The direct and interactive effects of web site delay length, delay variability, and feedback on user attitudes, performance, and intentions

Although its popularity is widespread, the World-Wide Web is well known for one particular drawback: its slow speed when moving from one page to another. This experimental study examined three speed-related factors and their direct and interactive effects on user performance on a search task, user attitudes towards a web site, and ultimately, to behavioral intentions toward returning to the site. The factors under investigation were delay length, delay variability, and process feedback to the user. Delay length (2 sec or 10 sec) is the amount of time that elapses from the user's click on a hyperlink to a complete page load; delay variability (0% or 60%) captures the constancy of delay length between page loads; and feedback (present or absent) defines whether the user is presented with partial page content during the download wait. The study was conducted as a laboratory experiment using 152 student subjects in a completely balanced, full factorial design (2x2x2). It was found that the most significant factor affecting user attitudes and performance was delay length: high delays were associated with more negative attitudes and lower performance, however, higher search efficiency was observed. No effects of delay variability were detected, possibly due to insufficient manipulation of this factor. Results associated with the feedback factor support the hypotheses that feedback generates more positive attitudes and higher performance than situations without feedback. In addition, users reported lower delays when feedback was present, suggesting that objective delay length and feedback are interrelated in affecting perceived delay length. Results also supported that attitudes toward a Web site are strongly positively correlated with users' intentions to revisit the site. This study extends the body of research on system response time to the World Wide Web domain.