Joi Ito's Web

Joi Ito's conversation with the living web.

Six Log
Support for nofollow

Recently, we’ve reached out to other blog tool vendors to try to coordinate information about comment spam techniques and behaviors. As part of these efforts, we’ve also begun to talk to search companies about enriching linking semantics to better indicate visitor-submitted content (like comments or TrackBacks).

The search team at Google approached us with the idea of flagging hyperlinks with a rel="nofollow" link attribute in order to alert their search spider that a particular link shouldn’t be factored into their PageRank calculations. The Yahoo and MSN search teams have also indicated they’d support this new spec, and we’ll be implementing and deploying this specification as quickly as possible across all of our platforms around the world.

This sounds like a good idea. Take a look at the whole post for more details, but your support would be greatly appreciated.


Sigh, even more "rel" attributes... <a href="" rel="vote-for nofollow boss blogger petowner tag">joi</a>

Installed, rebuilt and in full effect.

Here's a wiki page started by Sunir Shah that makes a good case that nofollow will dramatically reduce the power of blogs compared to static media:

A new way to game search engines, eh? When will ecto support this, so I can link little green footballs and democratic underground without giving them any of my precious googlejuice?

Before we all start congratulating ourselves, what exactly is the benefit of this? Hasn't this problem already been solved, at least twice?

I think this is a mistake... Comment spam is a problem, but who is to say that the comments might not be more valuable information or provide more relevant terms to a particular search than the main blog content? Only the owner of the site.

Maybe you could flag spam comments with this tag by default, but upon review, the owner should unflag useful comments so that they are properly indexed. The content of a comments page is not just the post, but the comments as well...

Removing graffiti quickly is key to combat the problem, which is the case with active Wiki communities. The spray can is applied elswhere since the artwork does not survive. The rel=nofollow parameter is one additional layer of protection.

Wikipedia and TWiki already support the rel=nofollow parameter. TWiki does that via the BlackListPlugin,

The Sunir Shah wiki mentioned earlier makes some good points. The key point, I think, and what this "no follow" thing is mostly about, is this:

Decapitation. Blogs, wikis, and other discussion fora that form the TwoWayWeb serve a critical function as a BalancingForce to traditional power centres. As the CluetrainManifesto asserts, the discussions amongst TheAudience are more powerful than the voice of TheAuthor alone. Yet implementing NoFollow or other techniques to remove discussions from the data sets of SearchEngines decapitates the very purpose of these fora in the global battle for attention. What's the point of blogs if they do not compete for power and attention in Google? Without making outward links count, how can criticism of say Union Carbide ever hope to raise the Bhopal disaster to the #2 position?"

This is the same point, essentially, that I made in my last comment here.
Glad to say I wasn't the only one to figure it out.

The "no follow" tag, as I see it, is less about spam control and more about damage control.

The effect this has may be limited. From my experience I would separate spam into two types: a) page rank seeking and b) plain advertising.

a) Some spammers post URLs of their websites to high-visibility blogs with high page ranks. This improves the page rank of the pages they link to. More people find them on the search engines, making the spamming profitable. This spam may include "information", but normally doesn't. This is more of an indirect approach.

b) Other spammers post "informational" messages into blog comment areas, to attract people to a website. While attracting a higher page may be a desired side effect, the main motivation is to direct people to a website. This is a direct approach.

The nofollow method will reduce page rank spammers. It will not, however, reduce type b spam. Type b spam may even increase, as spammers switch to direct "marketing", rather than increasing page ranks.

This does not mean the Google approach is useless. It is elegant and may cut down on spam, for a while. We should all support it. Unfortunately, it is not a comprehensive solution to blog spamming.

I agree that this attribute really only benefits Google, and won't do very much to slow comment spam. But helping Google isn't a bad thing.

In my opinion, nofollow is not the right way to limit comment spam. A simple use of captchas is a good enough deterrent.. while it may inconvenience genuine people, but there’s a fair price to pay. Also, it could be programmatically arranged that people on your blogroll would be able to comment without the need for validating captcha text….

yeah… those stupid wiki’s!
wonder what happen if we all use nofollow to them?

Have a good one.

7 TrackBacks

Listed below are links to blogs that reference this entry: Support for nofollow.

TrackBack URL for this entry:

==The search team at Google approached us with the idea of flagging hyperlinks with a rel="nofollow" link attribute in order to alert their search spider that a particular link shouldn’t be factored into their PageRank calculations. The Yahoo a... Read More

At 2NMC, you heard Dave and I talk about the importance of leaving comments on blogs so that your audience can triangulate you -- associate you with events, people and issues that impact your music. The side benefit has been Read More

Well, it's only a proposal from the search engines, but MovableType (and thus LiveJournal) will be implementing it -- with some work already completed and available for it. This is a good thing as there is even less a reason... Read More

Google's new rel="nofollow" attribute for links has people applauding, but I have two questions... Read More

Google has proposed a method to fight weblog comment spam that would dramatically decrease the influence of blogs. Comment spam... Read More

Imagine a Web with pagerank, Google created it. Pagerank is the value of a page, in the search results of Google, based on the number of pages doing an hyperlink to the page in question, and on their respective ranks. The more a page is successful, ...

Read More