SEO Lab: Testing Anchor Text Links


To be an effective and successful SEO, it's imperative that you're always testing, testing, testing. Why? Well, for starters, the variables that comprise and dictate the SERPs are totally out of our control. How do you know what truly works unless you test? Another reason is that's it's very unlikely you'll ever run into a search engineer from Google at a bar who's so hammered that he reveals Google's ranking algo to you.

But to be more specific, continuous testing is valuable and necessary because:

  • Algorithms are constantly being updated, so what worked yesterday may not work today
  • Without evidence or cold hard facts, you're operating blindly
  • Every test you conduct adds to your knowledge base
  • There's a lot of misinformation and SEO myths, so you can't believe in and rely on everything you read in forums

In sum, continuous testing and actual results are the only way you can support your contentions and actions.

SEO Test on Link Anchor Text

Now, there's always debate in SEO circles swirling around anchor text, like whether nofollow counts and does Google count different link text pointing to the same URL. With anchor text, I've pretty much relied on my own theoretical beliefs, but I've really never run any official tests myself. Now, I'm pretty sure I've seen other anchor text tests run before, but nothing recently. And since the algos and SERPs are always in flux (no more so than these past few weeks), I felt it was worth revisiting and testing myself. Plus, I've been challenged by other SEOs about my opinions on anchor text recently in a couple SEO communities. So rather than rely on assertions, I wanted to run my own tests so I could have some proof to stand behind.

So the two theories I wanted to test were that:

  • Google only counts the first instance of anchor text it encounters on a page to the same URL
  • Google does not count "nofollow" anchor text

To conduct my testing, I added a piece of text with unique anchors to our "Meet Our Bloggers" page, which is:

link text testing for seo

The first two text links (gootyackbootyack and misteropaki) both point to the same URL. The third text link (beenseesnoyl) points to a different page, but is tagged "nofollow." It took a week for the changes to index, and I waited two more weeks just in case anything was unresolved in the caching. Here are the anchor text test results.

The first instance of anchor text to the same URL indexed and is counted by Google:

anchor text tests for seo

The second instance of anchor text to the same URL was not counted by Google:

link anchor text testing

The "nofollow" link anchor text (which points to entirely different URL) also was not counted:

seo test on nofollow link anchor text

SEO Testing Results: Takeaways

These results verified for me that Google doesn't count "nowfollow" anchor text to a URL. Now, that doesn't mean that some amount of juice isn't leaking through, but that's another test for another day. Also, Google is only counting the first instance of anchor text it sees on a page (or more accurately the first instance it crawls in the code). Now, this doesn't mean you shouldn't use more than one link to an individual URL per page. Multiple links to the same URL are beneficial for usability or relevancy within context. But it does mean that if you do use more than one link on a page to the same URL, you should use your target anchor text in the first link.

Finally, as with anything related to SEO, these are my theories and my tests. And, as always, I highly recommend that everyone practicing SEO do their own testing and draw their own conclusions.

Find out how you're REALLY doing in AdWords!

Watch the video below on our Free AdWords Grader:

Visit the AdWords Grader.


SEO – google seo | Know Marketing Blog
Dec 10, 2009

[...] SEO Lab: Testing Anchor Text Links | WordStream7 hours ago by Ken Lyons  SEO Labs link anchor text testing: find out how Google treats anchor text for "nofollow" links and if multiple text links to the same URL are counted. – [...]

Jack Leblond
Dec 10, 2009


Just for fun, I used your three terms to search Bing. The only pages located are the two you use the words on and NOT the page they link to.

seo uzmanı
Jan 15, 2010

I’m reminded of the entirely different rendering engines in IE/Mac and IE/Win, and the problems those inconsistencies caused. And here we are again…

zaaylo creeper
Dec 10, 2009

interesting analysis and testing. i am wondering that with the search result we can get these much depth.

Dec 10, 2009

Interesting. Now what if you link to with "gootyackbootyack" and with "misteropaki", both from the same page. Would the same thing happen, or do you think that would be in the SERPs with "misteropaki"? I wonder if Google would treat it differently, since it isn't a link to the TLD, but to a page hosted on the same domain?

Dec 10, 2009

I did a few similar tests with the nofollow and alt texts on images. After more than 1 year and a few external nofollow links, the page that has been setup for the nofollow test is not even indexed. And the Alt text was passing linkreputation...

Dave Davis
Dec 10, 2009

Now to test if Google only uses the first anchor when you link to the page AND the root domain.

CJ Patel
Dec 11, 2009

Interesting test.. I'm curious how Google would react if you took this one step further - What if you kept the first link's anchor text as "gootyackbootyack", but gave it a title tag of "misteropaki"?

We have some links that have the anchor text purposely shortened for aestetics, but we use the full text in the title tag. Never really tested it like you have, but I think it'd be an interesting test.

Keep us posted on your findings!

Ken Lyons
Dec 11, 2009

everybody, thanks for commenting. lots of great ideas for future tests.
i'll begin running some of them now and report back in, say, three weeks.

Feb 23, 2010


Your experiment inspired me to do a follow up test, you can check it out here: