Research Methodology

---

The previous section presented the independent variables in the study. The dependent variable is website success. Various studies have investigated website success from different perspectives. Liu & Arnett (2000) investigate website success on the basis of outcome variables such as attractiveness, dependability, reliability, trustworthiness, meeting demand and pleasing customers. O'Keefe et al. (1998) investigated success on the basis of number of hits, visits and orders. This study investigates a specific success factor: the degree to which the website supports the marketing and sales of an organization. As such, it operationalizes on-line success in terms of the extent to which it is successful in supporting marketing and sales. Three measures along the customer-purchasing model were chosen-browsing, request for information and purchase. These measures are similar to those of O'Keefe et al. (1998). The reason that three measures were chosen was due to the fact than many customers initiate their contact with a company through a website but conclude the purchase via traditional sales channels: in-store, telephone, mail order. As such, a single measure such as website sales does not adequately capture the impact of a website. These measures were assessed as follows: (1) the number of hits per week to the website (browsing), (2) sales leads generated as a result of the website (request for information) and (3) revenue generated directly from the website (purchase).

These success measures were used to test two proposed hypotheses:

H1:

Organizational factors influence the success of the website. H1a: Early adopters are more successful H1b: Organizations with higher levels of IS/IT expertise are more successful H1c: Organizations with an e-business strategy are more successful H1d: Organizations who have a web champion are more successful H1e: Organizations who engage in planning are more successful H1f: Organizations who spend more on the development of their website are more successful

H2:

Retail specific factors influence the success of the website. H2a: The characteristics of the products being sold influence the degree of success achieved H2b: Organizations with larger on-line product catalogues are more successful. H2c: Organizations with an integrated "clicks and bricks" strategy are more successful.

The primary research was conducted by a postal survey. The selection of companies was done using purposive judgement sampling-companies registered to any of fifteen Irish web directories under the categories 'gifts' or 'arts & crafts'. Gifts are defined as items given as presents or articles suitable for presents. Arts and crafts are defined as products created using a particular skill and sold as gifts or products for personal use. This sampling technique identified 322 websites. In order to be eligible for inclusion in the study, the website had to have an on-line ordering capability. Each website was visited to ensure that the company sells online and this resulted in 280 suitable companies being identified (42 of the original 322 identified did not have on-line ordering capabilities).

The questionnaire was mailed to the 280 companies along with a personally addressed cover letter to the relevant person in the company responsible for the website. Of the 280 questionnaires mailed, 122 were returned yielding a response rate of 43.5%. Of these 122, 44 were returned after a second mailing of the questionnaire was sent to companies who had not responded after two weeks.

Non-response bias was assessed by treating responses received after the deadline given (two weeks after the first mailing of the questionnaire) as being representative of non-respondents. Chi-Square significance tests were used to test for significant differences between the first respondents and the late respondents in the study. The early and late respondents did not differ significantly (at p < .05) on the measures of success (hits, revenues and sales leads) and the other variables tested including e-business strategy, web champion, IS/IT expertise and product characteristics.

The results of the questionnaires were analyzed using SPSS version 10.1. The responses were first analyzed using statistical measures such as means, standard deviations and frequencies. For more advanced statistics, the Chi-square test for independent samples was used. The chosen level of confidence for the research was α = .05.

---