Overcounting Caused by Meta Refresh URL Redirects?
WebSideStory, Inc. recently announced internal findings that show ?meta refresh? URL redirects can significantly inflate Web site traffic figures. A meta refresh redirect is a common method used by Web site developers to automatically send a visitor from one URL to another. Because meta refresh redirects are difficult and time-consuming to filter out of Web server log file reports, Web operators and advertisers may get inflated traffic figures from companies using log file software. According to WebSideStory, which pioneered the outsourced method of visitor analysis more than five years ago, the use of meta refresh redirects has led to as much as 30 percent inflation of overall Web site traffic in cases it has examined. Currently, no filter exists to distinguish meta refresh pages from actual page views in log file reports, so they must be filtered manually for accuracy.
?Although this is not as familiar a problem for advertisers as robots and spiders, there can be a much greater potential for inaccuracy,? said Dick Bennett, senior vice president of audit services for ABC Interactive, an organization established by the Audit Bureau of Circulations to provide independent third party verification of online activity for advertisers. ?This is definitely an issue that needs to be addressed so that advertisers and other users of Web site activity metrics can feel confident that they are getting correct information. One of the ways to avoid overcounts from meta refresh redirects is by using an outsourced Web visitor analysis solution.?
There are two common ways to automatically redirect users from one URL to another. One way is to configure the Web server to do a ?302 direct?, which automatically maps one page to another when the initial page is requested. Although 302 redirects can be filtered out of log file reports because the result code (302) is stored with the recorded hit, many companies do not use 302 redirects because of the IT resources required to maintain the server software. The second way is to set up a meta refresh, which can easily be done by any HTML content creator. When the initial page is requested, the user is automatically redirected to the new page, and the server log files record both as normal page views, even though the visitor only requested and viewed one page. However, since both page views are recorded identically in the log files, meta refresh pages cannot be distinguished through any type of automatic log file filter. This incorrectly doubles the traffic recorded to a particular page.
Outsourced visitor analysis services such as WebSideStory?s HitBox Enterprise ( www.hitboxenterprise.com) typically ignore meta refresh pages and thereby provide more accurate reporting. These hosted analysis services collect data directly from the browsers of visitors to a site via ?tags? or ?beacons? placed in the content of the pages to be monitored. Unlike software methods, which record all requests made to the Web server, and must be filtered afterwards, the use of beacons give businesses total discretion in deciding which pages they want monitored and analyzed. As a result, they can easily avoid counting meta refresh pages, ensuring that all recorded page requests represent actual page views.