Research Support School of Business Administration October 1993 PROFILING TECHNOLOGY DIFFUSION CATEGORIES: EMPIRICAL TEST OF TWO MODELS Working Paper #723 James R. Taylor University of Michigan Edwin J. Amonsen Michigan Marketing Associates Eric G. Moore University of Michigan

I i I

Profiling Technology Diffusion Categories: Empirical Test of Two Models James R. Taylor University of Michigan Edwin J. Amonsen Michigan Marketing Associates Eric G. Moore University of Michigan October 20, 1993 ABSTRACT The authors provide an empirical test of two models which predict the technological adoption categories of small businesses for Hewlett Packard laser printers. The two models tested were a psychographic model as proposed in the book, Crossing the Chasm, and a more traditional benefit/price model. The adoption categories were defined by the sequence of laser printer model changes from the period 1985 to 1990. The results suggest that the benefits model predicts the buyer adoption better than the psychographics model. The study represents an exploratory phase of a future conclusive research project. INTRODUCTION Scope of the Study The research reported was conducted for the Hewlett Packard (HP) Peripherals Group in 1993. The purpose was to determine the relevance to the laser printer market of concepts presented in the book, Crossing the Chasm (Moore, 1991). The book applies the diffusion of innovations theory to technology driven products in general and does not claim that the propositions presented have been empirically tested with technology driven products. The thesis of the book is there are distinct, identifiable differences in the psychographic and behavioral characteristics of business people and organizations who can be classified as innovators, early adopters, early majority, late majority and laggards in the adoption of technology products. These differences impose different requirements on the nature of marketing activities to these five

market segments. This is especially true for the hypothesized chasm between early adopters and the early majority. The research was conducted by the Michigan Business School for Hewlett Packard as part of their effort to facilitate close relationships among faculty, students and HP management. The research project was to involve two phases. The first phase was exploratory research using focus groups composed of businesses who had purchased laser and dot matrix printers. The findings from the exploratory research would form the basis for formulating a questionnaire and designing the data analytic approach to be used in the conclusive research phase. The exploratory research results are presented in this paper. The authors argue that the importance of the exploratory research phase has not been fully recognized by academic and business researchers. A series of steps are proposed for exploratory research which includes coding the responses of focus groups participants and analyzing the association and directional relationship of the variables identified. This data matrix is used to evaluate the appropriateness of data analysis statistical tools for the management issues to be investigated in the conclusive research phase. Such analysis facilitates in-depth thinking as to the nature of the conclusive research results and provides a format to communicate potential scenarios of research results. These research scenarios allow management to more clearly evaluate the potential strategic usefulness of the conclusive research phase. The scope of the study was defined as the US laser printer market for small businesses. The small business segment was chosen for the study for two reasons. First, it is less likely the product purchase decisions will be made by a committee or a central purchasing agent. Thus, the small business decision maker should exhibit more psychographic traits in the purchase of a technology product like laser printers than larger businesses. Second, the small business segment represents the largest market potential for laser printers. 2

Hypotheses Because of the complex nature of the subject under investigation, HP did not think it to be realistic to address all the issues related to the technology diffusion cycle in a single study. The most important strategic management issue related to the testing of the following hypothesis: There are distinct, identifiable differences in the psychographics and behavioral characteristics of innovators, early adopters, early majority, late majority and laggards pertaining to the adoption of hard copy products. These differences are quantifiable and can be utilized by HP to more successfully market products based on their position in the product life cycle. The strategic marketing issues addressed by this hypothesis would allow HP to profile purchasers of various hard copy products within each category, profile products moving through the adoption process to identify timing issues for replacement products, categorize HP's current customer base, and identify ways in which they can take advantage of these characteristics to help in marketing products across the spectrum of technology adoption segments. To emphasize the strategic marketing issues of concern to HP, two competing hypotheses were formulated: HI: Psychographic variables are stronger predictors of when in the diffusion cycle a buyer will purchase a new technology than benefit/price variables. H2: Benefit/price variables are stronger predictors of when in the diffusion cycle a buyer will purchase a new technology than psychographic variables. Conceptual Model In Moore's book, Crossing the Chasm, he proposes a revised technology adoption life cycle (Exhibit 1). The five adopter categories are profiled in terms of their psychographic and 3

behavioral characteristics as they related to technological innovations. The model postulates that businesses adopt an innovation at different times after it becomes available. The order of adoption flows from innovators to laggards. While differences in the characteristics of the adopter categories are seen as gaps to be crossed by different marketing programs, a big gap is hypothesized between early adopters and the mass market of early majority. Exhibit 2 presents the research teams interpretation of variables underlying the five adoption categories. Research on adopters is nested in the diffusion literature (Rogers, 1983). Consumer research has viewed the diffusion process as being conceptualized in an S-Shaped logistic pattern reflecting an exponential growth pattern from innovators and early adopters to majority acceptance in the marketplace (Robertson, 1984). Studies of adopters in terms of diffusion categories have focused on psychographic profiles of individuals in each category (Gatignon and Robertson, 1989). These research paradigms have been challenged as derived from rural sociology (Rogers, 1983) and of being overly simplistic for most consumer and industrial products. Robertson notes that the original diffusion studies of hybrid corn (Katz, 1961) and medical treatments (Coleman, Katz, and Manzel, 1966) provided a weak basis for generalization to consumer and industrial products. These studies deal with innovations which where highly recommended by scientific experts and were of central importance to their users. Further, each innovation came with a clear and unambiguous measurement indicator which allowed adopters to quickly discern the benefits achieved from adopting the innovation (Robertson, 1984). Other criticisms were that S-Curve studies of innovation typically focus on a single innovation or product and ignore the role of making a designed pattern of controlled communication regarding product benefits to targeted audiences. Marketing literature has also been criticized for being too "consumer goods oriented" (Gatignon and Robertson, 1989) and needing more emphasis on the diffusion of high technology innovations as is consistent within the marketing field (Capon and Glazer, 1987). Clearly, 4

research is needed to extend the basic diffusion and adopter category paradigms to encompass decision patterns of adopters of high technology products. METHODOLOGY Laser Printer Market Evolution Hewlett Packard introduced the industry's first desktop laser printer in 1984. This LaserJet Classic Series revolutionized desktop printing, and customer demand for the printer catapulted HP from an insignificant market share to a position of industry leadership. The laser printer used a Canon engine that was a outcome of a joint research relationship the two companies had since 1975. In 1987, HP introduced the LaserJet Series II printer, which created a new industry standard for price and performance trade-offs. At a list price of $2,695 and widely available selling prices just over $1,600, it offered 300 dots per inch resolution, 8 pages per minute speed, and HP reputation for reliability and software compatibility. HP had surprised competitors by repositioning its entire product line in 1989, with greatly reduced prices on laser printers. The new LaserJet IIP printed 4 pages per minute and had a list price of $1,495, but sold in many retail stores for about $1,000. HP's pricing spurred industry wide price reductions averaging 25%. In 1990, Hewlett Packard again introduced an improved laser printer called the LaserJet Im. It had a list price of $2,400, printed 8 pages per minute, and produced the impression of print resolution of 600 dots per inch. Again, HP was able to leapfrog the competition with major technological improvements while lowering the price. Sample Design The sample was randomly drand drwn from HP's warranty card database of small business (under 50 employees), operating since 1985, which had purchased the Classic, Series II or Series 5

III laser printer model (Table 1). The sample of Laggards, defined as small business who had purchased a dot matrix or daisy wheel printer but not a laser, was randomly drawn from the telephone directory. The sample profile covered a wide variety of business types. Table 1 SAMPLING PLAN CATEGORY Innovators Innovators Early Adopters Early Adopters Early Majority Late Majority Laggards MODEL Classic Classic LaserJet II LaserJet II LaserJet 1-P LaserJet III Dot Matrix/Daisy PURCHASE DATE 1985-1987 1985-1987 1986-1988 1986-1988 1988-1989 1990 N/A SIZE 9 9 8 8 8 7 8 Total 57 FOCUS GROUP LOCATION San Francisco, CA Farmington Hills, MI San Francisco, CA Farmington Hills, MI Farmington Hills, MI Farmington Hills, MI Farmington Hills, MI The sampling frame for each adopter category was based on a sequential time range starting in 1985 when the Classic series warranty card data was first available. Table 1 also indicates the time frames sampled for each category. The research design assumes that businesses who bought a Classic LaserJet in the time period 1985 to 1987 are most likely innovators. Businesses who bought the LaserJet II series from 1987 to 1989 are potentially Early Adopters while business who bought the LaserJet IIP series from 1989 to 1990 are potentially in the Early Majority category. LaserJet HI purchasers would be classified in the Late Majority category. Data Collection Procedure The first phase of the project involved in-depth telephone interviews with small business owners/managers. These interviews were approximately 20 to 40 minutes in duration using open 6

ended questions in a semi-structured format. The purpose was to explore the purchase decision process of hard copy technologies and the adoption (or non adoption) of laser jet technology. These interviews, conducted nation wide, allowed the research team to "hear" the customer and formulate the issues to be explored in the focus groups. Based on these interviews and the objectives of the research project, a focus group interviewer's guide was developed. This first phase identified a potential source of non-response error in the phase two sample selection plan. The phase one interviews had a high proportion of Classic owners who were in the law, accounting and medical fields. Apparently, this type of business has greater longevity that other firms which characterize the small business profile. Consequently, the phase two sampling plan restricted the number of law, accounting and medical firms to 30% of the focus group sample. Coding and Operationalizion of Variables The variables were operationalized by coding the respondent's responses during the focus group sessions. An array of variables were measure including 1) type of business, 2) position of respondent, 3) business goals, 4) software applications, 5) adoption category assessment, 6) price sensitivity, 7) print quality importance, 8) productivity importance, 9) risk attitude and 10) importance of computers to the business. Exhibit 3 presents the variables selected for inclusion in the competing models. The Benefits/Price model contains three variables: print quality, productivity and price sensitivity. The Psychographic model includes the variables of risk attitude and the perceived importance of computers to business. The other variables were found to have weak relationships with the time of purchase and printer type. The correlation between the adoption category assessment and risk attitude was very high (r=.86) and consequently only the risk attitude variable was included. Our models are specific to printer technology adoption, and may not be completely applicable for all new high technology products. Exhibit 2 provides a summary of the expected directions of the variables for both the Benefits and Psychographic models. 7

BENEFITS VARIABLES Price Sensitivity measures the relative importance of the price criterion in the adoption decision. This variable is coded on three levels: 1) Low (Less sensitive) 2) Medium 3) High (More sensitive) Proposition 1: The less sensitive to price, the more likely a firm will adopt a new printer technology. Print Quality measures the importance of print quality of output to the user. This variable is coded on three levels: 1) Low (Less important) 2) Medium 3) High (Very important) Proposition 2: The more important print quality is to a firm, the more likely a firm will adopt a new printer technology. Productivity measures the importance of a tangible productivity gain to the user. This variable is coded on three levels: 1) Low (Less important) 2) Medium 3) High (Very important) Proposition 3: The more important productivity is to a firm, the more likely a firm will adopt a new printer technology. 8

PSYCHOGRAPHIC VARIABLES Risk Attitude toward New Technology is an aggregation of the risk perceptions about fax/modems, computers, printers, and software and measures the propensity to adopt new products in the specific product areas. This variable is coded on three levels: 1) Low (Risk Averse) 2) Medium 3) High (Risk Seeker) Proposition 4: The higher the level of risk seeking, the more likely a firm will purchase a new printer technology. Importance of Computers to the Business is a measure of the importance of computers to the operation of the business and is a proxy for the opportunity cost for adopting a new technology. This variable is coded on three levels: 1) Low (Less critical) 2) Medium 3) High (Very critical) Proposition 5: The more important PC's are to the business, the more likely a firm will adopt a new printer technology. ANALYSIS Verbatim transcripts from the focus groups were classified using protocol analysis. After the comments were organized by appropriate subject matter, the content was analyzed for each focus group member. The fundamental process of analysis is to compare each item with the previous incidents in the same and different groups coded in the same category. The participant psychographic adoption profiles were assigned through a systematic analysis of responses. A sample of the data matrix is found in Exhibit 4. 9

The hypotheses for each of the independent variables were tested, with the sign of the coefficient giving the direction for the relationship. The competing model's overall predictive ability were measured by how accurately each model classified subjects into their expected adoption category based on product adoption. A logit model involving an ordinal level dependent variable (McKelvey and Zavoina, 1975) of expected adoption category was applied to the data. Recognizing the size limitations of our sample and the requirements for the statistical techniques involved, the LaserII and m data was weighted double to more accurately represent the proportion in the buyer population. The actual coefficients are used in the graphical representations, but should be interpreted only for direction and not relative importance across variables. The models were tested with the appropriate variables coded as dummy variables. The medium condition was chosen as the 0 value, Level 1 of the variable was the Low condition, and Level 2 was the High condition. RESULTS AND DISCUSSION Benefits Model The statistical results of the benefits model are presented graphically in Exhibit 5. The strongest directional relationship is found with print quality. The companies that need a higher level of print quality would seek out new printer technologies earlier. This finding supports Proposition 2. The low level of productivity is highly associated with the late adoption or nonadoption of printer technology (p<.001). There is a weak positive relationship between the high level of productivity and early adoption and this supports Proposition 3. Price sensitivity fluctuates across the adopter and non-adopter categories. This result indicates an inverted "U"Shaped relationship which implies that both innovators and laggards have low price sensitivity. This result refutes Proposition 1 that price sensitivity would increase across the adoption cycle. Rather, the findings indicate that the middle adopters are the most sensitive to price. This implies the middle category demands higher product quality and productivity gains in a printer, yet is 10

more price sensitive than the laggard category. The laggard category appears to perceive few benefits from print quality and productivity, resulting in a lower intention of purchase. The benefits model correctly classified 72% of the focus group participants into the product adoption categories. The variables of print quality, productivity, and price sensitivity has a Chi-Squared statistic of 40.14, p<. 0002 (Table 2). Psychographic Model The directional results for the psychographic model are presented graphically in Exhibit 6. A high level of risk seeking has a strong positive relationship (p<.02) with early adoption. This supports Proposition 4 that earlier buyers are more willing to accept the risk associated with early adoption.. The high level of importance does not have any effect on time of adoption. However, the expected relationship from Proposition 5 holds in that the less important computers are to the business, the later new printer technology will be adopted (p<.001). The psychographic model correctly classified 56% of the focus group participants into the product adoption categories. The variables of risk attitude and importance of computers had a Chi-Squared statistic of 2.15 p<.71 (Table 2). Table 2 Results of Logit Analysis Benefits Model Low Level High Level Variable Statistic Direction Statistic Direction PRICE1 -1.37 p<.39 Negative PRICE2 -.87 p<.42 Negative OUTPUT1 -.65 p<.59 Negative OUTPUT2 1.78 p<.23 Positive PRODUCTIVITY1 -4.70 p<.000 Negative PRODUCTIVITY2.34 p<.76 Positive # Classified Correctly: 41/57 = 72% 11

Psychographic Model Low Level High Level Variable Statistic Direction Statistic Direction RISK1.50 p<.47 Positive RISK2 2.26 p<.01 Positive PC CRITICAL1 -1.73 p<.02 Negative PC CRITICAL2.002 p<98 Negative # Classified Correctly: 32/57 = 56% CONCLUSIONS AND IMPLICATIONS The central research question for this study was: which of two competing models is a better predictor of whether a person will adopt a laser printer? Seven focus groups of small business owners were held and then the transcripts were analyzed using protocol techniques. The participants were classified on multiple variables developed from the research question. The Chi-Square test of the hypotheses (Hi and H2), support the Benefits/Price model over the Psychographic model. While the purchaser's attitude toward risk and the importance of the personal computer (Psychographic model) plays a role in the new technology purchase decision process, the need for print quality, increased productivity, and price sensitivity (Benefits/Price model) are more important drivers of the adoption process. The Benefits/Price model also displays a high level of predictive power, correctly classifying 72% of the focus group members. In contrast, the psychographic model correctly classified only 56% of the subjects. The managerial implications of this research study are that companies which pioneer new technologies must focus on the benefits desired by purchasers. Early adopters look for the benefits which meet their needs better than current technologies. If the benefits are significant, these early adopters are less price sensitive and are willing to take risks to acquire the benefits. There appears to be nothing magical about the diffusion of new technologies. Businesses which focus their product development and entry marketing strategy on identifying and meeting key customer needs will enjoy the advantages of a pioneer. 12

Innovations have been characterized in terms of the extent to which the innovation requires buyers to acquire new knowledge and/or change their behavior. Three categories of innovations have been identified: continuous, dynamic continuous and discontinuous. Continuous innovations have the least impact on buyers in terms of changes in knowledge and behavior. Dynamic continuous innovations have moderate impact while discontinuous innovations have the greatest impact on knowledge and behavior. Laser printer technology would be classified as a dynamic continuous innovation. This category of innovation is typical of mainstream technological innovation and provides a significant test of the concepts presented in the book, "Crossing the Chasm". If there is a "chasm" in the diffusion process, it appears to be the high demands of the mass market. These demands are for superior benefits and lower price points. Designing a manufacturing and marketing strategy to lower the cost structure to meet the needs of the mass market is the challenge of the market pioneer. While risk attitudes are important in this process, they should not divert the attention of management from the key drivers of innovation: benefits and price points. The above conclusions are based on the measurement and analysis of respondent comments in focus group sessions. An important objective of this project was to demonstrate that this process adds important insight to the development of the conclusive research phase of the total project. As researchers, the process forces us to deal with issues of operationalizing variables, evaluating alternative data analysis approaches, and analyzing the exploratory data set plus addressing the key management information needs. For the management group, this process allows a "mock-up" of the potential conclusive research results. The constructive and insightful researcher-management interaction which results from this process can greatly facilitate the design of the conclusive research phase. For the conclusive research design, finding members of the laggard category will be the most difficult. Laggards are often defined as people who will buy in the category, but in the decline phase of the product life cycle. These potential buyers must be separated from the non 13

buyers as there may be fundamentally different views with respect to price sensitivity and risk attitude. The quantitative phase of the project must address this issue more directly than was possible in the qualitative phase. As a final observation, this study brings a new dimension to the diffusion of innovation literature. Prior research has tested the diffusion of innovation theory based on a single product innovation phase. Our project focuses on a series of innovations in laser printer technology across the five category diffusion curve. The results suggest that more research is needed to validate diffusion theory across high-technology innovations which come to market in rapid sequence. The comparison between the Psychographic and Benefits/Price models should also be explored. Hopefully, future cooperation between the academic and business community can close the gap between theory and practice in the field of strategic marketing management. FINAL NOTE: The authors wish to thank David Bufford and George Mulhern of Hewlett Packard for their support and assistance on this research project In addition, special thanks to my colleague Venkatram Ramaswamy for his valuable advice and assistance in the data analysis modeling. 14

REFERENCES Capon, N. and Glazer, R. (1987) Marketing and Technology: A Strategic Coalignment, Journal of Marketing, 51, 1-14. Coleman, J. and Katz, E. and Menzel, H (1966) Medical Innovation: A Diffusion Study. Indianapolis: The Bobbs-Merrill Company, Inc. Gatignon, H. and Robertson, T. (1989) Technology Diffusion: An Empirical Test of Competitive Effects, Journal of Marketing, 53, 35-49. Katz, M. and Shapiro, C. (1986). Technology Adoption in the Presence of Network Externalities, Journal of Political Economy, 94 (4), 822-41. McKelvey, R. and Zavoina, W. (1975) A Statistical Model for the Analysis of Ordinal Level Dependent Variables, Journal of Mathematical Sociology, 4, 103-120. Moore, G.(1991) Crossing the Chasm. New York: Harper Business. Robertson, T. (1971) Innovative Behavior and Communication. New York: Holt, Rinehart & Winston. Rogers, E. (1983), Diffusion of Innovations. New York: The Free Press. The Desktop Printer Industry in 1990. Case Study 9-390-173: Harvard Business School. 15

Exhibit 1 "Crossingthe Chas" - Psy(hogah<ByrPoie Sh~ ~ ~~~~~~~~~~~~~~~~~~." II.:~:':''' I' tl.:::''...:.iii:,: $^~ ^ ^^^^^ I I | ill | ^^^ N ~ %^^^~~~~~~~~~~~~~~~~~~~~~~~~i ^W^W^WAWW^WW^AV^^W?^^:?^^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~d ^A^^ W ^ fW::~::ii.X:::i";^^^^^~:c;~ ~~..::1~!:....~ Innovators * Technology is central interest in life * Appreciate technology for its own sake * Intrigued with advances in technology * First to realize potential of technology * Appreciate the architecture of new products * Like to debug new products * Don't need a whole product Early Early Adopters Majority * Use informed intuition * Characterized by to match new practicality technology to a ~ Seek references strategic opportunity before making * Driven by a dream for substantial investment a business goal * Prefer evolution to ~ Interested in revolution technology because it * Focus on offers competitive standardization rather advantage than specialization ~ Want large return on ~ Need complete system investment and plenty of support * Are builders, creaters * Prefer the market and visionaries leader ~ Expect radical ~ Generally risk averse discontinuity from old * Must live with their to new decision * Champion for a cause Late Majority * Not comfortable with their ability to handle high tech products * Very price sensitive * Buy just to get on par with the rest of the world * Against discontinuous innovation, don't like change * Wait for established standard, lots of support and then buy only from established company * Purchase whole solution Laggards * Probably will never purchase technology * Only purchase if it's buried in something else * Nostalgic traditionalists * Cheap

ci) O 0 s ~i: /C~; N Z~.~. s. .~.s s: X ii =r =r M — c= -r;.7 ~~~ :=: = = :... =. ==~ /~ s-~~ C::-5 r.t (I) t3) Cu a) -V 0) 0 0) 0) ~ I 3 f. O:.:.. * I:/-.' ^ O::"' ^:e.:"^:::.: > I ": =':-::i.,::,.:,.,,:,.,.,-:-:::....... w ^:-. - X., r,,I^,,. UJr.~~ ~ I2~':: Y;:X s 11 i -~~~~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~.:.. O -J 0J _0 -J sta -i ctt w-a Lit I c:3 -io 0 2 0 o -J I I a) 0) 0 212 E E D:3 CT 2 I 2E o a)..-j > I 0o 0 -J I 3 3 a) a) c 2 2 E E 2i 2 a) a) 0 I - 3..o I J cuo > _ > 3 D o -J >I z > c 0z < 0 Q) Uw: L C cD CC Q 0- ( U) L0 a ft Q) bI J-. 0) I DL r cc 0: 0 _J w 0O c z -J 0 82n 0

Exhibit 3 Variable Definition PRODUCT Classic LaserJet 11 LaserJet Il-P LaserJet 111 Non-Adopters ADOPTION PROFILE Innovator Early Adopter Early Majority Middle/Late Majority Laggards PSYCHOGRAPHIC MODEL Risk Attitude Importance of PC Growth Rate Payback Period Adoption Category BENEFITS MODEL Print Quality Productivity Price Quietness

Respondent Summary and Classification:;an Francisco Focus Group ',arch 30, 1993 rluesday 6:00 (Classic's) (Respondents ordered counter-clockwise from moderator) Name Member 1 Member 2 Member 3 Member 4 Member 5 Member 6 Member 7 Member 8 Member 9 Type of. Airport Importer Tech Power Printable Mkt. Word Legal Design Company Consulting Consult. Consult. Apparel Consult. Processor Non- Company Distrib. Profit Position Vice Pres. Owner Business Owner Mgr Info Owner Owner Associate Computer Mgr. Systems Manager Total 9 7 32 5 45 1 1 20 15 Employees. ________ Business Aggressive Grow Moderate Spinoff Aggressive Slow No, Grow if Reduce Goals Sales/New National Employee Company sales + (add one Retire possible employmen Markets Growth automation support) by....________ _ -... —_ _______ ___ __ __:__ _automatin Applications Tech Accounts WP, WP, WP WP WP WP, WP, Reports Graphics Graphics Legal Graphics First Laser Cl.ansic Classic: Classic Classic Classic Classic Classic Classic Classic Printer L Adoption IEarly Early Early Early Innovator Late Late Early Early Assessment Adopter Adopter Majority Majority Majority Majority Adopter Adopter it Ig 31

Exhibit 5 Directional Impact Benefits Model ' 1 O -1 -2 -3 -4 V -5 '~~~.:.~. ~.. ~::, s... - - - - - - -.- - - - -- - - - -. ---- -. -,..... -.-.... - _ - - - -. - N.~ ----- - - - - - - - - - - - - - - - - - -- - - - ---- - ---- --- - - - - -- - - - - - - -j LI~LI -,II -, I: - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - High (3) Medium (2) Level of Variable Low (1) ] Price 0q Quality -Ak- Productivity |

Exhibit 6 Directional Impact Psychographic Model 3 2 1 0 -1 -2 '' '''~~~~ -, - - - -- - I- - -------------- ---------------- j Hih (3) High (3) Medium (2) Level of Variable I Low (1) _ Risk ~P PC Importance I A., - - - =.