Tuesday, October 11, 2011

Blogger Magic - Third Party Visitor Logs / Meters, And Referer Spam

Blogger blog owners have been suffering from the onslaught of referer spam, in their Stats logs, for over 6 months now. Some blog owners, who use third party visitor logs / meters like FlagCounter, Sitemeter, and StatCounter, have started to wonder.
How do SiteMeter and StatCounter manage to filter out the referer spam, yet Google cannot do anything about it?
This would be quite a magic trick indeed - if this was happening.

When you setup a third party visitor information product, like FlagCounter, Sitemeter, and StatCounter, the installation process involves addition of a JavaScript code snippet, as part of the blog template. Any time a page in the blog is loaded, by a blog visitor, the JavaScript code references the FlagCounter, Sitemeter, or StatCounter server, and adds a visitor record describing the pageview.

That's such a simple way to gather information about your visitors - and for its simplicity, it is unreliable, in various ways.
  • Your visitors who use computers that filter content from domains - like BlogSpot, FlagCounter, Sitemeter, StatCounter, or possibly your non BlogSpot custom domain - won't be counted, reliably.
  • Your visitors who share a caching proxy server won't be counted.
  • Depending upon where in the page, the JavaScript code is added, your visitors may or may not be counted.

When Blogger / Google designed the Stats visitor information accessory, they used another technique for counting visitor activity. Blogger, unlike FlagCounter, Sitemeter, and StatCounter, has access to the server activity logs. When Stats was activated, for the Blogger blogosphere, they activated it on a blog by blog basis - and without requiring any updates to the blog code.

Every Blogger blog was given Stats access, and the statistics initially provided preceded actual Stats availability, because nobody had to install any accessory code to the blogs. That is because Blogger uses the actual server activity logs - and does not require add-on accessory code.

Unfortunately, server activity logs only record server activity - and this is how referer spam works - and why it is so hard for Google to block it. Referer spam is simply a normal access request from the spammer, which generates a server access record, and a Stats pageview. The spammer simply goes away after generating the initial access request. The Blogger server, as with any non Blogger server, has no way to detect when the spammer drops the connection and goes away.

Since third party products like FlagCounter, Sitemeter, and StatCounter use add-on code, they are not susceptible to referer spam techniques. Only products which work from the server activity logs are susceptible. FlagCounter, Sitemeter, and StatCounter, and other third party visitor information accessories, have no need to filter bogus activity.

And this is why I have stated that the only way that referer spam will ever go away is for us to make it unprofitable for the spammers, by not clicking on the links in the Stats logs.

>> Top

5 comments:

James said...

It's frustrating as Google Analytics is able to filter out the spammers yet their own stats page isn't!

tv-theme-songs said...

I am also being bombarded with referrer spam links.

Κωνσταντάς Χατζηκωσταράτσος said...

Well, how many spam referer adresses there are out there? Why Google can't have a list of them, and remove any reference of them from the statistics that go to each blog?
-.

Chuck Croll said...

There are an infinite number of refer addresses out there. Any address could be a spam address.

http://blogging.nitecruzr.net/2012/04/referer-spam-cannot-be-blocked.html

Elizabeth Anderson said...

I looked at the referring site to see who it was and haven't done so since but have been getting spam referrals for months :P