Within the past few months there have been two independent developments that may have a profound impact on measuring site traffic.
The most recent occurred on March 18, when the group
product manager for Google Analytics announced plans for the development of "a global browser based plug-in to allow users to opt out of being tracked by Google Analytics." The announcement and
(mostly) scathing comments are at http://analytics.blogspot.com/2010/03/more-choice-for-users-browser-based-opt.html.
There is no
shortage of speculation as to the motivation behind this, ranging from "who cares, it's a non-issue" to conspiracy theories that boggle the mind. Surely user privacy has been a much discussed topic
recently and perhaps this is Google's way to get out in front of the issue and take a leadership role.
The main point for analytics professionals is that GA traffic activity data will be
incomplete if/when this plan comes to fruition. To what extent, no one knows -- but it presents a problem for making business decisions that rely on hard data. This could lead sites to look for
another analytics solution. But what if the other analytic tools follow suit? Does an already imperfect measurement process become even more imperfect? And do those who rely on traffic data for
internal and external business purposes simply accept the imperfections and hope for the best?
The second development also seems driven by privacy and involves Flash cookies, aka local shared
pbjects (LSOs), to track unique users, a key metric used in the buying and selling of advertising. Without getting into the technical details, Flash cookies are more sophisticated than HTTP cookies,
are more difficult to detect and delete, and can "re-spawn" even if one does delete them. Further, the existence of Flash cookies is not very widely known.
Here at BPA we had been hearing
more and more about this technology and commissioned Eric Peterson, CEO, Web Analytics Demystified, to research and posit a point of view. The resulting white paper, (http://www.bpaww.com/Bpaww_com/HTML/flashcookieswp/index.htm ) released in January 2010, points out the pros and cons of using
Flash cookies and makes recommendations as to how to use this technology as a best business practice. The paper struck a nerve within the analytics community and there have been numerous posts about
how Flash cookies are used and misused. Further, Adobe issued a publicly available letter (http://www.ftc.gov/os/comments/privacyroundtable/544506-00085.pdf ) to the Federal Trade Commission on its support of user
privacy and the responsible use of its technology.
Which brings us back to measurement. The April 14th OnlineMediaDaily contains a story with the headline, "Report: More Users Delete Flash Cookies." That may bring a smile to privacy advocates -- but it has to send
a shiver up the collective spine of analytics pros. The article cites a 7% deletion rate over 30 days, which is more than double the rate from just 10 months ago. Whether the rate grows or levels off,
the accuracy of determining unique users is going to be off by a lot or a little. But it is going to be off.
So now we have two recent developments that can result in traffic counts being
either lower or higher. In a perfect world, they would cancel each other out and we'd arrive at a reliable number. That's not likely to happen. So what to do?
I would start with
standardization of metrics and measurement. At the Feburary OMMA Metrics & Measurement event in New York, a panel I was on got around to discussing "what's a big problem today" and "what's next in
measurement." On the issue of big problem,s everyone seemed to agree the lack of standardization in metrics definitions and measurement processes are huge hurdles to overcome. And while everyone
seemed to agree something should be done, I couldn't help but wonder who is going to lead such an initiative.
In fact some leadership has already been exhibited. The Web Analytics Association
has done terrific work defining metrics and the Interactive Advertising Bureau has done likewise issuing guidelines for measurement. This is a sizable start, but more needs to be done as
metric/measurement conditions are changing rapidly as illustrated by the two previous examples. More than any other media platform, online requires constant vigilance to stay on top of its rapid
evolution -- and may even need a small army of stakeholders to do so.
Any volunteers?