• Posted on: 12 June 2007
  • By: Brian Gilley

Google AnalyticsWe've been looking closely lately at what we call 'referral scoring' by using Google Analytics and another referral tracking tool. Referral scoring is more or less the score Google gives to your website from any refering source - a .com, .org, and yes, those .edu's.

Over the past two months weve been running a little experiment. We have several websites with between 30-80 .edu links on each of them. We have not built any links to these websites over the past few months nor taken part in any other linking campaigns. Many of the .edu links to these websites are on the resource pages of .edu's, such as pages where a student might recommend or list certain websites for other students looking for external sources of information.

Our experiment was simple. We wanted to see if having our more valuable .edu links clicked on more regularly would help our rankings in Google. Google is certainly tracking refering websites and counting their 'vote' for your website within their Analytics and rankings. After all, by using Google Analytics you are essentially giving Google real-time data to help their ranking algorithms - especially for very time-sensitive information like celebrity gossip or news.

Within the experiment we ran we already knew that we did not receive many direct referrals (by referring to Google Analytics data) from the .edu links that linked to us for any given month. So we decided to increase the click throughs from those sites by clicking on the links ourselves using varying IPs or anonynous IPs so that it really looked like multiple computers and people in different locations around the world clicking on the links. We did this only for a percentage of the total number of referring clicks for each website to not make things look so obvious or skew previously recorded Analytics data for our websites.

Example: If we had 1000 referrals in total from external websites, we increased the clicks from the .edu links pointing to our sites by a maximum of 15% of the total, where previously they had only been about 1% of the total referring URLs. We only tried to increase the number of clicks by about 8-10% from the .edu pages/links for another website we owned. Again, we did not want to click from the .edu pages in large quantities so as not to raise any flags and this domain of ours only had about 1% referring clicks from the authority .edu's.

What we received at the end of our experiment was several nice jumps in traffic which has since stabilized at the increased traffic levels for the keyword phrases used within the .edu links. We hypothesized that our increased search engine positioning on the keyword terms used within .edu's was not due to the number of occurances they were used on varying websites, but was due to Google's scoring of the more prominent .edu referral source to better qualify our positioning for that keyword term.

There was a direct correlation between the anchor text being used within the .edu links and the new and improved rankings for that term in Google. So, in short, it seems obvious to us that the quality and authority nature of the .edu referring traffic combined with our purposely increasing the clicks from the .edu's worked together and helped improve our rankings for those keyword phases. The increase in rankings for those keywords did not seem to be affected by us increasing any linking campaigns and the anchor text links on those .edu's have been there for a minimum of one year for all of the sites we tested.

Lesson Learned: Ranking better for keyword phrases does not always seem to correspond to the number of links increasing or link campaigns in general, but rather the popularity of the referring domain (.edu's) and the increased number of referral clicks from those authority domains. We think that Google Analytics gives preference and ranks great referring URLs accordingly, so by simply increasing the clicks from those authority sites, we've increased our positioning for those keyword phrases.

If anyone has other experience with measuring referral tracking URLs in Google Analytics or if you've noticed similar patterns of rankings, please feel free to chime in below and let us know what you experienced.


Hey Brian, Really good post mate. Opening up the old can of worms about G using analytics and toolbar data to push up sites getting clicks and traffic. A while back when I first looked into this I was quite sceptical about them actually using it, but now seeing your test results I think I'll take another look. Mike

Wow Brian,

this is a goldnugget IMHO which confirms observations I had myself but could never fully verify it in the past...

I appreciate you sharing this !

best regards

"We think that Google Analytics gives preference and ranks great referring URLs accordingly"...
That's what you can read everywhere these days. Interessting point of view... thanks for posting some details on it.

What you're actually saying is that Google uses analytics to mertic the popularity of your website based on hits.

Google simply cant know without analytics who clicked on what anchor text to hit your pages.

So how about putting up 2 websites, one with and one without google analytics?

This is great stuff. There's no shortage of opinion pieces out there about what effects search engine rankings and why, it's mostly anecdotal. I'd love to see more quantitative experiments with data to back them up like this one. Thanks for sharing, and please post more like this!

Edu sites are more or less only used in the USA so I'd guess that the value of links and referal trafic differs depending on wich tdl likes to wich language. Well I guess - you test and share, thanks for that.

Best regards

Edu Backlinks related Anchor Tags (keywords) will suddenly get Rank by Search Engines. Really Nice Post.


Very interesting post Brian.

The key sentence in the above is surely: "...by simply increasing the clicks from those authority sites, we've increased our positioning for those keyword phrases." -- this opens a serious can of worms.

One would like to see more evidence and examples before jumping on board, so I'm going to run my own tests and report back with the findings.



Add new comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.