Referral spam is the name for traffic that has no user, and is not sent to your web analytics data warehouse in the conventional sense. They are sent with the expectation that the user will check their referrers and see an address. It doesn't take much investigation to realize that the users that were logged were low value traffic, they typically have a 100% bounce rate and never convert (because they are not human).

Adverse Effects of Referral Spam

Referral spam can inflate the volume of traffic visiting your site and dilute your performance metrics. Since hundreds or thousands of robot users can pour into your website, load one page, then disappear it can make any point they enter seem low value.

Below is an example of how much referral spam can distort your traffic over time. Sometimes it bombards your site consistently, sometimes it only spikes on your website periodically. 

null

Below is an example of how your performance metrics like bounce rate, session duration and conversion rate can be polluted by worthless spam.

Comparing Metrics

The above metrics are segmented by all users, and users that are actually human. In this case about twenty percent of all traffic is false, this huge chunk of traffic reduces the pages per session, avg. session duration, and bounce rate by 10%. The websites distribution of new and return users shifts from a majority new users to a greater population of return users, this is important because yu would market to new users in a completely different way than someone that is already familiar with your brand.

If the first segment was presented to a board of decision makers and the referral spam coincided with an offline media buy or a newspaper ad then someone could draw the conclusion that the new advertising effort had an immediate lift for the website traffic. Hopefully the decision makers would investigate further before betting the farm on a traffic bump populated by robots that have no money and consume valuable website bandwidth.

How to Find Referral Spam - Look at Hostnames

There are several indicators for referral spam that are available in Google Analytics, it is a good idea to check multiple to catch most of it. There is one very strong method that is fairly easy to implement and will block a vast majority of spam traffic.

Not every analytics solution stores the entire url of a page, what many users do not know is that hostname is parsed out and stored as a dimension in most of them. In Google Analytics you can bring this information in as a secondary dimension to source/medium then filter anything that doesn't match your actual domain. If you have referral spam the results will look something like the picture below:

list of referral spammers

 

notice that all of the spam sources have 100% bounce rates, one pageview per session and zero duration, these are big red flags for spam.

Removing Spam = Important Housekeeping

While removing referral spam is not as exciting as creating new measurement events or creating new reports based on user behavior, it is important to create a clear and descriptive statistical image of your website user. If you dont get the basics right then it can definitely create problems for your more advanced website analysis.

If you would like to learn more about SEO, web analytics, or user experience as it regards to your website and Truvisibility you can learn more at http://www.truvisibility.com/en/blog. If you are interested in trying Truvisibility out today simply go to truvisibility.com and press the “Get Started” button, it is free and easy.