Monday, October 22, 2007

"How Many Site Hits? Depends Who’s Counting"

Read the full article by LOUISE STORY in The New York Times, October 22, 2007.

Organizations buying online advertising depend on services such as those offered by ComScore and Nielsen/NetRatings to measure traffic on websites. Organizations running those websites depend on their own measurement of site usage. According to this article, the internal measurements often exceed those of the external services.
"A main source of the discrepancies is over how to measure Internet use in the workplace. Nielsen/NetRatings and ComScore both track the Web use of representative panels of people, and use those traffic patterns to extrapolate the total number of visitors to a Web site. But online publishers say that their systems drastically undercount people who use the Web during work hours, particularly in offices where corporate software makes the wanderings invisible to the tracking systems. The issue is most pronounced at sites like and, which say that high numbers of people read them in the workplace. Mr. Spanfeller of says the ratings companies’ figures at times have 'no relationship to reality'; they in turn say that executives like Mr. Spanfeller are simply deceiving themselves about the popularity of their sites......the ratings panels still have problems. Condé Nast met with ComScore late last year to dispute the figures for 'They couldn’t really explain it, and they admitted as much,' he said. Condé Nast counts international readers and ComScore and neilsen/NetRatings do not, but that does not fully explain the discrepancies, Mr. McDonald said. He finds fault with the panels that both companies use, saying that they do not include enough of the wealthier people whom Condé Nast says frequent many of its sites.Complaints about the panels do not end there: some Web publishers say the panels lack representation from students on college campuses, Hispanics and other demographic groups. 'The results you get from a panel will reflect the choices you’ve made as you select the panel,' said Rob Grimshaw, advertising strategy director at The Financial Times. 'There’s a natural bias from panels. And on the Internet, we can have a genuinely more accurate system.'"

No comments: