Filtering Internal Traffic
It's a matter of excluding the outliers from your survey.
I'm surprised that there is ever any question about the value of filtering out internal traffic from web analytics reports. Imagine a Gallup poll that included responses from Gallup employees. The credibility of that poll would be dubious. Not because all Gallup employees would respond differently than the population at large, but because some of them might--and we wouldn't know which ones or whether they did. It would just be untrustworthy data. Because we wouldn't know to what extent it was impure, we wouldn't even be able to make good guesses about how to interpret that data.
Personally, I think the example above is clever enough that every skeptic should now be convinced that they need to filter out their internal traffic. But I'll give some more specific reasons to convince the skeptics.
First, just to clarify, by "internal traffic", I mean traffic from company employees, developers, marketers, stalkers, fan clubs, etc. Anybody related to the company should be filtered out.
Why Filter Internal Traffic?
What are some other reasons for filtering out internal traffic? Let's start by reviewing the reason we are using web analytics in the first place. We want to see trends and correlations. We want to spot weak points and strengths. In short, we want to know what to do to improve conversion and visitor experience.
It's this last point that is most relevant here: we need to create a to do list. What do we change to make things better?
I should underscore the fact that web analytics tools are lousy record keepers. If your software were an accountant, you would have fired it long ago. The reason you have a web analytics tool is not to keep a record of every order placed or how many times a video was viewed. Rather, it is to see what actions prevent users from doing those blessed activities, or it is to know what motivates them to do it.
Looking at reports that include the possibility of irrelevant traffic is just silly. They're called outliers. And here are some reasons that I wouldn't touch them with a ten foot pole--unless I was using it to hit them, those moochers:
- They interact with your site differently than other visitors. They look for different things. They probably use different buttons and searches.
- They don't use the same traffic sources to reach your site. In all likelihood they have bookmarked the site or they just type the URL into their address bar. They probably have not come to the site because of the new cpc campaign you started or because that new advertisement was so creative. In fact, they probably start their visits deeper into the site than a typical visitor would know to. In any case, their motivations for coming to the site in the first place are different.
- If you have some new or redesigned area of your website, they probably know about it first. If they are developers, they are going to those areas all the time to make sure they are working properly. Maybe they're going to get their employee discounts. Maybe they are going to see whether you used their suggestion for a background image.
- Internal traffic is more patient with your site than a typical visitor will be. They already know that what they want is there. So, if a page takes too long to load, if the video is broken, if the shopping cart deletes all their items, if the site puts a virus on their computer, they will still come back and try again. Typical visitors are not so merciful.
- They come to your site more frequently and for longer durations. Why aren't they working? I don't know, but they aren't. They're playing around on your website instead.
For all these reasons and more, we should purge the earth of internal traffic!
How to Filter Internal Traffic
Now that you feel silly even thinking about keeping the internal traffic in your reports, we should discuss methodology. When it comes down to it, there are only a few ways to do it.
Typically, we can filter the traffic by IP or network name. In this case, we would filter by the IP visible to the world. This might be the IP address of your firewall. It's the IP address that shows up when you visit any number of pages like www.whatsmyip.com.
In a case where we don't know or can't consistently pin down the IP or network name, we need to cookie internal traffic with a user-defined variable and then filter based on that cookie.
Creating the filter is easy. Just go to the filters list, click to add a new one, specify that it's an exclude filter and then put in either the IP addresses, network name (ISP organization, network location) or the user-defined variable.
