Stripping Out Referral Spam From Past Reports

[kirmeliux]

Hi,

I'm looking to confirm the best approach for retroactively stripping away referral spam (free buttons, SEMalt, etc.). Now to be clear, I already have filters in place to ignore them from current stats, so moving forward I'm fine. However, I'd love to go back and check untainted stats.

I've setup segments using a regex to strip the root words away and it seems to be working. I have a regex setup to strip out things like: social-buttons|seoanalyses|copyrightclaims|classifiedads|jobsense|free-share-buttons|e-buyeasy|acrobats.hol|cheap-online|amezon|search-help|qut-smoking and so forth.

I've been going through my referral data, noticing obvious spam, and adding their domains to my segment. Is this the optimal way for me to get a clear, untainted view of my past stats?

[kirmeliux]

Sweet, glad to hear our filters will suffice. Thanks for the input, Daniel.

[Daniel_Marks]

Hey, no worries and you're right that your filters should block them as well. Using .htaccess would be just an additional defense mechanism but may not be necessary.

[kirmeliux]

Hi Daniel,

Thanks again for the response. What would be the difference in Analytics data between my filters and going straight to .htaccess? If the data is the same, is there an additional benefit to .htaccess?

For regular users, I'd suspect less bandwidth since they can't load my domain, but I don't think these bots actually load the page or visit.

[Daniel_Marks]

I would use your .htaccess file to block them with the following code (this would for example block referrals from semalt.com and semalt.com subdomains):

RewriteEngine On

Options +FollowSymlinks

RewriteCond %{HTTP_REFERER} ^https?://([^.]+.)*semalt.com\ [NC,OR]

RewriteRule .* – [F]

You can also use .htaccess to block IP addresses associated with the spammy sources.

edit: just saw your edit but hope this helps nevertheless!

[kirmeliux]

Hi Daniel,

Thanks for the additional tips. I do have the bot filtering feature enabled as another point of protection. I checked my referral exclusion list and apparently set this up about a year ago for the initial wave of referral bots I noticed. I didn't know it added them to direct.

The majority of my spam referral hosts have been added to regular filters. I think with the combination of my retroactive approach and new filters, I should have reliable data going forward.

[Daniel_Marks]

Hi there,

You’re on the right track and the best way to retroactively remove spammy sources is through report filters and advanced segments.

A couple other notes:

• A good way to spot spammy referrers is to sort by bounce rate and eliminate any with 100% bounce and over 10 sessions.
• Avoid using the “referral exclusion list” since this will just count spam traffic as direct traffic instead.
• You should also enable the GA ‘bot filtering’ feature under ‘Reporting view settings’ as seen here