Note: Old code from this post has been removed – sorry blame WP and poor code support
When Blogcatalog was relaunched, the links from the directory were nofollow.
In my initial review I suggested that it would be a popular decision to remove the nofollow from the links, and also use a different method of tracking clicks – Blogcatalog have to track clicks because it is part of their blog rating system, and might also help with advertising sales for premium positioning.
Most blog directories do not provide direct linking.
Here is an example of the redirection code being used
<br /> <a onclick="return o('andy-beard-niche-marketing');" href="http://andybeard.eu">http://andybeard.eu</a><br />
This article will prove that onclick can pass PageRank – it will not prove that it always passes PageRank.
Some Background On Onclick and Pagerank
In the comments he was specifically asked about onclick for tracking, and didn’t respond.
As an example there was a comment on Matt’s blog by “Brian”, and based on the tone it was written in, I would suspect that was Brian White who works with Matt.
How are links you describe sneaky? I think Matt is referring to cases where a spammer wants one thing indexed, Google ranks the page based the spammy text (above in the example), and the user gets something much different. Google is reserving the right to rank against what the user sees and determine relevancy for themselves.
You see, Matt was referring to something totally different.
Patrick Cornwell in the comments said:-
I second PhilCâ€™s concern, but on the subject of clicked links: I have a couple of sites where link clicks are tracked â€˜onClickâ€™ but the eventual URL is exactly as displayed in the HREF. To me that would concur that nothing sneaky is going on, but again it makes me slightly nervous reading posts like this that Iâ€™m about to fall out of the index!
PhilC’s concern was actually due to using redirects for frames and that was the only question Matt answered in the comments.
PhilC, weâ€™re aware that sites have code to show frames in the correct way. I believe that such sites will not have any cause for concernâ€“this is a common idiom.
Matt didn’t answer the onclick question. Maybe there wasn’t a way he could a clear answer… one of those “It depends” situations.
Washington Post Blogrolls
In 2006 there a was a huge debate surrounding paid links on the Washington Post and their use on “onclick”
The clearest explanation or hint at what was happening, and how the links were interpreted by Google was on the Search Engine Watch forums a year ago.
That was dealing with links on the Washington Post where it was fairly obvious to a human that the links were sponsored.
Nowhere within that thread do Brian White or Matt Cutts state that every link that uses onclick automatically passes no juice.
Brian White said:-
These links will not count for PageRank value. For instance, gadgets-weblog.com is not receiving PageRank from washingtonpost.com. Neither will the links count from washingtonpost.com to finance-weblog.com, for that matter.
Matt has alluded to this many times in the past, and I came in to reiterate the point.
Matt Cutts said:-
Yup, we certainly noticed these a while ago. dyn4mik3, it may look like a clean link, but the fact is that the onclick behavior invokes a new page and different behavior from a typical hyperlink, and that’s visible to anyone viewing/analyzing the source code.
They do state that those specific links pass no juice.
Google rarely state everything, even if it would clear things up a little.
Google Use Onclick With Google Analytics
Patrick Altoft has been writing some great tutorials on how to use Google Analytics Click Tracking. Google use “onclick” for their click tracking.
Surely Google wouldn’t have an algorithm which allowed them to use onclick for tracking but prevented a 3rd party using their own custom solution?
Alternative To Nofollow
If onclick is looked on as an automatic indication that a link is paid for and shouldn’t be counted, it would have been mentioned as a way of disclosing paid links by now.
For many years, even before nofollow, you have been able to use code such as this for dynamic linking that is not counted by search engines.
You might also see this suggested in many places
3rd Party Testing Of Onclick
After the Webmaster World discussion, there was some testing done, such as on SEO Revolution
BlogCatalog – Blog Catalog Doesn’t Pass Pagerank?
There is currently a story on Sphinn claiming that it is a fact that Blogcatalog doesn’t pass PageRank.
As I have explained in the comments there, it is very hard to determine whether specific links pass PageRank, because Google in their wisdom have determined not to tell webmasters all of the links that are counted for ranking purposes.
As an example, at the time I responded whilst checking my own links, Google was not showing links from SearchEngineLand, Problogger, Techcrunch, John Battelle, Lorelle on WordPress, Blog Herald and many many others.
In fact to be honest most of my best links were notable in their absense
But I didn’t give up my search for facts rather than interpretation of inconclusive forum comments that were quite dated.
I eventually found a link from BlogCatalog that is being counted
That shows a link to pontotriplo.org/quickpicks/ very clearly
I also took a longer screenshot that shows a number of other blog directories such as MyBlogLog also listed.
Google’s Lack of Information & Tools
Your average mom & pop webmaster isn’t going to know whether one inplementation of a tracking script or another is allowed by Google’s webmaster guidelines.
Sneaky redirects is about cloaking, not click tracking. It is about indexing one page full of junk, but showing a user another.
Here is what the webmaster guidelines currently say
In Blogcatalog’s case, they are tracking clicks, which are applied to rankings which are useful.
It took a lot of time to determine that the links were being counted, though it may only be a minimal amount.
Maybe giving my profile on Blogcatalog a link will give it a boost and make it more visible though it does have a lot of links on the page going to other people’s profiles.
It would be possible to just use a PHP redirect
This was one reason why Alexa redirects to try to game Alexa were so bad, and why MyBlogLog banned their use. They are just ugly and confuse people.
Debunking a Few Other Points Raised
- The profiles on Blogcatalog are not being paid for. It is possible to have the profile listing be more prominent, but in no way is anyone paying for a listing of their blog and they have fairly high editorial integrity of the sites that are admitted, though a few have slipped through from previous owners. It does take time, and the community does help ensure that rogue sites are removed.
- There is no requirement to link to Blogcatalog – they now provide a way to authenticate a blog using a meta tag, and the widgets. Not every blog platform supports this.
You can authenticate using a small badge or a link, and I haven’t read anywhere that those links have to be followed and I have never seen a requirement to have them on every page.
- Nofollowed links from content snippets – die to the cyclic nature of ranking calculations, it is doubtful that having more extenal links on the page to your content would actually be a huge benefit, but that is very hard to determine with some experimentation. It might help newer content to rank slightly better due to the temporary anchor text, but it does take search engines some time to take that into account.
The links currently use a redirect, in this case a 302 redirect – it doesn’t really matter because they are nofollowed. It seems to be fairly standard industry practice for blog directories to nofollow links to the blogs they list, because they provide user generated content.
Overall, I think the primary objective of Blogcatalog should be to structure their SEO efforts around getting as many of their pages indexed in the search engines as possible, and providing a useful user experience.
How many pages a site has in the SERPs changes a lot, as does how it is being reported.
Currently Google is showing 729,000 pages indexed in total, and 179,000 in the primary index. It was something like 211.000 in the primary index a couple of days ago, so that shows how these things jump around a little.
Google may well be discounting the value of many of the links from Blogcatalog, just as they may discount the majority of links any site receives, and only retain a percentage.
They are more likely to discount links from deeper pages, thus unless Blogcatalog created a purely flat profile, some of those deeper listed sites do get less juice just like old blog posts on poorly optimized blogs.
Disclosure: I do some minor consulting with Blogcatalog and will potentially gain financially if they are ever “flipped” but I try not to let that affect my judgement whilst talking about them or their many competitors, many of whom I have helped for free in the past, and where I maintain good relations