Our blog has moved!

You should be automatically redirected. If not, visit
http://www.masterlink.com/blog/
and update your bookmarks.

November 28, 2006

"nofollow" - Does it Really Work Like Google Claims?

I am not one to usually try to stir up debate or call the search engines out for misleading people, but I still can't figure out what a "nofollow" tag REALLY does. That being said, I know what it SUPPOSEDLY does, but there is data to contradict this claim. After reviewing an experiment I did, it seems that Yahoo and MSN are obeying the rules, but Google, who was first to encourage the use of this tag is not doing what people think it is.

For those of you who do not know what the "nofollow" tag does, here is a quote from Search Engine Watch:

"If Google sees nofollow as part of a link, it will:

    1. NOT follow through to that page.
    2. NOT count the link in calculating PageRank link popularity scores.
    3. NOT count the anchor text in determining what terms the page being linked to is relevant for."
Now, I did a little experiment that anyone can do (some may call it blackhat - but to me it is just research) to test whether these benefits of the nofollow tag actually work in each of the 3 major search engines.

What I did was go to the AskDaveTaylor.com blog and found an article to which I posted a comment. I chose this site due to its strong page rank and its level of trust in the search engines as seen by his many rankings - thanks Dave :-). The post that I left is under the name of "Mark Warranty Peterson," just a random name with a keyword in it and linked to the site "theautoclub.com". Now, since the comments are on this blog are treated with the rel="nofollow" tag, you would assume that none of the 3 rules of nofollow as outlined above would be disproven.

Ok, so here it is a few months later and the page that I left the post on has a PR5 and the SE algorithms have all had time to fully evaluate this page. So, I went in and did an exact search in Google, Yahoo and MSN for "Mark Warranty Peterson". Now, one would think that the only page that would show up in the results would be for the page in which the post was left. Well, the only search engine that lived up to this belief is the infamous Microsoft Live Search. Microsoft also did not include the page with the post in a backlink check to theautoclub.com

Yahoo's results showed the page of the post and one spammer page. A backlink check of the domain (theautoclub.com) in Yahoo did, however, include the page from the post that had the nofollow tag.

What did Google show? Hmm.....this is where I was VERY SURPRISED. The results show the page with the post on it, AND the page to which the nofollow link points! They did not show this page in a backlink check, but when does Google ever show many links there anyways. Now, the other interesting thing theautoclub.com is now ranked #74 for "Mark Warranty Peterson" (edit: now showing at #70 on page 7). Notice that this is a broad match for this term and not an exact match query and the site does not have the word Peterson anywhere in the whole site. I then decided to check the cache of the page after finding it at #74. What did Google have to say about the search query?

"These terms only appear in links pointing to this page: mark peterson"


Now, if I read correctly, according to a Google post, ''when Google sees the attribute (rel='nofollow') on hyperlinks, those links won't get any credit when we rank websites in our search results." Well, this seems to be pretty good proof that this isn't true. I also have the server logs to show that Google has crawled through this link to the destination website in the link with the nofollow. That disproves that Google will "NOT follow through to that page."

Now, I am pretty sure that I am not the only person that has done this test and seen the results. This must explain why I haven't seen a decline in comment spam on my blogs as it looks like the "link condom" is broken...

So, can somebody help me out and explain exactly how Google REALLY treats the nofollow tag?

16 comments:

Overtone said...

Great research! Keep up the good work! It appears Google treats the "nofollow" parameter the way you hashed it out. They IGNORE it! LOL!

Well done...

Yuri said...

Well, I have commented on some popular blogs with my name and my other blog URL and the blog URL has PR5 after 3 months of existence, along with ranking in the 50s for my name (just the first name).

Kaj Kandler said...

As far as I know Google does not say they do not follow a link that is marked rel="nofollow", which makes the name of the 'relationship' rather misleading.

In any case, using a live website with other links and purposes does seam to me not helpful to prove anything. You write you saw in your logs that Google followed the link. I wonder how you can see that unless it is the only link on the web that is there. And to create a single link to a home page is close to impossible to do with spiders like whois.sc to crawl domain registrations.

However, I'd hope that Google would abandon this ill conceived mechanism anyway. It has not worked and will not work and is not helpful in any case. I think it is totally counter productive to the idea of Google's interest of a strong web. It is an approach that throws out the baby (the good linked sites, like on the point blog comments and Wiki) with the bathwater (the spam comments).

Captchas have resolved a large portion of this problem and even better would an e-mail based system that requires the poster to verify and sign every single comment with Pretty Good Privacy (PGP).

Mark Barrera said...

Kaj,

Thanks for your viewpoint. Google does say in a post "Q: Should I put rel="nofollow" on the link to my comments page?
A: Probably not, because lots of interesting discussion can happen there. Also, if other people link to your comments page, a spider can follow that link and find any spam that's lurking on the comments page."

This would lead me to believe that they are saying that a "nofollow" tag would prevent the spider from crawling the link.

I also agree that this tag has not and will not help any of the spam problems. I personally believe that both through captchas and moderation (both through software and then user review) is the best way to slow these spam techniques.

John said...

Domain root pages are treated differently on Google. It would be more interesting to test a node which was not linked from the root of the domain. That would also block crawling strategies of MSN / Yahoo, who cover newer domains directly (from the domain registration data).

Other than that, good test.

Geoffrey Faivre-Malloy said...

It's true...that's why people are still spamming the hell out of blogspot.

G-Man
Look What G-Man Found

Der Installateur said...

It's pretty exactly what i learned while placing links with noffollow despite the fact i knew se4rch engines wont honor it.

Wait a while and the link will become good. Simple and true.

my 2 cent

Der Installateur

Dirk Olbertz said...

I also never thought that Google (or any other search engine) should not follow such a link - which is, why this name is so misleading.

From what I learned, the attribute should tell the search engine only to not use this link to make up a relationship between the two sites. For instance: do not inherit the page rank of the source website.

Poczta Polska said...

if google spider no fallow by these link they lost a lot of information...

M.-J. said...

If I go to the source on what Google will or will not do with nofollows, I find a post (http://googleblog.blogspot.com/2005/01/preventing-comment-spam.html) on the "official Google blog" which reads:

"From now on, when Google sees the attribute (rel="nofollow") on hyperlinks, those links won't get any credit when we rank websites in our search results.

It doesn't say anything about no following and never has ... it is just about not passing PR.

incrediblehelp said...

Who said they would not follow it. All Matt Cutt's has been saying is it is not supposed to pass PR juice.

The whole idea is flawed anyway. I cover it more here:

http://www.jaankanellis.com/what-is-pr-scuplting-and-does-it-work/

Hartlepool said...

Excellent article - may have to do a few tests myself!

thanks

Mario said...

i am surprised about this, can we trust to google?

AXZM said...

Good stuff... There was a great article last year on Graywolf's SEO Blog called "Google’s Policy on No follow and Reviews is Hypocritical and Wrong" on a similar subject some of you may find interesting as well, if you haven't already read it. At the end of the day as long as you aren't posting blatant ads, and attempt to lend some relevant information to the conversation, I think comments are good... =P

Free Bingo Money said...

Hey that's really interesting and thanks for the post - strong research. Seems pretty convincing.

Cher K said...

The research seems interesting! Though this post is an old one, I would like to conduct research on the same subject.