Measuring Factors that Influence the Success of Internet Commerce

Abstract

Efforts to develop measures of Internet commerce success have been hampered by (1) the rapid development and use of Internet technologies and (2) the lack of conceptual bases necessary to develop success measures. In a recent study, Keeney (1999) proposed two sets of variables labeled as means objectives and fundamental objectives that influence Internet shopping. Means objectives, he argues, help businesses achieve what is important for their customers—fundamental objectives. Based on Keeney's work, this paper describes the development of two instruments that together measure the factors that influence Internet commerce success. One instrument measures the means objectives that influence online purchase (e.g., Internet vendor trust) and the other measures the fundamental objectives that customers perceive to be important for Internet commerce (e.g., Internet product value). In phase one of the instrument development process, we generated 125 items for means and fundamental objectives. Using a sample of 199 responses by individuals with Internet shopping experience, these constructs were examined for reliability and validity. The Phase 1 results suggested a 4-factor, 21-item instrument to measure means objectives and a 4-factor, 17-item instrument to measure fundamental objectives. In Phase 2 of the instrument development process, we gathered a sample of 421 responses to further explore the 2 instruments. With minor modifications, the Phase 2 data support the 2 models. The Phase 2 results suggest a 5-factor, 21-item instrument that measures means objectives in terms of Internet product choice, online payment, Internet vendor trust, shopping travel, and Internet shipping errors. Results also suggest a 4-factor, 16-item instrument that measures fundamental objectives in terms of Internet shopping convenience, Internet ecology, Internet customer relation, and Internet product value. Evidence of reliability and discriminant, construct, and content validity is presented for the hypothesized measurement models. The paper concludes with discussions on the usefulness of these measures and future research ideas.