In the interests of starting with a bold statement, I'll go with what I twittered earlier this morning:
My seo theory for the morning. Nofollow is dying.
During SES last week, I had the chance to talk to a few people about a trend we believe we are seeing in Google's treatment of two big trends on the web. You can't have failed to have spotted the growth of Twitter recently. Some people have started blogging less as a result of the ease and adoption of Twitter (and other similar "status update" micro-blogging platforms such as Facebook status updates). Many people have a greater reach on Twitter than they do via their blog and they certainly have more tools to make it easy to post thoughts, discoveries and questions.
This isn't another Twitter post though - it's about the impact on SEO of this evolving use of the web. In the same way that blogging, commenting, trackbacks (and associated spam) forced a change in the search engines' strategies, I believe we are starting to see similar changes caused by this new shift.
The effects that I intend to dig into a bit deeper are:
- Explosion of publishing on 3rd party platforms
- Increasing use of nofollow for a range of purposes
I believe that the combination of these trends has profound implications in the optimal way of constructing a search engine - something Google cares a great deal about. Not only that, but we are seeing the beginnings of the effects my theory predicts.
Publishing on 3rd party platforms
You will find advice in many places about best practices for blogging URLs. In order of preference for SEO from top to bottom, you would prefer:
- www.example.com/blog
- blog.example.com
- www.exampleblog.com (when your main website is example.com)
- example.wordpress.com (or .blogspot.com etc.)
- thirdpartydomain.com/example
- thirdpartydomain.com/examplepost1, thirdpartydomain.com/examplepost2 etc.
There are two main factors making #1 the right answer:
- Ensuring your link-worthy content acquires links to your main website
- Having all your content on a domain you control
Although we have seen many blogs created as sub-domains of 3rd party domains (such as *.blogspot.com), there is no overwhelming trend towards this and no fundamental reason why anyone would recommend it. With Twitter (as short-hand for any social service allowing micro-updates), however, not only is there no choice but to publish on a 3rd party domain, but with Twitter especially, many many people are creating link-worthy content scattered across that domain (like option #6 above). This is resulting in a massive number of links to twitter.com (see its position on the SEOmoz top domains list) and incredibly well-linked profiles (as internal links on Twitter such as @willcritchlow are nonofo).
[Incidentally, who would have thought that you could really create link-worthy content in 140 characters? Well, it turns out you can - anything that gets re-tweeted is essentially proving its linkworthiness - see below.]
Imagine for a second that instead of Twitter.com, the thing becoming hugely popular was micro-blogging software (a kind of Wordpress for Twitter). So we were seeing the current growth in usage, but instead of being www.twitter.com/stephenfry, Mr. Fry was posting his updates at www.stephenfry.com/twitter - but apart from that, the usage was the same. Individuals could choose to remove the nofollow from their links to their friends and other sites. We would be seeing an explosion in creation of hugely-interlinked small pages - a change to the layout of the internet as big the explosion in blogging a few years ago. We are seeing this change - but it's all happening at Twitter.com.
Increasing use of nofollow for a range of purposes
Google lists the three main intended uses of nofollow as:
- Linking to untrusted content
- Paid links
- Crawl prioritisation (typically linking to yourself with nofollow)
Leaving aside for a second the ability / likelihood of webmasters using nofollow correctly (which means that the search engines need to work even with broken implementations just as they often rank HTML code that doesn't validate), there are two big uses of nofollow that are breaking the model:
- Complete "silo-isation" of large sites
- Domain owner not trusting trusted content authors' links
"Silo-isation"
Disregarding the fact that I just made that word up, there is a very real trend of powerful sites nofollowing all (or nearly all) outbound links even though they are the very definition of editorial links. The site owners have presumably seen the ranking power achieved by Wikipedia nofollowing all outbound links and are trying to form their very own black hole.
Lack of trust
My understanding of the original intent of proffering nofollow as a solution to the problem of linking to untrusted places was that it was mainly intended for situations like blog comments, profile links, etc., where users of your site could create links to wherever they pleased.
This is definitely valuable (as anyone who has ever had to moderate blog comments can attest) but what about once you do trust the commenter? Since so many sites have no mechanism whereby that nofollow is ever removed, we end up in a situation where people are creating huge amounts of really valuable content and the links they create are nofollow.
In my opinion, some of the most "valuable" links on the internet at the moment are nofollow. Some examples:
- The average quality of outbound links from Wikipedia is incredibly high
- Many people are leaving their RSS feed readers untouched and getting their news via links their friends drop on Twitter
- We know many sites whose biggest sources of traffic after search are links which happen to be nofollow (leading to interesting discussion of the effects on the random surfer model)
After posting my provocative theory on Twitter this morning, I asked people to show me some great links via the #nofollowisdead hashtag and I found some things I really wanted to read, including:
- A mention of me I hadn't seen before! From Carlos del Rio (whose book, User Driven Change, is out now - I read it when it was still a draft and I highly recommend it)
- All customers are liars (warning: some strong language!) via Wiep
- Possibly the coolest new blog I've seen in some time - 1000 Awesome Things via Hannah
Obviously this is a dreadful test, but I think it is a strong example - I find many of the most interesting things I read every day via my network.
So what does this mean?
I believe that just as the search engines have acknowledged the limits of webmaster declaration of untrusted or paid content and often downgrade links they believe should have been nofollow, I believe they have to acknowledge the limits in the other direction as well. In other words, some nofollow links should be followed. In the interests of finding the best content for their searchers, search engines are increasingly going to have to use their own (algorithmic) judgment to disregard some nofollows and include those links in their link graph.
Theories are all well and good, but we have also seen the first signs that this is actually happening. We have seen tests where new sites with only nofollow links to them are ranking. The most powerful example was the speed with which a new site (admittedly an exact match domain name) whose first link was from Wikipedia was indexed and began ranking - even before acquiring any subsequent links.
Agree? Disagree? Think it's already happening? Think it's never going to happen? I'd love to hear your thoughts in the comments.
If you liked this, you might like link building is not just for SEO that Rob here at Distilled wrote last week.
I have always been under the impression that nofollow "silo-ing" and creating a domain black hole doesn't actually work in the way that SEO Black Hat says it does in that post. Yes, Wikipedia does a great job with its internal linking, passing acquired PageRank to obscure pages by linking to them wherever possible, but I thought we busted the "leaking PageRank" theory quite some time ago. From what I understand, a page has a certain amount of PageRank that it can give away, but if it doesn't give that away, it doesn't keep it for itself. It's like a gift card that you can't use yourself but that you can give to someone else. Keeping it doesn't help you.
I'm curious to learn more about orphaned pages, minus nofollowed links, that have been indexed and ranked, and I wonder about other theories behind their discovery. Given that Google (but neither Yahoo! nor Live) initially stated that it would not use nofollowed links for discovery, is it coming across this content through some other means? Passed PageRank is only part of the puzzle as to how Gogle ranks pages, so could we be confusing cause and effect?
If I were a search engine, I'd at least use nofollowed links for discovery for all the reasons you state. I'd place pages discovered via nofollow into a category of their own, looking for followed links to those pages after the fact, and developing specific on-page requirements to try and determine their worth. Google's need to stay at the forefront of search technology must mean that it recognises the value of content behind nofollow... it has to acknowledge that the tag hasn't really worked exactly as it would have liked, given how heavily it's used to point at good stuff.
And surely, given the headache caused by this tag being used for so many purposes, the search engines could agree on something that passes no authority but allows crawling? Something that a site like Twitter could use. How about we start using rel="nolinkjuice" and see what happens ;)
Yeah - you're right about the PR leaking thing. I like the post from SEO black hat but don't agree with everything said. One thing that *does* work though is having subject pages on your site that you link to in place of external links (until you upset people enough that they don't link to you).
It's always difficult to extricate cause and effect in SEO and I can't claim any of this is gospel, but I believe that it's not as simple as Google's claim that "we don't follow them".
I love the idea of a different tag - I was wondering about suggesting this but my post was getting long. The various uses of nofollow are *so* different that we kinda need a different option to flag each one.
Isn't that similar to the way that the current XFN protocol operates?
Eventually we may see the personification of relationships developed between websites.
"From what I understand, a page has a certain amount of PageRank that it can give away, but if it doesn't give that away, it doesn't keep it for itself. It's like a gift card that you can't use yourself but that you can give to someone else. Keeping it doesn't help you."
This is one of the best PR comments I've read for ages!
I'm glad that came out okay; I'm going on only one round of coffee ;)
That's impressive, I need sleeeeeeep.
Back on topic though, look what I found searching for 'death nofollow' just now:
https://jing.bloommedia.co.uk/capture/2009-02-23_1125.png
"From what I understand, a page has a certain amount of PageRank that it can give away, but if it doesn't give that away, it doesn't keep it for itself."
But can the page pass its PageRank only to other internal pages on the same site by sculpting with nofollow? Hording the PageRank on the page might not help the page itself, but if that page has nofollows on all external links and most internal links but not on the link to internal page B, does page B get 100% of the available PageRank? If so, hording might still make sense.
Great comment, Jane. I definitely agree on the 'if-I-were-a-search-engine-part'.
Yes, Wikipedia's outbound links are nofollowed, but if the link and content is valuable to Google (or any other search engine), why wouldn't it count it or at least crawl it?
I agree that Google must recognise the value of some content behind nofollow. But even if it is not, PageRank is just one of the many factors of the algorithm.
Will, I have a few criticisms for you (constructive ones, I think), but I don't want to be overwhelmingly-negative, so I will do my best to conceal said negativity... using a compliment sandwich. <bread> The only reason why I have just spent 83 years criticizing your post is because you're so smart and I hold you to a much higher standard than most people. </bread> <meat> I hesitantly thumbed this post down, for a couple of reasons. First, because I had a hard time staying focused while reading it. Your initial statement ("Nofollow is dying.") was very clear, but the post didn't seem to contain any strong support of that claim--instead, it kinda bounced around, talking about loosely-related topics. I think by the time I reached the section called "Increasing use of nofollow for a range of purposes," I had completely forgotten what we were even talking about. I know this may sound like a trivial thing to criticize (especially coming from a rambler like me), but it really felt like this post started out as an argumentative essay... but then fizzled into something more like a general discussion. Second, because you never offered any solid data to support your arguments. Obviously, webmasters have been skeptical about how search engines treat nofollowed links... since the moment the attribute was announced. That's nothing new. In order to justify another "Does nofollow really work?" post, I thought for sure you were going to present some new data. Instead, your strongest support was this: ...we have also seen the first signs that this is actually happening. We have seen tests where new sites with only nofollow links to them are ranking. The most powerful example was the speed with which a new site (admittedly an exact match domain name) whose first link was from Wikipedia was indexed and began ranking - even before acquiring any subsequent links. That isn't enough to convince me. It ignores important details like: 1.) You can never be certain that a site only has nofollowed links. 2.) You can never be certain that a site's only link is/was from Wikipedia. Something I've noticed over the last couple of years is that whenever I register a brand new domain, Google somehow finds it and indexes it. So in my personal experience, I've seen domains get indexed (just the home page) that have ZERO links to them (that I know of). Lastly, I don't think your post accurately quantified or depicted the scope of the nofollow issue. For example, you mentioned Twitter and suggested: "...many many people are creating link-worthy content scattered across that domain..." "We would be seeing an explosion in creation of hugely-interlinked small pages - a change to the layout of the internet as big the explosion in blogging a few years ago. We are seeing this change - but it's all happening at Twitter.com." But let's put this Twitter craze into perspective. Alexa estimates that less than half of one percent (< 0.5%) of internet users use Twitter. By comparison, Facebook is OVER 15.0%! And really... out of the 0.5% of people that use Twitter, how many of them publish ANYTHING link-worthy... EVER? And if they DO publish quality content, what is the ratio of link-worthy posts to crap-that-no-one-cares-about posts? My point is... when it really comes down to it, Twitter's influence on the entire internet (including search engine results) is relatively insignificant. One of the other examples you used is Wikipedia, stating: "The average quality of outbound links from Wikipedia is incredibly high." However, this argument ignores the possibility that following Wikipedia links would eventually cause a decline in their average quality--the very problem that convinced Wiki to implement the nofollow in the first place! </meat> <bread> Your grammar and punctuation skills are topnotch! </bread> So there you go. Those are my thoughts. Now, onto my "real" work...
No lettuce, tomato, mayo, or mustard... That sandwich sounds dreadfully British.
Well played, Slatten.
...dreadfully British. Now now... there's no need to get redundant on me.
You get a thumbs down for that though ;)
Come to think of it... I fully intended on thumbing down your post, but I never actually did. On the one hand, I guess it's not too late, but on the other hand... it's the thought that counts. We know in our hearts where I stand on this post.
Thumbs up for this sir:
"let's put this Twitter craze into perspective."
Twitter is not the second coming of Jesus.
And for the sterling points made in your verbal sandwich of course.
You get a thumbs-up from me because I'm a sucker for the praise bits of the sandwich. In fact, I don't even remember the bad bits any more.
Seriously though - to respond to your specific points:
Appreciate the feedback - thank you for the nice bits and for the honesty of the rest ;)
Are they and tags part of the next HTML spec? I hope so.
Sleep it off, Stephen.
Stupid software. Wanted to say bread and meat tags but it lost them somewhere between Safari and SEOmoz.
Lots of links occur through syndication
Thus there are sites that use Wikipedia content that give links
There are sites that use Twitter content that give links, especially for tweets that use hashtags
That being said, Google don't count all links anyway, even without nofollow.
As an example, from observation I am fairly sure Google treats links to Technorati tag pages with little or no weight, and that may also be true for tag links on Wordpress.com as well.
If they continue that thinking forward in the future, then the use of Wikipedia currently is very similar to using it as tag space.
Alternatively Technorati are working hard to improve the quality of their tag pages with unique content which may reverse the situation.
Twitter not adding juice isn't too much of a problem - a heavily retweeted link will pick up links. Maybe not as many as it would 1 year ago, but it will still get links if it is notable.
The idea of Twitter just being a syndication platform for tweets has legs.
In terms of algorithmic consideration from a search engine's point of view, I'd definitely think that the engines will use some combinations of factors to help identify nofollow links that are worth following and worth ascribing value to. I just look at the life cycle of a story on the web - something like:
I can't imagine the engines wouldn't want to have some listing and weight put on the URL before it starts earning those followed links, so yeah, totally agree that together w/ social media platforms, micro-blogging is making it very likely that engines won't just treat nofollow as "don't bother with this link" in the near future (if they don't already).
Rand,
what if none of those links were nofollowed?
what would the case be?
answer: blog post goes to #1 in 14 mins.
I was under the impresssion that, as Will points out in the post, that a "nofollow" is:
But, that did not necessary mean that it would NOT be followed, if for discovery purposes ONLY, and not pass PR to those pages. I use them only on internal links.
I've looked everywhere for where I heard/read about it, but to no avail. I believe it was a Jane C post last year about her image being "followed" even though it was nofollowed.
You're quite right that I wrote about this last year, here.
I noticed that links to images in SEOmoz blog comments (which are universally nofollowed) were ranking for things from the SEOmoz page that linked out. What was happening was that Google thought the images were embedded in the SEOmoz page. The interesting thing about this is that initially, Google stated that it would ignore nofollowed links altogether. That it saw the linked-to images in SEOmoz's comments proved that they weren't ignored altogether. It was, however, confusing the link for an embedded image. This led to some fun spam (which worked for a while, damn them!) :)
Glad you enjoyed the 1000 Awesome Things blog - I'm a little addicted to it :)
Re your question - we've been talking about it quite a lot here. It occurs to me that because (on the face of it at least) there's nothing to be gained from 'nofo' links - people tend to game them less. Hence rather than being less relevant - I think you could argue that over time, they may have become more relevant than followed links.
For example, I think that now, many site owners nofolinks because they don't want to have to deal with loads of comment spam - and for the most part if you nofo links the spammers leave you alone.
Exactly - though this effect only persists as long as the search engines don't follow them. The irony of these links being the most valuable ones for the search engines until they use them in their algorithm.
From an SEO perspective, I think that in the long run it's going to boil down to what it always has - you want high quality links (but I think it's going to become less relevant if those links end up nofollowed).
I am in complete agreement. I have no qualms with nofollow links, and am seeing increasing amounts of traffic coming through them. The user/crawler effect is inversely proportional by its nature, so I've been on this bandwagon for a while.
OK Will, I'm awake now so proper comment time.
you said:
"
In my opinion, some of the most "valuable" links on the internet at the moment are nofollow. Some examples:
"
Allow me to expand: This suggests to me that is it not necessarily the presence of the nofollow which Google is homing in on, but that sites themselves.
The example you picked are Wikipedia - a pure, unspoilt land of user oriented, user geenrated content which does not spill a drop of juice outside its own walls. This is useful to Google becuase Wikipedia is exactly the kind of site they want us to make.
And Twitter, a superfast social network, where various communities share links and news. As you point out re-tweets denote interest/importance without the need to pass linkjuice.
These two sources alone are much more useful than crawling millions of URLs, many of which have been SEO gamed, and attempting to guess which are important and which aren't.
I think you are right, that Google is taking more notice of these sites than they let on in an effort to take another large step toward their goal of 'quality websites, by the people, for the people'.
I like Rand's life cycle of a web story above, but I think there's another story in the history of nofollow.
1. Blogs get written
2. Blogs get comment spammed to high internet heaven
3. Which is where Google resides
4. And listens to some of the middle heaven blog platform hosts who say, "We're not Pornseos!"
5. Google says, "Fine den, we're most interested in what the real peeps are saying so we give to you the all-good nofollow tag to apply to comments by default. Go and multiply."
6. Comments get nofollowed
7. Other comment spam smack-downs catch up
8. A big one being vetting to go with all that UGC (way to go peeps!)
9. Most spam is neutered without nofollow
10. And "friend" based vetting systems really catch on
11. But still carry the auto-applied nofollow, just to be safe, and keep the "no link spam here" sign in the yard for the hosts
12. Eventually though, high internet heaven gets a little quiet
13. Cause a lot of the Top 10 are now using a lot of nofollow
14. And all the unwashed mass of users in that UGC are pretty much clueless about this other worldly interactions labeled "nofollow"
15. They still hate spam though
16. And only retweet and update and poke their friendz
17. So Google massages Backrub to get it to allow that poking and chicken throwing and tweeting to count as real peeps making low internet interactions
18. Which is what Google wants to know all about
Let me know as soon as the graphic novel comes out. I'll buy it! [:-)]
Good stuff Will. Reminds me of this post about links from Twitter that was published on YOUmoz last week, except I think you've followed the hypothesis through more thoroughly.
The author of that post simply showed that links from Twitter showed up as external links in Webmaster Tools, without looking into the deeper possibility of whether or not that was passing any PR value in spite of the nofollows.
The idea that Google may be looking at a way to utilize the link data for trusted nofollow sites, like Wikipedia, or even crowd-sourcing opinions from Twitter, is certainly an intriguing one. As andymurd suggests they may now be treating nofollow as a strongly dampened link rather than ignoring it outright and I'm tempted to agree.
As I mentioned in my comment on the YOUmoz post, it makes sense for Google to be trying to incorporate links from Twitter for discovery of new pages, especially when it comes to breaking stories as part of the QDF aspect of their algorithm. The speed with which news stories break and spread on Twitter makes it an ideal source for such information.
They may not provide a long-term boost to rankings but if Google finds lots of people talking about, and linking to, the same page that would seem like a pretty strong indication that the page has value, at least in the short term - lasting value would still need to be shown by gaining editorial, followed links over time.
(BTW I'm pretty sure I was the first one to coin the phrase "Conversation Rank" to describe this sort of use of Twitter links in my comment on the YOUmoz post, so if we can all use that from now on that'd be very cool ;-)
When i first starting using the nofollowing tag I thought it was a great idea, but after a while I realized that the primary reason that tag was needed was due to poor site hierachy to begin with.
Now when I build a client site I put alot more attention into the link structure to begin with, focusing link power in the proper direction (and reducing distracting clutter at the same time).
The result is a better ranking site, with increased average page views and higher conversion rates.
I still use the nofollow tag in many circumstances, but not as a tool to fix poor initial site layout.
No tag will ever replace a high quality site hierachy. Nofollow or canonical tag will not beat out a site that is build with a nice flowing hierachy.
We think alike, great post. I think the nofollow dilemma is going to cause more problems for search engines than it solves. The problems won't be technical, but rather philosophical.
It won't look good if/when the news breaks that they disregard nofollowed links from some sites, count the nofollowed links from others, and apply some sort of sliding scale to yet another subset.
I run a wiki for my hometown (richmondwiki DOT org) based on the MediaWiki platform and it does pretty well in organic search. However, I trust 99% of the editors and don't feel that a blanket "nofollow" policy on external links is appropriate. Yes, it does keep out spammers but so does an active community of editors that are eager to delete spammy links and blacklist the users that created them.
Given the active peer review that occurs in many wikis, I would propose that older links that have been reviewed and allowed to remain by my editors should have the nofollow removed after some arbitrary amount of time (14-21 days). If the link is still around after that time, it should become "nonofo" (I like that term).
This solution could work on a variety of sites and CMS platforms where users are allowed to submit content and other users or admins have the ability to edit or remove blatant spam.
While this is all useful info (and is generating very interesting discussion) I've always wondered: what is the context outside of blogs? Is NoFollow intended only for preventing link comment spam, or are the purposes people are putting it to now (trying to prevent duplicate content, trying to stop juice leak, etc.) valid?
(Warning social media tangent below)
The blog strategies I currently use are, essentially, informational and any SEO plus they can bring is all to the good but not required. Since we're the only source for this highly specialized internal information we don't have to worry too much about it (and are leveraging that uniqueness where we can).
Everyone is on about social media this, twitter that, and I'm still uncertain what they're supposed to accomplish. Let's say I'm asking from a staid corporate angle, and not a linkbaiting, juice-attracting one. Has twitter arrived in the enterprise space? Speaking personally I find it a bit informal yet, much like MySpace was 2-3 years ago (Bill Gates, Status: Happy I'm the richest man in North America). I'm not denigrating the utility many of you have found in these services but I'd like to know more about their relevance from a traditional corporate communications angle.
In my opinion: It depends on the size of your company and your target market. On rare occasions, large companies have successfully communicated with their market via Twitter, but it's usually within a "reputation management" context. Some also use it to promote new product releases. Either way, the common ingredient seems to be a brand that is already well-established. A company that is trying to establish their brand would have a much more difficult time seeing significant returns on the resources spent on Twitter. And really... that's just my fancy way of saying Twitter is a waste of time.
Twitter can be really good for tracking issues. In many cases, recently, technical issues have been appearing on Twiter well ahead of other channels. For startups and companies that are tech-dependent the oportunity for communication an problem-solving is great.
Also Twitter has a serious potential news value that isn't used often. TwitPict offers a speed of communication about emerging news that even outstrips blogging.
The simplicity of Twitter means it would be easier (if you are a big enough company) to either:
(a) build (or have someone build) your own version
(b) use a real internal issue tracker
If you are looking for external news, RSS is and continues to be the winner. Unless it's just news about/from your immediate Twitter circle.
Aww now that's not fair, I love twitter, I use it every day...admitedly it's to chat to a list of about 5 knitters who I'm really good friends with and we talk crap all day, but it's cheaper than a phone bill!
I'm sorry, did you say "knitters"? Is that... like... slang for something, or do you literally use Twitter to talk to 5 knitters, as in "people who knit things"?
Interesting post Will, definitely agree about the high-quality nofollow links on Wikipedia.
One of the main reasons I've been hesitant to nofollow internal links is because I believe pages such as About Us, Privacy, Terms, Contact etc help to build trust in the search engines. In some cases nofollow has definitely proven beneficial, login pages and blog comments for example. But if the SE's are blocked from too many important (but non-ranking) pages it may also have a negative effect.
We've always looked at nofollow as a marker tag rather than an exclusion tag. From inception we've viewed it as a warning or indicator to Google rather than some sort of noindex for links.
Nofollow is like saying "Hey Google, we're putting this link in here and playing by your rules . . . now you decide if it should provide any link juice to the link destination." Obviously, nofo currently passes less juice than nonofo, but I would expect it to becomre more of a flagging tool and less of an external link canceler as time goes on.
Also, an internal nofollow might as well be called something completely different. Maybe rel="ignore" or something that tells them we're just trying to provide an additional navigation path rather than influence the weight of a page.
Isn't that what they really wanted anyway?
Will, I think I might only owe you half-that-pint-we-talked about. Good arguement about how nofollow is likely to be disregarded in the near future.
I still don't think that it is going away. I bet the 101 crowd still uses nofollow scheme for several years. Also, it is possible that content factors are going to be come more important to offset the proliferation of nofollow.
The content behind nofollow is still visible in most cases and we have seen test that imply nofollow has always been able to pass anchor text. So, I think nofollow will recede to former levels, but I doubt it will just disappear.
Next time I see you, you buy me one, I'll buy you one and we'll nofollow both... ;)
You are so on! I am going to NOFOLLOW you under the table. Wait... that sounds dirty.
There's no way to follow a bad joke like that... hold on... what did I just say?
Will,
I actually started an exoperiment before the end of last year.. the participants who DON'T use nofollow are all still #1 for their terms after going to #1 and sticking in anywhere from 3 hours to 3 days. (a weekend).
the ones who use nofollow, didn't always get to #1 right away but most of them are on page 1 AFTER a month and a half..
we used 12 people.. and the only optimization done was body copy and an anchor tag within the body copy.
I think Wikipedia is an interesting one as, along with using nofollow, it has strong automated and manual spam detection. This means probably also leads to links being high quality. Wikipedia wants to make sure links from it's site are good for users. With this in mind I would have thought that search engines may be using them to help understand what sites are contextually relevant for a particular topic.
Yes, I think it does make sense to leave even no-follow links. I say that because if you go any webmaster you can see the number of sites that link to yours. So, I believe it's useful. Why else would webmasters list them?
I've been thinking the same thing for a while. From what I've seen of sites with terrible/non-existent SEO, a couple of links from wikipedia and an aged domain are sufficient to rank highly for some competitive terms.
My gut feeling is that Google treats nofollow as a factor that significantly reduces the link juice passed but that doesn't mean that a nofollow link from a high TrustRank site is worthless.
Surely the engines must take into account the trust of a website giving the link and pass some juiceregardless of the nofo. Otherwise it runs against the whole ethos of the web?
There must be some sort of filter otherwise websites not intended to recieve editorial endorsement would recieve it?
Untrusted sites probably don't give as much juice as a trusted site but i'm sure your right Will. The engines must adopt something to combat sectionised websites.
Except Google, I am not sure if other search engines honor nofollow in any case.
Nice Post - Will
I joined SEO moz & I think this is great post for ever. Because currently, to getting rank with do follow link is really aggrasive job for all SEO specialist.
Great Work!!
NoFollow tag is good for ordinary websites in order for webmasters to control what links go outside their websites. They can remove it for the quality websites they want to link to.
On the other hand, as for blogs, I think DoFollow is important to share links with quality readers who keep on following your blog and their blogs are worthy to read. Bloggers can always control who to be followed and whom not and comment moderation is they key to stop the unwanted comments.
good post..
I feel your blog really useful and inspiring me. Thank you.
Well through my own first hand experience Google does follow Twitter/no-follow links.
There is a page on my site that only has Twitter posts linking to it... According to Google Webmaster Central there are 14 links from either Twitter.com or m.Twitter.com with no other links pointing to the page.
"This is definitely valuable (as anyone who has ever had to moderate blog comments can attest) but what about once you do trust the commenter?
My say..
more readers + follow == very good
more readers + non follow == good
I agree
Nofollow has always been open to abuse and therfore slightly flawed...
I don't much look behind the scenes anymore unless the link is something that I'm paying for...
This absolutely has to happen. Siloing is going to kill PageRank otherwise.
But I'm not sure this is happening yet. As a previous commenter said, so many sites syndicate Twitter and Wikipedia content that those links have to be followed one of those sites.
I agree with your assessment that NOFOLLOW is not that important. IMO, NOFOLLOW was stillborn. Imagine if each individual could go an edit their own credit score. How much reliance would lenders place on it?
NOFOLLOW has poorly understood semantics, so even well-intended content producers will not use NOFOLLOW in a uniform manner. And of course, all content producers are not well-intended.
Search engines really have no choice but to treat NOFOLOW as just another variable, among hundreds of of others, that drive their own heuristic calculations.
You right friend.
Imho, 'nofollow' has from its initiation been an idea doomed to fail eventually.
First of all, asking webmasters to distinguish between different hyperlink types (nofo/dofo) for self-benefit purposes (eg. enhanced control of PR flow, and of course reducing blog link spam) naturally leads to SERP manipulation. Google's goal is to objectively rank. Any webmaster's or seo's goal is to rank targeted pages as high as possible for mostly commercial reasons.That's a conflict of interests, and therefore not a good basis for 'cooperation'.
Second, bloggers suffering massive comment spam are only faced with this phenomenon because of SERP objectives by the link spammers, a direct consequence of the flawed nature of basic link-based SERP algorithm our friends at Google invented. Asking webmasters to include link meta data such as nofo means admitting the current algorithm lacks intelligence in distinguishing between relevancy and irrelevancy. If Google was able to properly distinguish without the use of link meta data such as nofo, link spamming was of no use and would decrease in frequency/volume, since it would not lead to enhanced SERPs for link spammers.
And finally, suppose each and every website dofo's to internal pages, and nofo's to external pages. And suppose Google would actually totally ignore nofo links. Then all domains would be isolated, and the core of the PageRank algorith stops working because domains are not interlinked anymore.Hence, the author of this blog post is completely right regarding the end of nofollow hyperlinks, and all self-beneficial corrective link meta data corrective action is futile.
Very interesting post Will. As it's extremely early (for me) my brainpowers are at a low ebb and I'm going to have to read it again later however a couple of initial thoughts...
I agree with your Twitter/microblogging assessment; as a knowledge sharing tool it is excellent. I find Twitter and it's ilk have a bigger impact as part of a network of social media profiles that are all updated in tandem.
"We have seen tests where new sites with only nofollow links to them are ranking."
Are you able to share any of these?
I think on the surface I agree with what you're saying, and I have seen similar trends in the SERPs. I'm off to get a coffee - more intelligent commenting to follow ;)
Nice post. I have always thought the same thing. In my business I can advertise a property on a third party site and then have a link on that site that links to mine that shows mor detailed information that you can not fit on the area they give you to put your property. In my opinion, anyone surfing that would find that link valuable, but they have made it nofollow.
You reference Wikipedia links being nofollowed. While this is true. I've seen where third party sites syndicate wikipedia content and these links are followed.
For one, Squidoo users can add a wikipedia page to their lens.
Interesting - time will only show!
I agree this is already happening, at least to some degree. The Wikipedia example is interesting, because it does show that the links are being followed. However, we have no idea how those NoFollow links are being treated. On my blog, I posed some hypothetical ways the search engines might be treating all this NoFollow data. It's all speculation, of course, but it does give some insight into the possibilities: https://www.seo-writer.com/blog/2009/01/14/how-is-nofollow-data-used-by-the-search-engines/
Wow posting at 1:30 AM is called DEDICATION! Good post.
It's been a while since I went after links that have the nofollow tags on it. I see the links appearing in my Google webmaster tools but I doubt it carries much weight.
Search Engines should follow 'nofollow links' that are coming from authority websites. A lot of authority websites may choose the 'nofollow' tags because they look at it as a disclaimer. Basically, 'click at your own risk'.
The search engines need to come up with a new tag.
1. NoFollow 2. Dofollow
How about tags for
3. Affiliate 4. Reference
If i'm writing about a topic, and would like to 'reference' a site then I should be able to put it as a 'reference' so more juice is given.
An affiliate link is just an affiliate link, and shouldn't carry much weight. Placing links in the footer should carry this tag, but tags within an article should carry a reference tag.
1:30 am in Seattle is what, 9:30 am in London? willcritchlow probably posted this right after the "Full English".
Cheers!
Heh. Yeah - it was after coffee here, but unfortunately no full English this morning - had too many of those last week ;)
some nofollow links should be followed
yes, but unfortunately the theory is against our wishes...
I think we should trust standards and what SE says.
I've always thought nofollow wasn't so black and white. Great thoughts all around here.
Hey Will, "All customers Are Liars" is a blog post from Ittybiz and not Wiep. I don't know if you intented to link also to a post on Wiep.com site
I meant 'via' Wiep rather than 'from' Wiep - he pointed me to the post. I'll update my post for clarity.
I follow quite a few people on twitter, but half of it is spam. "picked kids up, now dropping them off..." like I care. I try to post snippet links with useful info. I also agree with that other poster who said the article is confusing, I'm not sure what he's talking about - title misleading, then woffles on about twitter and wikipedia.
A way to know if a nofollow link should carry weight is by analyzing the CTR of that link and by analyzing the user experience on that web, including bounce rate and time before user leaves it.
Also the SE can check the number of incoming links from others source or from the same source (assuming that the Spam Filter works).
I'm starting to think the whole thing is one unholy mess. Good intentions and all but a lack of foresight that paints everyone right into a corner.
Vern
I'd like to see if this wollofon knil is worth anything
The trouble is, because you're limited to a certain number of characters people will try to squish more actual content in a microblog message and the first thing to happen is the URL will be squished using a service like TinyURL or a.gd.
And while not everyone will want or need to use a short URL service, thanks to client-side apps like Twhirl and TweetDeck who put the facility a drop down menu or mouse-click away, they'll eventually become the defacto standard.
Granted we're only talking about Twitter here, but since it's the Big Cheese of the MB world and most other MB platforms are following the 140-character limit, what happens with *it* will have an impact on all the other platforms.
So for the most part, I think when it comes to nofollow links, it's a non-issue with Twitter and the other MB platforms out there.
AMENDMENT: Tweetdeck recently added automatic URL shortening according to this page. Even more reason to not think about nofollow.
This may sound like a dumb question, but how does TinyURL work in regards to linking? If the blog post I want to link to is OK for follow - is the TinyURL automatically a nofollow?
Or will the SE disregard the nofollow and go ahead?
https://www.seomoz.org/blog/the-benefits-and-pitfalls-of-url-shorteners
sounds to me like there's just too much information, data, personal opinion, crazy and wonderful ideas and such being kept from seeing the light of day with all these nofollows.
perhaps google is going to have to come up with methods to find and index a good part of all this valuable nofollow material or maybe it already does, but doesn't assign value to it? yet.
I'd say this calls for a good old fashioned test.
In my personal opinion, the nofollow tag was flawed from the very start, whether or not the intentions for its introduction were right or not.
As you mentioned above, the use of nofollow has gone way beyond what I think Google ever saw its implementation being used for, and the mis-implementation of nofollow on sites certainly doesn't help matters there. Like most things, overuse of a particular tactic doesn't neccesarily help things, in certain cases it comes back to bite you in the ... (as I think Kev put a littl better)
Nofollow used correctly has its place , however it is by no means fully fit for purpose in its current guise. As such, I would have to agree with you that perhaps some bending in the rules occurs as Google attempts to evaluate whats good and whats not....
nofollow was a band-aid fix, and no band-aids last forever. now rel="canonical" on the other hand....
Google shot itself in the foot with this no follow blanket, didn't that Catt Mutts bloke tell twitter to add it recently though? So it will become no-follow is really follow on certain sites, but we're not going to tell you about them because we are Google.
I think a lot of it is relative.
If a particular site has 1000 no follow backlinks from trusted sources and another site has 300 of the same then Google will inevitably still select the former as the most poweful. It just has to be that way. At the same time one with 1000 follows will bet the one with 1000 no follows in theory.
So I dont actually think it is as complicated or a major concern to search providers like Google.
I dont use no follows for external links, I think if people have to use this a high percentage if the time on their external links then they should probabaly find another solution. i.e do it by hand the old fashioned hard working way. If you get so many posts that you cannot keep on top of it, then employ more editors.
I do use this no follow tag on one link only, an internal page and that is simply due to layout and the notion that the first link to a given page is the one that anchor text is selected from. This about us page is very important closer for us and so we want it noticable, we dont of course want to compete for the keywords 'about us' The page is still able to be optimised for other terms so we want to link to it with those.
I hope that worst case scenario would be no ranking for that page because of this lack of trust tag, despitethere being other links to it but so long as it is still there forour visitors then so be it. nothing lost.
Im not 100% comfy at using this tag tho. The ideal solution for this would be to create a tag that said 'no follow anchor text' The perfect solution in this scenario. (from my perspective anyway)Don't ya think?
This is on Sphinn at https://sphinn.com/story/102635