I was wrong about “NoFollow”

I was a supporter of the nofollow link, which told search engines not to count that link in their popularity ranking algorithms.

But, after seeing Loren Baker’s post about how lame nofollow is, I find I’m changing my mind.

I’m very concerned, for instance, about Wikipedia’s use of nofollow. More and more of my friends say they now go to Wikipedia to search first before going to Google. Imagine a world where Wikipedia could lock Google out of key information. Wouldn’t it be ironic that something that Google invented could actually be used against it?

That said, I don’t go for the point that comment links should be judged by search engines the same as links in a blog. That sounds really elitist, I know, but I’m seeing all sorts of gaming going on in comments and it’s a lot easier to build authoritative Webs in blogs than in comments.

Comments

  1. > …but I’m seeing all sorts of gaming going on in comments…

    comment gaming is to quality comments what splogs are to quality blog posts with links

  2. > …but I’m seeing all sorts of gaming going on in comments…

    comment gaming is to quality comments what splogs are to quality blog posts with links

  3. Robert,

    I don’t know what the big deal is with nofollow. I actually think it’s a great idea.

    Wikipedia is doing this because every bit of info in Wikipedia is/can be moderated. Wikipedia cannot exactly moderate outside content or be responsible for it. Imagine, for example, a link to an artlicle on a vacation destination. What if that outside link was bunk? Or an evil ad site selling vacations for a huge profit?

    A little control is a good thing here. Wikipedia has the right to do what they want with their own site. If Google doesn’t like it, so what. Google needs to be taken down a notch or two anyway.

    nofollow may not be popular, but it works for what it is intended to do.

  4. Robert,

    I don’t know what the big deal is with nofollow. I actually think it’s a great idea.

    Wikipedia is doing this because every bit of info in Wikipedia is/can be moderated. Wikipedia cannot exactly moderate outside content or be responsible for it. Imagine, for example, a link to an artlicle on a vacation destination. What if that outside link was bunk? Or an evil ad site selling vacations for a huge profit?

    A little control is a good thing here. Wikipedia has the right to do what they want with their own site. If Google doesn’t like it, so what. Google needs to be taken down a notch or two anyway.

    nofollow may not be popular, but it works for what it is intended to do.

  5. I’ll tell you what a bad idea is…

    prefetch cookies from Google. That’s evil. The first returned result get added to the hardrive even though the site was never visited. This could cause people to get into trouble at school or at work.

    The fact that Mozilla put this in their browers as a default setting is shameful.

    The first two things I do when I download Firefox under Linux is

    1. Turn off prefetch
    2. Turn off referrer logging

  6. I’ll tell you what a bad idea is…

    prefetch cookies from Google. That’s evil. The first returned result get added to the hardrive even though the site was never visited. This could cause people to get into trouble at school or at work.

    The fact that Mozilla put this in their browers as a default setting is shameful.

    The first two things I do when I download Firefox under Linux is

    1. Turn off prefetch
    2. Turn off referrer logging

  7. I think that Wikipedia is using NoFollow incorrectly. The idea behind NoFollow is good and there are valid ways and reasons to use it. The only thing that is keeping it from cutting comment and trackback span is the fact that spammers or rather the people who buy spamming software and use it don’t understand how NoFollow makes much of what they are doing worthless.

  8. I think that Wikipedia is using NoFollow incorrectly. The idea behind NoFollow is good and there are valid ways and reasons to use it. The only thing that is keeping it from cutting comment and trackback span is the fact that spammers or rather the people who buy spamming software and use it don’t understand how NoFollow makes much of what they are doing worthless.

  9. Who says Google relies on nofollow exclusively? My bet would be that’s weighed in along with many other factors to determine spammyness of the link.

    Besides, Google could always decide to just ignore the rel=”nofollow” on wikipedia’s domain.

  10. Who says Google relies on nofollow exclusively? My bet would be that’s weighed in along with many other factors to determine spammyness of the link.

    Besides, Google could always decide to just ignore the rel=”nofollow” on wikipedia’s domain.

  11. A thought – not sure if any search engine has tried this, but here goes.

    Assumption – MOST (though not all) comments systems associate a URL with the commenter (and/or an email address).

    What if, instead of associating the link inside of a comment to the blog where it was posted (i.e. my comment here to http://scobleizer.com) any links which were posted in a comment were associated with blog of the commenter?

    Yes there would still be a LOT of room for gaming (though of a rather different type than presently) – since few comment systems do a great job of proving that the URL associated with a commenter is, in fact, associated with them.

    If I were designing a search engine algorithm today, I would probably weight a link inside of comments differently than I would weight links inside of blog posts (and inside of blog posts differently than other links in the same site – i.e. sidebars, ads etc). My intuition is that in most cases I would weight comment links lower than other types – but there might also be some edge cases to consider (for some searches the links in a discussion might be more relevant and useful than the links in the post that sparked it – at least as a human user of many sites I have frequently had that experience.

    A worthwhile topic to keep in mind and discuss further.

    Shannon

  12. A thought – not sure if any search engine has tried this, but here goes.

    Assumption – MOST (though not all) comments systems associate a URL with the commenter (and/or an email address).

    What if, instead of associating the link inside of a comment to the blog where it was posted (i.e. my comment here to http://scobleizer.com) any links which were posted in a comment were associated with blog of the commenter?

    Yes there would still be a LOT of room for gaming (though of a rather different type than presently) – since few comment systems do a great job of proving that the URL associated with a commenter is, in fact, associated with them.

    If I were designing a search engine algorithm today, I would probably weight a link inside of comments differently than I would weight links inside of blog posts (and inside of blog posts differently than other links in the same site – i.e. sidebars, ads etc). My intuition is that in most cases I would weight comment links lower than other types – but there might also be some edge cases to consider (for some searches the links in a discussion might be more relevant and useful than the links in the post that sparked it – at least as a human user of many sites I have frequently had that experience.

    A worthwhile topic to keep in mind and discuss further.

    Shannon

  13. Looking forward to when Google is just another search engine among many instead of the growing monstrosity it is now.

    nofollw should be obeyed like any other tag. Failire to abide by it would be a good reason for Wikipedia to ban Google from spidering it altogether.

    Wikipdia could easily live w/o Google.

  14. Looking forward to when Google is just another search engine among many instead of the growing monstrosity it is now.

    nofollw should be obeyed like any other tag. Failire to abide by it would be a good reason for Wikipedia to ban Google from spidering it altogether.

    Wikipdia could easily live w/o Google.

  15. I think a lot of people are missing the main point of the nofollow tag. Naturally SEOs will be opposed to it, because they get no juice when they put their links in the comments on your blog. But, preventing people from manipulating search rankings by posting irrelevant links all over the place is the whole reason nofollow was created. It doesn’t matter if it’s link spamming on Wikipedia, or comment spam on Scobleizer. If Loren, you, or Wikipedia feel that they have control over the quality of the links being posted, then of course there is no need for the nofollow. When you remove the nofollow, you’re telling the search engines that you’re vouching for the quality of the links being posted on your website.

  16. I think a lot of people are missing the main point of the nofollow tag. Naturally SEOs will be opposed to it, because they get no juice when they put their links in the comments on your blog. But, preventing people from manipulating search rankings by posting irrelevant links all over the place is the whole reason nofollow was created. It doesn’t matter if it’s link spamming on Wikipedia, or comment spam on Scobleizer. If Loren, you, or Wikipedia feel that they have control over the quality of the links being posted, then of course there is no need for the nofollow. When you remove the nofollow, you’re telling the search engines that you’re vouching for the quality of the links being posted on your website.

  17. [...] Great linkbait Loren!  I agree with much of what you’re saying.  While comment spam was the original problem that nofollow was supposed to solve, I think that the greater concept is still applicable.  Yes, it’s a bandaid that’s imposed by the search engines to make their jobs easier.  But, I do see the usefulness of it.  If you (or any webmaster) do not have good editorial control over what links are placed on your website, nofollow is a good tool to let the search engines know that you cannot vouch for the quality of the links.  If you, or Wikipedia, can take the time to give editorial review to all links posted, then of course the nofollow is irrelevant.  I think a lot of people have perverted the meaning of it, though.  But, that shouldn’t surprise anyone. I posted the following on Scobleizer; [...]

  18. I don’t agree with your position about wikipedia, and here’s why:
    Wikipedia has a high page rank and most of their pages are anonymously editable. Often, people post links that are not spam, but are basically related promotional links. For example, if I had a site about web browsers, I could just link to it from articles related to web browsers. Not identical to what I consider spam, but definitely promotional and worth fighting with nofollow. More importantly, I don’t think anti-spam measures will work against this sort of posting.

    The effect on google shouldn’t be that severe… they’ll still index wikipedia, of course. A very small percentage of the links on the internet are from wikipedia. Additionally, most of the links that aren’t on wikis or blog comments took more effort to create, which might even be a very vague indicator of value (just conjecture).

  19. I don’t agree with your position about wikipedia, and here’s why:
    Wikipedia has a high page rank and most of their pages are anonymously editable. Often, people post links that are not spam, but are basically related promotional links. For example, if I had a site about web browsers, I could just link to it from articles related to web browsers. Not identical to what I consider spam, but definitely promotional and worth fighting with nofollow. More importantly, I don’t think anti-spam measures will work against this sort of posting.

    The effect on google shouldn’t be that severe… they’ll still index wikipedia, of course. A very small percentage of the links on the internet are from wikipedia. Additionally, most of the links that aren’t on wikis or blog comments took more effort to create, which might even be a very vague indicator of value (just conjecture).

  20. I call BS on this.

    Linking is a vote. Linking with nofollow is a negative vote. Period.

    Anybody reading further into this without evidence is just jealous or ignorant.

    For instance, the guy over at the Search Engine Journal claims that nofollow does not help stop comment spam. Huh? This was never the intent. The intent of nofollow is to prevent such comment spam from ever getting ranked in search engine results, which is where everybody in the world go to find stuff.

  21. I call BS on this.

    Linking is a vote. Linking with nofollow is a negative vote. Period.

    Anybody reading further into this without evidence is just jealous or ignorant.

    For instance, the guy over at the Search Engine Journal claims that nofollow does not help stop comment spam. Huh? This was never the intent. The intent of nofollow is to prevent such comment spam from ever getting ranked in search engine results, which is where everybody in the world go to find stuff.

  22. Not all comment spam is automated – I’ve only posted this comment because I wanted a back link (joking – since nofollow is on here).
    It’s still a legitimate comment though so how would you filter this out?, would you want to filter it out?, do you want 1000 legitimate comments to each post you make?.
    I can think of a comment for any blog post on the internet – it’s not difficult.
    I could then post comments to thousands of blogs and get great google ranking to my own site even though it is junk.
    Also, what a surprise, an SEO blog is complaining about nofollow – duh!
    My replies to the 13 points
    1. Wrong – as stated by Stephane Rodriguez above nofollow is not to stop comment spam, it’s to stop it gaming google
    2. Wrong, I’m posting this purely to drive traffic but it won’t get filtered out as spam (or will it…)
    3. and?, that’s like saying virus protection being switched on by default is bad
    4. Yahoo doesn’t use nofollow and I assume has no plans to so this point is moot
    5. Why should a blogger trust someone posting comments?, do you trust me?
    6. The use for nofollow is clear, it means search engines don’t follow the links – no confusion.
    7. Wrong, nofollow is an excellent solution that is working well – hence all the hate from SEO people, if your site is worth linking to then email bloggers and tell them about it and they will link, if not then you don’t deserve any links.
    8. The only true point
    9. People like to complain about the accuracy of Wikipedia but in general it is a fantastic resource – a spam sandwich it is not. The format of Wikipedia is open to abuse but that’s simply how it works, this solution is a good one
    10.Anyone who does this would be a bit thick
    11.It may not have reduced comment spam but at least now the spam sites are not all sitting at the top of google – you have to look at the bigger picture.
    12.Some sites ignore nofollow – that’s their choice – if they want to fill up with spam then so be it.
    13.eh?

  23. Not all comment spam is automated – I’ve only posted this comment because I wanted a back link (joking – since nofollow is on here).
    It’s still a legitimate comment though so how would you filter this out?, would you want to filter it out?, do you want 1000 legitimate comments to each post you make?.
    I can think of a comment for any blog post on the internet – it’s not difficult.
    I could then post comments to thousands of blogs and get great google ranking to my own site even though it is junk.
    Also, what a surprise, an SEO blog is complaining about nofollow – duh!
    My replies to the 13 points
    1. Wrong – as stated by Stephane Rodriguez above nofollow is not to stop comment spam, it’s to stop it gaming google
    2. Wrong, I’m posting this purely to drive traffic but it won’t get filtered out as spam (or will it…)
    3. and?, that’s like saying virus protection being switched on by default is bad
    4. Yahoo doesn’t use nofollow and I assume has no plans to so this point is moot
    5. Why should a blogger trust someone posting comments?, do you trust me?
    6. The use for nofollow is clear, it means search engines don’t follow the links – no confusion.
    7. Wrong, nofollow is an excellent solution that is working well – hence all the hate from SEO people, if your site is worth linking to then email bloggers and tell them about it and they will link, if not then you don’t deserve any links.
    8. The only true point
    9. People like to complain about the accuracy of Wikipedia but in general it is a fantastic resource – a spam sandwich it is not. The format of Wikipedia is open to abuse but that’s simply how it works, this solution is a good one
    10.Anyone who does this would be a bit thick
    11.It may not have reduced comment spam but at least now the spam sites are not all sitting at the top of google – you have to look at the bigger picture.
    12.Some sites ignore nofollow – that’s their choice – if they want to fill up with spam then so be it.
    13.eh?

  24. There’s a difference between using nofollow in a blog and in wikipedia.

    In blog comments, links can be added by anybody but only removed by one person. If that one person is lazy or overworked or not paying attention, you can game the system. nofollow is helpful here

    In a wiki, links can be added by anybody and removed by anybody. It’s this power of anybody to remove that should keep bad links out. Nofollow isn’t needed. If a community decides not to remove it, it must be a good link.

    If you deem this system of checks and balances to work for bad content, you have to trust it for bad links. By using nofollow, wikipedia is saying that they don’t trust their own model.

    PS.. I started nofollowing wikipedia links in my blog as soon as they announced this.

    PSS, regardless of the nofollow.. my links on wiki pages bring me hundreds of targeted visitors / day. I’ll take that over pagerank any day.

  25. There’s a difference between using nofollow in a blog and in wikipedia.

    In blog comments, links can be added by anybody but only removed by one person. If that one person is lazy or overworked or not paying attention, you can game the system. nofollow is helpful here

    In a wiki, links can be added by anybody and removed by anybody. It’s this power of anybody to remove that should keep bad links out. Nofollow isn’t needed. If a community decides not to remove it, it must be a good link.

    If you deem this system of checks and balances to work for bad content, you have to trust it for bad links. By using nofollow, wikipedia is saying that they don’t trust their own model.

    PS.. I started nofollowing wikipedia links in my blog as soon as they announced this.

    PSS, regardless of the nofollow.. my links on wiki pages bring me hundreds of targeted visitors / day. I’ll take that over pagerank any day.

  26. I dont agree with the lack of support to nofollow. Remember that many ppl just look at the title and barely reads the entry.

    Nofollow is one thing. Wikipedia adopting it, is another thing.

    Do you really think that nofollow sucks at all?

  27. I dont agree with the lack of support to nofollow. Remember that many ppl just look at the title and barely reads the entry.

    Nofollow is one thing. Wikipedia adopting it, is another thing.

    Do you really think that nofollow sucks at all?

  28. I totally agree with Loren’s standpoint. I have used DoFollow on all my WordPress blogs and have no problem.

    Remember that nofollow is not only for comments, but also for trackbacks, and useful links that people may include.

    I am much more worried about a spam comment ending up in someone’s mailbox using subscribe to comments, and I need to spend some time hacking it to be CAN SPAM compliant in case a blue pills comment gets through my filters – it has never happened, but there is always the chance.

    Some of the discussion here reminds me of how much FUD there is about nofollow and the way it is being bent from its original purpose.

    Nofollow was never meant to be a negative vote, it was meant to be a neutral vote for user generated links.

    Matt Mullenweg has already written a few days ago about how the nofollow initiative didn’t curb comment spam. (he was glad the attempt was made)

    Nofollow is now being unofficially prescribed for all commercial paid links, but it seems to be something only the little guys have to worry about. There are commercial relationships in all kinds of links.
    The Yahoo directory is just paid links if you are a company, and it has always been used to prop up Google’s pagerank algorithms. No one would honestly buy a link there other than for SEO purposes. When was the last time you browsed the Yahoo directory looking for a service?

    The commercial links between blogs in the same blog network, being commercial should these day be nofollow.

    Roberts links to Podtech could equally be looked on as commercial. I am sure Podtech gain more than $10 worth of link equity and traffic from each time a link is made to new content.
    People using services like PayPerPost are being told that the links they give to companies for their $5 to $10 have to be nofollowed or they might lose their ability to pass on link equity.

    Only 85% of my comments are spam – maybe an indication that removing nofollow has benefits.

    Many forums still allow you to post links and gain link credit, and they did for years before nofollow was introduced. Spam quickly gets removed or never appears due to various security arrangements.

  29. I totally agree with Loren’s standpoint. I have used DoFollow on all my WordPress blogs and have no problem.

    Remember that nofollow is not only for comments, but also for trackbacks, and useful links that people may include.

    I am much more worried about a spam comment ending up in someone’s mailbox using subscribe to comments, and I need to spend some time hacking it to be CAN SPAM compliant in case a blue pills comment gets through my filters – it has never happened, but there is always the chance.

    Some of the discussion here reminds me of how much FUD there is about nofollow and the way it is being bent from its original purpose.

    Nofollow was never meant to be a negative vote, it was meant to be a neutral vote for user generated links.

    Matt Mullenweg has already written a few days ago about how the nofollow initiative didn’t curb comment spam. (he was glad the attempt was made)

    Nofollow is now being unofficially prescribed for all commercial paid links, but it seems to be something only the little guys have to worry about. There are commercial relationships in all kinds of links.
    The Yahoo directory is just paid links if you are a company, and it has always been used to prop up Google’s pagerank algorithms. No one would honestly buy a link there other than for SEO purposes. When was the last time you browsed the Yahoo directory looking for a service?

    The commercial links between blogs in the same blog network, being commercial should these day be nofollow.

    Roberts links to Podtech could equally be looked on as commercial. I am sure Podtech gain more than $10 worth of link equity and traffic from each time a link is made to new content.
    People using services like PayPerPost are being told that the links they give to companies for their $5 to $10 have to be nofollowed or they might lose their ability to pass on link equity.

    Only 85% of my comments are spam – maybe an indication that removing nofollow has benefits.

    Many forums still allow you to post links and gain link credit, and they did for years before nofollow was introduced. Spam quickly gets removed or never appears due to various security arrangements.

  30. From a WebBot/Crawler Design point of view, nofollow was a good idea. If anyone were to sit down and try to find a solution for the problem of comment spam done for the purpose of increasing pagerank, then the simplest to implement solution would be nofollow.

    Did it succeed? I *think* the question is pointless. Akismet, to me, has solved the problem pretty much for now, and I am happy with it.

    But then again, I also use the DoFollow plugin for WordPress to override the nofollow that wp adds automatically for comments…

  31. From a WebBot/Crawler Design point of view, nofollow was a good idea. If anyone were to sit down and try to find a solution for the problem of comment spam done for the purpose of increasing pagerank, then the simplest to implement solution would be nofollow.

    Did it succeed? I *think* the question is pointless. Akismet, to me, has solved the problem pretty much for now, and I am happy with it.

    But then again, I also use the DoFollow plugin for WordPress to override the nofollow that wp adds automatically for comments…

  32. Let’s take an example.

    For example, I am commenting here to point to you of my latest StatBot post analysing the blog of Raymond Chen. I could post a link here, like this: http://blog.yuvisense.net/2007/02/15/statbot-visits-the-old-new-thing/

    Theoretically, you are not endorsing what I am posting here. Sure, if you find it interesting enough, you would post about it, and then I’d get the Google Juice. Nothing wrong with this model, I’d guess….

    But still, I’d like you to link to it:D

  33. Let’s take an example.

    For example, I am commenting here to point to you of my latest StatBot post analysing the blog of Raymond Chen. I could post a link here, like this: http://blog.yuvisense.net/2007/02/15/statbot-visits-the-old-new-thing/

    Theoretically, you are not endorsing what I am posting here. Sure, if you find it interesting enough, you would post about it, and then I’d get the Google Juice. Nothing wrong with this model, I’d guess….

    But still, I’d like you to link to it:D

  34. On WordPress.com Robert wouldn’t be able to make an editorial decision to make that link live, because all links are nofollow in comments.
    The same is true for default wordpress.org blogs and MT, Typepad, Blogger.com etc etc

    I have links posted frequently in my comments that are highly relevant to the conversation on that post, but don’t warrant a separate post that will end up in all my readers feeds.

  35. On WordPress.com Robert wouldn’t be able to make an editorial decision to make that link live, because all links are nofollow in comments.
    The same is true for default wordpress.org blogs and MT, Typepad, Blogger.com etc etc

    I have links posted frequently in my comments that are highly relevant to the conversation on that post, but don’t warrant a separate post that will end up in all my readers feeds.

  36. Thanks for very interesting article. btw. I really enjoyed reading all of your posts. It’s interesting to read ideas, and observations from someone else’s point of view… makes you think more. So please keep up the great work. Greetings.

  37. Thanks for very interesting article. btw. I really enjoyed reading all of your posts. It’s interesting to read ideas, and observations from someone else’s point of view… makes you think more. So please keep up the great work. Greetings.

  38. I do-follow links now on my Aussie housewife blog. After a long time procrastinating I have seen the importance of dofollow and using the “do follow” principal to help keep the Internet better connected. http://www.reallyreally.net – Take a look at my dofollow blog and feel free to comment. Thank you, Regards Bree.

  39. I do-follow links now on my Aussie housewife blog. After a long time procrastinating I have seen the importance of dofollow and using the “do follow” principal to help keep the Internet better connected. http://www.reallyreally.net – Take a look at my dofollow blog and feel free to comment. Thank you, Regards Bree.

  40. [...] Great linkbait Loren! I agree with much of what you’re saying. While comment spam was the original problem that nofollow was supposed to solve, I think that the greater concept is still applicable. Yes, it’s a bandaid that’s imposed by the search engines to make their jobs easier. But, I do see the usefulness of it. If you (or any webmaster) do not have good editorial control over what links are placed on your website, nofollow is a good tool to let the search engines know that you cannot vouch for the quality of the links. If you, or Wikipedia, can take the time to give editorial review to all links posted, then of course the nofollow is irrelevant. I think a lot of people have perverted the meaning of it, though. But, that shouldn’t surprise anyone. I posted the following on Scobleizer; [...]

  41. I think the interesting part, google makes this job unknown so we dont know exactly what is happening.. if google is counting or not.. is nofollow links work for someting or not.. so there is always a hope :)

  42. I think the interesting part, google makes this job unknown so we dont know exactly what is happening.. if google is counting or not.. is nofollow links work for someting or not.. so there is always a hope :)

  43. Not entirely true. They DO generally follow the links (so the page gets indexed) but the page gets no credit for the link to it. As google heavily use incoming links to influence the search position and key words, these nofollow links provide no help in this way.

    However, the page still gets indexed, so unique content may make it show up in google, and some search engines ignore the nofollow attribute anyway.

    If you want a search engine not to follow any links in the page you have to use it in robots or the header.