MeatballWiki | RecentChanges | Random Page | Indices | Categories

Google has created a standard to semantically mark the quality of outbound links using tags.

Last updated April, 2021

LinkSpammers like ShotgunSpam flood open conversations on the Internet as a way of raising their profile on Google, MSN, Yahoo!, and other SearchEngines. Google's PageRank algorithm is famously about counting references to a page rather than keywords on it, and so it makes sense to increase the number of references to your client's page by throwing links to it as far and wide as possible on the Internet. There is an economic incentive to do this because a higher PageRank leads ostensibly to more sales or more traffic or whatever is of value.

Thus, in the vein of NotIndexed, SixApart? and Google have proposed [1] [2] the simple EconomicSolution of flagging all outbound links on blogs with rel="nofollow" which tells the SearchEngine spiders to ignore these links in their data sets. Without the economic incentive, spammers in theory will stop spamming since it's costly and pointless.

This strategy will be a failure on several grounds.

The outcome of this is not simply to say to bloggers and other amateur commentators buying service from cooperating providers or downloading software built by collaborating developers, "Retreat!" It's not just that poor third world spammers have decimate the vox populi of the rich first world latté set. The outcome is also to bias the SearchEngines towards the owners and controllers of the static web. The non-discursive, traditional power centres that have dominated the world since the Industrial Revolution. Brochureware and other non-critical messages will increasingly dominate the rankings, leaving the rest of us run over by the Cluetrain that we were supposedly on just a decade ago.

Admittedly, bloggers will continue to make links that count in their own postings. The loss will only be in the comments. Thus, the actual effect for blogs is not so great. For wikis, however, the distinction between TheAudience and TheAuthor is none, and thus the impact is much greater.

I strongly disagree with this pessimistic view. It is left to the implementations to use the feature wisely. Wikis will of course not use nofollow for interwiki links. It can't be the job of google to tell good from bad links. If we value the impact of our links in search engines it it our own job to check them and remove the (default) nofollow attribute again. The question is not why it can't work but how to integrate it into our AntiSpam? efforts. -- FlorianFesti?

The single strongest reason why the proposal will fail is that the spammers do not read the sites they spam, and never will. This has been demonstrated countless times, as sites have implemented spam-crippling technologies and announced them loudly to absolutely no effect. The spammers will keep spamming because they have the automatic systems to do so. Whether the spam links have NoFollow or not, they will still appear on the sites being attacked, and that is all the spammers want. And, as Sunir points out above, their attacks will always be successful on the GhostTowns. To suggest that this proposal will be a SilverBullet, as the scads of blogs TrackBacked on SixApart?'s site and other places are, is remarkably naïve. -- EarleMartin

The only silver bullet against spam is [this]. The success of the nofollow attribute can only be to limit the effectivity of spam - to make spamming as a whole less attractive. Of cause we have to continue all our other anti spam messures. But nofollow fits in well into a lot of things we already have.

c.f. MotivationEnergyAndCommunity

I'm half in agreement here. It's not going to hurt us to implement it. But spamming is a bulk activity which is already predicated on very little feedback. This isn't a big disincentive. -- PhilJones
``flagging all outbound links on blogs with rel="nofollow"'' - actually it flags all outbound links in *comments* with rel="nofollow", links in posts remain unaltered. -- GeorgeHotelling?
I agree with much of the above but you say "to ignore these links in their data sets" is that definitely stated? I thought Google said that it would give no credit for these links for the site linked to. I have not seen Google confirm it would not, nonetheless, count these links in diluting other links on the page. I find it bizarre that Google should offer such an obvious way for Websites to give a different impression to spiders than to human visitors and smell something fishy going on --AndrewCates

Moved from above.

Surely this not abuse, but just an alternative use. It's not the reason google provides the rel=nofollow facility, but it is a bonus. Now people can link to websites which they don't like, without google making its usual assumption, that linking is positive. -- HarryWood

I think there is public value in raising the PageRank of a nemesis. If you are criticizing someone, they become a subject of public interest, and thus should be accessible. But I defer to your point and moved this bullet down into the discussion. -- SunirShah

I think a better idea would be to use semantic attributes such as "ugc" for UserGenerated? content, and let the SearchEngine do its job of ascertaining whether a link is spam or not. The "nofollow" attribute was good in the time it was introduced, but nowadays it's no longer as useful as before since we have better technology these days (it's 2021 after all). Google could for example store an ExternalLink marked as "ugc" in some database, and keep watch of that link. If it doesn't get removed from the webpage in 14 days, then that link is good and should now influence PageRank. 14 days should be enough time for good blogs to remove spam. This, I believe is a much better approach than the indiscriminate nature of "nofollow". -- JobBautista

Google already may positively weight nofollow links.

CategorySpam CategoryAntiPattern?


MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions