Careful with Google Analytics filters

I made a change to the filters attached to our profiles in Google Analytics (GA) so that we would capture only relevant traffic. Namely, I was attempting to avoid capturing traffic to our development servers. Unfortunately, instead of a domain-based filter I created one that resulted in GA capturing zero traffic. Fail.

The problem stems from my attempt to use a predefined filter to block out the unwanted traffic. I tried to set up a “include only traffic from the domains that include” filter, thinking this was a filter on the server’s domain. What this filter actually checks against is the visitor’s domain. I guess a little more attention to the “from/to” part of the filter would have made this apparent, but Google doesn’t provide much documentation (or even better, examples) on their filters.

After a few days I noticed the sudden lack of data and realized my error. Turns out what I really wanted was a custom filter with the following settings:

  • filter type: include
  • fiter field: hostname
  • filter pattern: regex for the domain (e.g. www.project2061.org
  • case sensitive: no

Even after looking at this again, it’s still not clear to me why these two are different. The descriptive language is almost identical. Google needs to do some work here.

Of course, there are also other solutions to this problem, listed below:

  • Use an include to pull in the GA code and only populate the file on the servers where tracking should occur. The only significant problem here would be forgetting to set up the include.
  • Create advanced segments in GA that filters out traffic to anything other than the production domains. This would have the benefit of being able to track both production and development server use in the same report. But if there are any significant difference between the two sites then this wouldn’t be all that helpful.

References: