Look who follows NoFollow links!

Earlier this year, I speculated on how the search engines treat NoFollow links.  For those who might be a little green, NoFollow links are not totally ignored by the search engines.  For those who really, really green, NoFollow links are believed to be totally ignored by the search engines (because they have the rel=”nofollow” attribute in the link code).

So we ran a little experiment. 

A client of ours had a fully developed website that has never been used.  Not a single link points to this website, so in the eyes of the search engines, it should not exist. 

It was not indexed at Yahoo. It should go without saying that Yahoo displayed no backlinks.

The site was indexed at Google.  (How, why and whether Google should index orphan sites that have not been released to the public is a topic for another post.) Google showed no backlinks, but the site did rank #8 at Google for one very important search, based primarily on the name of the domain. It did not show up in the top 100 for a few other key searches. All searches are for local terms specific to a certain city, so they are moderately low competition.

For three weeks, we posted comments on NoFollow blogs (yes, intelligent comments reflecting the specific content of the blog posts) to create a steady stream of NoFollow links, without creating any DoFollow or “normal” hyperlinks.

Were the NoFollow links followed?

At the end of week 4, we found Yahoo had indexed the website and showed 51 backlinks.  All of these are NoFollow links. The more important searches were all showing in the top 20, one as high as position #6. Remember that these are moderately low-competition, local searches, but this is all on the strength of a few weeks of exclusively NoFollow links.

Google showed no backlinks after 4 weeks.  No surprise there; Google is very sporadic with if, when, how and which sampling of backlinks it chooses to display. The ranking at position #8 had not changed, but a couple other search terms were now ranking at Google, one of them as high as position #11. Again, this is exclusively on the strength of NoFollow blog comments.

What can we conclude about NoFollow links?

NoFollow links still obviously count at Yahoo.  Do they count as much as DoFollow links?  A more complicated experiment might help answer that question.  Anyone feel like taking up the challenge?

NoFollow links also appear to count at Google.  Or perhaps some do and others don’t, depending on other factors Google might use to rate links from specific domains. However, we can be sure that Google does follow at least some NoFollow links.

The conclusion I would draw from this is that people really should not focus on the NoFollow/DoFollow issue. Build links that are officially followable when you can, but don’t let a NoFollow attribute in a page’s links dissuade you from creating a link you would otherwise pursue.

Comments

  1. I'm glad there is some value in no follows, even if it's small. You never know what efforts will help your website so as many options as possible is great.

  2. Hey – fantastic article. I often have wondered about the value of no follow links and whether or not they in fact have any value. You've definitely given me the answer I expected and was hoping to hear. I wonder how the exact same site would have performed over 4 weeks if only do follow links had been added…thats a topic for another experiment!

  3. I had a business come to me to do some work, I never ended up working with them but they had me check over their current link profile.

    All they had done was nofollow comments, ranging from spammy ones to coherent, and boy I was not expecting to see what I did.

    They ranked top 10 for a half dozen *insanely* competitative keywords they were after in Google. I've been working on an ecommerce site myself to see what results I can come up with in regards to this.

    Link Building is a strange mistress my friend 😀

  4. did you login to Google Webmaster Tools to see how many links google reported there?

  5. Good research, David! I agree that you should not choose to comment on a blog according to whether the link is followed or not. Instead, remember that PEOPLE are reading posts and comments and if you say something intelligent, it may prompt them to visit your website or blog. By wisely choosing blogs with topics that are relevant to yours on which to comment, you may get targeted traffic and maybe even a link or two from other websites.

  6. This is great piece of experiment. This really proves that a link is still a link, no matter if it is nofollow or dofollow.

    I think, Big G still recognizes nofollow links but, not taking it much. However, if your nofollow links come from highly authoritative site, I think, It will make a really big impact.

  7. This site is fresh and has no Webmaster Tools or Analytics accounts attached to it. But that might be something to look at.

  8. That is pretty amazing. I expect there were more than a few high-authority sites in there.

  9. Wow this is a very interesting experiment. I thought that NoFollow means NoFollow and that it doesn't count as backlink at all. I have to rethink my backlink strategy now.

    Thanks a lot for sharing
    Dan

  10. Nope I would forget Nofollows if you are looking to imporove ranking. There are too many other factors that could have caused those kwywords to start ranking. Nofollow for traffic and online public relations only.

  11. I think it's important to understand how Google and Yahoo define and treat nofollow links first.

    1) Just because the links have the "nofollow" attribute, does not mean that they are not crawled and/or used to find new sites/pages or used in any other way. 2) According to Google (yes i know we can't trust everything they say), the nofollow attribute is mainly used to prevent PageRank from flowing to the linked page… I would be more interetsted to hear if the site's PageRank has moved at all. I think that would be a much more accurate measure to test. 3) We can't rule out that the blog comment links aren't used in some other ranking factor (ie. mentions, buzz). I believe there was a paper produced by one of the search engines at a WWW conference recently (sorry, I can't find it right now) that looked at using links not only to pass PR, Trust, etc, but also looking at total volume of links and unlinked domain mentions over time as a measure of "buzz" regardless of whether they are nofollowed or not. I don't think we can rule something like that out in this case.

    Finally, IMO, I don't think it's possible to conduct accurate SEO tests like this. How do you know someone didn't link to your site (not unfollowed) without you knowing about it? How do you know for sure that no other ranking factors have played a role in the site's increases in ranking?

  12. One thought is that many blogs will display a 'Recent Comments' widget in the sidebar of the blog. Even though they nofollow comments the links in the widget are usually followed links.

  13. This is interesting, I've always subscribed to the "make it look natural" belief when linkbuilding going after dofollow and authority nofollow links. If all of your links are dofollow that really doesn't look very natural to the Big G. Works for me anyway.

  14. Simon, your points are well-taken, especially the last one. This is the problem with almost any SEO test. This test is not perfect, but I think it is good enough to establish that NoFollow links do feed into a website's rankings somehow. No other links were showing up as backlinks in Yahoo or Google; this does not rule out 100% that there were none, but it's pretty likely.

  15. Simon is right, nofollow really means notrust. Google and most other search engines follow all links they just don't pass PageRank when crawling a nofollow link.

  16. This is getting WAY too confusing for me. I may just give up on SEO altogether. I thought I understood it, but lately I've found myself completely perplexed and humbled, as everything I thought I knew has now been thrown into question.

    It may well be time for me to just hang up my boots and find someone smarter than me to do it for me…

  17. Google has many ways to spot URLs … Did you have pagerank turned on on google toolbar? Click on links from the site to google (so it could capture referring URL)? Etc …

  18. for nofollows, what counts is the anchor text value of the link thats it.

  19. I have been running my first website for around 8 months now. I, like most website owners, have done loads of reasearch on building backlinks. To be honest I am no clearer now on the issue as I was back when I first started.

    Follow, no follow, page rank off the page you have the link, etc. I suppose just gotta keep plugging away and see what happens. I don't think we will ever understand the true nature of link building 🙂

  20. Interesting test though not sure it's conclusive. nofollow links work for discovery and that's what they did for ranking in Yahoo.

  21. Despite the debate going on whether Google counts nofollow links or not, you still need nofollow links to make sure that your link building is balanced. If you only use dofollow links this may be a sign to Google that you're building your links artificially.

  22. RT: @mattcutts: If you read http://bit.ly/PUNLm carefully you can spot at least one possible (but huge) flaw with http://bit.ly/ZieAQ

    (matt cutts' response)

  23. What about people using browser embedded page-rank apps? Has anyone ruled out Yahoo for user installed tracking applications for users that when to the new site from the blog link? Could Google have reference registrar data for the URL search result? There still seems to be other variables that may not have been eliminated before coming to the conclusion that nofollows are arbitrarily followed.

  24. OK, I'll buy that…but what possible flaw is he referring to. I'm not sure I see it.

  25. 1. Google has stated many times that they use nofollow links for discovery
    2. Pages = PageRank
    3. If Google discover a page, it has some minimal PageRank even without any links
    4. Internal linking can concentrate PageRank on specific pages, even on a domain that apparently has no followed links
    5. Internal linking can pass anchor text
    6. First link priority

    Ultimately too many variables to prove whether links with nofollow pass PageRank

  26. I think there may be other factors at play to account for the SERPS. For example, don't search engines give more value to sites early on after the domain is indexed?

  27. I'm pretty sure that Google indeed does follow and pass linkjuice on certain sites with nofollow attribute.

  28. Where does the "nofollow means notrust" came from? I think this is just a huge speculation.

  29. David, thanks for showing us the truth behind the Nofollow links. Honestly, I wouldn't care too much about the dofollow or nofollow stuffs, as long as the comments are useful and no sounds like a spammy post to me!

  30. Was a pleasure to read the article, i believe the same, nofollow links count for Yahoo, my site is ranking for the search term i want in the first 50 pages using many nofollow links but is not ranking in Google in first 100 pages.

  31. Andy has some good points, though I am not sure about the first one. I have heard Matt Cutts say in his Webmaster Videos that if a link is Nofollowed, it is a NO follow for google and no SERP benefits are to be expected. May be what Andy says is just that google uses them to find new URLs and follow them separately.
    Anyway, I do have real doubts about how ignoring nofollowed links will affect the basic concept of PageRank (that links are like editorial votes), in this age of twitter et al, where all links are nofollowed. Not to forget that Wikipedia and blogs use nofollow and these are places where some serious interaction happen.
    But as for this experiment, it might be possible that On Page optimization might have done the trick. After all, it was a moderately low competition niche.

  32. Even webmaster tools doesn't show every single link that Google counts, though. Even so, I've seen nofollowed links show up time and time again for various sites I've worked on in webmaster tools.

    Haven't cared about nofollow for quite some time now…at best, I'd say they have somewhat less value than dofollow. For authority sites, their nofollowed links probably count a lot more than numerous dofollows.

  33. I think this debate will go on for a long time without anybody ever actually knowing whether nofollow works or not.

  34. I think Google is not counting nofollow links in our total backlinks list and Yahoo have been including all links in our backlinks list whether they are nofollow or dofollow.

  35. As I understand it, Google often follows NoFollow links, for the purpose of indexed the page found at the link, but they do not pass any PR to the site, because you're basically saying with NoFollow that you don't endorse the link.

  36. You know you never really know until you test,test and re-test. Thank you for this, it is really a starting point for everyone to follow up. I know I will be.

  37. Thanks for sharing this little experiment. I find it very interesting, and many of the other comments helpful as well.

  38. I believe that every link counts, but count more the ones from sites with high page rank. For example wiki links are nofollow, but because they come from Wikipedia will be indexed even by Google.

  39. I never really know where i stand with no follow links; there are so many conflicting views that I have yet to make my mind up about them.
    Its an interesting point however that no follow links should be used to convince google of genuine link building rather than artificial.

  40. I always knew that Google didn't follow nofollow links, but I was never sure who did follow them. I knew that some places did though. Thanks for the research and information.

  41. I agree with Andy (and others) that it is very difficult to conduct real research when it comes to SEO and experiments like this as it's almost impossible to have a control environment. However given all the circumstances, you did a nice job of presenting a decent look at NoFollow. In our opinion, website owners should not obsess over it. We wrote a post a while back on the subject if you would like to allow the link here it is: http://www.verticalmeasures.com/link-building/no-

  42. commenting on blogs with nofollow still brings traffic. and if those visitors have the google toolbar or if website has google analytics, or google adsense, well… google knows the website is getting traffic. so i'm not sure this test can be "definitive".
    but surely it's stimulating test 🙂

  43. This made for an interesting read and is good to know. There has always been a lot of fuss about do or don't follow, so thanks for sharing this.

  44. Wow I love your experiment that cool action to do…

  45. Success of a website depends on both no follow and do follow link. I think you have now made no-follow links popular by your post. Thanks for letting us know the importance.

  46. Great article and clearly written. Personally I think that some "no folllow" links will pass on some link juice, depending on what type of site they are from.

  47. I've noticed something similar, mostly through Yahoo!. I always figured that Google would have to read nofollow links purely because they are there. The weight that follows these links is a whole other story, but I'm willing to bet that comments on related blogs might just help enhance the theme of your site…

    Then again I could be way off the mark.

  48. I suspect that nofollow links pass pagerank. Why? Because I have seen nofollow links recognized by google when searching for 'link:http://www.domain.com' and google does not easily recognize links. It probably also depends on the environment in which the links are placed in.

  49. Nice post, thanks for that info. I had been searching for dofollow ones, checking the following of the blog. I think I should not care about the following anymore

  50. Thanks that great tips about no follow link, now I understand the easy way…