NOFOLLOW: A good try, poorly executed

Google, other search engines and blog software writers are co-operating to implement a new method of stopping some links from granting page rank.  Unfortunately, the “nofollow” proposal is poorly thought out.

The newest tool in stopping blog comment spam was discussed and debated and praised today after being announced by Google and Six Apart (though not before bloggers got the scoop).

The idea is that by including “rel=nofollow” in an HREF tag, like so:
<a href=“http://www.hopstudios.com/” rel=“nofollow”>
Google, and soon other search engines, will ignore that link for the purposes of pagerank. The idea is, any comment or trackback or other user-submitted content would pick up a “nofollow” and thus it makes no sense for spammers to try to stuff their URLs into the blogs of others.

Many writers of blogs are thrilled at the idea of another tool against comment spam.  I can understand why: I spend a few minutes every day, and a lot of minutes some days, erasing spam from my blog.  It’s wasted effort for me and the spammers both.

But this fix falls short in a number of areas, even once it’s widely adopted.

For one: links are links, and if spammers spam, people will still click.  Page rank won’t grow as much for the destination URLs, but they’ll still get traffic.  And for spammers who are trying to simply deface or plaster their message around, this does nothing.

For two: A lot of the current power anf benefit of blogs is in their linking, and in their user-generated content.  If you remove all search-influencing mojo (and I’m avoiding the phrase Page Rank because this applies to more than just the Google search engine) from non-moderated links, you remove the ability of bloggers as a community, to direct the focus of search engines to pages that are important.

Links provided in comments by non-spammers have fairly high value that search engines ought to note.  On an indivudal level, bloggers who often comment and ping other blogs are usually high value bloggers who deserve the search engine status that comes from their participation in the blogosphere.  It’s a virtuous circle, and bloggers are the baby getting tossed with this bathwater.

Finally, from a picky naming point of view, “nofollow” already has a meaning: in the robots meta tag of a page or in http headers, it means, don’t FOLLOW any links on this page.  But now, in the context of an HREF, it appears to mean, don’t VALUE this link (but follow it if you want).  So why didn’t they call it “novalue” or “norank” or “nocount”?

While most bloggers posting on this have given their support to the idea, the folks at the Search Engine Watch discussion boards seem less thrilled, and you can count me in with them.

Posted by Travis Smith on 01/18 at 10:43 PM • Blogging Tools
Tracker Pixel for Entry

Comments

  1. You knew I was gonna disagree with this one: here’s why. First, we have a *lot* of comment spam to look at to judge their behavior. Believe me, we have data. And the reality is, they don’t make links hoping a reader will click through. (If they were aiming for that, they wouldn’t spam abandoned blogs, but they do.)

    Second, this *doesn’t* affect the ranking conferred by someone linking to another person’s post. Not one bit. This only affects whether I can come onto your site, link to my own, and see a benefit.

    Finally, nofollow was chosen specifically *because* it has a specific meaning to spiders that crawl the web, and this is used with the exact same meaning in another context, making it simpler and more consistent for people who make software that needs to understand the value.

    Finally, it’s quite likely that people will not apply the “nofollow” value to links that are legitimate. But a site owner should be able to choose which links get their ranking increased, and now they have a tool to do so.

    Posted by Anil on 01/19/05 at 10:46 PM
  2. Anil,

    You’re a smart guy, and you definitely have more data than me available, so let me raise the points I’d like clarified, and maybe others can tell us who’s right:

    1) NOFOLLOW as implemented by your software applies to every comment link made on a blog, whether to the commenter’s own site or to a third party—the code applies to any href, right?  So if I try to link to a cool little charity from a comment where I talk about giving to that charity, that link (and that charity’s Web site) gets no mojo.  The Google blog says: “only the links within comments and the link immediately after “Posted by:” would get the rel=“nofollow” attribute.” Or am I wrong and this only applies to the URL field?

    2) You say:

    nofollow was chosen specifically *because* it has a specific meaning to spiders that crawl the web, and this is used with the exact same meaning in another context, making it simpler and more consistent for people who make software that needs to understand the value.

    What is the specific and consistent meaning of NOFOLLOW? I’m only aware of the Robots NOFOLLOW and the Rel NOFOLLOW.

    ROBOT NOFOLLOW:
    http://www.robotstxt.org/wc/meta-user.html
    “The [Robots Meta Tag] FOLLOW directive specifies if a robot is to follow links on the page.” [Therefore, NOFOLLOW means do not follow any links. The spec says nothing about page rank, but it’s fair to assume that they give no Page Rank mojo as well.]

    REL NOFOLLOW:
    http://www.google.com/googleblog/2005/01/preventing-comment-spam.html
    Google: “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.”

    http://www.ysearchblog.com/archives/000069.html
    Yahoo says: “[The blogger] can tell search engines that the links are effectively untrusted”

    http://blogs.msdn.com/msnsearch/archive/2005/01/18/nofollow_tags.aspx
    MSN says different and agrees with you: “Any link with this tag will indicate to a crawler it is not necessarily approved by this page and shouldn’t be followed nor contribute weight for ranking.”

    3. You say

    Finally, it’s quite likely that people will not apply the “nofollow” value to links that are legitimate. But a site owner should be able to choose which links get their ranking increased, and now they have a tool to do so.

    Small quibble: NOFOLLOW says which links get ranking decreased, not increased.  Moving on to the main point, and I’m in the minority on this, I know, I disagree that, beyond spam control, a site owner should be able explicitly to determine which links give ranking and which don’t.  There’s a good way to do this already: don’t link.  If you link, you’re saying, here’s how to get to this thing I’m talking about.  Search engines benefit from knowing what people hate as well as what people love. 

    For example, if I hate Cameron Diaz (I don’t), Google still wants to know which Cameron Diaz I’m talking about and be able to correctly update how significant her “mindshare” online is. Semantic tagging (“This part of the page is a blog roll, these are comments, this is navigation”) would allow different search engines to provide different values to parts of the page—that is far more useful than a generic, mass-applied and unexplained NOFOLLOW.

    Posted by Travis on 01/24/05 at 09:13 AM
Commenting is not available in this channel entry.

<< Back to main