analysis of variance was conducted to determine whether there was a difference in concern about money between the four clusters. The result showed a statistically significant difference at the p<.001 level in money concern for the four groups: F (3, 1794) = 9.3, p=.01. In a scale of 1 to 10 with 1 meaning least important and 10 meaning most important, the mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=7.56, SD = 1.93) was significantly different from the mean score for traditional routiners (M=6.84, SD = 2.16). The average degree of money concern for traditional routiners was also significantly different from the mean score for skeptical latecomers (M= 7.28, SD = 1.96) and tech friendly open consumers (M= 7.32, SD = 1.94). The optimistic trendsetters are more concerned about money than traditional routiners. Traditional routiners are less concerned about money than all the other groups. Concern about saving time The result showed a statistically significant difference at the p<.001 level in time saving for the four groups: F (3, 1814) = 66.64, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=3.02, SD = 1.07) was significantly different from the mean score for traditional routiners (M=2.22, SD = 1.04), and from skeptical latecomers (M= 2.43, SD = 1.08). The average degree of concern for

36

saving time among traditional routiners was also significantly different from the mean score for tech friendly open consumers (M= 3.11, SD = 1.15). Optimistic trendsetters are more interested in saving time than the traditional routiners and skeptical latecomers. Tech friendly open consumers are also willing to save more time then the traditional routiners. Lack of spare time Respondents were also asked a slightly different question, whether they felt that they lacked spare time. The result showed a statistically significant difference for the four groups: F (3, 1754) = 17.3, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=6.28, SD = 1.86) was significantly different from the mean score for traditional routiners (M=5.43, SD = 1.82), and from tech friendly open consumers (M= 5.7, SD = 1.87). The average degree of lack of spare time for traditional routiners was also significantly different from the mean score for tech friendly open consumers and skeptical latecomers (M = 6.08, SD = 1.82). Optimistic trendsetters have the least amount of spare time, skeptical latecomers are the group with the second least amount of spare time, and traditional routiners and tech friendly open consumers are the least busy of the clusters. Pay more for quality The results showed a statistically significant difference at the p<.001 level in their willingness to pay more for better quality for the three groups: F (3, 1805) = 9.92, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=7.99, SD = 1.86) was significantly different from the mean score for skeptical latecomers (M=7.72, SD = 1.95). The average degree of willing to pay more for higher quality for skeptical latecomers was also significantly different from the mean score for

37

tech friendly open consumers (M = 8.15, SD = 1.60) and skeptical latecomers. Optimistic trendsetters and tech friendly open consumers are more willing to pay for higher quality than skeptical latecomers. Pay more for convenience The results showed a statistically significant difference at the p<.001 level in their willingness to pay more for convenience for the four groups: F (3, 1807) = 24.06, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=5.3, SD = 2.68) was significantly different from the mean score for traditional routiners (M = 4.33, SD = 2.56) and skeptical latecomers (M=4.31, SD = 2.62). The average degree of willing to pay more for convenience for traditional latecomers was significantly different from the mean score for tech friendly open consumers (M = 5.46, SD = 2.58).The average degree of willing to pay more for convenience for skeptical latecomers was also significantly different from the mean score for tech friendly open consumers and skeptical latecomers. Optimistic trendsetters and tech friendly open consumers are more willing to pay more for convenience than the traditional routiners and skeptical latecomers. Buying expensive things to impress friends The result showed a statistically significant difference at the p<.001 level in buying expensive things to impress friends for the four groups: F (3, 1813) = 17.82, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=2.44, SD = 2.30) was significantly different from the mean score for traditional routiners (M = 1.69, SD = 1.43) and skeptical latecomers (M=1.82, SD = 1.66). The average degree of buying expensive things to impress friends for traditional latecomers was

38

significantly different from the mean score for tech friendly open consumers (M = 2.42, SD = 2.04).The average degree of buying expensive things to impress friends for skeptical latecomers was also significantly different from the mean score for tech friendly open consumers and skeptical latecomers. Optimistic trendsetters and tech friendly open consumers are more likely to buy expensive things to impress their friends than the traditional routiners and skeptical latecomers. Overall privacy concern Each cluster was analyzed for how concerned they are about privacy overall. The result showed a statistically significant difference at the p<.001 level in privacy concern for the four groups: F (3, 1811) = 14.89, p=.00. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=8.66, SD = 1.80) was significantly different from the mean score for traditional routiners (M = 7.74, SD = 2.43), skeptical latecomers (M=8.3, SD = 2.04) and tech friendly open consumers (M = 8.28, SD = 1.89). The average degree of privacy concern for traditional latecomers was significantly different from the mean score for tech friendly open consumers and skeptical latecomers. Optimistic trendsetters are more concern about personal privacy than other groups. Tech friendly open consumers are more concerned about personal privacy than the traditional routiners, but less than the optimistic trendsetters. Traditional routiners are least likely to be concerned about personal privacy. Importance of relationships with people The results showed a statistically significant difference at the p<.001 level in the importance of the relationships with people for the four groups: F (3, 1794) = 9.9, p=.00. The

39

mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=7.16, SD = 2.32) was significantly different from the mean score for traditional routiners (M = 6.53, SD = 2.43) and skeptical latecomers (M=6.67, SD = 2.40). The average degree for traditional latecomers was significantly different from the mean score for tech friendly open consumers (M = 7.28, SD = 2.08).The average degree of importance of relationships with people for skeptical latecomers was also significantly different from the mean score for tech friendly open consumers and skeptical latecomers. Optimistic trendsetters and tech friendly open consumers are more likely to consider personal relationships as extremely important than the traditional routiners and skeptical latecomers. Privacy and computers and technology A one-way between subjects ANOVA was conducted to test mean difference on Feelings about invasion of privacy concerning computers and technology among the four clusters. The means of this variable differed significantly across the four clusters, F (3, 1814) = 3.47, p = .015. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 2.60, 95% CI [2.50, 2.69]) show significantly higher concern about invasion of privacy by computers and technology than the tech friendly open consumers (M = 2.39, 95% CI [2.27, 2.51]), p = .046. Comparisons between the optimistic trendsetters and the traditional routiners, the skeptical latecomers and were not statistically significant at p > .05. Effect computers and technology have on control of life A one-way between-subjects ANOVA was conducted to test the mean difference on Effect computers and technology have on control of life among the four clusters. The means of this variable differed significantly across the four clusters, F (3, 1814) = 22.56, p = .000. Tukey post-

40

hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 1.42, 95% CI [1.37, 1.48]) think that computers and technology have significantly more control of their lives than the traditional routiners (M = 1.62, 95% CI [1.53, 1.71]), p = .001 and the skeptical latecomers (M = 1.69, 95% CI [1.62 1.75]), p = .000. Comparisons between the optimistic trendsetters and the tech friendly open consumers were not statistically significant at p > .05. 3) Behaviors ATM and debit cards Respondents were asked if they currently had an ATM card and if they currently had a debit card. The responses to these questions are important in two ways. First, debit cards are a relatively new technology at the time and are a sign that respondents are more open to new means to pay for goods and services. Second, ATM cards can only be used to make physical transactions, whether it be at the bank or a store. There is no way to use an ATM card over the phone, by catalog, etc. This means that respondents would need a debit or a credit card to pay for goods while home shopping. It is more important to look at the purchasing behavior with debit cards, since that is more closely related to the way customers would be purchasing products through ESIs new technology. Fifty three percent (SE: 2.1, 95% CI) of respondents in the survey have ATM cards, while only 17.6% (SE: 1.6, 95% CI) have debit cards. Chi-square analyses were run to see if there were differences between groups in who had debit cards. There is no statistical significance between optimistic trendsetters (our target group) and traditional routiners (p > .05). There is also no statistical significance between optimistic trendsetters and skeptical latecomers (p > .05) meaning that our target group and the other clusters are similar. There is also no significant

41

difference between behaviors while using debit cards between our optimistic trendsetters and the traditional routiners. Getting cash at ATM, transferring money from one account to another, making deposits, obtaining account balance, ordering tickets or commuter passes, gas purchases, grocery shopping, retail purchases and restaurant purchases were all not significant after running chi-squared tests (p > .05). Shopping and television Survey respondents answered a number of questions directly related to their previous viewing and purchasing behavior in the past month as related to infomercials, home shopping channels, radio and television commercials and on-line computer services. While there are some significant differences between groups, it is important to remember that overall, the statistics that describe past behavior with home shopping are low across all clusters. When asked whether they have ever watched infomercials, 42.9 percent (SE: 3.3, 95% CI) of the optimistic trendsetter cluster responded yes as compared with only 26.9 percent (SE: 2.7, 95% CI) of the traditional routiners. A chi-squared test of significance was used to compare the two groups and the test was significant at the level 2 (1, N=857) = 21.87, p< .001. Respondents were asked if they ever watch home shopping channels and there was a statistical difference between optimistic trendsetters and traditional routiners. Thirty six percent (36%) (SE: 3.3, 95% CI) of optimistic trendsetters have ever watched an infomercial compared with 22.6% (SE: 2.8, 95% CI) of traditional routiners. The difference was significant at the level 2 (1, N=833) = 16.48, p< .001. Look at behavior in the past month, there are significant differences between optimistic trendsetters (our target group) and traditional routiners (least likely to adopt ESIs technology). 42

Optimistic trendsetters are more likely to have bought merchandise from a home shopping channel on television, bought merchandise or services from a brief radio or television commercial where you call an 800 number to order and bought merchandise through an on-line computer service. All of these differences are statistically significant and are detailed in Table 15. Table 15: Confidence intervals and chi squared for differences in shopping between optimistic trendsetters and traditional routiners

Go out for a romantic evening The result showed a statistically significant difference at the p<.001 level in willingness to go out for a romantic evening for the four groups: F (3, 1785) = 19.24, p=.000. The mean difference between each group was compared using Tukeys HSD test. The average degree for optimistic trendsetters (M=6.77, SD = 2.87) was significantly different from the mean score for traditional routiners (M = 5.7, SD = 3.03) and skeptical latecomers (M=5.71, SD = 3.03). The average degree for traditional latecomers was significantly different from the mean score for tech friendly open consumers (M = 6.73, SD = 2.75).The average for skeptical latecomers was also significantly different from the mean score for tech friendly open consumers and skeptical

43

latecomers. Optimistic trendsetters and tech friendly open consumers are more likely to go out for a romantic evening than the traditional routiners and skeptical latecomers. Number of televisions in households A one-way between subjects ANOVA was conducted to test the mean difference on the Number of televisions in households among the four clusters. The results showed a significant difference between the number of televisions in the households among the four groups, F (3, 1814) = 9.65, p = .000. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 1.11, 95% CI [1.09, 1.14]) have fewer televisions at home than the traditional routiners (M = 1.18, 95% CI [1.14, 1.22]), p = .025 and the skeptical latecomers (M = 1.22, 95% CI [1.18, 1.25]), p = .022. Comparisons between the optimistic trendsetters and the tech friendly open consumers were not statistically significant at p > .05, suggesting that those two clusters have similar numbers of televisions in their households. Media use among different clusters For the sake of reaching the target consumers more effectively, the researchers explored the types of media regularly used by the different clusters. In the survey, a general description of the media usage frequency distributions by clusters was given first. Then each medium (TV, Radio, Magazine, Newspaper) was discussed in detail. Since most of the relevant variables are either nominal or ordinal, cross tabulations and chi-square tests are conducted to examine whether significant differences exist across the four clusters in these four areas, especially between the target cluster and the other clusters. Overall media usage

44

Table 16 shows the media usage frequency distribution by clusters. According to this table, among the four main media, optimistic trendsetters prefer television to radio, magazines, and newspapers. To test whether the optimistic trendsetters use a specific medium more than the respondents in other clusters, cross tabulations and corresponding chi-square tests between the optimistic trendsetter and other clusters with respect to the usage of TV/radio/magazine/newspaper were conducted. Among the four media, no significant differences across the clusters with respect to TV/Radio/Newspaper usage were found. However, for magazines, there is a significant difference on whether optimistic trendsetters read magazines compared to the respondents in other clusters 2(1, N=1818) = 14.36, p<.01. As shown in Table x, among the optimistic trendsetters, there are 88.9% (95%CI [86.2%, 91.6%]) of the respondents read magazines. While among the combination of the other three clusters, only 82.0% (95%CI [79.6%, 84.4%]) of the respondents read magazines on a regular basis. The media popularity ranking for the other three clusters is similar to that for optimistic trendsetters. Among the four clusters, the optimistic trendsetters and the tech friendly open consumers are more likely to use all kinds of media than the other two clusters. Table 16: Overall Media Usage of the Four Clusters Whether use the following media or not TV Radio Magazine Newspaper Optimistic Trendsetters % of 95% CI the (%) cluster 99.2 98.5,99.9 96.3 94.8,97.8 88.9 86.2,91.6 85.3 82.2,88.4 Traditional Routiners % of 95% CI the (%) cluster 98.2 96.8,99.6 92.4 89.4,95.4 81.9 77.3,86.5 83.4 79.0,87.8 Skeptical Latecomers % of 95% CI the (%) cluster 98.1 96.9,99.3 94.8 92.9,96.7 78.3 74.4,82.2 84.6 81.3,87.9 Tech Friendly Open Consumers % of 95% CI the (%) cluster 98.9 97.8,100.0 98.0 96.5,99.5 87.6 83.9,91.3 87.9 84.3,91.5

45

Cluster 1 versus the other clusters * MAG-Q.1 Whether Read Magazines Crosstabulation MAG-Q.1 Whether Read Magazines yes Cluster 1 versus the other clusters optimistic trendsetters Count % within Cluster 1 versus the other clusters others Count % within Cluster 1 versus the other clusters Total Count % within Cluster 1 versus the other clusters 527 88.9 no 66 .1 Total 593 1.0

1004 82.0

221 .2

1225 1.0

1531 84.2

287 .2

1818 1.0

TV Respondents to the telephone survey were asked a series of questions about their access to cable television. Analyzing the different behaviors in television access and programming choices between the four different clusters revealed differences among them. Comparing how respondents answered whether or not their household has cable TV, there was a significant difference between the target group of optimistic trendsetters and skeptical latecomers. Sixty nine percent (69.1%) of optimistic trendsetters have cable compared with 60% of skeptical latecomers. The difference was significant at the level 2 (1, N=1131) = 10.25, p< .001. Out of those respondents who do not have cable TV, but it is available in their area, there are no significant differences between optimistic trendsetters or tech-friendly open consumers and traditional routiners or skeptical latecomers (p > .05) in the likelihood they get basic cable or premium cable in the next year. Among all four groups, the responses for not likely at all are very high with 51.3% (SE: 9.1, 95% CI) of optimistic trendsetters not likely at all to get basic

46

cable and 62.9% (SE: 8.8, 95% CI) of optimistic consumers not likely at all to get premium cable within the next year. The top ten most popular TV programs were Local news [News/Information](54.90%), Home Improvement (45.60%) [Comedy], Public Television (43.50%) [Educational/Cultural], 20/20 (43%) [News/Information], Sports (38.4%), Discovery Channel (37%) [Cable Channels], Network Evening News (36.60%) [News/Information], CNN (34.10%) [Cable Channels], Oprah Winfrey (31%) [Talk Shows], Roseanne (31%) [Comedy]. Among the ten programs, the respondents in the target cluster were significantly more likely to watch the following programs than those in other clusters: Home Improvement, Roseanne, 20/20, Discovery Channel, Oprah Winfrey show. There were some significant differences in television viewing tendencies among the optimistic trendsetters and both traditional routiners and skeptical latecomers on watching several specific TV programs. Optimistic trendsetters are more likely to watch Oprah Winfreys talk show,2(1, N=1462) = 26.4, p< .01. Optimistic trendsetters are also more likely to watch the cable channel A&E than the traditional routiners,2(1, N=924) = 4.72, p< .01. Optimistic trendsetters are more likely to watch the Discovery Channel than both the traditional routiners and skeptical latecomers,2(1, N=924) = 12.4, p< .01. The tech friendly open consumers are more likely to watch CNN than the traditional routiners and skeptical latecomers 2(1, N=1225) = 7.4, p< .01.

47

Radio To explore how clusters differ in the number of hours they listened to the radio, we used a one-way between-groups analysis of variance. There was no statistically significant difference between the three groups: F (3, 1817) = .95, p =.42. We also analyzed the types of radio shows that the different clusters reported listening to. The most popular type of radio programming across all four groups is music programming. The tech friendly open consumers were the most likely to listen to music programming on the radio (80.1% SE:1.8, 95% CI), the optimistic trendsetters were slightly less likely (78.8% SE:1.9, 95% CI), 70% of skeptical latecomers listened to music (SE:2.1, 95% CI), and 69.8% of traditional routiners listen to musical programing (SE:2.1, 95% CI). A chi square test of independence evinced a significant difference between listening patterns, 2(1, N = 924) = 9.24, p <.01. Traditional routiners were less likely to listen to music on the radio than were optimistic trendsetters. Very few people in any of the clusters reported listening to childrens shows, comedy shows, and sports, and talk shows, and slightly over 20% of all four groups listen to news shows. Magazines A one-way between subjects ANOVA was conducted to test mean difference on Number of magazines read per month among the four clusters. The means of this variable differed significantly across the four clusters, F (3, 1814) = 13.87, p = .000. Tukey post-hoc comparisons of the four clusters indicate that the optimistic trendsetters (M = 2.75, 95% CI [2.49, 3.01]) read more magazines per month than the traditional routiners (M = 1.94, 95% CI [1.68, 2.20]), p = .000 and the skeptical latecomers (M = 1.85, 95% CI [1.65, 2.06]), p = .022. Comparisons

48

between the optimistic trendsetters and the tech friendly open consumers were not statistically significant at p > .05. As for the preference of magazines read on regular basis, generally, the top ten most popular magazines among the respondents are Readers Digest (24.2%), Better Homes and Gardens (16.3%), Good Housekeeping (14.2%), TV Guide (13.5%), National Geographic (13.2%), People (10.6%), Ladies Home Journal (10.4%), Sports Illustrated (9.6%), Newsweek (9.5%), and Consumer Reports (8.1%). Among the ten magazines, cross tabulations and chi-square tests were conducted to examine whether the target cluster (optimistic trendsetters) were more likely to read particular magazines. According to the results, the respondents in the target cluster were more likely to read the following magazines than those in other clusters: Better Homes and Gardens, Good Housekeeping, TV Guide, and People. A chi-square test was used to determine whether optimistic trendsetters or the other clusters were more likely to read the magazine Better Homes and Gardens. The result was statistically significant, 2 (1, N=1818) = 17.46, p<.01. A total of 20.2% [95% CI (13.0%, 27.4%)] of the optimistic trendsetters read this magazine, while only 12.7% [95% CI (7.5%, 17.9%)] of the respondents in other clusters read this magazine. Therefore, the optimistic trendsetters are more likely to read this magazine than the respondents in other clusters. A chi-square test between optimistic trendsetters/other clusters and whether the respondents read the magazine Good Housekeeping was statistically significant 2 (1, N=1818) = 5.23, p<.01. A total of 16.5% [95% CI (9.2%, 23.8%)] of the optimistic trendsetters read this magazine, while only 12.6% [95% CI (7.4%, 17.8%)] of the respondents in other clusters read

49

this magazine. Therefore, the optimistic trendsetters are more likely to read this magazine than the respondents in other clusters. A chi-square test between optimistic trendsetters/other clusters and whether the respondents read the magazine TV Guide was statistically significant 2(1, N=1818) = 4.73, p<.01. A total of 15.7% [95% CI (8.3%, 23.1%)] of the optimistic trendsetters read this magazine, while only 12.0% [95% CI (6.7%, 17.3%)] of the respondents in other clusters read this magazine. Therefore, the optimistic trendsetters are more likely to read this magazine than the respondents in other clusters. A chi-square test between optimistic trendsetters/other clusters and whether the respondents read the magazine People was statistically significant 2(1, N=1818) = 4.79, p<.01. A total of 12.6% [95% CI (5.1%, 20.1%)] of the optimistic trendsetters read this magazine, while only 9.3% [95% CI (4.0%, 14.6%)] of the respondents in other clusters read this magazine. Therefore, the optimistic trendsetters are more likely to read this magazine than the respondents in other clusters. Newspaper Reading In the sample, 85.3% (SE = 2.7%, 95% CI) of the optimistic trendsetters read the newspaper. Most of the optimistic trendsetters read newspaper for less than 40 minutes. To determine whether there is a relationship between the length of time spent reading newspaper and cluster membership, we conducted the chi-square test for independence. We found a significant relationship when comparing the optimistic trendsetters and the traditional routiners on their length of reading newspaper 2(1, N=712) = 4.2, p< .05. Optimistic trendsetters spend less time reading newspaper than the traditional routiners in a week.

50

Most of the tech-friendly open consumers (56.6%) (SE = 2.7%, 95% CI) spent more than one hour reading newspapers in a week. We found a significant relationship when comparing the optimistic trendsetters and the tech friendly open consumers on the length of time they spent reading newspaper 2(1, N=713) = 5.18, p< .05. We also found a significant relationship when comparing the tech friendly open consumers and both the traditional routiners and skeptical latecomers on their length of reading newspaper 2(1, N=713) = 5.2, p< .05. Tech friendly open consumers spend more time reading the newspaper than other groups. Table 17: Newspaper sections most frequently read by optimistic trendsetters
Sections Local News World News Advertisement Classified Comics % 74.7% 56.5% 56% 48.4% 42.8%

Different clusters read different sections of the newspaper. Table 17 shows the most popular sections of the newspaper from the overall sample. The optimistic trendsetters and the traditional routiners both read the technology section, and the relationship is significant 2(1, N=924) = 7.4, p< .01. The tech friendly open consumers and the traditional routiners also both read the technology section, and also have a significant overlap 2(1, N=1225) = 15.6, p< .01. Optimistic trendsetters and the tech friendly open consumers are more likely to read the technology section of the newspaper. Optimistic trendsetters are also more likely to read advertisement sections than the traditional routiners 2 (1, N=924) = 26.4, p< .01. We found a significant relationship when comparing the tech friendly open consumers and non-target groups on reading the fashion section 2 (1, N=1225) = 16.3, p< .01. Tech friendly

51

open consumers are more likely to read the fashion section than the traditional routiners and the skeptical latecomers.

IV.

Cluster Description

Optimistic trendsetters Optimistic trendsetters are the top target group for ESI because they are the group most likely to buy items from home and also the group most likely to try new things. Optimistic trendsetters like having choices, they have an optimistic view of the information superhighway, and they are slightly more likely to buy from the mail than other groups. Optimistic trendsetters tend to be between 30-39 years old, have graduated from college, be married, have children at home, and be working full time. They are most likely to work from home or all of the four groups. They are most likely to live in a single family home. This group is very busy, and would like more spare time. When they have any free time, they prefer to stay at home. Optimistic trendsetters are both price conscious and quality conscious, they are willing to pay more for convenience, and they willing to spend more if they are assured of high quality goods, but they are not as willing as other groups (such as the tech friendly open consumers). Optimistic trendsetters are influenced by their peers. They place a high value on their relationships with other people, and they are likely to buy expensive things in order to impress their friends. Optimistic trendsetters are more diverse ethnically and racially than the three other clusters.

52

Tech friendly open consumers Tech friendly open consumers are the second most likely group to adopt this new shopping technology. They like having choices available to them, they have a positive attitude towards technology and like to try new things. They like to work outside of the home, and are open to shopping in the store or by mail or phone. They have the highest percentage of overall employment among the four groups, but the lowest percentage of full time employment. The people who are employed are more likely to be part time or have some flexibility. There are more people with post college educational level than the optimistic trendsetters. Of all of the clusters, they were the second most likely to shop from home and were equally likely to adopt new things as the optimistic trendsetters. Although they are similar optimistic trendsetters, but there are some key differences. First, they tend to be slightly older (40-49) or slightly younger (20-29) than the optimistic trendsetters. Second, they are less likely to be married, and also have fewer people living in the household. They are less likely to live in a single-family house, and more likely to live in an apartment than the optimistic trendsetters. Similar to the optimistic trendsetters, Tech friendly open consumers are very concerned about saving time, and are willing to pay more for convenience. They are the most willing of all the clusters to pay more for higher quality. Tech friendly open consumers are extremely social, when they have free time, they like to go out, and they place a very high value on personal relationships. They are most concerned among the clusters about their own personal privacy, but least concerned about invasion of privacy related to technology use. They are very likely to be an early adopter.

53

Skeptical latecomers Skeptical latecomers are the second least likely group to use the new technology. They are not as likely to try new things as the top two groups, but they are more likely to try new things than the traditional routiners. They have a low attitude about the information superhighway. They are more likely to shop from a store, and are neutral about having choices. They would like to earn a living from home, if possible. The highest percentage of skeptical latecomers is between 40-49. They are more homogenous racially and ethnically than the optimistic trendsetters. They tend to live in single-family homes that they own and tend to be employed full time. They are extremely price conscious, are unlikely to pay more for high quality, and less likely to pay more for convenience than the other two groups. They do not buy expensive things to try and impress their friends. They are less social than the two previous groups, being less likely to value personal relationships or go out for romantic dinners. Although they report that that they are very busy and do not have a lot of time, they also report a low interest in adopting time saving measures. They have a less positive attitude towards TV. Computers are less important to them than to the optimistic trendsetters or the tech friendly open consumers. They are very concerned about potential privacy invasion through technology. Traditional routiners Traditional routiners are the least likely group to adopt the new technology. They are neutral about choices, they like to work outside the home. They are the most likely cluster to shop at a store. They have the lowest attitude towards the information superhighway, and the lowest likelihood to be an early adopter. The majority of people in this group are in the 60-69 age category, and it is not surprising that they report fewer numbers of people in the household,

54

as any children they may have had have moved out. They have the highest percentage of people reporting they live in a single-family home. They tend to be homogenous both racially and ethnically. Many of them are employed full time. They like to stay at home in their free time. They have a less positive attitude towards TV than the two target clusters, and they report the least number of hours of TV of all the clusters. They are the least likely to sue all types of media (TV, radio, newspaper, magazines). Computers do not play an important role in their lives. They are concerned about privacy invasion related to technology. They are not concerned about the price of goods, but are less likely to be willing to pay more for higher quality items. They report the most amount of spare time, have a low interest in time saving measures, and are unwilling to pay more for convenience. V. The Story and the Takeaways for the Client After analyzing the overall data and the patterns among the clusters, a clear story emerges about potential users of the new home shopping technology, and possible strategies for reaching those users. Practical communication and marketing strategies are provided for the two target clusters, optimistic trendsetters and tech friendly open consumers, and opportunities and barriers are provided for the client. Opportunities and barriers are also provided for the two non-target groups, to enable the client to devise strategies for reaching those groups in the future. The target cluster Optimistic Trendsetters Appealing messages: 1. Develop marketing promotions in different languages and take cultural differences into consideration to appeal to different racial and ethnic backgrounds.

55

2.

Focus the marketing and promotion on the convenience and timesaving characteristics of the new home shopping technology.

3.

Stress the benefits to customers of home shopping technology in the areas of value and price, since the target cluster is both price consciousness and quality consciousness.

4.

Incorporate social interaction into the home shopping technology to appeal to the social nature of the target groups.

How to reach them: 1. Optimistic trendsetters read magazines. Better Homes and Gardens, Good Housekeeping, TV Guide and People are especially popular among the group. It would be efficient and effective if put the ads of the new home shopping technology on these media platforms. 2. Optimistic trendsetters watch television. Some popular TV programs or TV channels among them are Home Improvement, Roseanne, 20/20, Discovery and Oprah Winfrey talk show and A&E. Opportunities: 1. Appeal to their desire to stay at home and find ways to save time. The new home shopping technology would allow them to shop without leaving home and it would also save their time since they dont need to go to real stores. 2. Highlight the novel, technology-powered characteristics of this technology. Optimistic trendsetters have positive attitude toward new technology and electronic devices. Barriers: 1. They tend to have fewer TVs at home.

56

2. While most people in the cluster have cable at home, about 30% still do not have cable and therefore cannot use ESIs product. 3. They have high concern of privacy invasion concerning new technology. 4. Most of them dont have debit cards, which might be a necessary paying tool for the new home shopping technology. 5. The Second Target Cluster Tech Friendly Open Consumers Appealing messages: The same strategies developed for the optimistic trendsetters can be used for tech friendly open consumers How to reach them: 1. Tech friendly open consumers watch television and listen to the radio, so those are effective channels to reach this group. CNN is popular TV channel among them. They tend to listen to music programs on the radio. 2. Tech friendly open consumers spend more time reading the newspaper than the other groups, especially the fashion and technology section. The newspaper would be a good place to put ads for the new home shopping technology. Opportunities: 1. Tech friendly open consumers are extremely social people. A word-of-mouth marketing strategy can be used to persuade them.

57

2.

Tech friendly open consumers also have positive attitudes toward new technology and electronic devices.

3.

Although they have a high concern of personal privacy, when comes to the new technologies, they are captivated by new technologies, so would be willing to take a risk to experience them.

Barriers: 1. 2. They prefer going out instead of staying at home at their free time. Most of them dont have a debit card, which might be a necessary paying tool for the new home shopping technology.

The Cluster of the Latecomers Skeptical Latecomer Opportunities: 1. Skeptical latecomers are most price conscious group. If they can be convinced that the home shopping technology can help them save money on shopping, they might be interested in trying this new technology. 2. Skeptical latecomers tend to be busy and have less spare time. If the home shopping technology can help them save time on shopping, they might be interested in trying this new technology. Barriers: 1. Skeptical latecomers are less likely to use media, including TV. They tend to spend less time watching TV.

58

2. Skeptical latecomers are highly concerned about possible privacy invasion concerning new technology. 3. Most of the skeptical latecomers dont have a debit card, which might be a necessary paying tool for the new home shopping technology. The Laggards Traditional Routiners Opportunities: There is almost no opportunity to drive those people to use the new home shopping technology. Barriers: 1. Traditional routiners hey tend to be older people who are not open to try new things, including new technologies. 2. Traditional routiners spent the least amount of time of all groups watching TV. 3. Traditional routiners have high concern of privacy invasion concerning new technology. 4. Many traditional routiners dont have a debit card, which might be a necessary paying tool for the new home shopping technology.

VI.

Limitations This analysis has a number of limitations. The primary limitation is the use of an existing

dataset, rather than the opportunity to design a data set specifically tailored to this analysis. The existing dataset gathered a great deal of data, but many of the variables were categorical rather than interval or ratio level. This significantly limited the types of statistical analysis that could be conducted. The second limitation is the lack of metadata on the dataset itself. The researchers

59

had to divine the multiple item measures from the survey and test whether those groupings were correct. In an ideal situation, the researchers would have access to the intended multiple item groupings. It is likely that some of the items may not have been grouped optimally, but the researchers were limited to working backward from the survey item by item, instead of starting from the intentions of the original survey designers.

60

You might also like