Richard Croome, Meredith Lawley and Bishnu Sharma, University of the Sunshine Coast
How and why consumers use the Internet is important to understanding retail change in the online context. A mail survey approach was used in completing this study. One hundred and forty six Australian consumers participated in the survey. The results suggest that consumers have a preference for purchasing digital and non-bulky analogue products online, specifically tickets, books, travel & hotels, music and cameras. Product attributes play an important role in choosing the initial retail channel. Pricing and product information are potential switching drivers and trust is associated with online purchasing. The implications for pure-play and multi-channel retailers are discussed.
Online retail sales as a percentage of total retail sales are increasing (Kau et al. 2003), albeit coming off a small base and in relative and absolute terms are still small considering the Internet was touted as the “new economy” during the 1990’s (Shabazz 2004; Finkelstein 2001). This lack of online purchasing activity presents an interesting research problem for academics, technologists and retailers alike, since consumers now have the opportunity to shop from the comfort of their own home in their own time from a far greater number of retailers located around the globe instead of only just locally, as before (Hamel 2001; Merrilees 2001).
From a theoretical perspective, neither the theories of retail change or the online consumer behaviour literature adequately explain the low conversion rate from search and evaluation to outright purchase on the Internet. Recently, several academics (Balabanis & Diamantopoulos 2004; Gehrt & Yan 2004; Koyuncu & Lien 2003) have explored the potential of integrating the consumer behaviour literature with what is known about retail change to progress the understanding of why and how consumers use the Internet to purchase goods and services. Their results, and the results of Keen et al. (2004), suggest that much can be learned about the factors influencing retail change through a better understanding of consumer behaviour during the buying decision process. In particular improving the understanding of the underlying motivations that influence consumer purchase behaviours. This paper builds upon this research by examining how and why value is created for the consumer in the context of the online buying decision process.
Two major research streams and three distinct research phases are identifiable in the online consumer behaviour literature. One research stream explores the possibility that product attributes, such as the product’s ability to be evaluated and delivered via the Internet, and the level of purchase involvement are important drivers of online purchases (Thornton & Marche 2003; Zeng & Reinartz 2003; Deveraj et al. 2002). The other major research stream explores the importance of the consumer’s situation as an important driver of online retail sales. Within this stream the consumer’s situation is taken to include access issues, product availability, technological familiarity, experience, trust, brand and customer service concerns (Tan & Sutherland 2004; Gefen et al. 2003; Argawal & Venkatesh 2002).
Within these two research streams, there are, broadly speaking, three identifiable phases of research. Whilst they overlap chronologically, and there is no distinct cutoff for each, the three phases appear to be broadly grouped into the useability phase from 1995 to 1997 (Alba et al. 1997; Jarvenpaa & Todd 1997; Boling 1995). The e-service quality phase from 1998 to the present (Argawal & Venkatesh 2002; Feinberg & Kadam 2002; Zeithaml et al. 2002; Hamel 2001), and the product attribute phase, which enjoyed some popularity from 1995 to 1997 (Peterson et al. 1997; Stern 1995) and has been pursued by a limited number of authors since including Zeng and Reinartz (2003), Vijayasarathy (2002), Phau and Poon (2000).
Within the two streams, six key dimensions are identifiable as potential online buying decision drivers. They are tangibles, trust, enjoyment, differentiation, price and content.
Content is a multifaceted dimension that refers to the information within the web site being relevant, current and timely, with the appropriate depth and breadth of information that allows the consumer to readily comparison shop and to customise and personalise their online retail experience (Zeng & Reinartz 2003; Janda et al. 2002; Vijayasarathy 2002).
The term tangibles, has two different meanings depending on whether it is approached from the product attributes or e-service quality perspective. In the former, tangibles refers to the information richness of the product (Zeng & Reinartz 200; Kolesar & Galbraith 2000). In the latter case it refers to website design and useability (Song & Zhang 2004; Tan et al. 2003).
Research to date indicates that trust on the Internet is multifaceted (Grabner-Kraeuter 2002). Trust in relation to online retailing primarily relates to the consumer’s confidence in the medium and the online vendor (Chen & Dhillon 2003). Trust of the medium, in relation to online shopping is primarily related to a belief that the Internet is trustworthy in terms of the information and the transaction (George 2004). Security in relation to the medium and the vendor is often cited as being one of the reasons why consumers will not use the Internet to make retail purchases (Doolin et al. 2005; Ziethaml et al. 2002; Lee & Turban 2001). Thus, from the literature, trust in its aggregate form includes aspects of ease-of-use, perceived usefulness, site aesthetics, assurance, brand, service recovery / returns, safety and security.
The cost of goods is also considered to be a key determinant of why and how consumers use the Internet to make buying decisions (Zeng and Reinartz 2003; Torkzahdeh and Dhillon 2002; Peterson et al. 1997). These authors focus on the absolute cost of the product and refer to whether the cost of the good is low, moderate or high, suggesting that goods high in dollar value terms that are infrequently purchased will sell well on the Internet. Other authors contend that as price increases so does financial uncertainty (Chen & Dubinski 2003) and the findings of Phau and Poon (2000) do not support the notion that only high priced goods will sell well on the Internet. They found that low cost goods that are competitively priced have a high purchase intention among Internet consumers. Thus from the literature, it is uncertain whether absolute cost is an important online shopping driver.
In terms of relative or competitive pricing, Figueiredo (2000) notes that all retail goods on the Internet, except for unique and unusual goods, suffer from commoditisation and therefore suffer pricing pressure. The severe discounting during the latter part of the 1990's (Porter 2001) supports this. Additionally, Zeithaml et al. (2000) measured “price knowledge” which refers to the consumer knowing the range of prices offered by online vendors for a particular product such as a specific model of camera, and both Burton (2001) and Kalakota and Robinson (1999) have argued that a competitive price is essential to online retailing. Absolute, competitive and transparent, pricing are thus the three major components of pricing that may influence the online buying decision.
The fifth key factor in the literature is the notion that product differentiation is an online shopping driver. Peterson et al. (1997) suggest that highly differentiated products will sell well on the Internet. Eastlick and Lotz (1999) refer to specialty products whilst Figueiredo (2000) hypothesises that unique and unusual products will sell well. In all three cases the authors suggest that products that have a limited number of substitutes or competitors and/or are hand made and/or are one-offs and/or can only be purchased from a particular web site are products that will succeed. This has yet to be established quantitatively and it may be that the opposite is true.
The sixth and final key factor relates to the concept of shopping enjoyment. Enjoyment has been identified as an important determinant of why consumers shop (Doolin et al. 2005; Mathwick & Rigdon 2004; Monsuwe et al. 2004; Mathwick et al. 2002; Jarvenpaa & Todd 1997; Hoffman & Novak 1996). Holbrook and Hirschman (1982) identify fantasy, feelings and fun as motivators for shopping. Later articles by Holbrook (2000) extend the shopping enjoyment concept to entertainment, and still later to evangelizing and exhibitionism (Holbrook 2001a; 2001b). Koufaris (2002) found shopping enjoyment to strongly predict intention to return to a website and measured the dimension in terms of the experience being fun, exciting, enjoyable and interesting.
Underlying the product attributes research stream is the notion that only products that leverage the basic condition of the shopping channel will succeed. In this scenario the consumer chooses a retail channel to obtain content and price information. The type of content and price information the consumer desires is a function of the product under consideration, which can be classified in terms of the product’s tangibles and differentiation attributes. If the consumer is satisfied with the content and price information they receive, trust and enjoyment develop and the consumer concludes the purchase. Thus, the model in Figure 1 asserts that, product attributes determine initial retail channel selection, and that content and price are the main retail channel-switching drivers and trust and enjoyment are the purchase drivers.
Figure 1: Product attributes model of consumer value
Underlying the situationalists’ literature is the notion that once the consumer trusts a retail channel, they will enjoy shopping within it, and thus the development of trust and enjoyment are necessary before the consumer begins looking for price and product information. The situationalists’ thus contend that there is potential for any product to be successfully sold online.
Figure 2: Situationalists’ model of consumer value
To examine the online buying decision process, depth interviews, focus groups and a field survey technique was employed. Depth interviews with four experienced and knowledgeable people were conducted. This provided a variety of informed responses. The four people selected included an owner of an online retail business, a specialist academic in the field of e-commerce, a senior government official responsible for public online information dissemination and payment facilities and an experienced online consumer. These individuals represented the spectrum of knowledge and experience relevant to the research and were considered adequate for this stage of the research, since no new information was gathered after the first three interviews.
Three focus groups, consisting of between five and seven people were convened drawing from a pool of university staff members invited to participate via the all-staff email list. In all three sessions there were at least two males and two females, and overall there were nine females and eight males giving a percentage split of 53 / 47 thereby minimising the possibility of gender bias in the responses.
The field survey respondents included a proportional sample of the Australian electoral roll and staff of a Queensland university. Respondents were initially asked to identify a product they had purchased entirely online within the past 12 months. Respondents were then asked to rate their agreement or disagreement in relation to the nominated purchase to 44 statements representing the six key dimensions identified in the literature. A modified five point Likert scale ranging from strongly disagree (1), through neutral (3) to strongly agree (5) including the option of “not-applicable” was used. The survey also collected demographic data. The 44 items and their origin within the literature are presented in appendix one.
Three rounds of pretesting were completed using 15 of the university staff members that participated in the focus groups to check the survey wording and layout. Initially 2000 surveys were mailed to a proportional random sample of the Australian voting population. One hundred and eight-eight mail surveys were returned. Sixty-nine were returned unopened, marked as “return to sender”. A further 31 were unusable since they were only partially completed or not completed at all. This reduced the possible sample to 1900 and represented a response rate of 4.6 percent. A follow up mail out survey was not possible since the mailing list purchased limited its use to one time only. Due to the low response rate a top up online survey of 412 university staff was conducted. This online version of the survey returned an additional 80 useable responses, giving a total of 146.
The results of the depth interviews confirmed the six key dimensions identified in the literature. All depth interviewees thought that the main products consumers would purchase online would be digital or non-bulky and that digital products would dominate. The findings of the depth interviews also indicated that not all six dimensions identified in the literature are equally important, and not all dimensions are important at the same time.
The key dimension of differentiation was thought to be the weakest of the six dimensions, and tangibles the strongest. Additionally, respondents thought the key dimension of tangibles strongly influences the type of content and price information sought, which in turn were prerequisites for the formation of trust and shopping enjoyment. It was also noted that trust might also be a primary online shopping deterrent in instances where the consumer does not trust the medium. In general however, the depth interview results supported the product attribute stream of research in the literature.
The results of the focus groups suggest consumers will make a pure online purchase if they think it will be faster, easier and they will get a better price. Product type underpins this decision as does the perception that the consumer will obtain better information about the product on the Internet, will have more choice in terms of comparison shopping and outright purchase options as well as access to this information whenever they want it. They would also consider the Internet to make purchases for products that are not locally available.
The focus group discussions extended the findings of the depth interviews by identifying a pure online shopping scenario where the Internet retail channel is selected as a last resort in instances where a local shop has closed and the only way to purchase the product is online. Respondents in this cohort, expressed anxiety at having to purchase the product online, citing credit card fraud as the main reason.
Another extension identified as part of the focus group discussions is the idea that the Internet offers a “no-pressure” shopping environment. When probed, many of the respondents said one of the things they dislike about traditional shopping is “pushy” sales people. The Internet overcomes this problem in that it allows consumers to shop at-theirown-pace and make their purchase when they are ready. The consensus was that “pushy” sales people are more interested in the sale than serving the customer’s interests. This causes uncertainty in relation to the product they are buying and the vendor selling the product, and introduces an element of distrust, which is a disincentive to shop.
Thus, the focus group results provided further support for the product attributes model with two improvements, a link from differentiation to tangibles to cater for remote or time poor consumers, and a link from content to enjoyment in recognition of a shopping experience where trust does not or is only poorly developed.
Figure 3: Conceptual model of consumer value
The demographics and the responses to each of the two field surveys were analysed for differences with 95 percent confidence. Some potentially significant differences existed for some of the demographics and five of the 44 item statements. Potential differences in the two demographics were expected since the university sample was unlikely to be representative of the Australian population particularly in terms of their level of education, income and age. These differences should not affect respondents shopping motivations and owing to the small individual sample sizes of the two surveys, the five items with potentially significant differences were taken to reflect the variation possible in the population and the two datasets were merged. The means and standard deviations for each of the 44 items for the combined surveys are reported in appendix two.
For the combined surveys 70 percent were female, significantly higher than the Australian population average of 50.4 percent (ABS 1996). The average number of hours per week respondents usually used the Internet to find information about goods and services they would like to buy was 1.35 with an average of 0.7 purchases a month. The products respondents reported buying is presented in Table 1.
Table 1: Types and counts of products survey respondents reported having purchased in the past 12 months
|
Note: The question asked respondents to think of one Internet purchase they had made within the past 12 months where they had searched, evaluated and bought the good or service online and write that purchase in the space provided. A number of respondents from the university survey listed multiple products. These multiple product listings are included in the table. Digital or high information products are bolded.
In Table 1, the most popular purchases were tickets, books, travel & hotels and music, all of which can be considered digital or high information purchases. The most popular analogue products were clothes, computer hardware, household goods and mobile phones. The computer hardware category consisted of notebooks, laptops, pocket PC’s, graphics cards, printer cartridges and other non specified “computer parts” and “small computer parts”. The household goods category included a frypan, cutlery and other non-specified “household goods”. Thus, all of these products can be considered non-bulky except for two motorbikes, two suitcases, two instances of wine purchases one vehicle and one bicycle. For the combined surveys, the proportion of digital and high information products purchased is 68 percent. This proportion is significantly different at p< 0.001 (z= 9.31) when compared to the proportion of analogue / low information products reported purchased.
The combined sum of the four most popular digital / high information products (134) and the four most popular analogue / low information products (38), which are non-bulky products, for the combined surveys, is 172. This proportion is 74% of all the products purchased. These results suggest that consumers principally search, evaluate and buy digital and high information products online and will also buy non-bulky, analogue products, however the proportion of analogue / low information purchases is much smaller. These results support the findings of Zeng and Reinartz (2003), Vijayasarathy (2002), Phau and Poon (2000), Peterson et al. (1997) and Stern (1995).
AMOS was used for data analysis. The advantage of structural equation modelling over traditional parametric analysis is that hypothesised behaviour can be modelled with non-experimental data collected from a survey (Maruyama 1998). Additionally, structural equation modelling overcomes the limitations of traditional statistical analysis techniques where the components contributing to an independent variable are summed, and it is assumed each indicator is measured with equal error and that each indicator contributes equally (Anderson & Gerbing 1998). The review of the literature and the qualitative results suggests that this is not the case and that several of the key dimensions identified may be correlated. Given this, the most appropriate analysis technique is to use exploratory and confirmatory factor analysis (Holmes-Smith et al. 2005).
Analysis was conducted using a two-step approach. Firstly the measurement models were tested for each of the one-factor congeneric models to simplify analysis of the three competing models (Anderson & Gerbing 1998). Measurement testing was used to establish the reliability of the constructs and their appropriateness for inclusion in a structural model.
The first step in the data analysis was to establish the convergent and discriminant validity of the six key dimensions. Convergent validity was verified with a principal components factor analysis, by running a separate analysis for each construct to establish a single eigen value above one. For the dimensions of content, trust and enjoyment convergent validity was verified when all items were included. For the dimensions of price, tangibles, and differentiation, which consisted of the sub dimensions, absolute and competitive & transparent pricing, differentiated and undifferentiated product attributes, and digital versus analogue items, convergent reliability was established for all of the sub constructs with the competitive & transparent pricing, undifferentiated and digital sub constructs loading on the first factor of each analysis respectively.
Further exploratory factor analysis was run in AMOS. The measurement models for each of the six key dimensions were initially tested with only enjoyment meeting acceptable levels as hypothesised. The measurement models for each of the remaining five dimensions were revised. Using an iterative process of reviewing the t-test scores, multiple correlation scores, standardised residuals and the modification indices, items were dropped that made sense from a theoretical perspective. The resultant six acceptable measurement models are presented in appendix three. Several of the congeneric models exhibit weaknesses in the measurement indices (appendix 3) including Tables A3.3 and A3.5, and several models are over fitted (Tables A3.4, A3.6, A3.8 & A3.12), however in all cases where this occurs, the models are theoretically based with good validity probabilities, and were therefore accepted.
The next step in the analysis was to let the six acceptable congeneric models co-vary to determine inter scale reliability, and potential structural coherence. The standardised loadings and the latent factor correlations for each of the six dimensions are presented diagrammatically in appendix four. Discriminant validity analysis indicates several of the dimensions are not entirely distinguishable empirically. The dimension content has several cross loadings with tangibles, differentiation, price and trust. Tangibles and differentiation also exhibit cross loadings as does differentiation and trust.
The Cronbach inter scale reliability of the post-hoc congeneric models (appendix 2) was .748 for content (Table A3.1), .498 for tangibles (Table A3.3), .591 for differentiation (Table A3.5), .727 for price (Table A3.7), .765 for trust (Table A3.9) and .915 for enjoyment (Table A3.11). The inter scale reliabilities for tangibles and differentiation are low. This explains the lack of discriminant validity when combined with the high factor correlations of -.81 between differentiation and trust, .77 between differentiation and content and .77 between tangibles and content. The lack of discriminant validity between content and price (.82) and content and trust (.87) appears to be due to the high inter dimensional correlations, as is the correlation between trust and price (.72). The key dimension of enjoyment is the only dimension exhibiting discriminant validity.
There is no theoretical evidence to suggest that the dimension content is measuring similar underlying aspects of the dimension trust or visa versa. To confirm this, further exploratory factor analysis was undertaken constrained to the items representing only these two theoretical dimensions. The results of this analysis identified the two pure factors as theorised. This analysis was repeated for each of the theoretical dimensions where discriminant validity was an issue, yielding the same results for each. This suggests each of the theorised dimensions is unique, albeit highly correlated.
Confirmatory factor analysis was conducted to test the two theoretical models and the conceptual model developed during this research. Factor score regressions for each of the six key dimensions were used to compute a composite score for the latent variables under consideration (Holmes-Smith et al, 2005; Joreskog & Sorbom 1989). These composite factors were then used for testing the three models in AMOS. The findings indicate that there is no support for either the product attributes or the situationalists’ models, and some support for the conceptual model. Post-hoc testing of the conceptual model was required to achieve a statistically valid result.
For the product attributes model, all of the structural pathways are significant as indicated by the critical ratios (Table 2) however only two of the paths contribute moderately to the model. This is evidenced by betas of .395 (price to content), and .395 (price to trust). None of the other structural links contribute significantly and dimensional reliability is generally poor, indicating a poor fit. The goodness of fit indices presented in Table 3, confirm this with only the GFI having an acceptable fit (.941). Therefore the product attributes model is not supported.
Table 2: Standardised estimates for the product attributes model
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Table 3: Goodness of fit estimates for the product attributes model
|
The results for the situationalists’ model are presented in Tables 4 and 5. Only the path between price and content contributes moderately to the model as indicated by a beta of .418. The path between enjoyment and content has a non-significant critical ratio and the other two paths, whilst having acceptable critical ratios, do not contribute significantly to the model and none of the goodness of fit indices are acceptable. The situationalists’ model is therefore also not supported.
Table 4: Standardised estimates for the situationalists model
| ||||||||||||||||||||||||||||||||||||
Table 5: Goodness of fit estimates for the situationalists’ model
|
The results for the conceptual model are presented in Tables 6 and 7. Two paths, price to content (.359) and price to trust contribute moderately to the model as indicated by betas above 0.3. The paths between trust and enjoyment and content to enjoyment have nonsignificant critical ratios. All of the goodness of fit indices for the conceptual model except the TLI (.882) and the probability score (.035) are however within the acceptable range. Despite some of the goodness of fit indices meeting the criteria, the conceptual model is not supported. Of the three models tested however, it is the one that is closest to an acceptable fit.
Table 6: Standardised estimates for the conceptual model Table 7 Goodness of fit estimates for the conceptual model
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
The results of the conceptual model suggest the dimension of enjoyment should be removed due to its non-significant contribution to the model. Further support for the removal of this dimension is provided in the results for the situationalists’ model where the path between enjoyment and content is also non-significant. The AMOS modification indices however suggested the addition of a structural path between differentiation and trust, which is plausible. Adding this path improves the model, however the path from content to enjoyment remains unchanged and non-significant. Removal of this path significantly improves the goodness of fit indices and standardised estimates.
The results for the post-hoc conceptual model are presented in Table 8 and Figure 4. All of the goodness of fit indices are acceptable, with a moderate probability (.340) that the model is valid. The goodness of fit indices mask some weak measurement scores for the model (appendix 3), however the post-hoc results are a close approximation of the conceptual model and should be accepted recognising that the dimension of enjoyment is a weak contributor and requires the indirect influence of tangibles and differentiation via trust, for the path between trust and enjoyment to be significant.
Table 8: Goodness of fit estimates for the post-hoc conceptual model Figure 4 Standardised estimates of the post-hoc conceptual model
|
The most frequently reported purchases for the pure online shopping scenario were tickets, books, travel & hotels and music, and overall 68 percent of purchases were either digital products, in that they could be purchased immediately online with no need for delivery, or high information products such as books, which whilst needing to be delivered, were non-bulky items. This provides a strong indication that consumers have a preference for purchasing digital or non-bulky goods online. This finding lends support to the product attributes stream of online consumer behaviour research.
The findings of this research support and extend the previous findings of Vijayasarathy (2002) and Phau and Poon (2000), in that there is a strong preference for actually purchasing digital products online, rather than just a strong intention to purchase these types of products. The results also cast doubt on whether highly differentiated products will sell well online as hypothesised by Peterson et al, (1997), since none of the products reported purchased by respondents were unique or unusual. With respect to cost, the reported purchases also show no indication of a preference for low, moderate or high cost items as hypothesised by Peterson et al, (1997) and Stern (1995).
The results indicate that the dimension of enjoyment may be a component of the pure online shopping experience, however the path in the structural model is weak. The path is also not necessary to achieve a statistically valid result. Thus, whilst there is some support for the findings of Koufaris (2002) who developed the item scale used, the empirical evidence in this research, suggests that shopping for pleasure online, if it is an important shopping driver, was not adequately captured.
The testing of the three models developed during this research suggests that the dimensions relating to product attributes (tangibles and differentiation) directly or indirectly influence the key dimensions of content, price and trust. This suggests that the refined product attributes model, successful as the post-hoc conceptual model, is a better approximation of reality than the situationalists model since the path between trust and enjoyment is non-significant without the indirect effects of tangibles, differentiation, price and content. Additionally, whilst the path between trust and content is significant, although somewhat weak in the situationalists’ model (cr = 2.659, β .226), and in the post-hoc conceptual model (content – trust cr = 2.785, β .227), a statistically acceptable outcome is only achievable when the path originates from content.
The research thus facilitates the development of a model of online consumer value as it relates to the buying decision process that incorporates the initial online shopping channel drivers and the key online buying decision drivers (Figure 9).
Figure 9: Model of the online buying decision process
Theoretically, the identification of the initial shopping drivers clarifies where each of the product attributes and the situation research streams can potentially contribute to the understanding of the online consumer value creation process. Knowing how consumers use the online retail channel is fundamental to understanding the initial step in the consumer value creation process. The results from this research suggest that the nature of the product, expressed in terms of tangibles and differentiation determine initial shopping channel selection. This points to the consumer evaluating the basic condition of the Internet to arrive at their first decision within the online buying decision process. This knowledge contributes to a better understanding of the potential impact of a discontinuous innovation in retail and can contribute to understanding the process of retail change.
The identification of the key decision drivers once the Internet channel is selected identifies the potential vendor and channel-switching drivers. The results from this research suggest that once the Internet has been selected, consumers seek out product and pricing information before arriving at the purchase decision. This finding highlights the significance of content and price as a component of the online consumer value creation process, and importantly identifies when these two dimensions are important within the buying decision process. This knowledge contributes to modelling consumer value and establishing the structural path from product attributes to product information in Figure 9.
The finding that the dimension of trust (and possibly enjoyment) is the key purchase decision driver completes the process of online consumer value creation and establishes where the situationalist literature can make its biggest contribution. The findings indicate that the development of trust in the online context is reliant upon the consumer’s expectations in relation to content and price being met and that the product information sought is a function of the nature of the product under consideration. This finding suggests that research in the area of online consumer trust can be significantly advanced if viewed as a function of the nature of the product under consideration.
In addition to the implications for theory addressed above, this research has implications for practitioners in the retail industry. With respect to incumbent retailers, the results suggest that those currently operating bookstores, music stores, travel agencies and general ticket offices offline face a bleak future and should consider migrating their business online as a matter of urgency. Retailers of other product types that are, or can be, easily digitised, such as electronic games, movies, radio broadcasting, magazine and newspaper publishers, television licence holders and telecommunication companies should include the Internet as part of their distribution strategy or risk losing market share in the near future.
Offline retailers of non-bulky consumer durables should also consider the same, particularly where miniaturisation in the industry is occurring, such as in the area of consumer electronics. Manufacturers of these non-bulky products should also consider the strategic impact of pursuing the traditionally popular policy of imposing geographic selling restrictions and / or exclusive distribution rights over a geographic area since artificial trading barriers will encourage new entrants, particularly from technologically emerging nations such as China.
Retailers of tangible bulky products not suited to shipping either because of cost or potential breakages are unlikely to be seriously impacted by the Internet. Retailers of such products however, can take advantage of the informational potential of the Internet to disseminate product and company information.
For aspiring entrepreneurs who wish to contribute to the success of the Internet as a retail channel the possibilities are many. Entrepreneurs can assist in the adaptation of existing products suitable to digitisation using current technology. Where necessary they can develop hardware and software to adapt existing products not yet suited to the channel or they can develop entirely new products. Digital books, magazines and newspapers are examples of potential adaptations using current technology. Television and telecommunications are examples of hardware and software development requirements, and new product development possibilities seem bounded only be the requirement that they be digital or readily digitised.
Despite attempts to ensure that the findings of this research are both reliable and valid, a number of limitations exist. Firstly, the original mail out survey had a low response rate and a subsequent top up survey was required. However, the initial mail out sample was probability based and appears representative of the population and variation between the two field samples is generally within acceptable limits.
Secondly, this research addresses theoretical gaps in the online consumer behaviour and retail change literature for which limited precedent has been set in terms of combining the six key dimensions within a single study. The scales used were largely adapted from the online consumer behaviour literature. The ability of these six scales to reflect the complexities and dynamics of the entire online purchasing influences has not been sufficiently explored.
Next, it is recognised that the survey instrument measured perceptions at one point in time. It is a risk to suggest that this model of online consumer value will stand the test of time. There is a further possibility of respondent bias of the results, as it is possible that respondents with computer and Internet experience may be more likely to respond to the survey than respondents with little or no exposure to computers or the Internet. Additionally, the majority of respondents to both field surveys were female, and thus the results may not reflect the true perceptions of males.
Finally, in any modelling, favourable results are relative and not absolute. Good model fit does not imply that the chosen model represents a valid reflection of reality. All that can be expected from good model fit is that the model indicates a good representation of relationships between factors. The real possibility of specification error, that is, not specifying an important construct or not specifying enough measurements exists (Baggozzi & Baumgartner 1994). For this research the standardised measurement estimates for the dimensions of tangibles and differentiation are examples of this.
These limitations do not necessarily undermine the findings of this research. They are outlined to acknowledge their existence and stress the need for further research. The results provide opportunities for further research with respect to the delimitations of scope, the further testing and validation of the scales and models developed, the extension of the models to the incorporation of demographic variables, and / or segmenting the population on the basis of age, culture, income, education, gender and computer and Internet usage. The research could also be extended to an examination of the pure offline buying decision process and shopping scenarios where consumers switch channels during the buying decision process, and finally, the methodology used.
In summary, this theory building research has synthesised existing research and theories, combined the information gathered with the results of exploratory research and confirmatory factor analysis to contribute to existing knowledge in the prediction of online consumer behaviour. This research has also contributed to the field of retail change and has set a foundation for further research about how and why consumers make shopping decisions and the likely impact of continuous and discontinuous innovations in retail such as the Internet.
ABS, (1996), ‘Census of Population and Housing’, Basic Community Profile, Canberra, ACT, Catalogue No. 2020.0.
Alba, J., Lynch, J., Weitz, B., and Janiszewski, C. (1997), “Interactive home shopping: Consumer, retailer, and manufacturer incentives to participate in electronic marketplaces”, Journal of Marketing, v 61, i3; pp38-53.
Anderson J.C. and Gerbing D.W., (1988), “Structural equation modelling in practice: A review and recommended two-step approach”, Psychological Bulletin, v103, pp411-423.
Argawal R., and Venkatesh V. (2002), “Assessing a Firm’s Web Presence: A Heuristic Evaluation Procedure for the Measurement of Useability”, Information Systems Research, v13, i7, pp168-186.
Bagozzi R.P. and Baumgartner H., (1994), “The evaluation of structural equation models and hypothesis testing”, in R.P. Bagozzi (ed.) Principles of Marketing Research, Blackwell, Oxford, pp.386-423.
Balabanis G and Diamantopoulos A, (2004). “Domestic Country Bias, Country-of-Origin Effects, and Consumer Ethnocentrism: A Multidimensional Unfolding Approach”, Academy of Marketing Science. Journal, V32, i1; pp80.
Boling E. (1995), “Useability testing for Web Sites”, Proceedings of the 7th Annual Hypermedia ‘95 Conference, available at http://www.indiana.edu/lirg/ARTICLES/usability/usability.main.html (accessed Jan 2003).
Burton T. (2001), “Cisco systems best practice customer care”, available at www.ecustomerserviceworld.com/articles/article_523.rtf (accessed November 2002).
Chen C. and Dubinsky A. J. (2003), “A conceptual model of perceived customer value in e-commerce: A preliminary investigation”, Journal of Psychology and Marketing, v20, i4, pp323-347.
Chen S. C. and Dhillon G. S. (2003), “Interpreting Dimensions of Consumer Trust in E-Commerce”, Information Technology and Management, v4, pp303-318.
Devaraj S., Fan, M., and Kohli, R. (2002), “ Antecedents of B2C Channel Satisfaction and Preference: Validating e-Commerce Metrics”, Information Systems Research, v13, i3, pp316-333.
Doolin B., Dillon S., Thompson F. and Corner J. L. (2005), “Perceived Risk, the Internet Shopping Experience and Online Purchasing Behavior: A New
Zealand Perspective”, Journal of Global Information Management, v13, i2, pp66-88.
Eastlick M. A. and Lotz S. (1999), “Profiling potential adopters and non adopters of an interactive electronic shopping medium”, International Journal of Retail and Distribution Management, v27, i6, pp209-223.
Feinberg R. F., and Kadam R. (2002) “E-CRM Web service attributes as determinants of customer satisfaction with retail Web sites”, International Journal of Service Industry Management, v13, i5, pp432-451.
Figueiredo J. (2000), “Finding sustainable profitability in electronic commerce”, Sloan Management Review, Cambridge, v41, i4, pp41-53.
Finkelstein S. (2001), “Internet startups: So why can’t they win?” Journal of Business Strategy, V22, i4, pp16-22
Gefen D. Karahanna E. and Straub D. W. (2003), “Trust and TAM in Online Shopping: An Intergrated Model”, MIS Quarterly, V27, I1, pp51-90.
Gehrt K. C. and Yan R. (2004), “Situational, consumer, and retailer factors affecting Internet, catalog, and store shopping”, International Journal of Retail & Distribution Management, v32, i1, pp5-18.
George J. F. (2004), “The theory of planned behaviour and Internet purchasing.” Internet Research, v14, i3, pp198-212.
Grabner-Kraeuter S. (2002), “The role of consumers' trust in online-shopping”, Journal of Business Ethics, v39, i1/2, pp43-51.
Hamel G. (2001), “Is this all you can build with the net? Think bigger”, Fortune, New York, v143, i9, pp134-138.
Hoffman D. L. and Novak T. P. M. (1996), “Marketing in hypermedia computer mediated environment: conceptual foundations”, Journal of Marketing, v60, pp50-68.
Holbrook M. B. (2001a), “The millennial consumer in the text of our times: Evangelizing”, Journal of Macromarketing, v21, i2, pp181-199.
Holbrook M. B. (2001b), “The millennial consumer in the text of our times: Exhibitionism”, Journal of Macromarketing, v21, i1, pp81-95.
Holbrook M. B. (2000), “The millennial consumer in the text of our times: Experience and entertainment”, Journal of Macromarketing, v20, i2, pp178-192.
Holbrook M. B. and Hirschman E. C. (1982), “The experiental aspects of consumption: consumer fantasies, feelings and fun”, Journal of Consumer Research, v9, i2, pp132-141.
Holmes-Smith P., Coote L. and Cunningham E. (2005) Structural equation modelling: From the fundamentals to advanced topics, SREAMS, Melbourne.
Janda S., Trocchia P.J. and Gwinner K.P. (2002), “Consumer perceptions of Internet retail service quality”, International Journal of Service Industry Management, v13, n5, pp412-431
Jarvenpaa S. and Todd P. (1997), "Consumer Reactions to Electronic Shopping on the World Wide Web", International Journal of Electronic Commerce, v1, i2, pp59-88.
Joreskog K. G. and Sorbom D., (1989), “LISREL 7 User’s Reference Guide”, Scientific Software, Chicago.
Kalakota R., and Robinson M. (1999), “e-Business: Roadmap for Success”, Addison-Wesley, Harlow.
Kau A. K., Tang Y. E. and Ghose S. (2003), “Typology of online shoppers”, Journal of Consumer Marketing, v20, i2, pp139-156.
Keen C., Wetzels M., de Ruyter K. and Feinberg R. (2004), “E-tailers versus retailers: Which factors determine consumer preferences”, Journal of Business Research, v57, pp685-695.
Kolesar M.B., and Galbraith R.W. (2000), “A services-marketing perspective on e-retailing: implications for e-retailers and directions for further research”, Internet Research, v10, i5, pp424-438.
Koufaris M. (2002), “Applying the Technology acceptance model of Flow Theory to Online Consumer Behaviour”, Information Systems Research, v13, i2, pp205-223.
Koyuncu C. and Lien D. (2003), “E-commerce and consumer’s purchasing behaviour”, Applied Economics, v35, i6, pp721.
Lee M. K. O. and Turban E. (2001), “A trust model for consumer Internet shopping”, International Journal of Electronic Commerce, v6, i1, pp75-91.
Maruyama G. M. (1998), Basics of Structural Equation Modelling, Sage, Thousand Oaks.
Mathwick C. and Rigdon E. (2004), “Play, Flow, and the Online Search Experience”, Journal of Consumer Research,I v31, i2, pp324-333.
Mathwick C., Malhotra N. K. and Rigdon E. (2002)”,The effect of dynamic retail experiences on experiental perceptions of value: and internet and catalog comparison”, Journal of Retailing, v78 i1, pp51-60.
Merrilees B. (2001), “Do traditional strategic concepts apply in the E-marketing context?” Journal of Business Strategies, v18, i2, pp177-190
Monsuwe T. P., Dellaert G. C and de Ruyter K. (2004), “What drives consumers to shop online? A literature review”, International Journal of Service Industry Management, v15, i1, pp102-121.
Peterson R. A., Balasubramanian, S. and Bronnenberg, B. J. (1997), “Exploring the Implications of the Internet for Consumer Marketing”, Journal of the Academy of Marketing Science, v25, i4, pp329-346.
Phau I., and Poon S. M. (2000), “Factors influencing the types of products and services purchased over the Internet”, Internet Research: Electronic Networking Applications and Policy, v10, I 2 pp102-113.
Porter M. (2001), “Strategy and the Internet”, The Harvard Business Review, May
Shabazz D. (2004), “Toward a Better Understanding of e-Marketing Strategy: Past and Present”, Services Marketing Quarterly, v26, i2, pp117-130
Song J. H. and Zhang J. Q. (2004), “Why do people shop online?: Exploring the quality of online shopping experience”, American Marketing Association. Conference Proceedings: 2004 AMA Winter Educators' Conference 15, (January) pp154-160.
Stern J. (1995), “World Wide Web Marketing: Integrating the Internet into Your Marketing Strategy”, John Wiley and Sons, New York, NY.
Tan F. B. and Sutherland P. (2004), “Online Consumer Trust: A Multi-Dimensional Model”, Journal of Electronic Commerce in Organizations, v2, i3 pp40-58.
Tan K. C., Xie M. and Li N. Y. (2003), “A service quality framework for Web-based information systems”, The TQM Magazine, v15, i3, pp164-173.
Thornton J., and Marche S. (2003), “Sorting through the dot bomb rubble: How did the high-profile e-tailers fail?” International Journal of Information Management, v23, I2, p121.
Torkzadeh G., and Dhillon G. (2002), “Measuring factors that influence the success of Internet commerce”, Information Systems Research, v12, i2, pp187-204
Vijayasarathy L. R. (2002), “Product characteristics and Internet shopping intentions”, Internet Research: Electronic Networking and Policy, v12, i5, pp411-426.
Zeng M., and Reinartz, W. (2003), “Beyond Online Success: The Road to Profitability”, California Management Review, v15, i2, pp107-130.
Zeithaml V.A., Parasuraman A., Malhotra A. (2002), ” Service quality delivery through Web sites: A critical review of extant knowledge” Academy of Marketing Science Journal, v30, i4, pp362-374
Appendix 1: List of questions included in the survey
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Appendix 2: Means and standard deviations for the forty-four questions for the combined field samples
|
Table A3.1 Standardised estimates of the post-hoc dimension content one factor congeneric construct
| ||||||||||||||||||||||||||||||||||||||||
Table A3.2 Goodness of fit estimates for the post-hoc dimension content one factor congeneric model
|
Table A3.3 Standardised estimates of the post-hoc dimension tangibles one factor congeneric construct of the theoretical model
| |||||||||||||||||||||||||||||||||||
Table A3.4 Goodness of fit estimates for the post-hoc dimension tangibles one factor congeneric model
|
Table A3.5 Standardised estimates of the post-hoc dimension differentiation one factor congeneric construct of the theoretical model
| ||||||||||||||||||||||||||||||
Source: developed from this research
Table A3.6 Goodness of fit estimates for the post-hoc dimension differentiation one factor congeneric model
|
Table A3.7 Standardised estimates of the post-hoc dimension price one factor congeneric construct of the theoretical model
| |||||||||||||||||||||||||||||
| ||||||||||||||||||||
Table A3.8 Goodness of fit estimates for the post-hoc dimension price one factor congeneric model
|
Table A3.9 Standardised estimates of the post-hoc dimension trust one factor congeneric construct of the theoretical model
| ||||||||||||||||||||||||||||||||||||||||
Table A3.10 Goodness of fit estimates for the post-hoc dimension trust one factor congeneric model
|
Table A3.11 Standardised estimates of the dimension enjoyment one factor congeneric model
| |||||||||||||||||||||||||||||||||||
Table A3.12 Goodness of fit estimates for the dimension enjoyment one factor congeneric model
|
Appendix 4 – Structural coefficients and latent factor correlations for the six post-hoc congeneric models
The missing correlation between content and enjoyment is 0.26.