Converting browsers to buyers: exploring what drives consumer choice in internet e-commerce

From WWW/Internet 2005 Proceedings


Why do internet users behave as they do, are their activities solely determined by website design? Alternatively do they create their own pathways as a response to designated systems? For many, internet design is about the imposition of schemas, predetermined flows and consumer motifs, allowing the shepherding of an understood and mapped user towards buying products and services. However if this were true then every browser would also be a buyer. The underlying concepts of current website interface design rely on a number of pretexts which, when reviewed in relation to human activity and activity interpretation, become questionable in their veracity.

In order to define and reference human interface interaction this study has considered the measurement and intent of interaction. Aspects of measurement have been acquired to establish incremental dimensions and operational processes. Intent is used as the control element of a driver and offers a way to observe changes in interaction by the imposed environment of a website. Theories relating to commercial website developments, human computer interaction (HCI) and aspects of shopping are referenced in this project. HCI is also considered in relation to its scientific basis relative to this project.

Data has been acquired under controlled conditions to reduce anomalies associated with the use of the incumbent populations in University based research at this level. This study combines both diagnostic and reflective methodologies through the acquisition of primary data in three studies. Data has been captured in pilot, main study and counterpoint studies enabling a non-deterministic (by initiation or limitation) based review of activity. They establish a complex multi-level interaction process that is constantly modified by aspects of one of three domains and the interplay between initial drivers, trajectories and choice gateways. These components have been annotated to establish the ground work for a formal interaction mapping process.

Further research needs to be conducted into how data is acquired and reported. Methods underwriting reporting have a direct affect upon data’s holistic value. Studies devoted to a singular proposition may exclude much useful data while seeking a specific resolution. Further research needs to be conducted into the way that consumers reference their activity is required. Further research is needed into avatars that transpose interaction into specific dimensions, adaptations and conventions. These could provide an insight of parallel and concurrent mechanisms in interaction.

This was first published in 2005 but the questions raised have yet to be answered and the insights gained have never been developed outside of commercial projects.

The full paper is available on request.


1.1      Overview of Chapter

There are three clear aspects that underwrite internet shopping under review. Firstly that many companies believe by converting a process using technology that it becomes more successful. Automatic acceptance that establishing internet presence achieve sales, without in-depth research is a prime example of technological determinism. Secondly the relationship to how a consumer behaves and interacts within internet shops, what elements effect their decisions and define their final results (O’Cass, 2002). Finally there is an aspect based upon social communication systems, in how they are transformed by their use on the internet. An understanding of the foundation of the internet is also required to set the scene into which these various aspects are reviewed.

1.2      Foundation of the Internet

In order to understand internet shopping the foundations of the technology and the reasons for its development need to be understood. While the World Wide Web (WWW) is a relatively new idea, commencing in the 1980’s the foundations of it can be traced to the work of people like Vannevar Bush in the 1930’s and 1940’s (Bush, 1945). Bush was mainly concerned with the logistics of vast amounts of statistical data and saw his role being overwhelmed in the future. His idea was to extend human capacity for the storing and retrieval of information through mechanistic augmentation. This information was to have a linking / indexing technology and was later developed into hypertext. Douglas Engelbart attributed his work on hypertext to Bush’s concept of “Memex” (Engelbart, 1962). Using hypertext Tim Berners-Lee was able to develop linking pages of information at CERN which was the first internal net (Berners-Lee, 1989). Throughout its development and application the internet provided ways to store, disseminate and share information which supported various centred human activities.

1.3      Technological Determinism

There is a notional level where technology creates an absolute route to a desired conclusion. This determination surmises a benign, logical process or system which creates processes and control mechanisms free from random, illogical and uncontrolled influences (McLoughlin, 1999). However while an idealised perfect use of technology may create such an operation its use by people automatically creates random, illogical and uncontrolled influences.

1.3.1      Foundation of Commercial Computing

During the initial development of computers certain decisions were made that had a profound influence upon all future computer based technologies (Naughton, 1999). Land (1999) describes how the cost of transaction processing time was very high due to the initial data having to be aggregated to suit the formats of available computer languages. The cost of initial data preparation was absorbed into human time which was then far cheaper than computer time. In this relationship humans were subsumed to basic data conversion for input/output tasks, while the computer carried out complex data analysis tasks. In this definition of interaction, each component operates by predefined tasks which enable a cost effective mode of computing. This was set and operated within one social, hardware and software timeframe. However, this deterministic philosophy has, since the 1950’s, become embedded in computer design and commercial usage culture.

1.3.2      Commercial Computing and the Internet

Commercial companies have since the 1940’s, with the development of the J. Lyons computer (Land, 1999), been using technology to augment business processes. Their priority was the analysis of data to cut costs and maintain a competitive advantage, through supply chain management (SCM) systems. Multiple supply routes, manufacture, distribution and sales are responsive along a closed cyclic system with each other. Further developing of this cyclic process involved localised data entry and the development of network solutions (Laudon and Laudon, 2004). Because of this previous experience companies developing into the open market system of the internet were unprepared for a sophisticated user group able to shop elsewhere.

1.3.3      Internet Statistics

Statistics are a cornerstone of many commercial activities especially product sales which are augmented by market forces studies, demographic and product focus groups. These studies create past, present and forecast data for statistical and descriptive strategic planning (Evans, 2000). It is to prove return on investment (ROI) that many deterministic data review methods have been used in relation to internet shopping. In internet shopping the use of retrospective data is uniform and is based upon past clicks, page views and cookies that sample user information.

1.3.4      Data Mining and Information Foraging

Existing retrospective statistic based methodologies produce results that use data mining or information foraging techniques to reflect upon what has happened on websites. They do this in an attempt to forecast which elements of a website lead to users making purchases and which elements do not. The resultant data suggests ways to increase conversion from browsers to buyers by improved linking between websites elements. These systems rely upon information foraging and web data mining techniques, to predict optimal relationships between data and page locations. They often use complex mathematics to create algorithms intended to describe patterns of use, access routes and cue to activity (Chi, et al. 2000). The concept of Information Scents, suggests that users decide on their course of action based upon cues, which derive behavioural patterns of interaction then form guide routes. This formalised process lacks input from either cognitive processes or cultural human contexts.

1.3.5      Use of Data and Information

The Cognitive Walkthrough of the Web (CWW) has attempted to marry information scent with cognitive processes (Blackmon, et al. 2002). Recognition is however given that different user populations produce different results and no testing was done in this area. The Bloodhound Project sought to create automated, usability and accessibility reviews based upon InfoScent™ (Chi, et al. 2003). The project attempts to formulate a clear method showing consistent, measurable elements that provide benefit in the form of usability inspection tools for designing websites. These ways of interpreting website data have not been supported by the commercial usability community. They rather believe that the results are synonymous with a specific type observation of users seeking information. This deceptive description of informational routes is stated regardless of where the user finds what they are looking for (Nielsen, 2004). However User Centred Design (UCD) also uses the notion of mapped routes of activity which are intrinsic to inspection and usability developments in by consultants (Lazar, 2001). While there is a conflicting view of how to interpret this data, clients continue to have the problem with accessing “actionable statistics” (Foley, 2001) for businesses decisions. Additionally while consultant’s methods of measurement remain opaque, they will have their veracity questioned (Rosenfield, 2001). Both of these are highly limited expressions of human interaction as they lack any deep understanding of why these actions were taken. No understanding of human intention, for product purchase can be observed through this data.

1.4      Consumer Behaviour and Interaction

A common description of interaction is, we are what we do, but this depends upon what is measured. This limited view of people is supported by those who use website statistical data to determine interaction, yet is it this simple to understand people? An understanding of cognitive functions, desires, experiences and the purpose of activities offers characterisation to establish measurable dimensions in internet shopping.

1.4.1      Cognitive Functions in Interaction

Human cognitive functions in the area of interaction considered in this project rely upon several types of input. These include desire or drivers on a contextual basis unaffected by usability issues but defined by consumer traits (Perea, et al. 2004). Inputs also come from experiences in personal / social and commodity / product relationships which create biographical templates (Kopytoff, 1986). Inputs are derived from current activity as a form of self narrative (Flanagan, et al. 1998) and related to a specific time frame. Finally there are inputs that describe aspirations and goals (Hutchins, 1995). These inputs drive choices and decision activity prior to new actions and act as a description of cognitive engineering (Long and Dowell, 1998). Additionally they establish a personalised framework for the characterisation of success, error and failure in complex tasks. Ultimately a formal measurement of interaction is needed, as a process rather than just a destination (Green and Petre, 1996). In the interim the term narrative enables an interpretation and review of real time activity data.

1.4.2      Desires and Drivers

To start activity some form of catalytic reason, desire or drive is required. Not only does this define the activity but it determines aspects of how it progresses and describes a condition of success. This initiator can be based upon environmental factors including other people, society, places or environments and systems which may be attributed or conferred upon the user. These catalysts then operated in a multi-dimensional framework that can be influenced by many factors.

1.4.3      Biographical Templates

User’s experiences inform their attitude and response to stimuli. In the case of gender, women have in the past twenty years created their own digital divide gaining on and overtaking men in the accessing and utilising of internet shopping (Ono and Zavodny, 2003). This can be seen as an iteration of female shopping experience accessing a new channel. However as the process evolved over considerable time additional factors should be considered. Human physical, emotional and experiential activity maintains a biographical element through significant moments or indices. These indices create biographical competencies that relate success, failure, frustration and many other emotions to activity in social, mechanical and technological environments. As a construct that determines choice, a biography (Appadurai, 1988), (Kopytoff, 1988) is superimposed upon objects or commodities defining where they have been, how they have been changed by external factors and proposing trajectories and possible blockages. Activity is obscured by many external factors including historical, political or social conventions. Biographical notation enables the salient understanding of information that would otherwise be lost. The understanding of human interaction can be viewed as participation in the creation of personal historical elements having both biographical and active elements. The cultural disposition of technology, interactions and resultant pathways remain difficult to interpret without recourse to an activity framework. A method is then required to relate the electronic media habitus to external attributable counterpoints.

1.4.4      Temporal Narratives

Narratives allow the recording of active elements in internet shopping, which describe responses to information in numerous potential trajectories (Jennings, 2005). While this narrative can be characterised through a think aloud protocol (Ericsson and Simon, 1980) representations of this discourse establish the foundations of individual drives towards action (Nakhimovsky, 1988). Effective mapping can be achieved using a lexical approach (Gulrajani, 2003) as associated with recovering endangered languages. This would allow the use of rational linguistic descriptions of dimensions including orthography, morphology, syntax and semantics. The creation of a lexical basis (Pustejovsky, 1991) makes individual actions expressible aspects of groups of actions (Flanagan, 1998) with related compound, processed and adaptive meanings.

1.4.5      Goals

Goals can be a descriptor of predetermined final destinations or may offer a general context rather than a specific, “I’m looking for a book” as opposed to “I’m looking for this book”. The general interpretation of an open and untamed (Benyon, et al. 2005) source of information like the World Wide Web (WWW) requires a systematic review of actions. Actions and user activity in relation to an observable world require a common representation to determine navigation, related target acquisition or goals (Jul and Furnas, 1997). These goals can subsequently be reduced to a form of knowledge morpheme. As an inter-related sub-rationale unit “the item I seek”, the goal then would have a distinct and finite form. In seeking to achieve these goals, adaptations have been established by reduction or addition “the item I seek is not available in red” so to gain my item, “I will take it in black”.

1.4.6      Adaptations

Adaptation allows the extension of narratives creating alternative experiences on the same object or situation (Nakhimovsky, 1988). Further modifications can be made in a process of use, where adversity produces redirection. Often activity adversity is characterised by choices in terms of “satisficers and maximizers” (Schwartz, 2004) too little or too much information causing sensory deprivation or overload. In these cases activity may cease through this adversity or be directed to an alternative source of information (Hudson, 2005).

1.4.7      Conventions

Conventions allow the creation of index points in a narrative activity where rule systems have affected internet activity (Flanagan, et al. 1998). Agreement of conventions in social, emotional and commercial arenas for completion, enable a measurable resolution to tasks. Social conventions are considerably more complex that is possible to iterate in this project. It is difficult to externally characterise an individual’s success, error and failure except by some imposed system. Error and failure to fulfil planned system objectives in human related systems is inevitable (Hutchins, 1995). Commercial conventions have been integrated into internet shopping design.

1.5      Social Communication Systems

Gaining cognition of interaction relies not only upon what is done but to discern what is intended. The study of shopping is associated with sociology, cultural theory and research (Miller, 2001). Miller discloses a highly complex process and is reactive to social, environmental, ethical and economic contexts. There are similarities between the results of offline shopping and online shopping, in the transfer of a product or services for payment. However these processes are not exactly the same and operate differently from each other.

1.5.1      Social Conventions and Signs

Conventions create a common acceptable process for activity allowing social constructs like chatting with friends to transcend their normal ecosystem. They can then interpose themselves, with modifications upon new environments and media. Obvious modifications in chat rooms involve not seeing people’s faces, observing intonation in vocal patterns and confirmation of identity. These elements allow the building of a picture explaining no just where activity is, but also what is acceptable in this area.

1.5.2      Social Mediation

Mediation is developed as a process to amalgamate and morph interfaces. In the case of a chat room, mediation is approved form of language, including slang like “lol” meaning laugh out loud. Changes operation by talking through a keyboard, how the experience is visualised and environment of public conversation from a private location are mediated by software and existing social constraints. Where the mediator gains a pivotal role in any transaction its affect is continually present, thought how it mediates is not necessarily visible (Dourish, 2001).

1.5.3      Commodities, Circulation and Exchange

Our cultural values have been modified by access to information sources through the internet that were not previously available. The “way things are used” has been transformed into commodity in its own right rather than just what is used (Appadurai, 1988). For example; buying chocolates for a friend in another country over the internet allows a purchase from their local store. This shows the commodity process of purchase, extraneously from seeking information and looking at products. Where new aspects of commodities are implied there are changes in the ways thing circulate. Who has access to these items? What is their intention? Questions like this then become critical as recognisable terms of social regulation and acceptable behaviour become opaque. This further influences society redefining aspects of exchange, as now anyone can get an item, regardless of social standing or perceived suitability.

1.5.4      Consumption

There is a complex relationship between consumption and “individual choice set within a market structure” (Miller, 1998). Simply because an item is available as an option or choice it is not implicit that it will be consumed. Rather other external factors describe consumption. Before the internet simple elements like consumption and redemption changed from barter with avatars like money or credit. These symbols of value have the potential to be interpreted or misinterpretation, unless linked to specific results. Other factors enable the distortion of the process by intervention through perceived or imposed social or legal conventions.

1.5.5      Trust and Legalities

The internet has created a dynamic relationship with users through immediate access to information; however the normal or primary social mannerisms that can be tested through a face to face transaction do not exist. Without the capacity to determine the truth of information gained, secondary clues are required (Chong and Liu, 2000). The nature of trust and legal recourse on the internet is a matter of great concern yet there is limited research to determine what factors create these aspects in internet shoppers. Other factors that engender trust in offline shopping include privacy (Miller, 1998). Privacy concerns will be a defining aspect of consumer confidence and company profitability (Prabhaker, 2000) in the future.

1.6      Human Computer Interaction

HCI has its basis in sociological philosophies and academic research (Benyon and Imaz, 1999). HCI has evolved around interface design in effect dealing with problematic interfaces that were produced due to subsuming of humanity to a secondary element in the computation processes. In a recent keynote lecture Alistair Sutcliff from University of Manchester suggested that HCI should “pillage any discipline with theories we can use” (Sutcliff, 2005) this underlying focus on importation poses questions regarding the state of HCI theories and their capacity to comment upon research.

1.6.1      HCI Theories

A variety of theoretical approaches exist that attempt to characterise HCI (Sutcliffe, et al. 1991) but a full definition remains elusive as its context involves evolutionary elements. HCI is constantly developing as new research becomes available for review. HCI as a product of research (Long and Dowell, 1989) is defined by its own title and constantly seeks to define this relationship by various forms of computer interface. However this project is more focused on the interaction element. Interaction is a combination of human and computer where the starting point is humans. As with most other HCI studies consideration is made to human factors for initial affects upon the character of a study population. However this project goes deeper into aspects of user demographics in relation to activity.

Diagnostic methods of review (Smith, 2005) like grounded theory do not predispose research data acquisition or review method. It suggests an emergent form of theory choice, by what fits the results (Glaser, 1967). Emergent review of data fundamentally differs from a hypothesis testing as it creates a theory from the available result data. This kind of theoretical base is quite compelling when faced with so many potential theories to examine and review human interaction data. Interaction mediates in the process where human physical, emotional and experiential activity correlates to a machine world (Jul and Furnas, 1997). This counterpart machine world displaces recognised liner actions (Wegner, 1997) shown through data with parallel activity in new systems. This further requires a need to understand both constructs relative to one another. Recent ideas of embodied interaction suggest this context of real and mechanical / electronic world as a unifying characteristic of separate domains. “Embodied interaction” poses the idea that embedded tangible and social computing systems define their meaning from their context rather than the parts that create them. This is a central theme in understanding interaction. Thematic approaches afford an elemental view of interaction drawing from many disciplines rather than seeing hierarchies in theoretical relevance (Dourish, 2001). Rather than a forced acceptance of one methodology over another a unification of activity theories and practice offers a valuable and functional appraisal of activity. While embodied interaction aligns many disciplines there remains a lack of relevant interlaced underpinned conceptual and philosophical work (Sutcliffe, et al. 1991). Several associated theories have attempted to capitalise upon this situation by trying overwriting HCI with more recognisable commercial theories (Kuutti, 1995) like Activity Theory. This situation has evoked a process of reflection upon the fundamental basis of this arena of science.

1.6.2      Scientific Foundation of HCI

HCI attributes its basis to scientific process, theoretical forms and systematic methods. The primary underlying philosophy of science is logic through progressive selection, review and interpretation of data. Formalised modern logical philosophy is derived through Aristotelian logic as defined by Plato’s review of Socrates understandings of his mentor (Plato, 372BC) in western societies.

1.6.3      Socratic Logic

Socratic logic or the Socratic Method involves a dialogical process operating through reduction. An initial hypothesis is made then contradictions are observed which are used to steadily undermine and modify the original hypothesis. An example of Socratic logic is statistics where data that falls within specific boundaries or fences is accepted and outer fence or outlier data is removed. Provided a justification for removal is stated little consideration is given to the potential loss of perspective and the morphing of results through distorted windows of interpretation. While this logical process creates definitive results the capacity to relate this information back into real world environments is questionable in the same way that a picture once reduced, when re-enlarged cannot be exactly reconstituted but loses its smoothed edges. This is an example of data captured being lost by the application of a Socratic logical philosophy which has been transformed into a process.

1.6.4      Dialectic Logic

In “Phänomenologie des Geistes” The Phenomenology of Spirit or The Phenomenology of Mind in English (Hegel, 1807) an alternative basis for logic is offered by Hegel where a threefold approach of hypothesis is stated, an antithesis proposed and a synthesis is created. While these terms are not directly attributed to Hegel they are his legacy. This evolutionary process seeks to consider a counter-proposition in relation to the proposed. By creating a question the desired trajectory of research is set, however if the immediate response is to consider the fallibility of the hypothesis then a more considered question, based upon combination is possible.

1.6.5      Mechanistic vs. Humanistic Logic

Both Socratic and Hegelian logic have derivative philosophies which have affected perceptions of humanity and create datum’s that guide theoretical developments. These theoretical developments can then be associated through the philosophies of Socrates / Bourdieu, deterministic subordination and Hegel / Marx open system materialism. Socratic / Bourdieu logic and social theoretic forms observe deterministic absolutes. Bourdieu determined that a society’s structure is defined by cultural and social inequality, which pervades all forms of interaction (Bourdieu, 1990). This would be exemplified by an elite class controlling access to new technology. Dialectic logic seeks to see a complex relationship between concurrent processes. By maintaining a pragmatic view of these theoretical ideologies and philosophies it is possible to present inter-related propositions relating HCI to logical activity within a whole unedited context. While Hegel / Marx reserve an organic logical form specific to each interaction. Trotsky, Marx, Engels and Lenin modified Hegel’s dialectic, though the only specific note of this was published after Engels death in “Nature of Dialectics” (Engels, 1883) by the removal of its idealistic orientation. Under Marxism Hegel’s dialectic logic evolved into dialectic materialism, retaining its essential focus upon contradiction. This would be exemplified by the recognition that immigrant populations provide valuable assets in society, where the social atmosphere might be chauvinistic. In an internet context, research that revealed outlier data and contradictions would be included in the results. Methods of measurement are intrinsic to scientific developments and enable provable and repeatable theoretical works. In this manner survey based web metrics could be considered to be the thesis; observed ethnographically derived data an antithesis and a combination of these results the synthesis. The inclusion of contradictions and outlier data in both forms of research would therefore maintain a holistic view of interaction.

1.7      Summary & Conclusion

The integration of informational aspirations and commercial functionality in the WWW has created a melange of vying purposes which remain prevalent in shopping website design. The change in status associated with a shopper who becomes an internet shopper revolves around a process of modification by action, environment, methods of acquisition and appropriation. While similarities are used as procedural points or avatars technological modifications produce diverse and untamed results more associated with wicked problem solving (wicked and untamed) than logical progressions. There is a direct cultural relationship to the way people utilise and react to internet technology.


The continuing battle between mechanistic technological determinism and humanistic evolutionary theoretical forms remains a central focus of our society, setting its aspirations and defining its potentialities. These two competing forces have produced developments in both our understandings of technology, its social impact and delineating the boundaries of developments. The involvement of people in this process undermines any deterministic view of interaction. If modifications in interaction are characterised using Socratic logic, results that contradict the hypothesis modify it, but they are lost in progressive changes and outlier data is not reported. If both the acquisition and review of data follows dialectic logic the initial hypothesis is kept. Each contradiction and outlier is noted as antithesis, and then both hypothesis and antithesis are combined to create a synthesis. It may be that dialectic thinking has been sidestepped due to its association with Marxism. This affords an organic melding of ideas as opposed the reductionism of the Socratic Method.

1.8      Design & Implementation

In the process of generating this review a lack of previous research about consumers thinking when using internet shops has been observed. As a consequence of this, a pilot study has been produced to find the questions that will elicit that information. Additionally who is asked is as much of a priority as what is asked. In this respect students are considered to have a technologically superior experience than the average consumer. This population group has not been the main source of data, but rather an open demograph has been employed. Extending the pilot study into a main study has depended upon this literature review and the pilot study results. Both research procedures have informed on key characteristic of structure, content and method of review. A counterpoint to these studies is needed to offer results that are balanced showing both local and remote data capture. The counterpoint study is based upon observation and recordings of interaction scenarios and interviews. Scenarios are often used in UCD to establish a user’s experience of an interface (Lazar, 2001). However in this case they were used to observe participants accessing both prior experience and in gaining new experiences, so that these processes could be compared. Secondly they were used to view differences between consumer users and heuristic user shopping. In this way consideration in given towards the technological experience of the participant and the effects that it may produce in the results. Dialectic logic has been used as the basis of data capture and review.

Research Methods and Results are in the full paper.

4.0        CONCLUSION

4.1      Overview of Chapter

The combination of these diverse data acquisition and review methods has far ranging results. However the expression of this information requires consistent reference to both known forms of review so that the potential of comparative results is not lost.

4.2      Dialectic Enquiry Process

It has been a key operational activity throughout these studies to maintain all the data acquired even if it appeared as contradictions or outlier data. This was done as a direct response to dialectic logic which maintains all aspects of a possible solution, objections to that solution and then creates a synthesis from all available data. This process operates in a similar way to grounded theory as the final hypothesis and antithesis is developed from the study results. However the processes used to review the data also followed a dialectic process by comparing opposing methodologies of data logs and interviews to maintain a holistic view of what drivers exist, there influences and how they correspond to their context. The description of drivers, trajectories and choice gateways has been derived both by positive enforcement and negative exclusions creating synthetic results.

4.3      Measuring Interaction

A major element of this project has been to establish both elements and pathways of interaction. Linguistic dimensions have been used as a starting point. In this way the relationship between elements was established at the outset. The linguistic term of a sub-rational element is used to describe drivers of interaction. Drivers in this sense do not of themselves have meaning except by adaptation or addition. In the situation of this project environmental influences create an initial context for interaction. Upon accessing a website various stimuli including text, images and navigation create choices. These choices are directly linked to contextual influences which render a direction and a point to point trajectory. This recognition of elements further suggests a mapping process.

4.3.1      Interaction Annotation

This project has revealed elements and descriptors of interaction, referencing components of lexical dimensions. These descriptors follow conventions and are adaptable enabling modification, new trajectories and are goal focused. In the course of activity trajectories are influenced by choice gateways where varying degrees of local influences modify both velocity (time ratio) before next choice gateway and or direction. These modifications are shown by the proximity of the line towards each set of influences having both positive and negative aspects. The following map shows a female user, with little time, using a known site; uses browse functions to locate products, then reviews product picture and comparable text to confirm purchase. The alternative would be that they are overwhelmed by too much text and unable to verify a product through a recognisable picture so make no purchase. This example has been derived from two separate shopping attempts one successful the other not.

Choice Gateways in E-Ccommerce Karl Smith

Figure 4 – Shopping activity process map overview

Note:       Choice Gateway construction is not indicative of degrees, but rather shows a general location of choice decision; this requires additional specific research to describe.

While the notation shown in Figure 4 offers a visual map of a single driver, its domain and a number of choice gateways the process is shown in an overview mode for the simplification of this process. Additionally an alternative outcome is shown by the inclusion of a broken line at a key choice node.

4.3.2      Trajectory and Choice Gateways

The trajectory avatars created in this study show a morph-able entity which is modified and restructured at each choice gateway. These contexts project a trajectory or avatar showing a temporary manifestation or aspect of a continuing entity prior to a choice gateway. These trajectories afford themselves relative contexts for activity by comparison of offline and online experiences. This notation is in its inception and requires considerable development beyond the scope of this project.

4.4      Discussion of Results

This project has produced results on several levels. There is characterisation of interaction through labels of drivers, trajectories, choice gateways and modifiers. Secondly a notation has been developed to recognise of the changes that interaction goes through. These two elements give credence to the notion of process but also through discovered contradictions suggest that further in-depth study is required. Additionally research has been conducted using a dialectic approach maintaining the integrity of all data that has been captured.

4.4.1      Understanding the Problem

There has been much written about the benefits of user centred design, (Nielsen 2000, Lazar 2001) however in its current form it is directed from a corporate or website designer perspective rather than a user. Although considerable work has been carried out in UCD in other domains of HCI it has not migrated to website design. In website design heuristic users test websites after determinations on structure and content are decided, to establish how they interact with existing structures. Consumer users are the last testers prior to site launch or are used in a discovery process when problems arise. This design process is vastly different from those employed in other areas of design for example furniture design. In furniture design the purpose of the end product is to provide a service for users and a transfer of finances from users to suppliers. Its dimensions are determined through anthropometrics based upon predetermined tested standards fitting the proposed user group. How it interacts with the user is based upon ergonomics explaining the established boundaries of activity for the product. These ergonomic boundaries are created both through considerations of the specific project and of vast quantities of testing both prior to commencement and at progressive stages during product development. In undertaking research primarily from a user perspective the final product is usable and accessible rather than modified to fit after construction. In the same way this study has attempted to map and create dimensions that describe activity, its boundaries and relevant user groups or populations.

4.4.2      Online Shopping

There is currently very little research into the online shopping process. Mainly studies are orientated around what an interface does to a persons information gathering (Chi, et al. 2000). How they are changed by them or the difficulties derived from this imposition. This study has discovered what a user imposes upon the internet shopping process. Acquiring data from a user centred basis to see how a user interprets various interfaces. The initial interface of the computer was not a focus as the participants were existing online shoppers. Also operation systems or browsers were not considered for the same reasons, even though data on these things was acquired. It is in the electronic world that there is a lack of shopping related studies. No comparative study was found except in the domain of sociology where Daniel Miller describes shopping as viewed from a dialectical process (Miller, 2001). This process is highly complex and is reactive to social, environmental, ethical and economic contexts. Further the mechanisms of online shopping were considered in relation to how people think and store their experiences. How people access their experiences when shopping and verbalise them. Also consideration was given to the relative experience of offline shopping and how users compare and combine the two experiences.

4.4.3      Comparative Results

The main study produced a snapshot of United Kingdom e-Consumer views about their own interactions. Several startling opinions were suggested; that internet shopping would replace high street shopping, that consumers had more confidence in internet security than high street shops and that personal convenience, lifestyle or time outweighed price as a key motivator in internet purchasing. Conversely the ethnographic study uniformly showed price as the as the key motivator in internet shopping. Other considerations of lifestyle were stated but these related to a lifestyle of bargain shopping. Only two participants expressed and operated differently, they were both heuristic users who considered access to information their primary goal that may lead to a purchase online or direct them to an offline shop. Consumer users were uniformly manipulated by systems of navigation and online marketing often unable to recognise paid for links to sites. Heuristic shoppers were able to adapt their online experiences, based upon computer competencies and personal will to suit their desired shopping experience. In comparing these two studies there does appear a clear contradiction between the need to project a more civilised sophisticate shopper in the main study. Whereas in the counterpoint study participants expressed their opinions and interactions in a relatively real observed shopping experience. Their purpose in internet shopping is not to meet basic needs and desires but is rather responding to social change in their appropriation methods. While both sets of ethnographic study participants clearly wanted their basic needs and desires met and gave little credence or time to sites that confused or did not seem to have information in a format that they could manipulate. Survey, ethnographic scenario and interview results produced contradictions on an individual basis. The results show a considerable difference between perceived interaction and actual interaction by users. If they themselves are limited in their characterisation of their own interaction it is so much harder to observe and may require the type of complex review methods used in discourse analysis (Bauer and Gaskell, 2000).

4.4.4      Conclusion

This project has revealed complex relationships and processes associated with converting browsers to buyers. The foundational elements of these processes and their relationships to one another have also been investigated. In this study the statement that consumers are not a mapped and understood component of an internet shopping process is verified. Internet shopper’s interactions have been shown to be primarily based their ability to gain control over interaction and relates to technological competency. Secondly personal experience guides interaction before choices are made regarding any website about to be used. Other factors in the results suggest that there are contradictions in survey based questioning. They do not reveal a true picture of shopping interaction but rather describe aspirations of shopping behaviour. This is not because participants chose to mislead, but rather because they themselves characterise their interactions differently from what they actually do. In this context the methods of data acquisition to describe e-Commerce interaction must come under review. Additionally the use of heuristic users to describe problems in website design also becomes questionable as they operated in a different way to consumer users when seeking to make online purchases. To empower consumers and view interaction from a true user perspective therefore requires considerably more complex study than that undertaken in this project.

4.5      Project Evaluation

The project answered it objectives and has produced results that describe current United Kingdom consumer experience of internet e-commerce. Also a set of dimensions that describe current user decisions influences and fulfilments have been created through both the literature review and experimentation. The project was far more complex and time consuming than had been expected. Data evaluation to answer the specific questions required intensive work, yet only a small portion of the total data captured was utilised. This additional data could then form the basis of other studies in interaction or further work using discourse analysis.

4.6      Strengths and Limitations of Project

This project has fulfilled its purpose of exploration and has produced several insights into internet shoppers. However due to the complex nature of the information exposed only initial elements have been described. This project has been limited by the time available to both develop the series of studies and analyse the resultant data. Several possible elements can be further investigated and lead to a more concise understanding of internet shopping interactions.

4.7      Further Work

This study has produced a number of interesting results which require reproduction and extensive additional investigation.

4.7.1      Reporting HCI data

Further research needs to be conducted into how data is acquired and reported. Methods underwriting reporting have a direct affect upon data’s holistic value. Studies devoted to a singular proposition may exclude much useful data while seeking a specific resolution. The use of dialectic enquiry in data acquisition and reporting enables diverse and unexpected results and offers new unexplored avenues of research.

4.7.2      Interaction Mapping

Further research needs to be conducted into the ways that consumers reference their activity is required. Consumer drivers have been observed yet the trajectory of these drivers has only been annotated on an individual basis in this project. This poses questions regarding concurrent trajectories and methods to characterise key identifiers of groups of people. Demographic organisation has been used throughout this project to determine the nature of participant populations. This provides a set of common linking descriptions accessible to both scientific and commercial domains. Choice gateways have been used to describe complex nodes of divergence and require considerable further investigation to define operational mechanisms of choice. In this aspect eye tracking would be very useful to determine visual inputs as initiators to the mental process.

4.7.3      Interaction Language Avatars

Further research is needed into avatars that transpose interaction into specific dimensions, adaptations and conventions. These could provide an insight of parallel and concurrent mechanisms in interaction.

Full Data Set is from 2005.

10.0        REFERENCES


Adriaans, P., Zantinge, D. (1997). Data mining. Addison-Wesley Longman Publishing Co., Inc.   Boston, MA, USA.


Appadurai, A. (1988). The social life of things. Commodities in cultural perspective, p.64-91. Cambridge University Press: Cambridge, UK.


Bauer, M., W. Gaskell, G. (2000). Qualitative Researching with text, image and sound. Sage Publications Ltd: London, UK.


Benyon, D. Imaz, M. (1999). Metaphors and Models: Conceptual Foundations of Representations in Interactive Systems Development. Human-Computer Interaction.14 (1) p.159-189 [Electronic version]. Retrieved July 1st, 2005 from


Benyon, D., Turner, P., Turner, S. (2005). Designing Interactive Systems. Pearson Education Limited: Harlow, UK.


Berners-Lee, T. (March, 1989) Information Management: A Proposal. The original proposal of the WWW, HTMLized. Retrieved July 1st, 2004 from


Blackmon, M., H., Polson, P., G., Kitajima, M., Lewis, C. (April, 2002). Cognitive walkthrough of the Web. Conference on Human factors in Computing Systems: Proceedings of the SIGCHI conference on Human Factors in computing systems: Changing our World, changing ourselves. ACM Press: Minnesota, USA.


Bourdieu, P. (1990). Distinction: A Social Critique of the Judgement of Taste. Harvard University Press: Harvard, USA.


Bush, V. (1945). As We May Think. Atlantic Monthly, July. Retrieved July 1st, 2005 from


Chi, H., Pirollie, P., Pitkow, J. (2000). The Scent of a Site: A System for Analyzing and Predicting Information Scent, Usage and Usability of a Web Site. Xerox Palo Alto Research Centre: Palo Alto, USA


Chi, H., Rosien, A., Supattanasiri, G., Williams, A., Royer, C., Chow, C., Robles, E., Dalal, B., Chen, J., Cousins, S. (April 2003). Web usability: The bloodhound project: Automating discovery of web usability issues using the InfoScent™ simulator. Proceedings of the SIGCHI conference on Human factors in computing systems. p. 505-512. ACM Press: New York, USA.


Chong, S., Lui, K. (2000). The Social Aspects Neglected in e-Commerce. Ubiquity Published: ACM Press: New York, USA


Dourish, P., (2004). Where the action is: The foundation of Embodied Interaction. The MIT Press: Cambridge, Massachusetts, USA.

Engelbart, D. (1962). Letter to Vannevar Bush and Program On Human Effectiveness. Stanford University. Retrieved July 1st, 2005 from


Long, J., Dowell, J. (1989). Conceptions of the Discipline of HCI: Craft, Applied Science, and Engineering. In: Sutcliffe, A., Macauley, L. (ed.): Proceedings of the Fifth Conference of the British Computer Society Human Computer Interaction Specialist Group – People and Computers V. August 5-8, 1989, University of Nottingham, UK. p.9-32.


Engels, F. (1883). Nature of Dialectics.


Erricson, K., A., Simon, H., A. (May 1980). Verbal Reports as Data. Psychological Review. Volume 87, No 3, p.215-251. Published: American Psychological Association, Inc. Washington D.C, USA.


Evans, P., Wurster, T. (2000). Blown to Bits: How the new economics of information transforms strategy. Harvard Business School Press: Massachusetts, USA.


Ferguson, J. (1970). Socrates: A Source Book. p.35-50 Published: Open University, UK


Flanagan, M., Clanton, C., Marks, H., Murray, J., Arble, F. (1998). Interactive narrative: stepping into our own stories. Conference on Human Factors in Computing Systems: CHI 98 conference summary on Human factors in computing systems. p.88-89. ACM Press: New York, USA


Foley, P. (2001) Internet and e-commerce statistics. European Business Review. Volume 13, No. 2. Published: Emerald Fulltext.


Fowler, B. (1997). Pierre Bourdieu and Cultural Theory: Critical Investigations. Theory, Culture & Society, Nottingham Trent University. Published: SAGE Publications Ltd.


Glaser, B., G., Strauss, A., L. (1967). Discovery of Grounded Theory: Strategies for Qualitative Research. Published: Aldine De Gruyter, Hawthorn, NY, USA.


Green, T., R., G., Petre, M. (1996). Usability analysis of visual programming environments: a ‘cognitive dimensions’ framework. Visual Languages and Computing. 7, p.131-174.


Gulliksen, J. Boivie, I., Bannon, L., Oshlyansky, L., Thimbleby, H. (2005). Lost or Liberated without a Theory. Proceedings of the 19th British HCI group Annual Conference. p.299-301, 344. Published: The British Computer Society, UK.


Gulrajani, G. (August 2003). SHAWEL: Sharable and interactive Web-Lexicons. Max-Planck-Institute for Psycholinguistics. Max-Planck-Institute: Nijmegen, Germany.


Hegel, G., W., F. (1807). Phänomenologie des Geistes. Translated by. J. B. Baillie. Published: University of Idaho, Department of Philosophy

Hudson, W. (2005). The cost of More: Psychology of Choice in Interaction Design. ACM Press: Minnesota, USA.


Hutchins, E. (1995). Cognition in the Wild. The MIT Press: Cambridge, Massachusetts, USA.


Jennings, P. (2005). Constructed Narratives a Tangible Social Interface. Creativity and Cognition: Proceedings of the 5th conference on Creativity & cognition. Pages: 263 – 266. ACM Press: New York, USA.


Jul, S., and Furnas, G., W. (1997) Navigation in Electronic Worlds: A CHI 97 Workshop. SIGCHI Bulletin. Vol 29, No 4 October.


Kopytoff, I. (1988). The cultural biography of things: commoditization as process. In: Appadurai, A. (ed.) The social life of things. Commodities in cultural perspective. p.64–91. Cambridge University Press: Cambridge, UK.


Land, F (1999). A Historical Analysis of implementing IS at J. Lyons. In: Currie, W., Galliers, R. Rethinking Management Information Systems. p.310-325. Oxford University Press


Laudon, K. & Laudon, J. (2004). Management Information Systems. (8th ed.). Upper Saddle River, NJ: Pearson Education Inc.


Lazar, J. (2001). User-centred Web development. Jones and Bartlett Publishers: Sudbury, Massachusetts, USA.


McLoughlin, I. (1999). Creative Technological Change: The shaping of technology and organisations. Routledge: New York, USA.


Miller, D., (2001). The Dialectics of Shopping. The University of Chicago Press: Chicago, USA.


Mumford, E. (1999) Routinisation, Re-Engineering and socio-technical design. Changing ideas on the organisation of work. In: Currie, W., Galliers, R. (eds.) Rethinking Management Information Systems. Oxford University Press.


Myers, M.D. (1999), Investigating Information Systems with Ethnographic Research. Communications of the Association for Information Systems. Vol. 2, No. 23,


Nakhimovsky, A. (June 1988). Special issue on tense and aspect: Aspect, aspectual class, and the temporal structure of narrative. Computational Linguistics. Vol14 Issue 2 Pages: 29 – 43. MIT Press:   Cambridge, MA, USA


Naughton, J. (1999). A Brief History of the Future: The origins of the internet. 8th Edition. Orion Books Ltd: London, UK.


Nielsen, J. (August 2, 2004). Deceivingly Strong Information Scent Costs Sales. Alertbox Retrieved July 8, 2005 from


Nielsen, J. (March 19, 2000). Why You Only Need to Test With 5 Users. Alertbox Retrieved July 8, 2005 from


O’Cass, A., French, T. (2002) Web retailing adoption: exploring the nature of internet users Web retailing behaviour. Journal of Retailing and Consumer Services. Vol 10. Elsevier Science Ltd.


Ono, H., Zavodny, M. (March 2003). Gender and the Internet. Social Science Quarterly. Vol 84, No 1. Blackwell.


Office of National Statistics. (n.d) Data collection methodology: Optimising information gathered by surveys. Retrieved July 8, 2005 from


Oppenheim, A.N. (1966). Questionnaire Design, Interviewing and Attitude Measurement. p.210-228. Pinter Publishers: London, UK.


Perea y Monsuwe, T., Dellaert, B., G., C., Ruyter, K. (2004). What drives consumers to shop online? A literature review. International Journal of Service Industry Management. Vol 12, No 1, p.102-121. Emerald Fulltext.


Prabhaker, P., R. (2000). Who owns the online consumer? Journal of Consumer Marketing. Vol 17, No 2, p.158-171. Published: MCB University Press.


Pustejovsky, J. (December 1991). The Generative Lexicon. Computational Linguistics. Volume 17, No 4, p.409-441. The MIT Press: Cambridge, Massachusetts, USA.


Rosenfield, J. R. (November 2001). Lies damned lies, and internet statistics. Direct Marketing. Published: Garden City. Vol 64, No 7 p.61-64.


Smith, K., (2005). Converting Browsers to buyers: exploring what drives consumer choice in internet e-commerce, Interim Poster. Proceedings of the IADIS conference WWW/Internet 2005.


Sutcliffe, A., (2005). Grand Challenges for HCI. Proceedings of the 19th British HCI group Annual Conference. p.2. Published: The British Computer Society, UK.


Sutcliffe, A., Carroll, J., Young, R., Long, J. (1991). HCI theory on trial. Proceedings of the SIGCHI conference on Human factors in computing systems: Reaching through technology. Conference on Human Factors in Computing Systems. ACM Press: New York, NY, USA


Schwartz, B. (2004). The Paradox of Choice. Published: HarperColins, New York, USA.


Woolrych, A. and Cockton, G., (2001). “Why and When Five Test Users aren’t Enough,”. In: Proceedings of IHM-HCI 2001 Conference, eds. J. Vanderdonckt, A. Blandford, and A. Derycke, Cépadèus. Éditions: Toulouse, Vol 2, p.105-108.