19 replies to this topic

[qwerty]

I was under the impression that Google, Yahoo and MSN had all agreed to refrain from following links with the rel="nofollow" attribute. Yahoo said they'd do that back in a January post on their blog, but apparently that's not the case.

I just noticed that Yahoo considers this page from Matt Cutts' blog to be a backlink to my site. I commented there, and my name links to my site, but the code behind those links (three of them, I think) is

CODE

<h3 class="commenttitle"><a href="http://www.raisemyrank.com/" rel="external nofollow">Bob Gladstein</a> Said,</h3>

So, is anyone not following nofollow?

[Jill]

Doesn't surprise me. I don't think anyone ever said they definitely wouldn't follow it. More that they would keep in mind that those were not trusted links (or something like that).

[Alan Perkins]

Yep, "nofollow" is a bit of a misnomer. "notrust" is closer to the mark.

[Scottie]

Last March, Yahoo said they hadn't decided how to treat "nofollow" yet and Google said at that time, they would not follow the link but reserved the right to change that at any time as they decided how best to work with the attribute.

Everyone has clearly said, nofollow is a vote of no confidence for a link. Nothing more, nothing less. Draw your own conclusions.

[qwerty]

Well, feh. I mean, I don't suppose it makes sense to complain about the fact that I actually get a backlink when I comment on someone's blog. But I'd kind of like people to know that I only do so when I actually have something to say. Now it looks like I'm just posting for a link, so I may as well just tell them about my mesothelioma site and forget about giving my opinion on anything.

[esoos]

I ran a bunch of test on this about 2 months ago, and everything I saw indicated that none of the major search engine were following nofollow links at all. No link pop passed, the links themselves weren't followed, and the pages they linked to weren't indexed.

It's possible that my methodology was flawed. It basically involved using css to hide a few nofollow links on fairly popular and regularly crawled pages. Nothing sinister, I just wanted to make sure that no one else linked to those same pages, thereby getting them indexed and screwing up part of the test.

I took the test down after a few months of the nofollow links not getting crawled. I now kind of wish I left it up to see if things have changed, but those few hidden links were making me nervous, and I didn't want to sacrifice a perfectly good page in the name of science.

Anyway, that's what I found. Things could have changed recently. Another possibility is that the fact that Matt Cutts uses 'external nofollow' as opposed to 'nofollow' attributes on his site could be confusing some search engines. It would be a bit ironic if Matt's blog is one of the last remaining major blogs that's still easily blog-spammable.

Anyone else have 'nofolllow' followed or not followed?

[Randy]

Welcome esoos !

I've never tested it since I could never fathom having to use it. I'm not sure if anyone else has or not, but I'm sure they'll chime in if they have.

One thing on your test...

You introduced another variable into the test by hiding the links via css in addtion to the nofollow. It could be that the engines didn't follow because of that reason too. At least some of them seem to be actively seeking out hidden content for action lately.

[esoos]

Hi Randy, and thanks.

It's possible that the CSS could have munged the experiment, though I'm a bit skeptical.

The nofollow links were placed in h6 tags, which were styled in an external style sheet to have the same color as the background they were on. I tried to avoid some of the more obvious CSS tricks a search engine might pick up on, but it's possible they detected it by reading the CSS file and comparing the color of the text to the background-color of the div it was in.

Reading what I just wrote, I guess it doesn't seem that incredibly hard, but it seems like a lot of work to detect a few silly links, and might be giving the search engines too much credit. And there was no penalty handed down (not that there should have been, as I was just linking to a few articles on my own site).

I'm looking into some of those reports you mentioned about search engines detecting CSS tricks. I'm no advocate of shadowy haberdashery, but it seems like there's so many ways to structure a CSS file that detecting most of those tricks would take a huge amount of processing power. But perhaps not, if you've got the right algo.

For me, the hidden link was just because I wanted to test whether search engines were following nofollow links at all, and if the link was publicly viewable, someone else might have linked to that same page without nofollow.

At the time, nofollow wasn't being followed (at least in this trial). Of course, this is something search engines can turn off and on at will, so any testing really needs to be ongoing. I'll probably run the test again with a few modifications and see what turns up.

Taking a quick look at backlinks of some popular blogs that use nofollow, it does look like Yahoo is current showing nofollowed links in its backlinks, MSN is not, and, with Google, who can tell?

For example, a nofollow link on this page:

www.seroundtable.com/archives/002595.html

show up as a backlink here:

search.yahoo.com/search?p=link%3Ahttp%3A%2F%2Fwww.bigpromotions.com&vs=www.seroundtable.com
But just because a site shows up in a link: command, does that mean it's getting link pop? Not sure that's a question any of us here can answer.

<live links removed>

Edited by chrishirst, 16 November 2005 - 06:38 AM.

[Jill]

But just because a site shows up in a link: command, does that mean it's getting link pop?

I think that is definitely the question. Which is why I'm not really sure the thing can even be tested. It's all conjecture, no matter how you slice it.

[Randy]

Agreed. The important part simply cannot be tested easily or effectively. You could set up a nofollow link that had nonsense anchor text to a page that had never been active before, wait a few weeks and see if it gets picked up.

But since it's a nonsense word with zero competition does that really tell us much? Probably not since there was no competiton in the first place.

esoos, on the hidden text/hidden links front, Matt Cutts talked about it a few weeks ago on his blog when discussion reinclusion requests. If memory serves, he didn't specifically mention CSS. Just generically talked about hidden stuff. Though CSS is definitely the easiest way to do that.

[Remi van Beekum]

I recall Tim Mayer saying they will follow the nofollow links because most links that are not ‘trusted’ are actually valid links, at the SES in NY February/March. They just weight less than links without the nofollow tag.

Tim even said the nofollow-tag was a temporary way to fight comment-spam and proposed an other approach by using the following tags:

< div class=‘content-public’ > … < /div >
< div class=‘content-nav’ > … < /div >
< div class=‘content-default’ > … < /div >

‘content-public’ for all content (and links) that was not made by the owner of the website. (like most comments)
‘content-nav’ for navigation
‘content-default’ for comments made by users that are logged in (and trusted)

Edited by Remi van Beekum, 18 November 2005 - 08:18 PM.

[jaynaud]

I thought nofollow was more a way to prevent losing pagerank? I could be wrong though.

[Jill]

I thought nofollow was more a way to prevent losing pagerank? I could be wrong though.

Nothing to do with that.

[qwerty]

Welcome to HR, Remi

[jaynaud]

Nothing to do with that.

You sure..

You add the attribute rel="nofollow" when linking to sites that comment on your content. Search engines largely consider such links to be "unapproved" by your site. These links, therefore, don't lend additional weight to the pages you're linking to.

I also cam across this at wikimedia ( I would link to it but you I am not allowed to give links yet

nofollow is a possible value of the attribute rel in the <A> markup of HTML. Many search engines, such as Google, rank pages based on the number of other websites that link to them, but ignore links that contain rel="nofollow". This one attribute determines whether Wikipedia has any influence on the search engine ranking of any page linked to by Wikipedia.