When Google announced it would support the “nofollow” HTML attribute back in 2005, I was pleased. Blog spam was (and still is) a major problem. The invention of “nofollow” took a lot of bite out of the usefulness of comments for spammers. It seemed like the right move at the time, and looking at it now; it still makes a lot of sense. I can’t imagine how bad things would be without it.
For the unfamiliar, “nofollow” is an attribute that can be assigned to links on websites. This attribute, when recognized by Google, causes the link to be ignored in Google’s index, thus preventing it from contributing to the sites “PageRank” score on Google. Therefore, in theory, leading to more relevant search results.
Here is my gripe, when Google announced this feature, it did so under the stated guise of “Preventing Comment Spam,” but it’s turned out to be abused. There are those who would argue that “nofollow” has become a tool that creates an unfair balance where higher trafficked sites don’t share traffic with lower-ranked ones. Now I’m not saying that “nofollow” should go away but that it should be used less. Why am I complaining? Because three of the sites I use most often implement “nofollow” in places that, while well-intentioned, work out as unfair. These sites?
For starters, Facebook’s usage makes no sense. Facebook has several checks to prevent spammers from joining the service and multiple ways to report it when it happens. I am not saying they are perfect, but it has very much maintained a strong and effective walled garden. Considering that Facebook has pushed users to make more of their content public (while also having major issues with privacy), it is downright unfair that they will not allow a PageRank incentive for this.
At first glance, however, the inclusion of “nofollow” does seem logical for Flickr and Twitter. The ease of access to these sites makes them obvious targets for spammers. In the case of Twitter, a platform with a substantial spam problem, “nofollow” seems like a practical solution to de-incentivise spamming. At least in the short term.
Despite this reasoning, I still believe it to be the wrong solution.
Flickr is a publishing system, and some users use it as a blog with very long written posts to accompany their photography. Yet Flickr automatically attaches “nofollow” to any link posted anywhere on its site, even in the user’s written content.
With every photograph I post on my Flickr account, I include a link to a related blog entry whenever possible. These things directly relate and semantically should be linked. My analytics also clearly show this has been effective for increasing my readership. A significant portion of my traffic comes from these Flickr links. Yet, Google ignores them because of the “nofollow” attribute. The same with Twitter.
Doesn’t this go against the whole spirit of “nofollow”? In my opinion, it is lazy and unfair, especially as a Flickr Pro user. I’m paying for a service that is going out of its way to prevent me from getting PageRank from it. That’s a bunch of crap.
There is another way to look at this, though. One could argue that PageRank is doing the opposite of its intention and hurting Google’s relevance. The three biggest traffic sources to this site are Twitter, Facebook, and Flickr; these account for almost 50% of my traffic, however, Google’s mysterious PageRank algorithm ignores these sources, thus making it harder for my site to gain relevance in search results. Isn’t this leading to an inaccuracy in what’s “relevant?”
“nofollow” is used almost ubiquitously by any service that has an easy sign-up system. Therefore it is creating a tiered system on the web that takes away power from user-generated content and gives more strength to those who have a greater technical knowhow. There are many out there who will never understand how to set up a blogging platform such as WordPress or MovableType but can easily get a Twitter, Flickr, or Facebook account. Why should their voice matter less to PageRank? And on the flip-side, why should a major media outlet matter more? Isn’t this working against the democracy of the web?
How do we decide what is signal and what is noise? Surely just because something is easier to do doesn’t make it less relevant, does it? Can’t we come up with a better technological solution that empowers users not punishes them for their lack of tech-savvy?
5 replies on “‘nofollow’, I Don’t Follow”
A great post . . .
Keep in mind a few things.
1. PageRank is one of many factors now that Google uses to position a URL in their search results. Google has stated they have HUNDREDS of factors that they now use. Don’t get too hung up on ‘PageRank’. Know that other factors exist such as Google Toolbar data, clickthrough data from the search results, previous searcher activity (did they click on that domain before, do they typically look for ‘Golf’ the sport or ‘Golf’ the car, etc.).
2. Google is very good at launching things and making a big to-do about them when they launch and then change something later without making any mention of it. Just because that is the way nofollow worked in 2005 doesn’t mean that is the way it works today. Personally, I feel it’s pretty close to the way it works today but I have some pretty strong evidence that Tweets from large twitter accounts still have an impact on SEO.
3. Just because you may not get a benefit directly from Facebook or Twitter doesn’t mean you don’t get a secondary benefit. I did a post about this on my own blog (well, really just a Posterous, shorter more random thoughts). Tweets get fed via API to lots of locations on the web. Blogs, websites, etc. When that is done, the links are ‘followed’ links and thus an advantage is had at that time (it is also why ‘good’ spammers are still pounding the twitter stream with links).
4. People on Twitter and Facebook are early adopters (well, less so now than a year ago, but . . . ). They are well connected on the internet in comparison to others. When they see your tweet, they are more likely to link to it from their own blogs, websites, professional sites, etc.
5. Lastly, pulling PageRank directly from Twitter.com (with a HIGH number of links and overall domain PageRank) would really screw up the algorithm and be really beneficial for spammers. Keep in mind, I used to be a spammer. I’d setup 35 profiles online and drop links on newsgroups, forums, blogs, etc. I even went further than that and I figured out how to automate Amazon’s So You’d Like to . . . and Amazon’s Listmania lists . . . later learned how to automate product reviews. All of those things REALLY helped me to do well in the search results for the products I sold. Removing the significant advantage spammers have (to your point) makes it less worth a spammer’s time invested to figure out how to crack the code. As for Facebook . . . I work in the Search Engine Optimization industry. Make it possible to gain significant PageRank advantages by setting up a fan or profile page that has 50,000 fans or friends and . . . the SEO industry will figure out how to exploit it.
There is a TON of money in getting to the top of Google. How much? Do a search in Google for ‘Google Keyword Research Tool’ (while logged out) and click on the first result. Then type in a fairly common word. Export the data to CSV and open it in Excel. Do the math between the traffic volume for the different terms and the ‘cost per click’ to pay for a top position in Google AdWords for that term. You will see it is often times hundreds of thousands of dollars per month. Do a search for a 3Ps term (porn, poker, pills) and that skyrockets even further.
This stuff makes the livelihood of many people and organizations possible. Not just the spammers. But the legit SEOs (like I have become sense 2005), the legit advertising firms, legit marketing departments, and the people that support them from the janitor to the CEO.
Again, great post . . .
Brent D. Payne
SEO Director, Tribune Company
Thanks for the info, very useful
well said John, I can certainly see your point, everyone hates spam and honest seo’s wont use it. That said checking through the code of this page shows that your links carry the ‘nofollow’ attribute. That however is probably not your fault, as most blog software carries this attribute as a default in any case. Other than that, Well Said I say!
I can’t agree more with you: nofollow is useless!!!
Have a look at the paradox I explained on Matt Cutts blog: http://www.mattcutts.com/blog/quick-comment-on-nofollow/#comment-243379
Google should make its own way to discriminate what is spam and what is not and not rely on nofollow that immediately has been abused by webmasters so much that now the same Matt Cutts came out with another page on his blog saying that webmaster should not abuse nofollow, an that they should let PR flow, and blah, blah: http://www.mattcutts.com/blog/pagerank-sculpting/
So before we should use nofollow, now we should let PR flow. That really sounds weird! It sound like Google algorithm is getting either overloaded and not bale to follow all links and assign the proper PR that’s why PR is not so important anymore (according to Matt Cutts) that we should let it flow.
The funny part is that Matt Cutts blog itself is packed with nofollow links, even many smart comments form smart authors still get a nofollow link.
Have been doing some research on ‘nofollow’ and what you have said makes perfect sense. Some of these rather large sizes need to re-think how they treat links to other websites, otherwise before long the internet is going to be dominated by few companies that can instantly have an advantage. Nicola