Benchmarking web site functions

Purpose – To measure, classify and compare web site functions' development. Design/methodology/approach – The objectives were achieved by developing a methodology to measure, classify and compare web site functions development. The measurement was based on the presence (or absence) of 91 web site components. The classification was achieved using an applied correspondence analysis. The comparison was performed with respect to 4,485 company web sites from two provinces in Canada. A formal procedure involving 50 assistants was developed to collect data over 5,000 sites within a month period and a conceptual model was developed to interpret results. Findings – Findings show that web site functions development could be described on a three dimensional space, the three axes corresponding to directions of development. The development status of the whole sample appeared as a cone in which five classes (or categories) of web sites could be identified and described with respect to their development profile. The development of 18 web site functions was also benchmarked with respect to observations within a class and with respect to some other characteristics like the industrial sector and the province where the company was located. Research limitations/implications – Web site functions development was evaluated only with respect to components accessible to the general public. Development of intranet and extranet were not evaluated and hence taken into account for overall web sites development. Practical implications – This research project of web site functions development was part of a more comprehensive project aimed at evaluating and documenting the impact of using a web site on business processes. So, combining the results of the two projects, allowed the authors to advance suggestions of how web sites should be developed to generate value for companies. Originality/value – The approach to web site functions development evaluation is original. The methodology could be replicated anytime. The typology that emerged from the analysis is unique. Moreover, given the large sample of 4,485 web sites, results are statistically valid.