Website benchmarking: a tropical tourism analysis

Chapter 1: As more companies move on-line and establish their presence on the World Wide Web demand exists for an efficient benchmarking tool that can suitably assess their measurable on-line features and/or can allow comparison with website competitors. Website benchmarking approaches within organisations vary, and a majority of researchers, in-part capture website aspects using subjective techniques to obtain a result.\ud \ud This study finds website benchmarking approaches in the literature are generally time-consuming, survey-based, with little agreement on what and how to measure website components. Currently no broad-based, systematic, measurement method or website benchmarking theory exists.\ud \ud This research aims to rectify the above problem, by developing a theoretical approach to website benchmarking, and by establishing a website analysis method (WAM). This approach is literature supported, hierarchical, objective, and facilitates the evaluation of a website against a universal set of website component measures.\ud \ud Design Science Research Methodology (DSRM) is employed at the research level. In the development of 10 published (and one submitted) papers, and in the layout and development of this thesis. The DSRM contains 6 activities: problem identification; solution objective; artefact design and development; artefact demonstration; artefact evaluation; and communication.\ud \ud Chapter 2: This chapter consists of paper one, which identifies the DSRM problem. Current website benchmarking approaches in the literature are sourced, classified by type, and from both a customer perspective and a business perspective are type-collated as a framework. When combined, these approaches deliver a multi-level, website benchmarking approach, offering a broad-scale analysis pathway, capable of delivering at level website benchmarking. However, individually none of these approaches delivers a comprehensive approach to website benchmarking.\ud \ud Chapter 3: This identifies a solution for the above DSRM problem across papers two, three, and four. In paper two, components identified in the literature are built into a three domain framework to test a Likert survey approach against a dichotomous survey approach. Although results demonstrate the dichotomous survey is much quicker and delivers more accurate results than the Likert survey, respondents still make errors. Paper three now looks at the theory and mathematics behind website benchmarking.\ud \ud As no definitive theory of website benchmarking exists, this section of chapter 2 proposes the initial steps towards this theory's development. The mathematics behind website measurement are built, via a proposed three level domain, group and item set-theory approach. This offers a multi-level, interpretable measurement solution to managing websites, while shedding light on developing a website benchmarking theory. Paper four delves deeper into the areas of subjectivity and objectivity in relation to website benchmarking.\ud \ud Subjective techniques, introduce the vagaries of human opinion, along with a variety of acknowledged limitations, such as survey length restrictions. Hence, participant interpretation of various questions related to a website can create bias. To eliminate this subjectivity I introduce a quantitative website benchmarking approach offering objective website ratings at the website, domain and function levels, and suggest at these website levels comparisons between websites and competitors can be useful.\ud \ud Therefore the solution to the problem is development, of an objective website benchmarking approach. However before a website benchmarking approach can be fully developed a definitive theory is required. The development of this theory and the design and development of the DSRM artefact follow in chapter 4, in paper 5.\ud \ud Chapter 4: The artefact design and development stage of the DSRM begins with confirmation that literature-identified website benchmarking approaches have altered little since my first review of the literature. I establish a theoretical approach to website benchmarking and deliver a comprehensive website analysis method (WAM), the DSRM artefact.\ud \ud This approach employs rough-cut set-theory, and mutual-exclusivity of 230 components allowing causal-summing of typological website components into the objective WAM artefact. These typological website components deliver interpretable website benchmarking scores at website, domain (aesthetic, marketing, and technical), and function levels.\ud \ud WAM is not static allowing new/emergent components to be included (and redundant components removed) as upgrades to WAM's three domains, and 28 functions. Such modifications help keep the latest benchmarking comparisons (and/or website upgrades) optimised. As this DSRM study employs a dichotomous present/absent component approach, the WAM artefact's components can be software programmed, and merged at three different levels, delivering a useful website benchmarking tool for corporates. Demonstration of the artefact is the next step in the DSRM, this follows in papers six, seven, eight, and nine comprising chapter 5.\ud \ud Chapter 5: In paper six different website components are tested in a variety of ways to ensure validity. First, I consider how website quality is measured in the literature. I typologically classify 27 empirical studies into 6 functions. This approach captures the broad divisions of website quality as a tabulation from which a comprehensive website model is developed. The results suggest breadth is not adequately captured when considering a business' websites quality. Paper seven goes further, looking at location of components on websites.\ud \ud Here, the location-specific placement of 22 literature-identified homepage web objects is analysed within 5 services-industries websites, and across 10 countries. This study applies content analysis with a 3D fine-grid approach and considers if web object placement corresponds with consumer mental positioning models. This approach differs from previous studies, where consumers subjectively located various objects on a 2D grid representation of a homepage. Results show regardless of country, several web objects have industry-specific consistency in their actual grid placement and a generic placement of five web objects exists. The placement of some web objects also corresponds with consumer mental models, therefore meeting their expectations.\ud \ud Paper six has shown website components can be captured and typologically classified, paper seven shows some have specific locations on websites. Now papers eight and nine demonstrate the effects on visitor traffic when new components are added to a website. A tourism informational website is progressively developed, through static, then interactive, and then into dynamic formats.\ud \ud Here, six stages of increased website richness shows differential offerings can contribute to growth in first-time visiting-consumer website traffic. When sequenced, and released in close succession, Facebook posts boost first-time visiting-consumer website traffic. Return-visitor traffic into a website arises when the consumer is in pursuit of website offerings that are sufficiently in line with their motivations and/or with their desired outcomes. At each website change, the levels of component presence, interactivity and/or external post are altered to determine their effect on website traffic. By including, and maintaining, currency across all six stages of website development, successful consumer-targeted destination tourism websites can be delivered.\ud \ud Chapter 5's papers confirm a computer programmable software solution should be considered. Therefore in papers 10 and 11 in chapter 6 the evaluation of the success of WAM is discussed.\ud \ud Chapter 6: Here, in paper 10 the pilot study employs content analysis and the presence/absence of 47 website quality components on 30 tropical tourism websites. Results demonstrate the majority of websites are lacking in e-service quality and security quality. Several websites do not operate in all versions of internet explorer, while others are not mobile device ready. Hence all tropical tourism websites in this study fail to reach high ratings for website quality. These results provide the impetus to increase the size of the study and to use the hierarchical WAM artefact.\ud \ud The main study, paper 11, evaluates WAM using content analysis and the presence/absence of components. An abridged WAM with 77 (easily viewable) components from 28 functions, across 3 domains of a website is employed to test 280 tropical tourism websites from 4 tropical continental areas. A focus group then ranks 20 of the websites from best to worst from their perspective. They then try to identify the presence/absence of 12 components on the same websites.\ud \ud Results demonstrate website scores at level are comparable, and there are significant differences in website scores, domain score, and function scores when related to continental area. Focus group results indicate it is much easier for participants to distinguish between poorly scoring websites than it is to distinguish between higher scoring websites.\ud \ud Although focus group participants located a limited number of components on a website they did not always locate all components. Therefore for website owners/managers to be able to easily, effectively, accurately, and cost effectively assess their websites at a point-in-time WAM is the logical website benchmarking approach.\ud \ud This study has identified a problem in the literature and the motivation to solve this problem. The objective for a solution to the problem has been identified, and the artefact to solve the problem has been designed and developed. The artefact has been demonstrated and the results evaluated.\ud \ud Chapter 7: This chapter communicates the findings of this website benchmarking thesis. It encapsulates thesis outcomes delivered across 10 published papers and 1 submitted paper. It delivers theoretical, empirical, and real world contributions for website benchmarking. It also provides key limitations, recommendations, and future research opportunities