Death Match: Cookie/JS trackers VS. Log Analysis Software

And the winner is...

[martinibuster]

On examination of the stats from an expensive WebTrends log, I felt suspicious that certain SE search terms were not being accounted for, which could have explained a dramatic upswing in traffic. Which got me suspicious about Hitbox Pro.

Am I imagining this, or are JS/Cookie trackers in fact inferior for tracking SE referrals and keyword phrases used?

Is log analysis the way to go for tracking these?

[dingman]

Just speaking as a single consumer, it's highly likely that if you're trying to track my visit to your site with js and cookies, you won't get any info on me until and unless I decide to buy. I've got a nice top-level "setting" menu in my browser with little checkboxes so that I can turn those things on and off in two clicks, and I turn them on at the point that I try to buy but get a screen telling me that the site requires one or the other. I don't know how many users do that, but if it's many it could screw up such tracking systems.

[Mikkel Svendsen]

Both tracker scripts and logfiles track referrers fine - that is, IF there is any referral information. If the referrer information is filtered out or the agent simply do not pass it (like robots) there is no way for either the server monitor log or a tracker to record it.

Trackers and log files collect the data at two different points: Trackers count when the page with the script is loaded (no matter if it's loaded from your server, a browser or any other cache). Logfiles count when the webserver recieves a request for a file (which, in the case of cached pages never happen)

There is one more way to collect you visitor data: By network Packet Sniffing. It gives you access to more data than what you get from trackers and logfiles (like http stop requests, user connection speed and detection of how much of each file the user got loaded before the page was abandoned).

If you compare the numbers from these three methods they are all going to be different - even if they are varified to be correct.

[fathom]

Hand down no question. BOTH.

Each has merits the other doesn't, and both start out as extreme inexpensive.