I'm using google analytics and recently i've noticed some weird trends going on. I have a lot of visits that last mere seconds but mark several page views... more than a normal human can see in that range of time. A specific case is that the only visitor from Ireland i've had until now recorded 11 pageviews in a 3 second visit. Are these crawlers? Shouldn't google analytics filter those out?
2 Answers
Google Analytics won't pick up crawlers (since it uses javascript, and most crawlers do not run that).
Do you have iFrames or ajax or something? Perhaps a single page load is loading a few other files and those are being logged (like in an Iframe or something) Or are you calling the .trackPageview()
method anywhere in your javascript?
EDIT: Also, 11 pages in 3 seconds isn't that hard. If I held Command/Alt and clicked on all the links on this page, they would all open in tabs in the background. I could easily open up 20+ pages in 3 seconds

- 419
- 4
- 11
-
Hi @webdestroya. nope, I have no iframes, no call to .trackPageview(). Ajax is being used to handle contact form submission only and jquery is used for simple animations. Other than that the site is standard Html and CSS. – May 01 '10 at 23:21
-
Also, 11 pages in 3 seconds isn't that hard. If I held Command/Alt and clicked on all the links on this page, they would all open in tabs in the background. I could easily open up 20+ pages in 3 seconds. – Mitch Dempsey May 01 '10 at 23:22
i don't know what caused those 11 hits.
I do know that javascript-executing bots are a routine and well-documented issue for Web Analytics practitioners.
One significant source of js-executing bots (and the most well-documented) is load-testing tools such as Gomez, Keynote, et al. This isn't an incidental or de minimis source either--these bots can and often do generate a lot of hits in GA. (Several web analytics practitioners have remarked, in a forum threat i linked to below, that over the course of a few days, "Gomez Advisors" suddenly showed up at the top of their "Top Service Providers" Report, then disappeared from it a few days later.) This issue is discussed frequently in the Web Analytics Boards/forums and techniques to filter the data they generate are easy to find (e.g., create custom filters in GA using published list of IP addresses, based on service provider, etc.)
Given that it's easy to exclude this traffic using fine-grained filters in GA, my view is that it's a good idea just do do it, rather than first it's probably a good idea to do so, then check if you continue to see those patterns in your data.
(A few more sources that contain discussions of js-executing bots in the context of Web Analytics:

- 245
- 2
- 10