Web Analytics Industry Confronts Cookie-Deletion Trend
The tempest caused by the Jupiter Research report showing widespread deletion of cookies was not surprising considering how important cookies are to tracking Web users. Without them, data on how many unique visitors a site was getting could be called into question, which in turn casts doubt on marketing moves and other decisions based on that data.
05/25/05 5:00 AM PT
Few analyst reports hit home as squarely as the recent Jupiter Research study on cookie deletion struck at the heart of the Web analytics industry.
Released in March, the report said 39 percent of Web users acknowledged deleting cookies -- code stored on a user's hard drive so that a site they click to in the future will recognize them as return visitors -- at least once a month, with smaller but still significant numbers eliminating them from their computers daily or weekly.
The rise of spyware -- which often identifies third-party cookies as problematic code -- new browsers that alert users to attempts to place cookies and increased awareness of how to manually delete cookies had given rise to the trend, Jupiter said. "Consumers tell us they are concerned about cookies and actually doing something about it," Jupiter Research analyst Eric Petersen, who wrote the report, said.
The ink had barely dried on Jupiter's report when the ripple effects began to be felt around the Web analytics industry and a debate ensued about how serious the problem really is.
The Atlas Institute released a counter study suggesting that Web users exaggerated how often they actually removed cookies, saying that survey respondents might have been giving what they thought was the right answer rather than the honest one. However, Nielsen//NetRatings released data showing the rate of deletion could be even higher than Jupiter estimated.
The tempest was not surprising considering how important cookies are to tracking Web users. Without them, data on how many unique visitors a site was getting could be called into question, which in turn casts doubt on marketing moves and other decisions based on that data.
Since then, experts have stepped forward with still more data to suggest that Web-tracking is probably suffering less than originally thought. The analytics industry, recognizing that the issue of cookie deletion needs to be addressed, has begun to roll out new solutions.
In a report released Monday, analytics firm WebTrends focuses attention on the issue of cookie rejection, in which users refuse to ever accept cookies when they first visit a site.
The rise of alternative browsers, notably Mozilla's Firefox, as well as more robust firewalls and enhanced security settings in Windows XP SP2, have given users more options when it comes to cookies, with pop up windows asking users if they want to allow the code to be placed on their machines.
WebTrends said that the percentage of users saying no to third-party cookies has risen four-fold in the past 18 months, from 2.8 percent in January of 2004 to 12.4 percent in April of this year. However, the firm also said that growth seems to have peaked. Some industries are being hit harder than others, with a nearly 17 percent refusal rate in retail, more than 15 percent in telecommunications and 12 percent among media firms.
"Cookie rejection is a more serious issue" than after-the-fact deletion, Jeff Seacrist, vice president of product marketing for WebTrends, told the E-Commerce Times. "If users are blocking cookies altogether, you are missing out on insight about that visit from the start."
Seacrist said the true impact of cookie denial or deletion is hard to judge and likely varies based on how the analytics in question are being done. Some vendors offer a backup system that uses IP addresses and browser information to categorize users, which preserves some data.
However, Seacrist said WebTrends was working on ways to solve the cookie-deletion issue before the Jupiter report hit. The firm rolled out version 7.5 of its hosted and software program Monday, with an upgrade that enables companies to use so-called first-party cookies instead of those installed by the analytics firm.
First-party cookies are not as likely to be tagged as spyware and are therefore far less likely to be removed or denied by the user. Seacrist said the WebTrends products will provide the same level of analytical data -- and in fact include a number of upgrades of its user interface designed to make it easier to sort, store and share key data.
In its report, the Atlas Institute said cookie-based measurement will "continue to provide reliable metrics and optimize online campaigns." Young-Bean Song, director of analytics, said many transactions are completed so quickly by Web shoppers that sites can gain a full spectrum of information on where they came from, how they navigate a site and so on, before the user ever has a chance to delete a cookie.
Still, he acknowledged that "technologies and methodologies need to continually evolve in order to keep pace with industry dynamics."
It's clear that firms are eager to offer alternatives. In addition to the latest WebTrends release, Affiliatetracking.com quickly followed the Jupiter Research report by "reintroducing" its "cookie-less" affiliate tracking software option, which monitors, without using third-party cookies, how traffic is directed to a site.
Another alternative being eyed is the use of Flash to work around cookie deletion with a process called Local Shared Objects.
However, Seacrist believes that anti-spyware tools will soon be updated to detect and remove that code as well. He said other solutions that disguise a third-party cookie as those installed by the primary site also come close to or even cross an ethical line in the industry.
Jupiter's Petersen said the solution shows that the Web industry is seeking creative alternatives to the cookie dilemma. For his part, the analyst believes the best long-term solutions will be those that are "user-friendly" and "describe to the consumer what the site is trying to do and why."