Cookie-based metrics inflating website audience figures

Data released today by comScore shows the degree to which cookie-based measurement of unique visitors is skewed by users deleting cookies from their PC’s.

by Helen Leggatt

It is estimated that $19.5 billion will be spent on internet advertising in the U.S. this year, and that figure could nearly double to $36.5 million by 2011. With such large amounts being invested in online advertising, accurate measurement of a site’s reach and frequency is essential.

comScore’s latest report comes out of a December, 2006, study. The study was based on the analysis of 400,000 home PC’s in the U.S. It found that almost a third (31 percent) of users are “serial resetters”, clearing cookies from their PC’s once a month. These serial resetters can account for 10 or more “unique” visits over the course of a month, claims comScore’s president and chief executive, Dr Magid Abraham.

And “serial resetters” aren’t selective about which cookies they reset. First-party cookies, those left by a website when visited, aren’t as invasive or sinister as third-party cookies, left by tracking and ad networks, yet they are deleted at the same rate. As well as those who manually reset their cookies, there are those whose cookies are reset by security software, which also deletes spyware and other malicious programs.

The result of counting unique cookies, concludes comScore, is an exaggeration in size of a website’s audience by around 150 percent. Experts agree that the study further underscores the importance of panel-based measurement and linking visits to individuals.