Companies advertising on the World Wide Web have no accurate way of measuring who, or even how many people see their ads according to a study by the University of Southern California, Los Angeles. Researchers at USC's Marshall School of Business have found that current methods for determining the number of "hits" an Internet ad receives can be wildly inaccurate.
The researchers studied thousands of hits on five major Web sites (news, education, entertainment, and two information databases) and tabulated those hits on the basis of Internet protocol addresses alone. They then tabulated the hits more accurately by imposing mandatory log-ins and other identification methods on the same visitors to the sites -- measures that most Web sites are reluctant to use for fear of alienating their visitors. The resulting disparities were striking.
Using Internet protocol addresses alone as a means of identifying hits led to a 39% underestimation of visits, a 64% overestimation of the number of pages seen by each visitor, and a 79% overestimation of the time spent on each visit. "While print and broadcast media have long-established standards of advertising effectiveness, such variables as reach, frequency, and gross rating points are difficult to measure on the Internet," says study co-author Fred S. Zufryden, of the University's Marshall School.
One problem is that users often lack unique online identities. So when marketers try to measure user requests, or "hits," for a particular Web page, they can't tell whether one person requests the page five times or five people request it once.
In addition, caching prevents market researchers from measuring how often a Web page is seen during a user session. Once a surfer clicks on a page, his (or her) computer temporarily stores the page for the duration of the session, in case he wants to see it again. This caching feature prevents researchers from knowing how many times a user is exposed to the ad.