SEOmoz Whiteboard Friday - Dangers of Nofollow from Scott Willoughby on Vimeo.
By: Scott Willoughby
Whiteboard Friday - Dangers of Nofollow
Whiteboard Friday
The author's views are entirely his or her own (excluding the unlikely event of hypnosis) and may not always reflect the views of Moz.
We're big fans of using 'nofollow' for linkjuice sculpting around your site. If you know what you're doing, you're careful, and you're considerate, it's an incredibly powerful strategy that can have a big payoff. But what if you make a mistake? If you don't pay attention, or you go about it willy-nilly, site-sculpting with 'nofollow' can cause some major problems...and that's what we're looking at in this week's Whiteboard Friday.
A great look at how to appropriately use a 'link condom' as Rand put it. Thinking through to the effects of all pages beyond the target of no-follow is obviously critical, but something I've seen overlooked.
Even though, having a WBF up on Thursday made me a little dizzy :P
I totally dug the "link condom" refernce as well. :) Nofollow is one of those techniques that is easy to screw up. Lots of great info in this video, and I'm definitely going to look back at a few sites I've worked on to make sure I didn't block something important by mistake. Great info as always!
Great vid Rand.
Quick question. If I have multiple links to page Y on page A. But only one of those links use the correct anchor text while the others say something like "more" or "view details" etc... Will the SE's ignore the nofollowed links but still use the followed link to pass juice through?
Is there a benefit to this with regards to keyword rich anchor text or am I misunderstanding the whole thing :)
Thanks
Are you asking if anchor text would flow through a nofollowed link? If so, I think the answer is no (at least from our testing).
Rand is correct!
Google Guidelines
How does Google handle nofollowed links?
We don't follow them. This means that Google does not transfer PageRank or anchor text across these links. Essentially, using nofollow causes us to drop the target links from our overall graph of the web. However, the target pages may still appear in our index if other sites link to them without using nofollow, or if the URLs are submitted to Google in a Sitemap. Also, it's important to note that other search engines may handle nofollow in slightly different ways.
Not exactly.
An example is a category page listing products. If each product has three links to the product detail page:
- Product Title Link - "Big Widget"
- Product Image Link - "Big Widget" (alt tag)
- View Text Link - "View..."
If I nofollow the Product Image link and View Text Link. Is that a good thing for passing the right anchor text as there is only one link not nofollowed.
Thanks
I believe the bots used the first Link/Anchor text combination they find that isn't nofollowed
So if you have 3 links...but the view text link is at the top..then the bots will use that link first (if it's dofollow) and ignore the other 2links.
If you have 3 links that are going to the same link...and nofollow 2 of 3..then the bots should use the 3rd link which is dofollow.
Nofollow is used to stop pagerank/link juice AND anchor text.
The purpose of nofollowing the 'view text link' and 'product image link' will be to show the bots that the correct combination of anchor text/link is the Product Title Link 'Big Widget'
hope that makes sense
Rand, care to weigh in on the whole "3 links on the same page that point to the same url" thing?
I agree with the point that the Contact page can potentially have useful links that should be considered. However, there's an added benefit to nofollowing a contact page. Namely, removing it from searches can make it less likely to be spammed.
Granted, that's a non-search-related benefit, and there are better ways to prevent spam, but I'd say it's worth considering.
Yes I agree on some part but I have to say I would never nofollow a client’s contact page.
I want the contact page to be ranking when the email, telephone number or address is searched. To me that’s an important one..if the customers cant find the companies contacts details easily or worse they are directed to some big business directory instead of the actual website is bad news :{
But deffo a big boo to spam...!!
Well, to play devils advocate here, nofollowing the contact page will probably not have much effect on spammers, because they likely look for hosts from search engines (domains) then scan each domain for regular expressions that resemble email addresses or forms... You might be better off using a javascript concatenation function to "break" pattern recognition.
Rand, this video post is another confirmation IMO that anyone wanting to implement nofollow site-wide needs to first construct a data/web page model or at least use a database modeling tool. this would include the full construct and visual of how pages link to each other.
Visualizing a 10k, 50k, or 300k page website and constructing how to use nofollow effectively can only be done wisely with this type of implementation.
In fact, I will not do this for clients for large scale websites unless we/they first do the above. it's just too risky handling it otherwise.
Excellent thinking, Brian. If you can't model your link flow structure intelligently and in an easy to understand way, you probably shouldn't be applying nofollow to link sculpt in the more "risky" areas.
great - would u recommend any crawler+modeling tool?
This sounds tempting! Can you give some more info on automated way to do that? Any technique?
I'm pretty comfortable using nofollow on some of the "low risk" pages that Rand mentioned, where appropriate. But seeing as how most of what I do falls in the "consulting" category and I don't have very much control over who is doing what to various sites, I would be terrified to go about using any of those "high risk" ones! I can't imagine all the unintended effects that could cause if an uneducated client/developer/someone messed with it. Yeesh.
If you were to Nofollow your site's "utility links" (ie warranty, my account, etc) should you also add Noindex to keep them out of the index or does it matter?
Great video btw.
It might be very cool to see how a real life example website or even a fake website can use the nofollow, meta noindex/index values to craft benefits for one page... with a look into typical warning signs and issues that someone should take into account when sculpting link juice flow.
I don't think it matters. I have our site map page as a nofollow footer link with no other links to it and it doesn't rank for "company site map" - our homepage and another page show up #1 #2 for that.
Nezzo - I don't think you want to use noindex unless you have a very, very good reason for wanting that page out of the search results. The problem is that if someone links to those pages, or they start to gain in importance/attention, they won't appear at all. You never know when something might be relevant/valuable/desired. If you're pretty sure it's unimportant, nofollowing links to it is OK (to conserve for more valuable pages), but noindexing is drastic and I wouldn't recommend it unless you actively want that page out of the results.
Thanks for the clarification Rand.
What if you choose to noindex and nofollow duplicate content? isnt it better to noindex duplicate content pages in order to know how much of the good content is indexed?
Hello,
Regarding pagination, in order to give more link juice towards product pages, would it be a good (risky) idea to add more product links on the first page when the user agent is a search engine ?
i will be watching this video later in the morning, but i think someone was a bit too keen to submitting this vid. [april 30th??? GREAT SCOTT!]
I usually try to post them around 11pm Thursday night so the new WBF will be available for our readers in Europe...11pm in Seattle is 7am GMT. They're from the future.
Rand, I have a few points to add about your second diagram:
1) The three lost pages linked from Y are only lost if they are only linked from Y. Otherwise if they are a part of site navigation or linked from other places they are likely to be indexed and found. For that not to happen, our nofollowed link to page Y would have to be a very important one (when it was followed).
2) Page A links to page Y and page B, page B essentially gets a direct link from A and also three links from C,D,E which are getting their link juice from page Y which is getting its link juice via page A. Unless other pages link to C,D,E isn't that just a two level filtered and split link A juice that flows to page B already anyway?
I realise that there could be some additional factors of benefit here such as viariety of context in linking and also anchor text. But....
If the answer is "YES, they are still beneficial", that would mean that every new page adds a bit of its own fresh PR value. Essentially this could mean that a website with content alone and no external PR could generate some link juice on its own.
I am very curious to see what you think about this.
Interesting video!
What about using nofollow attribute as a link dilution control technique?
For example if you have 2 or more links from Page A to Page Y, you could put nofollow on all links except one –the one which includes the keyword you need to target- in order to maximize this link’s value and the ranking of the target page for that specific keyword. On the other hand you could leave 2, 3, or more keyword targeted links as dofollow, however the actual link juice will suffer from a dilution of 50%, 66%, or more for each individual link.
I guess it’s the topic of how to fine-tune your internal page optimization with nofollow...
really have never understood the big hoohaaaa about this unless it is desired to be used to totally keep private pages out of an index. In those cases there are better ways of dong it anyway.If u dont want to endorse spammy links on your site no follow them, if you dont want unneccassary high pr on privacy notices etc that may lead to poor traffic then no follow them.pr sculpting is what we have always been doing anyway, 'popularity sculpting' The more important pages have more links to them naturally both hopefully internal and external.It was brought out to combat the first situation, spammy links or pointing to bad neigbourhoods for various reasons. Then it opened up a large debate because 'people' implied that perhaps a website with product pages 6 layers deep may not be able to flow much pagerank to these deep pages and by no following certain paths they would be able to flow a little more juice to these areas.Again, it is quite obvious that there are better ways to do this, if a product page is popular and useful point more links to that product page internally and externally. External links in particular don't give a flying f about site architecture so you can get higher pr without doing anything at all on your own site.I use the no follow on my terms and conditions type notices. No 'value' in search for my target audience. simple as that. (actually it is also plain text so doesnt point to any heirarchy)So why didnt I just block them via other methods? Well, altho they are not important in a search sense something tells me that is sensible to let 'engines' know that they exist. (hey, it may be one of 200 algorithms nowadays) From all accounts it seems that using this tag 'does' show that they exist so it suits me at this point. I would suggest that it is a very tiny percentage of website owners that use any sort of robots text. Thats scarey stuff for some of us without initial codeing backgrounds. To me, the only other 'concern' that people may have is probably related to whether they should bother aquiring external links that have a no follow attribute. With the general belief that it will slow down a lot of spam posts because they belive this is true then that can only be a good thing. To us that get links in useful places have always been doing so because of the relevant traffic, absolutely nothing to do with page rank.So ermm, yes. I just thought it was quite straightforward basically but maybe Im missing something, wouldnt be the first time.
Very interesting video.
But, I'm still wondering why I absolutely wouldn't that my Terms of service pages being indexed by search engine? I think that I'm not understanding all the inconvenients...
Anybody can explain?
It's more a question of why would you want to give linkjuice from every page on your site (assuming your ToS is linked to in persistant nav, as is common) to the ToS page? Is it a valuable conversion page for you? Does it contain content that will get you good search traffic? Which of your keywords--even longtail--is your ToS optimized for?
ToS isn't normally a valuable page from a search or conversion perspective, thus the argument that you shouldn't waste linkjuice giving is followed links.
Thanks a ton, Rand. Your posts are always thought out well and flow very nicely. Have a nice weekend!
Here's Matt Cutts Video released today about using Nofollow
https://www.youtube.com/watch?v=unEML5n5vTo
'nofollow' in links to a disclaimer, privacy statement ...
What is "Secret Page Rank Scaling Factor is 42"?
It's the answer to life, the universe...everything. https://budurl.com/rgmw
Very interesting topic. I ran into huge problems using this inappropriately on our site which saw my site almost de-indexed entirely until we removed all these. So I would ask -
I'm glad you showed low/medium/high risk. I think too often pages fall into the medium/high risk category that I have no followed attempting to link sculpt.
I think less no-following is definitely more here!
Good video, thanks.
Great video Rand.
Firstly to anyone interested in getting their heads around PageRank sculpting and how internal PageRank flow works and where PR comes from in the first place, I would strongly recommend reading Darren Slatten's "Professional's Guide to PageRank Optimisation" over in the SEOmoz guides section. Seriously, one of the best SEO documents I've ever read.
Secondly, a question for Rand. You mention towards the end of the vid that nofollowing "About" and "Contact" pages can be risky because people may want to find those pages. Can you be more specific on a scenario for that? First,, I'd say those pages will get indexed anyhow if you link to them from an HTML sitemap and include them in your XML sitemap, and they don't need much link juice anyway because the terms used to search for them would be pretty long tail anyway (eg "How do I contact SEOmoz?"). Also, how many people would look for that kind of info directly via search? I'd say most people would probably go to a homepage and look around for an "About" or "contact" button. It would be great if you could elaborate on this, as I generally go ahead and nofollow these links too.
Edit: forgot to include link to PR optimisation guide...
Good topic. It would be nice to see some more not as simple of examples as you showed in beginning, but I guess that would be giving your secrets away.
Great video! I just made it through the Professional's Guide to PageRank Optimization and this video answers a couple of the questions that I had. I am still curious about real world application.
Let’s say you have a page structure like this:
Widgets > Round Widgets | Square Widgets | Triangular Widgets
Round Widgets > Little Round Widgets | Medium Round Widgets | Large Round Widgets
You find that the Widgets page is scoring higher for the search "round widgets" than the Round Widgets page, and you would like the Round Widgets page to be driven up in the results. The Widgets page just houses a small summary of all three kinds of widgets, while the Round Widgets page has all of the good info about your product. Your navigation is setup as an Include using Unordered Lists and every page links to every other page on your site (via CSS-enabled dropdowns).
Knowing that there are other followed links to the Widgets page other than the Nav (for instance textual on-page links, the HTML and XML sitemap) and there are the direct Nav links to the Round Widgets page... Does it make sense to NoFollow the link in the navigation to the Widgets page?
Perhaps you can provide some scenarios and examples of how you use the Nofollow tag beyond Terms and Privacy pages?
Rand, great video,
An e-commerce site I am optimizing had many filtering possibilities resulting in far more pages indexed than the actual number of products available in the website. The solution was to nofollow these combinations of filters, and I am waiting to see the result. From your experience how much gain can be made when you start making it easier for search engine to focus on you main content? I am talking about a website with about hundreds of thousands of products.
Thanks for the post Rand.
Ik do realize, that in the given time you just can't cover such a complex topic. You're doing fine in the guides though :)
But I missed a few good things (low hanging fruit as you like to call it) IMO.- What about nofollowing from most but not all pages? Great possibilities there... It lowers the risk as well.
- What about a nofollowing-scheme for small sites with keywords only on the main page thus: Home links to all, all other pages link to none except home.
Thanks.
As for nofollowing most but not all links to a page - if enough juice is passed the page will rank.
We nofollow our sitewide privacy policy footer link. On our About Us (which has juice) I have an in-text follow link to the privacy policy. That was enough to get it to rank for "company privacy policy".
Right. That's exactly what I mean.
Great possibilities to leave the bulk of the link juice for the more difficult ranking pages.
Good video Rand.
Scenerio: I own a site that primarily makes money from affiliate links. So all my outbound links to these other sites from my page go through an internal tracking peramiter say /tracking/outboundurl.
Do I need to No-Follow all of these links even if I have a disallow /tracking/ in my Robots.txt?
If you do that then you will pass no "Link Juice" to their sites. If you setup your tracking code to redirect using a 301 the juice will get sent to thier site.
Rand is really only discussing the internal links from page to page on your own site. Not linking from your site to another external site.
Rand, You listed the Privacy Policy as a low-risk page (and I completely agree), but I've known a few SEOs who recommend keeping those as followed links, because search engines might use the Privacy Policy as a trust signal of some kind. Have you heard anyone speculate about this? Do you know of any patents or research papers that support/debunk this theory?
A privacy policy would be a domain indicator though wouldn't it. So nonofo (ie "follow") linking from every page to you privacy policy is over the top as long as the privacy page is indexed then it should show as a plus, IMO. That said, I can see how linking it from every page might be better for users.
Just the presence of a privacy policy doesn't mean anything of course your policy could be "We will sell all your details to the highest bidder, and then resell them to the next highest, etc.; this site will install software to your computer including trojans, keyloggers and web trend reporting mechanisms.".
It would take human intervention (possibly even lawyers) to establish whether the privacy policy was benevolent.
Never heard that before, Darren, but it's an interesting theory. Seems like a long shot, but I suppose you never know what sorts of correlations the engines might find with content/links on the web.
That seems like something that Google would add in webmaster tools if it were true...very interesting theory
This is the exact question i posed to a focus group of mine. Would the engine still read the anchor text to see that you do in fact have site credibility indicators and pages. Or would it be overlooked entirely??
Nice video, I always love watch your explainations ;)
Question 1 - Have you done any tests, that aprove your statement about not crawling page with "nofollow"?
Question 2 - Why pages C,D,E are "bad" because of linking from page Y? I can't understand that. Moreover, what if I linking to those pages from another pages?
I've done some tests about nofollow, but your video is telling me something new about that and I want to better know your thinking ;)
well said Rand, I am looking forward to put some of low risk into practice
Interesting video, thanks.
I'd also like to know about the use of nofollow on external links and if it's there any risk with it if you systematically put it for all of them. If theoretically must be used like "I don't trust that website" and you put a nofollow on a link to a website that Google considers trustworthy, might be suspicious.
If you puts nofollow to the link, you are telling that page is not thematic with yours. That's all. If that page is not thematic, Google won't give any benefits by your link (high rank).
shpyo - I disagree - it's not about "theme" and I'm not a big believer in "themed" links as a major ranking factor anyway (too many examples of sites/pages ranking with theme-agnostic links). Nofollow is just a way of telling the engines not to use the link as part of their link graph metric calculations (PageRank, TrustRank, anchor text, etc). Essentially, you're telling the engines to pretend there's no <a href> tag attached to the text/image.
Excellent explanation, Rand, of external nofollow! Worth repeating:
Internal nofollow gives webmasters control over link metrics at a granular level. As Brian Gilley and Rand comment, do not sculpt w/o visual "control" or full evaluation of site hierarchy and page relationships. Don't go granular until the bigger picture is clear.
That was a good one Rand. Excellent explaination. No space left for further comments. :)
How about a nofollow to an Internal Link that is penalized? :)
example
Subpage 1 has an internal link that points to Subpage 2.
Subpage 2 is penalized.
We should try to look for why Subpage 2 is penalized, and try to fix it before doing any major changes...
but..
What if we can't figure out why Subpage 2 is being penalized..
Should we
1. Put a nofollow on the link that is pointing to Subpage 2 from Subpage 1
2. Remove the link from Subpage 1
3. You dont believe that interlinking with penalized pages is going to affect the website or other pages on the website.
Hi Rand,
I think today everybody puzzled a little bit.
You said that using nofollows in category page links are high risk nofollows and can make big problems. But which problems? why?
In the beginning of the video you mentioned that even a page has many links, at the same time having nofollowed links can make this page be removed from the indexes of SEs. Here I was confused. Most of SEOs say that having links from good sites ( like Wikipedia ) is a good link even if they are nofollowed. What do you think about it?.
Actually my strategy is nofollowing some of low risk pages as you wrote. I also nofollow some of links - Logo in header, Navigation Links, etc. Instead of these links, I make a dofollow links with the anchor texts in the footer. Rand, I use this strategy, because I read it from you in one of your posts (or maybe it was Q+A, i cant remember). Can you explain this situation more clear please?
zevu - if you employ nofollow as a link sculpting strategy (essentially, a way to maximize the amount of link juice that flows to particular pages on your site), then this logic applies. If you're talking about external link building, it's not particularly relevant.
What I'm saying here is that certain places on your site can be dangerous to employ nofollow because you might have unintended consequence due to the second and third order effects of a recursive algorithm like PageRank calculating link popularity on pages downstream from your intended target. When you're using nofollow to sculpt, make sure you don't accidently cause a page to lose juiec that you actually wanted to gain.
Hope that helps!
I hope, SEs will not count nofollows on my navigation links as a negative votes when I will try to sculp the PR.
Thanks Rand!
SEs don't count nofollows as a negative vote, actually I think Matt Cutts addressed this somewhere and stated that when a link is nofollowed, it's just ignored.
Rand's Post With Matt Cutt
"There's no stigma to using nofollow, even on your own internal links; for Google, nofollow'ed links are dropped out of our link graph; we don't even use such links for discovery. By the way, the nofollow meta tag does that same thing, but at a page level.)"
edit: Added Link
Hey Chenry,
Thanks for the link,
I have just found the answers to the questions that were flying in my brain, those I got after watching Rand's video.
--Vusal
I have noticed this myself and I think some companies are over doing it when it comes to the nofollow.
Personally I think you have hit the nail on the head and agress with you on pretty much every aspect however something that Darren was saying about the privacy policy does make me think google could find that useful for perhaps trust issues.
just another thought on that....
There are too many factors that have nothing to do with the content of the page the helps a site rank.
Content (privacy page) is great, but you can still have a lot of quality links point to you w/out having 'quality content'.
There are many websites that rank w/out the privacy page. Having one or not having one shouldnt affect rankings. IMO
If anything...I believe the 'About Us' page may help since it will help the robots determine what the website is all about.
Either way..the robots still crawl the link..so they'll know if you have a privacy page or not.
When it comes to using the nofollow tag internally I preach "KISS" Keep It Simple Stupid... like you mentioned I have seen people go crazy overboard with the nofollow tag .. just keep it simple kids.
Where is the enbed code for these great weekly videos?
Should be a button for it in the corner of the vid when you play it.
Interesting post. I recentely wrote down a small articile about link building and nofollow tag on my italian blog.
Is it not strategically better to manage the robot's journey by utilising robots.txt file, this contains all rules in one place.
How about "nofollow" on your logo in the upper left? It's just an image...why not create in your site footer a descriptive text link to your home page?
I too have thought about this question...
Why have a link to your homepage be followed on every single page of your site? With the sites that I manage, I have seen no ill effect by nofollowing the logo or home image, but these sites are larger ecommerce sites that have lots of incoming links to the homepage.
I would love to get some more info on this question to see if it would affect a small to medium site.
Right, more info is needed for this issue
The logo link is not just an image, it appears to be using a variation of the most popular image replacement technique FIR (see my review of image replacement techniques, https://alicious.com/2009/new-css-image-replacement-jir/ , and suggestion for the best, one modified by me I term LLJ).
FIR requires an extra element, usual span so that the title text can be hidden and an image shown instead. This technique is used extremely widely despite having some flaws.
Here both the seomoz and the "looking for talent" titles are display:hidden and the background is applied to a wrapper element (#banner).
FWIW.
Nice technique, but one thing wonders me. If I will use it in a logo image (hide a link with anchor text under it) must I make tha image link nofollowed?
Quite simply, because it seems that if you nofollow the first link to a particular page, the rest of them will also be counted as nofollow and you could end up with major problems of your homepage not getting the juice it deserves from your internal links.
More info and testing on this at:
https://www.marketingfan.com/revisited-google-rates-only-first-link-text
https://smackdown.blogsblogsblogs.com/2007/10/25/single-source-page-link-test-using-multiple-links-with-varying-anchor-text-part-two/
https://www.seomoz.org/blog/results-of-google-experimentation-only-the-first-anchor-text-counts
Edited to make the links clickety. BTW, whats up with that? URLs used to turn into links automatically in SEOmoz comments!
video doesnt work, loads but doesnt really play and now is just showing still images but no sound... will try it later today