There’s Something Rotten In… The Data
Over the past couple weeks, our digital marketing department started to notice a big increase in overall traffic. Assuming one of our SEO strategies was rapidly paying off, we started digging into site data to capitalize. OH…how sad we were to discover we were being spammed in new and exciting ways.
Referral Spam (Anagrams to: Rear Slap rrmf!)
This new kind of spam has a name, and it shall be known as referral spam, or the anagram above… possibly. In short:
Shady people have devised an even shadier way to make your Google Analytics tracking code report page views without even visiting your site.
Yup, the only information they need to start spamming is your Google Analytics Tracking/Property ID. Most of the time, they don’t even have your specific ID, they just guess and run with it.
Does It Hurt My Site?
While your site has NOT been hacked, referral spam can cause a host of detrimental effects. Mainly:
- Over Reporting: We follow the correlation between our SEO efforts and overall site traffic closely. When there is a bump in traffic we want to know what to attribute it to and then attempt to replicate that success across all clients. Referral spam gives us a bloated picture of these efforts, sometimes reporting traffic up to 20% higher than it legitimately was. Supremely irritating and bad for ongoing SEO strategies.
- Server Overload: An overloaded server means slower load times. Slower load times lead to higher bounce rates. Higher bounce rates lead to lower SERP rankings. It’s a vicious cycle. With SEO being one of our Digital Marketing Team’s many hats, we’ll take a leg up in ranking factors wherever we can get it.
How Do I Fix It?
Well, it depends. A Google search will yield various and sundry tactics. In our case we wanted to not only focus on finding a broad way to limit referral spam without sacrificing potential real traffic, but also keep the solution safe enough to prevent long term damage from potential mistakes. Our answer came in the form of a question, Jeopardy style, from Andreas Veithen’s blog:
Use Filtered Reporting In Google Analytics
He postured that, since most of these shady people (with pencil thin mustaches… probably) are running requests for your homepage “/” but not actually visiting, why not filter out all homepage requests then reintegrate the people that actually visit? It’s a good question. To which we answered: Yes, we would very much like to do that.
Before we get started, I want to mention that you should be following Google Analytics best practices. This means having a master view (that filters out office traffic) and a testing view (to make sure your new filters work correctly, among other things.) If you don’t have a testing view, set one up… now!
First, create this filter to remove all traffic requesting the homepage:
Second, create this filter to make sure folks that actually make it to the homepage get reported:
Third, give yourself a high five or hearty pat on the back.
You work hard and deserve to be recognized, even if only by yourself ;).
Oh! And the order matters. Click Assign Filter Order so your filters are arranged like this when you’re finished:
Google recently started offering a Bot Filtering option in the view section of GA accounts:
FYI, we tested these filters before switching it on. So our filter plan definitely works… but it couldn’t hurt to check this box as well. They give a pretty vague answer as to what they’re doing to keep the bots out, but it’s a relatively new problem and they’re a pretty big organization. It’ll take some time to get a solution fine tuned, but we trust ’em to get it right eventually!
Does my take not “do it” for ya? Try checking out the resources I used to build this post: