PageRank sculpting

People think about PageRank in lots of different ways. People have compared PageRank to a “random surfer” model in which PageRank is the probability that a random surfer clicking on links lands on a page. Other people think of the web as an link matrix in which the value at position (i,j) indicates the presence of links from page i to page j. In that case, PageRank corresponds to the principal eigenvector of that normalized link matrix.

Disclaimer: Even when I joined the company in 2000, Google was doing more sophisticated link computation than you would observe from the classic PageRank papers. If you believe that Google stopped innovating in link analysis, that’s a flawed assumption. Although we still refer to it as PageRank, Google’s ability to compute reputation based on links has advanced considerably over the years. I’ll do the rest of my blog post in the framework of “classic PageRank” but bear in mind that it’s not a perfect analogy.

Probably the most popular way to envision PageRank is as a flow that happens between documents across outlinks. In a recent talk at WordCamp I showed an image from one of the original PageRank papers:

Flow of PageRank

In the image above, the lower-left document has “nine points of PageRank” and three outgoing links. The resulting PageRank flow along each outgoing link is consequently nine divided by three = three points of PageRank.

That simplistic model doesn’t work perfectly, however. Imagine if there were a loop:

A closed loop of PageRank flow

No PageRank would ever escape from the loop, and as incoming PageRank continued to flow into the loop, eventually the PageRank in that loop would reach infinity. Infinite PageRank isn’t that helpful 🙂 so Larry and Sergey introduced a decay factor–you could think of it as 10-15% of the PageRank on any given page disappearing before the PageRank flows along the outlinks. In the random surfer model, that decay factor is as if the random surfer got bored and decided to head for a completely different page. You can do some neat things with that reset vector, such as personalization, but that’s outside the scope of our discussion.

Now let’s talk about the rel=nofollow attribute. Nofollow is method (introduced in 2005 and supported by multiple search engines) to annotate a link to tell search engines “I can’t or don’t want to vouch for this link.” In Google, nofollow links don’t pass PageRank and don’t pass anchortext [*].

So what happens when you have a page with “ten PageRank points” and ten outgoing links, and five of those links are nofollowed? Let’s leave aside the decay factor to focus on the core part of the question. Originally, the five links without nofollow would have flowed two points of PageRank each (in essence, the nofollowed links didn’t count toward the denominator when dividing PageRank by the outdegree of the page). More than a year ago, Google changed how the PageRank flows so that the five links without nofollow would flow one point of PageRank each.

Q: Why did Google change how it counts these links?
A: For one thing, some crawl/indexing/quality folks noticed some sites that attempted to change how PageRank flowed within their sites, but those sites ended up excluding sections of their site that had high-quality information (e.g. user forums).

Q: Does this mean “PageRank sculpting” (trying to change how PageRank flows within your site using e.g. nofollow) is a bad idea?
A: I wouldn’t recommend it, because it isn’t the most effective way to utilize your PageRank. In general, I would let PageRank flow freely within your site. The notion of “PageRank sculpting” has always been a second- or third-order recommendation for us. I would recommend the first-order things to pay attention to are 1) making great content that will attract links in the first place, and 2) choosing a site architecture that makes your site usable/crawlable for humans and search engines alike.

For example, it makes a much bigger difference to make sure that people (and bots) can reach the pages on your site by clicking links than it ever did to sculpt PageRank. If you run an e-commerce site, another example of good site architecture would be putting products front-and-center on your web site vs. burying them deep within your site so that visitors and search engines have to click on many links to get to your products.

There may be a miniscule number of pages (such as links to a shopping cart or to a login page) that I might add nofollow on, just because those pages are different for every user and they aren’t that helpful to show up in search engines. But in general, I wouldn’t recommend PageRank sculpting.

Q: Why tell us now?
A: For a couple reasons. At first, we figured that site owners or people running tests would notice, but they didn’t. In retrospect, we’ve changed other, larger aspects of how we look at links and people didn’t notice that either, so perhaps that shouldn’t have been such a surprise. So we started to provide other guidance that PageRank sculpting isn’t the best use of time. When we added a help page to our documentation about nofollow, we said “a solid information architecture — intuitive navigation, user- and search-engine-friendly URLs, and so on — is likely to be a far more productive use of resources than focusing on crawl prioritization via nofollowed links.” In a recent webmaster video, I said “a better, more effective form of PageRank sculpting is choosing (for example) which things to link to from your home page.” At Google I/O, during a site review session I said it even more explicitly: “My short answer is no. In general, whenever you’re linking around within your site: don’t use nofollow. Just go ahead and link to whatever stuff.” But at SMX Advanced 2009, someone asked the question directly and it seemed like a good opportunity to clarify this point. Again, it’s not something that most site owners need to know or worry about, but I wanted to let the power-SEOs know.

Q: If I run a blog and add the nofollow attribute to links left by my commenters, doesn’t that mean less PageRank flows within my site?
A: If you think about it, that’s the way that PageRank worked even before the nofollow attribute.

Q: Okay, but doesn’t this encourage me to link out less? Should I turn off comments on my blog?
A: I wouldn’t recommend closing comments in an attempt to “hoard” your PageRank. In the same way that Google trusts sites less when they link to spammy sites or bad neighborhoods, parts of our system encourage links to good sites.

Q: If Google changed its algorithms for counting outlinks from a page once, could it change again? I really like the idea of sculpting my internal PageRank.
A: While we can’t ever say that things will never change in our algorithms, we do not expect this to change again. If it does, I’ll try to let you know.

Q: How do you use nofollow on your own internal links on your personal website?
A: I pretty much let PageRank flow freely throughout my site, and I’d recommend that you do the same. I don’t add nofollow on my category or my archive pages. The only place I deliberately add a nofollow is on the link to my feed, because it’s not super-helpful to have RSS/Atom feeds in web search results. Even that’s not strictly necessary, because Google and other search engines do a good job of distinguishing feeds from regular web pages.

[*] Nofollow links definitely don’t pass PageRank. Over the years, I’ve seen a few corner cases where a nofollow link did pass anchortext, normally due to bugs in indexing that we then fixed. The essential thing you need to know is that nofollow links don’t help sites rank higher in Google’s search results.

535 Responses to PageRank sculpting (Leave a comment)

  1. Matt, I’m sorry but allowing nofollowed links in blog comments to detract from the PageRank flowing within my site is absolutely moronic. Yes that was the way that it worked before the nofollow attribute but people also bought links without concern of the nofollow tag back then. Obviously things change.

    You say that Google encourages links to good sites, but the whole point of the nofollow tag is that you’re not vouching for the links, and not checking whether the links are to good sites or not.

    Basically you’re saying, hey, you know the ORIGINAL reason we created the nofollow tag? Yeah well now it’s only really good for disclosing paid links to us.

  2. Wow! Great post, Matt. Thanks for clearing this up. Question, though. For those of us SEOs who have been utilizing this strategy with our clients and given that it’s not worth the time to do it…is it worth the time to go back and undo what we’ve done? Or, should we just let it ride? Thanks!

  3. thanks for this wonderful entery, it’s helping webmasters to be better

    I’ll read more from ur blog 🙂


  4. Ohh Matt,
    Once again I am confused, what to use in my blog for links, “nofollow” or not to improve my PageRank. In your previous you encouraged using “nofollow” And now it goes to 3rd place… Oh Matt wat to do.

  5. Wow. Was awaiting for a long time such a clear explanation from the Lion’s mouth, clarified major doubts, especially that of no following feeds and internal links. I even nofollowed my feedburner chicklet link!

  6. Are you worried that people will play games with links now so they don’t “consume” pagerank? By that I mean adding links via obfuscated javascript or some other method (flash).

  7. Thank God I followed you on Twitter! It leads me here! Thank you for sharing

  8. Matt Grenville

    It’s funny, just the other day some guy approached me about page rank sculpting. “It’s all the rage and done wonders for my website, you should try it.” he said, and he gave me a link to a post on page rank sculpting.

    Ever since then I have heard nothing BUT page rank skulpting, page rank skulpting this, page rank skulpting that. And after all this, it has not worked for a year anyway and yannow what? I think i’ve given up caring on what works and what doesn’t.

    I can have my little website, and I can stroke it and feed it words here and there, she will sit there, and look pretty, and be nice to me. But low and behold I try and promote it and hey, i’ll be told NO, don’t do this, don’t do that, we advise you not to do this, we advise you not to do that.

    Best bit of advice anyone can give is…. “Write good content and people will link to you”


    Cause if I do that, If I write good content, whilst my 100+ competitors link build, article market, forum comment, social bookmark, release viral videos, buy links, I’ll end up the very bottom of the pile, great content or not and really I am just as well taking my chances pulling off every sneaky trick in the book to get my site top because, everyone does it anyway and if I don’t what do have to lose?

    So I don’t care if the magical Mr Cutts is telling me page rank sculpting don’t work no more. I’m gonna bloody do it anyway.


    Know why? Natural organic link building only happens for those with million buck advertising campaigns.

  9. Matt, would you also not recommend using nofollow for pages like “policy & terms”, “contact us”, etc.

    Like the RSS feed, it doesn’t make sense to have these pages in Google results.

  10. Awesome post. This is the kind of clarity I’ve been wanting on the subject for a long time!

  11. So what happens to the PageRank that belongs to those nofollowed links? For example you have a page with 50 “points” of PageRank, 50 links, and 25 of them are nofollow. So that page passes 25 points of PageRank. What happens to the other 25? Does it get discarded? Redistributed to the rest of the web?

  12. So just for clarification: No-follow links do not pass PageRank, but, they do act as part of the equation that calculates the amount of PageRank that do-follow links pass on the same page?

  13. Perhaps, my first comment on your blog.
    Thanks for informing.

  14. Do you think Matt’s using the Thesis theme for it’s SEO capabilities =oP.

  15. Correct me if I’m wrong, but “parts of our system encourage links to good sites” gives me the idea that a site can be “rewarded” for not only creating good content and good site architecture, but for also providing relevant outbound links to other (perhaps more authoritative) sites. Do I have that correctly?

  16. Hi Matt,

    Thanks for the clarification, but I do have a further question. I might be missing something here, but you say the reason Google changed how it counted the links was:

    For one thing, some crawl/indexing/quality folks noticed some sites that attempted to change how PageRank flowed within their sites, but those sites ended up excluding sections of their site that had high-quality information (e.g. user forums).

    Lessening the amount of PageRank passed to the pages that weren’t nofollowed won’t help this situation, will it? The links to the sections with high-qulaity information would still be nofollowed. I don’t see how that helps the situation at all.

    Everything else you say makes perfect sense, but I don’t get this part.

  17. Thanks for the clarification, Matt. We were just wondering today when we would hear from you on the matter since it had been a couple of weeks since SMX. I think we’d all be interested to know the extent to which linking to “trusted sites,” helps PageRank. Does it really mitigate the losses incurred by increasing the number of links? I ask because it seems pretty conclusive that the total number of outbound links is now the deciding metric for passing PageRank and not the number of DoFollow links. Any thoughts from you or others?

  18. @Stephan you make a fantastic point. I think Matt is trying to tip toe around the fact that they basically don’t want people “manipulating” the way PageRank flows around their site. Sure it was using a tool they pushed on us for the very purpose they created it, but my guess is that it was being used too effectively.

  19. I take it using noindex, nofollow on your date based archives isn’t deliberate then? 😉

    Yes that is nasty, possibly enough to knock you a few rankings for Matt

    Matt, in my recent post on this topic I finished it with:-

    “I would love a much clearer indication of page size that Google will index as there are just vague notions that it can be more than 100 links per page.
    If a size is specified, is that gzipped? ”

    My recommendation will be to increase the amount of internal linking plus a change in the way comments are handled (not what people might assume).

    So real indication of how large a document can get, e.g. will Googlebot give up on 1000 links?

    The problem comes when comment links come before sidebar and footer navigation. It is better for the user to get the conversation first, but not if Googlebot will give up before the navigation.

  20. Wow, what surprises me most about this is that its been happening for over a year, and no one noticed. Another item I saw addressed in this was the 3 way linking, and how that has become less effective due to the decay factor to fight the Infinite PageRank loop. Good to know..
    So, after everything is said and done, don’t cut corners, create great linkable content, and mind your site architecture.

  21. will Google ever consider going open about such technologies as Pagerank? will there ever be such thing as Google Pagerank API?

  22. Ah jeez, it gets clear and then complicated again all at the same time. I just wish Blogger had a magic “SEO” button that you could click on to magically optimize Blogger blogs that are unique content rich. Call it the “That was Geezy” button. The more I learn about how search engines work, the less I understand.

  23. Hi Matt,

    Thanks for clarifying this issue. I’ve read different versions/explanations about this issue but it’s finally good to hear it from the lion’s mouth as raj put it.

  24. THANK YOU!
    I thought I’d had the nofollow deal figured out, then I started to doubt myself, hearing so many SEOs touting it’s value and the be-all end-all.
    From what i read here, my original mind-set was bang on. It has value, but not worth spending your life on. There is more value, researching and developing good content on a site.
    I’ve usually say if you don’t want Google to follow a link I just leave out the tag lol..

  25. abilitydesigns

    Just wanted to say Thank you Matt for a detailed post on this topic.
    ( and saving us some unproductive work!)

  26. Regarding why Google changed how it counts nofollow links, you say:

    For one thing, some crawl/indexing/quality folks noticed some sites that attempted to change how PageRank flowed within their sites, but those sites ended up excluding sections of their site that had high-quality information (e.g. user forums).

    If the assumption here is that webmasters will remove the nofollow attributes in response to this change, then why did take “more than a year” for someone from Google to present this information to the public? It seems that if this logic had anything at all to do with the decision to change the nofollow policy, Google would have announced it immediately in order to “encourage” webmasters to change their linking policies and allow access to their pages with “high-quality information.”

  27. Thank you for a clear explanation of no-follow and page rank. I still have a few questions, which are mentioned by people above. I hope to see a reply from you soon! 🙂

    Thanks Matt

  28. Duh! I think I “got” about a half of that. Now can we have the “for dummies” version in plain English?

    Plus a question. So why does Google which owns Blogger add a ‘no follow’ to the category labels in Blogger?

  29. Matt, that’s an excellent, in-depth discussion of the evolution of the PageRank algorithm and the flow of outbound links and internal links. I particularly liked the “decay factor” built into the algorithm and your succinct explanation of how it works.

    The PageRank Citation Ranking Report from ’98 cites “150 million web pages” on topics as diverse as information retrieval to “‘What is Joe having for lunch today?'” (Ev & Biz must have had an A-Ha moment leading to the brilliant simplicity of Twitter asking a simple question: “What are you doing?”)

    With 150 million pages, the Web had 1.7 billion edges (links). Any estimate on what the ratio is today in the age of link building?

    Sad to see that one of the co-authors of the report was Rajeev Motwani, who recently passed away. That his genius lives on in perpetuity in Google’s hybrid intelligence is a tribute to his visionary mind.

    What surprised me most: power-SEOs had overlooked or completely missed a number of crucial changes in the Google ranking algorithm:

    “At first, we figured that site owners or people running tests would notice, but they didn’t. In retrospect, we’ve changed other, larger aspects of how we look at links and people didn’t notice that either, so perhaps that shouldn’t have been such a surprise.”

    As a former executive editor of Search Engine Watch, I plead guilty. We missed those changes, too. Who was watching the Watchmen? 🙂 Look forward to reading more in-depth info on previous changes to PageRank now that they’re part of search engine history.

    btw – didn’t mean to give you a hard time on Twitter today about not having subscribed to the Consumer Reports WebWatch blog. That was one pub SEW didn’t miss citing. Chris Sherman and Danny Sullivan wrote some great SearchDay articles on the CR watchdog blog in the first half of the first decade of the 21st century.

    Plus, you had a speaking gig in NYC when I first started in the search industry: a panel at Consumer Reports WebWatch’s 2003 National Summit on Web Credibility. So I missed the irony and humor in your tweet that you never subscribed … now I get the gag.

  30. Matt, you spend most of this post talking about PR Sculpting on internal links, then you have this question:

    Q: If Google changed its algorithms for counting outbound links once, could it change again? I really like the idea of sculpting my internal PageRank.

    Did you err in putting “outbound links” in that question? Internal linking and outbound links are two different things, at least to most of us. 🙂 Would be helpful if you’d clarify that this is about PR sculpting on internal links. Thx.

  31. Halfdeck, my hope/guess, and this is purely speculation, is that Google uses something like “DDD” Dynamic Domain Dampening.

    It is something that is possibly needed to handle hanging/dangling pages effectively anyway, whereby rather than giving this part of the dampening factor to the whole web, it is redistributed with the domain instead.

  32. A good start at explaining, Matt. Thanks. Not looking forward to the wave of flash, form, and script-based hacks that are sure to follow but I guess the dofollow people just got a nice boost.

    Nofollow _was_ a nice solution for those who didn’t want to choose between users and SEO at the margins – but the Googlers comments @ SMX left me wondering if they really get what the user experience issues are for many types of sites – especially larger Ecommerce stores.

  33. Thanks for the post Matt.

    I would tend to agree with you re not using nofollow to sculpt PR.. In my experience, less tricks the better results.


  34. Thanks for this great post!
    Now I made my final opinion about NoFollow

  35. The Web professionals are worried with the Search Engines and forgetting the most important part, the users. As Eisenberg said:“SEO should not be an acronym for Search Engine Optimization, but for Search Experience Optimization.” This is one of the best sentences that I ever read in my short career!

    Thank you for clarifying this information, Matt! The “our world” stopped for a few days after this question you leave on the air! Hahaha…As always, relevance and facility are the rules! 😀

  36. I’m calling this the “use it or lose it” version of PageRank. Thanks for keeping it interesting, Matt 🙂


  37. Thanks for this vital and informative post. Now, I will know what to do with my blog.

  38. Matt, thanks for this article. I’m relatively new to SEO myself and have been trying to find good information about it but there seems to be an awefull lot of contradicting stories about it. I realize that it’s not in Google’s best interest to tell us about the exact algorithm but articles like yours sure give me a good insight.

    From my newbie SEO point of view I never even realized that I could flow pagerank through my site, so I guess everybody reads something different in your article.

  39. Ben Finklea, this is a change that’s been live for well over a year; if you’ve got a site that works for you and you’re happy with, I wouldn’t worry about going back to change a lot of work. If you were starting on a fresh new site then this info is something to bear in mind.

    Hemanth Kumar, a good rule of thumb is: if a link on your website is internal (that is, it points back to your website), let it flow PageRank–no need to use nofollow. If a link on your website points to a different website, much of the time it still makes sense for that link to flow PageRank. The time when I would use nofollow are when you can’t or don’t want to vouch for a site, e.g. if a link is added by an outside user that you don’t particularly trust. For example, if an unknown user leaves a link on your guestbook page, that would be a great time to use the nofollow attribute on that link.

    btmorex, I just think there’s usually better ways to spend your time as an SEO. If you have a pretty good site architecture to start with, you normally don’t need to think much about PageRank sculpting (e.g. WordPress does quite well at the mechanics of SEO by default).

    Matt Grenville, I’ve seen a lot of sites and bloggers do very well with great content, insightful analysis, or by providing a useful service. In my experience, the better your underlying content is, the more of a head-start you have when trying to get people to notice your site.

    Amit Agarwal, great question. Given the way that Google works since this change, I would let PageRank flow even to your privacy and terms-of-service type pages. Even those sorts of pages can be useful for more searches than you would expect.

    Halfdeck, it’s a bit complicated, esp. since Google doesn’t view pages exactly in the framework as “classic PageRank” any more. You can think of that PageRank going into the reset vector without being too far off.

    Joe Hall, correct.

    Donovan, I tried Thesis more for the customizability and ease-of-use, plus I was getting really dang tired of having a green blog. 🙂

    Michael D, that’s a good way to interpret it.

    Stephen Cronin, the hope is that sites will allow PageRank to flow more within their own sites, which yields more pages ranking in the search results.

    Scott Cowley, sorry it took a while. I wanted to check on a few things when writing the post, and then run it by a few colleagues to have them check for any glaring errors. I didn’t say that linking to high-quality sites helped your PageRank, but rather other parts of our system would encourage/reward those links.

    Andy Beard, I was only talking about the nofollow attribute on individual links, not noindex/nofollow as a meta tag. But I’ll check that out. Some parts of Thesis I really like, and then there’s a few pieces that don’t quite give me the granularity I’d like. As far as page size, we can definitely crawl much more than 101KB these days. In my copious spare time I’ll chat with some folks about upping the number of links in that guideline.

    youfoundjake, those would definitely be the high-order bits. The fact that no one noticed this change means (to me) even though it feels like a really big shift, in practice the impact of this change isn’t that huge. By the way, I have no idea why CFC flagged you, but I pulled your comment out of the Akismet bin. Maybe some weird interaction of cookies with WordPress caching? Sorry that happened.

    Roshan Joshi, I think Google talks a lot more about PageRank than lots of search engines talk about how they rank things. But I wouldn’t expect an official “PageRank API” any time soon.

    Nashville Grant, here’s the mental model I’d employ: search engines want to return great content. If you make such a fantastic site that all the web has heard of you, search engines should normally reflect that fact and return your site. A lot of bad SEO happens because people say “I’ll force my way to the top of Google first, and then everyone will find out about my site.” Putting rankings before the creation of a great site is in many ways putting the cart before the horse. Often the search rankings follow from the fact that you’re getting to be well-known on the web completely outside the sphere of search. Think about sites like Twitter and Facebook–they succeed by chasing a vision of what users would want. In chasing after that ideal of user happiness and satisfaction, they became the sort of high-quality sites that search engines want to return, because we also want to return what searches will find useful and love. By chasing a great user experience above search rankings, many sites turn out to be what search engines would want to return anyway.

    Jeff Starr, we started saying it in various ways softly at first, but the direct question at SMX (where the audience is more advanced to start with) seemed like a really good opportunity to address the topic head-on.

    “With 150 million pages, the Web had 1.7 billion edges (links).” Kevin Heisler, that ratio holds true pretty well as the web gets bigger. A good rule of thumb is that the number of links is about 10x the number of pages. I agree that it’s pretty tragic that Rajeev Motwani was a co-author of many of those early papers. I got to talk to Rajeev a little bit at Google, and he was a truly decent and generous man. What has heartened me is to see all the people that he helped, and to see those people pay their respects online. No worries on the Consumer WebWatch–I’m a big fan of Consumer WebWatch, and somehow I just missed their blog. I just want to reiterate that even though this feels like a huge change to a certain segment of SEOs, in practical terms this change really doesn’t affect rankings very much at all.

  40. Hi Matt,
    I was eager to read clear explanation about link sculpting as there was active discussion running in internet industry after your reply on link sculpting at SMX and things was blurry due to lack of sound explanation.
    Most of the points discuss in given post are clear but it would be really helpful if you give more details on the query asked by Matt McGee.

  41. Hey Matt,

    I was waiting for your post in hope that the changes will become clearer for me but… 🙂 Here are two scenarios I would like to ask you about:

    Scenario 1: I have page with PR 10 that has 5 links (doesn’t matter whether internal or external). 2 out of those 5 links are “nofollow”. Here is how PR is distributed:
    – My page loses no PR from the links, which means it remains with PR 5
    – The 2 “nofollow” links each get PR 1
    – The remaining links get each PR 2
    In general from the PR 10 from my page I distribute 6 to sites I endorse, 2 to sites I DO NOT endorse and I lose 2.

    Scenario 2: I have the same situation but the PR is distributed as follows:
    – My page loses no PR from the list (still PR 10)
    – The 2 “nofollow” links get each PR 0
    – The remaining links get each PR 3.3(3)
    In general from the PR 10 from my page I distribute only to sites/links I endorse and I don’t lose PR.

    From your explanation above I think what is happening is Scenario 1. In general I get the reasoning behind that but one question that remains is: how do I REALLY say that I REALLY don’t want to pass any PR to a site I think is useless and deserves no PR at all?

  42. Hi Matt,
    So what about huge database driven sites, (of which I can give you many examples) that have used nofollow to:
    1. stop spiders indexing piles of effectively duplicated content (e.g where the database returns different URLs based on specific query structure – but these URLs provides basically the same content as any other number of URLs where the query is structured slightly differently);
    2. to help spiders avoid session IDs appended to the query string (i.e. for cookie rejecters). Many large reputable sites have built clean crawl paths to structured content, and have effectively kept the complex dynamic content (and spider confusion) out of the index using nofollow.
    Should they be using obfuscated javascript to keep that effectively dynamically generated “worthless” cookie appended content out of the index – or continue using nofollow?

  43. Thanks for the clarity on nofollow Matt,

    It sucks that I have to give pagerank to any blog post I link to though, even if I am saying that blog post x should be avoided because it is garbage. It seems to me from reading this post that Google will think that my page somehow wishes to feed pagerank to that page I am ranting about and link to with a nofollow.

  44. But Matt. I have don’t want the anchor text of my “more….” links counted. So a rel=”nofollow” them, as there is a link on the article name anyway. Its not page sculpting, its common sense. Why would I want a search engine to give my “more..” anchor a vote? I don’t want my pages to rank for “more”.

    Now, I find out I have been killing my page rank as a result, which explains why i have dropped so many inner pages, whilst keeping the same rankings on my main pages.


    For a year now, I have effectively been penalised for using a rel=’nofollow’ in a responsible way.

    I can think of many other reasonable uses that do not amount to page sculpting, where sites will have been penalized.

    The damage this has likely done to my sites Matt, …. the horror of it to think I was doing the right thing, telling search engines to ignore unimportant duplicated links…. only to find out google has penalized me for it?

    I am blown away Google made this change without giving notice. The value in hurting page sculptors, could not possibly have been as great as hurting those using it sensibly.

  45. Michelle you probably haven’t been penalised (my current interpretation not Matt’s or Googles), because the more link most likely appears after the title link. Google will take the first meaningful link, and probably ignore the more link as it is going to the same location, and isn’t unique or useful anchor text)

  46. Matt McGee, good point that the wording of that phrase could be interpreted in different ways. I changed it to “counting outlinks from a page” to clarify–thanks!

    Dan Thies, your video was a much funnier response to the issue. If folks haven’t watched the video that Dan did, it’s a hoot. Just remember to keep your sense of humor intact and that he’s joking about the wild conspiracy theories at the end.

    Matt Hodgson, it’s definitely a much simpler approach and leaves more free time to make new, linkable content.

    Gradiva Couzin, that’s not a bad way to think of it.

    Fred Leeflang, I’m glad that it was helpful.

    ToddySN, I’m worried your example might confuse people. PR10 and PR5 normally refers to values in the toolbar. That’s why I talked about “points of PageRank” so that people wouldn’t think about it as toolbar values. Regardless, if you really really don’t want to pass PR to a site, the nofollow attribute on a link does exactly that.

    Chris_D, great question. If you have a single product page that can have multiple urls with slightly different parameters, that’s a great time to use a rel=canonical meta tag. You can use rel=canonical for pages with session IDs in a similar fashion. What rel=canonical lets you do is say “this page X on my host is kinda of ugly or otherwise isn’t the best version of this page. Use url Y as the preferred version of my page instead.” You can read about rel=canonical at Bear in mind that if you can make your site work without session IDs or make it so that you don’t have multiple “aliases” for the same page, that’s even better because it solves the problem at the root.

    Eric Vold, if you really don’t want to pass PageRank to an external site (e.g. because you think it’s garbage), then a nofollow attribute on the link will do what you want.

    michelle, the case that you mention probably wouldn’t have made that much of a difference in practice. The fact that no one noticed this change means that the overall impact is on the smaller side. But I’m sorry that you’ve been doing that work when you didn’t need to.

  47. Matt, as you know, I was kind of annoyed when you suggested sculpting to a room full of SEOs back in 2007. We’d been told over the years to do things for humans, not to overly worry about having to do stuff for search engines — and suddenly, here you were suggesting that SEOs could flow PageRank to their most “important” pages. I’d figured Google had long since been smart enough to decide for itself what percentage of a page’s PageRank spend to assign to a particular link. That assumption didn’t just come out of the blue — it came from things Google had hinted at over the years. So being told to start overtly flowing around the PageRank? It seemed counter-productive.

    But some people loved it. Some people swore it worked. And some people still swear it works despite the change you say happened a year ago. Meanwhile, I’m sorry — but this post doesn’t clear a lot of things up for me. It just makes things even more messy. So some specific questions and observations.

    First, let’s talk about life before nofollow. A page has $10 of PageRank to spend. It contains 10 links. Simplistically, the money is divided equally — each got $1.

    Next, nofollow comes along. The same page puts nofollow on 5 of the 10 links. The PageRank is divided equally among the remaining 5 — each gets $2.

    Next, in 2008, there’s this change you’re now disclosing. The same page still has $10 to spend, still has 10 links, still has 5 of them nofollowed. Google looks at ALL the links, divides PageRank equally and is ready to give each link $1. Those without nofollow get their $1. Those with nofollow don’t get any spend at all.

    Simplistically speaking, if that’s what’s happening now — what a nightmare will now be unleashed. It means that people who still want to ensure that some links get the most PageRank cash will be encouraged to still sculpt by simply not having as many links on their page.

    This is the argument that quickly emerged about blog comments recently. Say I have an article on a blog with 5 links in the editorial copy — some of those links leading back to other content within the blog that I hope to do well. Then I get 35 comments on the article, with each comment having a link back to the commenters’ sites. That’s 40 links in all. Let’s say this particular page has $20 in PageRank to spend. Each link gets 50 cents.

    With nofollow before the change, I could have (if I were worried about flowing PageRank), kept any of those comments from getting some of my PageRank spend. Nofollow them all, and the 5 remaining links each get $4.

    With this change, I can still get the $4 if I simply don’t allow comments. Or I show comments, but I use an iframe, so that the comment actually reside on a different page. In either case, I’m encouraged to reduce the number of links rather than let them be on the page period, nofollow regardless. If I’m worried my page won’t seem “natural” enough to Google without them, maybe I allow 5 comments through and lock them down after that.

    Rather than clarify things, I feel like this is what your post is going to do — cause people to consciously reduce the number of links they allow on their pages. We’re going to see an increase in iframe usage or other techniques to reduce links and flow more PageRank to the remaining links, for those who really worry/believe in such things.

    To make matters worse, none of this matter, right? Because we’re being simplistic in all this. The reality is that Google is deciding how to divvy up the PageRank spend.

    Google might see 10 links on a page that has $10 of PageRank to spend. It might notice that 5 of those links are navigational elements that occur a lot throughout the site and decide they should only get 50 cents each. It might decide 5 of those links are in editorial copy and so are worthy of getting more. Maybe 3 of them get $2 each and 2 others get $1.50 each, because of where they appear in the copy, if they’re bolded or any of a number of other factors you don’t disclose.

    PageRank isn’t spread equally among links and hasn’t been for years and years, right? As I said at the beginning, this is what I long understood to be the case from things Google hinted at before 2007. And at SMX Advanced, you confirmed it to be the case.

    So really, isn’t the better explanation simply this?

    “Google itself solely decides how much PageRank will flow to each and every link on a particular page. The number of links doesn’t matter. Google might decide some links don’t deserve credit and give them no PageRank. The use of nofollow doesn’t “conserve” PageRank for other links; it simply prevents those links from getting any PageRank that Google otherwise might have given them.”

    As for the use of nofollow as a way to keep pages that shouldn’t be indexed out of Google (as with your feed example) is terrible advice. Your use of it on your feed link does nothing. If anyone links to your feed without nofollow, then it’s going to get indexed. Things that shouldn’t be indexed need to use either robots.txt or meta robots blocking. Nofollow on links to those items isn’t a solution.

    That also means this part of what you said:

    “Those sites ended up excluding sections of their site that had high-quality information (e.g. user forums).”

    Isn’t correct. Very few people are going to be able to block the high-quality information you or the crawling team are worried about with nofollow. If it’s high-quality, then people outside the particular site will link to it, you see that link, you crawl. You only don’t crawl if — again — meta robots or robots.txt is employed.

  48. Matt, why would you nofollow the above link to Dan’s video?

    Clearly you think that this is useful content. You’re just confusing the issue even further….

  49. Does Wikipedia fit in the “Infinite PageRank” model.

  50. Thanks so much for writing this. I still vote that the more google is totally transparent about the way it does things the better for all of us long term.

    Please clarify if this is wrong, but I use nofollow if I link from one of my sites to another one of my sites, or to a paid sponsor of one of my events.

    You gave me advice publicly a long time ago to get a blog and write quality content. I have done that and it has worked amazingly well for me and my clients. It is also amazing how few people bother to take the time to do it.

    That advice may benefit from a slight addition. Write quality content that is beneficial for other sites to link to. It is slightly different than simply writing quality content. It is writing quality content that gives the other webmaster/blogger something that is benificial to them to link to your page.

    The above with simple coding and Cha Bing Bada Bing!

  51. Hi Matt
    Thanks as always to share the info with us.

    @All I really don’t understand why you are blaming with Matt. He simply reported some new behavior on Google. Sure, he takes part on decision, but have you ever though when you started sculpting that probably artificial modifying the SE understanding of your web page wasn’t properly a good thing?
    Concentrate yourself on doing good web site and content instead of blaming with Google for respecting the users.

  52. Matt,

    Thank you for taking the time to explain this, although I must say I’m still a bit confused on the best way to implement this.

    It almost seems like the best thing to do is abandon nofollow altogether, implement a superb anti-spam plugin, limit users to posting one link each, and moderate if necessary.

    Maybe it’s time to just kill off nofollow altogether.

  53. “what a nightmare will now be unleashed.” That’s what’s been running quite well for over a year now though. I mentioned it in passing, but there are also parts of our system that trust and encourage sites to link out well.

    I did this post because I wanted people to understand more about PageRank, how it works, and to clarify my answers at SMX Advanced. Yes, I would agree that Google itself solely decides how much PageRank will flow to each and every link on a particular page. But that’s no reason to make PageRank a complete black box; if I can help provide people with a more accurate mental model, overall I think that’s a good thing. For example, from your proposed paragraph I would strike the “The number of links doesn’t matter” sentence because most of the time the number of links do matter, and I’d prefer that people know that. I would agree with the rest of your paragraph explanation–which is why in my mind PageRank and our search result rankings qualifies as an opinion and not simply some rote computation. But just throwing out your single paragraph, while accurate (and a whole lot faster to write!), would have been deeply unsatisfying for a number of people who want to know more.

    Regarding nofollow on content that you don’t want indexed, you’re absolutely right that nofollow doesn’t prevent that, e.g. if someone else links to that content. In the case of the site that excluded user forums, quite a few high-quality pages on the site happened not to have links from other sites. In the case of my feed, it doesn’t matter much either way, but I chose not to throw any extra PageRank onto my feed url. The services that want to fetch my feed url (e.g. Google Reader or Bloglines) know how to find it just fine.

    “I use nofollow if I link from one of my sites to another one of my sites, or to a paid sponsor of one of my events.” purposeinc, that’s still true.

    Amit Agarwal, do you mean that if you have a loop, eventually the PageRank goes infinite? The decay factor in the PageRank model is what prevents that from happening. People beat up Google for showing Wikipedia so often, but I think at least one other search engine shows Wikipedia even more often than Google does.

  54. I was exactly thinking the same thing what Danny Sullivan had said. If comments (even with nofollow) directly affect the outgoing PR distribution, people will tend to allow less comments (maybe usage of iframes even). Is he right? Maybe, Google should develop a new tag as well something like rel=”commented” to inform spiders about it to give less value and wordpress should be installed default with this attribute 🙂

    And my vital question about Amazon affiliate links. I think many people also wonder about it as well. I have several blogs where I solely write unique content reviews about several Amazon products, nothing more. As you know, all these links are full of tags, affiliate IDs whatsoever (bad in SEO terms). Should I nofollow them all or leave as they are?

  55. Hi Matt,

    Excuse my ignorance, so does this banish those stories I heard about it being dodgey to link to “bad neighbourhoods” without using “nofollow”?

    My main reason for ever adding “nofollow” was if I wasn’t convinced the site being “vouched for” was fully legit i.e. it might have a penalty imposed on it for paid links, and I don’t want to be seen to give it the thumbs up.

    So are you saying that now I shouldn’t worry about who I link to, regardless of their practices?

  56. Stu, that’s a quirk of WordPress. Any link in the comments section gets a nofollow. Earlier on, the admin of a blog could link and leave off the nofollow attribute. That’s probably something that I could drop an email to MattM or MarkJ about, but I’ve never gotten around to it.

    Sean Weigold Ferguson and David Airey, there are times when nofollow is useful. Paid links remains one of those reasons. When someone you don’t necessarily trust adds a link to your site is another good reason. It turns out that links with nofollow prevent quite a few blog spam sites from getting links, for example.

  57. Okay, I’m going to bed. Don’t take it the wrong way if I don’t approve or reply to comments for a while. 🙂

  58. Danny Sullivan, you just saved me 30 minutes of writing questions. Fortunately, that means 30 more minutes of deleting less-valuable links (that used to be nofollowed) from my website.

    Next up: reading about iframe.

  59. Thanks for answering, Matt. Enjoy your sleep.

  60. Danny covered most of the key points, but I do want to take a quick moment to address this:

    A: If you think about it, that’s the way that PageRank worked even before the nofollow attribute.

    Suggesting that this change is really just the equivalent of “resetting” things to the way they were is absurd. nofollow is still be using on outbound links in mass by the most authoritative/trusted sites on the web. Allowing us peons to have a slight bit of control over our internal juice flow simply allowed us to recoup a small portion of the overall juice that we lost when the top-down flow was so dramatically disrupted.

    If you really want everyone to forget about sculpting, then either ditch support for nofollow completely, or at a bare minimum, implement some type of real filter that demotes sites with excessive levels of external nofollows. The idea that the sculpting mom & pop struggling to compete is somehow a spammer, yet sites like the wiki are algorithmically rewarded for systematically cutting off the flow of juices to thousands of sites that are in no way close to the kind of sites nofollow was developed to combat, is simply insane.

  61. I think that this comment from Matt:

    I didn’t say that linking to high-quality sites helped your PageRank, but rather other parts of our system would encourage/reward those links.

    Is very telling and an important thing to consider. Taking the model of a university paper on a particular subject as an example, you would expect the paper to cite (link to) other respected papers in the same field in order to demonstrate that it is couched in some authority. As PageRank is based on the citation model used in university work, it makes perfect sense to incorporate a “pages linked to” factor into the equation.

    If I’m writing a page about the use of the vCard microformat on a page, it absolutely makes sense for me to link out to the definition where it was originally published, and improves user experience as well as lending authority to my arguments. Often as SEOs we get obsessed with the little things, claiming that its hard to get links on particular subjects, and that is pretty true, but its mainly our own selfishness in linking out to authority content that prevents other people giving us the same courtesy.

  62. Woohoo, here’s another nofollowed comment to siphon away pagerank from this site!

    I guess we can expect to see a lot more javascript hacks and flash links from the SEOs out there who want to hoard and sculpt their pagerank. Yes, having quality external links matters, but I’m pretty sure the SEOs out there are going to optimize the hell out of this to the detriment of users.

  63. @Danny Sullivan I totally agree with you. It also encourages the so called “power-SEOs” to use old fashioned ways to sculpt their pagerank…

  64. What I want to know is, what happens to all that PageRank that ‘evaporates’ as a result of all those nofollowed links? Is it the equivalent of the stash of confiscated drugs and weapons at a police station? Can Google set up an ‘evaporated PageRank’ lottery every month where a lucky site can win it all (a bit like Free Parking in monopoly)? 🙂

  65. Link manipulation… isn’t it black hat? I don’t care much on links, if they put comments and put link on it, as long as the comments are relevant and informative I approve it.

  66. Matt,

    One of the things why Google was recommending “nofollow” had to do with link buying… since that was a big thing for Google I am quite surprised that I do not find anything about it in this blog post… what is Google’s vision about buying links and “nofollow” now?

  67. I saw that you already answered that one a couple of comments back (…sigh). I need some coffee I guess…

  68. Thanks for clarifying the nofollow change.

    Firstly I think it’s important to avoid panic, this change happened a long time ago, and if you are doing fine – then perhaps you can do even better without nofollow on these links.

    Secondly I agree with Danny Sullivan, I think this change could cause people to link less, if they think about protecting and channelling the PageRank.

    I was never a big fan of using nofollow in PageRank sculpting, and the change is not that important in my work. I have of cause been using nofollow on links to archives and log-in pages as everyone else working with WordPress

    For me PageRank sculpting is about promoting the most important content on the website, thus channelling PageRank to these pages. This will not change, and should not work against Google’s rules and recommendations – if you want to present the best and most relevant content on my site.

    The change will in my opinion happen on the links to less important pages. Should you link to these in the main navigation, or should you “move” these so they collect less links and PageRank.

  69. Matt, does Google still count nofollow links as links when it’s working out how many links it’s willing to “notice” on one page (used to be 100, now an indefinite number >100) or does the robot just skip over that link code entirely and not take any notice of it?

    Because I can see a scenario if nofollow links don’t count as links where it’s still handy to be able to legitimately “conceal” some of the links e.g. a long and complex navigation menu to ensure that the other links on the page will all get noticed and crawled.

  70. Interesting details in this post!

    Danny writes:
    > Simplistically speaking, if that’s what’s happening
    > now — what a nightmare will now be unleashed. It
    > means that people who still want to ensure that some links
    > get the most PageRank cash will be encouraged to still sculpt
    > by simply not having as many links on their page.

    If the 5 links that were previously nofollowed are not meaningful, than they shouldn’t be on the site in the first place, or should they? Googlebot should not be treated substantially different from visitors. If you’re exposing humans to the links, they should be meaningful, and why should meaningful links not be evaluated by a search bot?

    I think the same about links in blog comments, by the way, even though Google seems to approve of nofollowing them there. At the Blogoscoped forum, if a link in the comment will not get deleted by an admin, then its nofollow attribute will be removed after some days. A useful link in a comment should count just normally.

  71. Matt, in practice, a page on my site might consist of

    15 main navigation links
    5 article headings with article title being links, with more.. links
    20 classified ads with titles as links, all with more… links.

    so, 65 links, with 25 of them killing pr.

    i am not sure, why you would say this would have no major effect?

  72. @Michelle… I think that they simply not look at interal nofollowed links… I think that nofollow will only be reckoned with if they are external of nature.

  73. Excellent! I was wondering when Google would finally release information regarding this highly controversial issue. I have always agreed with and followed Matt’s advice in having PR flow as freely as possible, natural linking is always the best linking in my experience with my search engine experience and results. I am very glad that you have addressed the topic of nofollow links having no effects in the Google SERPs, I was getting tired of telling the same topics covered in this article to my clients and other “SEOs”.

    Thanks for releasing this Matt!
    Whitney Segura

  74. This still encourages larger, more professionally run websites not to ‘share the wealth’ of PageRank to other sites. How is a small niche website ever going to get decent results in SERPS if when a bigger site does link to it, it nofollows the link? Not passing any PageRank benefit, or anchor text through the link does nothing to encourage smaller more specialised sites – it encourages the bigger, broader sites to hoard the goodness.

    If the web is about getting the most relevant content to the searcher, how is an information site about a niche subject less relevant than an Amazon product page?

  75. So if I get 50 comments on my blog post, all nofollowed, and I made in my post a link to a friend of mine because I think he is really cool, did this change means that the “juice” my link can pass to my friend is divided by 51 instead of not beeing divided at all?

    Or does that mean that the numbers of links present in a page doesn’t affect anymore the link popularity one link can pass by itself? That’s really important, as it can radically change the way people would externally link to others. I think I would be much more generous about that.

  76. Well what I took away from this is to stop worrying so much about the flow of link juice “out” of my site and to use more effective site architecture techniques than nofollow to direct the flow within my site.

    I have a pretty large directory site that nofollows almost all of its outgoing links. I will be removing the nofollow tags in the near future.

    I’ll let you all know what effect (if any) doing so has on my rankings.

  77. Seth, its been done to stop internal pr sculpting, so its looking at internal links

  78. I work on SEO at a large group of websites, market leaders across europe for our industry, and the change in nofollow attribution was a great event for us actually, but not in the way that you would imagine.

    We have long employed the traditional pagerank scultping techniques to funnel strength to the pages that were most important, yet we had an outdated on site navi which I basically left well alone, because I had ensured that it was all completely nofollowed, ergo did not create “me” any problems on site.

    With the massive changes, we have completely rebuilt our on site navi – with SE’s in mind, but the net result is also a navigational structure which in my opinion is far cleaner and of more use to human visitors, not just bots.

    This is most likely a great side benefit of these changes, but modifying your site so that HUMANS get the best possible experience is really a cornerstone of good on site SEO, and this nofollow change has reinforced that

    from me at least, THANKS for the changes!


  79. When explaining SERPs to my clients, I always try to boil down what we understand of the algorithm to “authority” + “relevance”.

    One thing I’d never considered before is that by linking to established sites with authoritative and informative content, we’re increasing the “relevance” part of that equation.

    It’s quite a big mental shift to think that that content we’re linking TO could essentially be seen to add to the “relevance” part of the equation by being considered – in some way – as part of the overall content of the page we’re linking from.

    From a usable index POV, which is at the end of the day what Google is there to provide, it also indicates that a site is more focused on providing value to their visitors than keeping them corralled inside their own content.

  80. Here is another interesting article on PageRank Sculpting & Blog Comments –

  81. Matt,

    Before I start this, I am using the term ‘PageRank’ as a general term fully knowing that this is not a simple issue and ‘PageRank’ and the way it is calculated (and the other numerous methods Google use) are multidimensional and complex. However, if you use PageRank to imply ‘weight’ it make it a lot simpler. Also, ‘PageRank sculpting’ (in my view) is meant to mean ‘passing weight you can control’. Now… on with the comment!

    As I have always said, Google makes the rules and needs to make those rules fit with what it wants to do and also change them when needed to fit with the changes that happen on the web.

    Just like the new structure on JavaScript links and them now carrying weight and being crawlable, the PageRank sculpting change is understandable. Google now can and wants to index more of the web (JavaScript link change). Google wants to reverse a method that can only help people in the know (PageRank sculpting change). Logically, all is very understandable.

    However, where the JavaScript link change is evolution, the PageRank sculpting change in not. Let me explain.

    Using ‘nofollow’ on untrusted (or unknown trust) outbound links is sensible and I think that in general this is a good idea. Like wise using it on paid links is cool (the fact that all those people are now going to have to change from JavaScript to this method is another story…). I also believe that using ‘nofollow’ on ‘perfunctory’ pages is also good. How many times in the past did you search for your company name and get you home page at number one and your ‘legals’ page at number two. Now, I know that Google changed some things and now this is less prominent, but it still happens. As much as you say that these pages are ‘worthy’, I don’t agree that they are in terms of search engine listings. Most of these type of pages (along with the privacy policy page) are legal ease that just need to be on the site. I am not saying they are not important, they are (privacy policies are really important for instance), but, they are not what you site is about. Because they are structurally important they are usually linked from every pages on the site and as such gather a lot of importance and weight. Now, I know that Google must have looked at this, but I can still find lots of examples where these type of pages get too much exposure on the search listings. This is apart from the duplicate content issues (anyone ever legally or illegally ‘lifted’ some legals or privacy words from another site?).

    In my view there is nothing wrong with saying ‘hey Google, these pages are not important from a search engine perspective, let me not give them so much weight’. Regardless of how Google now views these type of pages from a weight perspective, doing the above as a webmaster should be logical and encouraged. You have said this yourself at least a few times in the past.

    Likewise, ‘nofollowing’ your archive pages on your blog. Is this really a bad thing? You can get to the pages from the ‘tag’ index or the ‘category’ index, why put weight to a page that is truly navigational. At least the tag and category pages are themed. Giving weight to a page that is only themed by the date is crazy and does not really help search engines deliver ‘good’ results (totally leaving aside the duplicate content issues for now).

    To finish, I guess I want to make two points (which do have some embedded questions too), namely:

    1. Now that we know that weight/PageRank/whatever will disappear (outside of the intrinsic wastage method that Google applies) when we use a ‘nofollow’ link, what do you think this will do to linking patterns? This is really a can of worms from an outbound linking and internal linking perspective. Will people still link to their ‘legals’ page from every page on their site? Turning comments ‘off’ will also be pretty tempting. I know this will devalue the sites in general, but we are not always dealing with logic here are we? (if we were you (as head of the web spam team) wouldn’t of had to change many things in the past. Changing the PageRank sculpting thing just being one of them).

    2. Was there really a need to make this change? I know all sites should be equally capable of being listed in search engines without esoteric methods playing a part. But does this really happen anyway (in search engines or life in general)? If you hire the best accountant you will probably pay less tax than the other guy. Is that really fair? Also, if nobody noticed the change for a year (I did have an inkling, but was totally and completely in denial) then does that mean the change didn’t have to be made in the first place? As said, we now have a situation where people will probably make bigger and more damaging changes to their site and structure, rather than add a little ‘nofollow’ to a few links.

    All in all, PageRank sculpting (or whatever we should call it) didn’t really rule my world. But, I did think that it was a totally legitimate method to use. Now that we know the ‘weight’ leaks, this will put a totally new (and more damaging) spin on things. Could we not have just left the ‘weight’ with the parent page? This is what I thought would happen most of the time anyway.

    Still, I guess all of this keeps us all in a job, so we should not complain too much! However, I think you guys have got this one wrong and we will see in the next weeks and months how people jump on this.

    P.S. having turned on the ‘nofollow’ indicator plug in on Firefox a long time ago, I have seen some of the abuse on this. However, I still don’t think that this way is the best method to combat this. You could of just ‘downgraded’ the trust score on sites that had abused the ‘nofollow’ thing to silly levels.

  82. Thanks for the info Matt –
    I appreciate that you wrote out and answered questions on this. I think I have a much better understanding..

    BTW Just a quick comment to Matt Grenville…
    “Matt Grenville
    Best bit of advice anyone can give is…. “Write good content and people will link to you”
    Cause if I do that, If I write good content, whilst my 100+ competitors link build, article market, forum comment, social bookmark, release viral videos, buy links, I’ll end up the very bottom of the pile, great content or not and really I am just as well taking my chances pulling off every sneaky trick in the book to get my site top because, everyone does it anyway and if I don’t what do have to lose?”

    I have had sites rank WITH no links in to it. Eventually we had to add them, but we ranked in the top 20 for a single highly competitive word within 5 weeks of launch with just content, on page SEO, proper code (for code text ratios) and architecture.

    We did no page rank sculpting. I think the point being. You can rank better with the help of these types of techniques and off-site techniques always are the cream to the coffee, BUT if you do follow the idea of content, content, content, good site structure and on page SEO you would be surprised how well you can do.

    Thanks just wanted to add it because I see it so often!

  83. Matt, you don’t mention the use of disallow pages via robots.txt. I’ve read that PageRank can be better utilised by disallowing pages that probably don’t add value to users searching on engines. For example, Privacy Policy and Terms of Use pages. These often appear in the footer of a website and are required by EU law on every page of the site. Will it boost the other pages of the site if these pages are added to robots.txt like so?

    Disallow: /terms.htm
    Disallow: /privacy.htm

  84. “There may be a miniscule number of pages (such as links to a shopping cart or to a login page) that I might add nofollow on, just because those pages are different for every user and they aren’t that helpful to show up in search engines” – it doesn`t make much sense. If a page isn`t helpful and should not show up on search results, the best option is to meta-noindex the page and disallow it on robots.txt.

    Nofollow is not a safe way to not let a page get indexed. If you use nofollow on login or subscribe to feeds links, seems like PageRank Sculpting or dupe content prevention. Which makes more sense than using nofollow to prevent page indexing.

  85. Matt,

    One of the exceptional uses of nofollow was when you have printer friendly, send to friend, etc versions of the same page. In some CMMs, these pages are created with funky URLs where they could not be blocked with robots.txt files (or were 3rd party hosted where you didn’t control the robots.txt file).

    On to the actual question: If you have a page with a score of 10 that contains 10 links. 5 of those links lead to pages that are disallowed in the robots.txt file (printer friendly, etc versions). Do those remaining 5 links that are suppose to be crawled get 1pt or 2pts?

    Before nofollow, we use to toss these user functionality links into java. However, we found that we were isolating some iPhone/Storm users who could not always use the java functionality on the site. So we moved to an html link and nofollowed the links (and usually block them in robots.txt). However, now Google is crawling some java.

    What is the recommended way to add functionality to the page such as printer friendly, send to friend, etc to pages? It almost seems like adding hard-to-crawl flash is the next step.

  86. I had some questions before I read you post ! Thanks and Google is awesome!

  87. @Matt cutts
    Don’t mistake me I have not enough knowledge on SEO,I ask u one question so the more number of outgoing links pointing to other blogs or websites from my blog means am I loosing my page rank to these sites.

  88. “Matt Grenville, I’ve seen a lot of sites and bloggers do very well with great content, insightful analysis, or by providing a useful service. In my experience, the better your underlying content is, the more of a head-start you have when trying to get people to notice your site.”

    Matt, in almost every example you have given about “employing great content” to receive links naturally, you use blogs as an example. What about people that do not run blog sites (the vast majority of sites!), for example an E-Com site selling stationary? How would you employ “great content” on a site that essentially sells a boring product? Is it fair that companies that sell uninteresting products or services should be outranked by huge sites like Amazon that have millions to spend on marketing because they cant attract links naturally?

    I think Matt Grenville’s comment is a very valid one. If your site, for whatever reason, can not attract links naturally and all of your competitors are outranking you by employing tactics that might breach Google’s TOS, what other options do you have? As well as this people will now only link to a few, trusted sites (as this has been clarified in your post as being part of Google’s algorithm) and put a limit on linking out to the smaller guys.

    It seems to me like strong sites and corporates will get even stronger, and the little guys will get weaker 🙁

  89. Hi Matt,
    I’ve been advising my clients about adding no-follow links to their least important pages. There are few rumours floating around about nofollow links.
    According to rumours, no follow links wastes the PRs.
    Can you clarify the affect of the no follow please? Shall we keep using it or we should not bother it at all?

  90. Thanks Matt,
    So it’s hard to differentiate what you are saying, saying between the lines or not saying 🙂
    So I have to ask:
    Does a nofollowed link to an outside website also “evaporate” linkjuice? This question has not been asked yet as far as I can see, but is on everybody’s mind it seems …

  91. So what you are saying is that you don’t apply the ‘nofollow’ to comments because you like page rank to flow freely. Which means I have got a back link from this comment and potentially if this blog has a big page rank then my blog will benefit. Or at least thats what I think you said?

  92. Hi Matt, thanks for the post. Still trying to get my head round this.

    You’ve stated that any page with nofollowed links on it passes less PR than it would otherwise through any followed links. Its voting ability is reduced.

    This means that PR is lost
    (a) within the site (through any internal links) but also
    (b) on the web as a whole (through any external links, and then onwards from there).

    If (a) is correct that looks like bad news for webmasters, BUT if (b) is also correct then – because PR is ultimately calculated over the whole of the web – every page loses out relative to every other page. In other words, there is less PR on the web as a whole and, after a sufficient number of iterations in the PR calculation, normality is restored. Is this correct?

  93. I don’t get it, it seems Google is constantly making rules & regulations as they see fit. I don’t try to “manipulate” any links we have on our site or any clients we work for. Links take time period. No way around it. But, now this explanation gives more fuel to all the Google bashers out there. I recently read an article about Guy Kawasaki has been “loaned” one, two, three cars in three years & is still within Google’s guidelines? Makes me wonder how many rules and regulations are broken. My take is do your job right, and don’t worry what Google is doing. If content is King then everything will fall into place naturally.

  94. Interesting article. I don’t think the Nutch algorithm currently takes nofollow into effect. I will have to see about adding that.

    Also hadn’t thought about decreasing the rank value based on the spammyness of sites a page is linking into. My guess on how to do it would be determining the spammyness of individual pages based on multiple page and site factors, then some type of reverse pagerank calcuation starting with the those bad scores, then overlaying that on top of the “good” pagerank calculation as a penalty. This is another thing which would be interesting to play around with in the Nutch algorithm.

  95. Matt, this is all getting too much now: A slight crack of the whip from “your” hand reverberates through the tether (if that’s the right word) over months, if not years, distorting the web through each iteration.

    It costs money, time and headache pills not just for developers, but for content writers too. If Google decides what link/website gets the juice, then why not just leave it there?

    We now have a situation where CMS-installers toggle the default no-follow button because authors don’t understand its use; is it any wonder Google is missing out on the high quality pages?

    I know in my niche (a two year old new industry) that getting any linkjuice at all is ridiculously hard; and I’m not talking from a selfish POV here, but from the POV of many new businesses without the time, resources, nor geeky inclination to navigate this maze of rules.

    I think the dial has swung too far.

  96. Correct me if I am wrong but what Matt is actually saying is that the nofollow thingy is already going on for at least a year… if nothing shocking happened to your website within that year what are you guys all worrying about?

    I got a email from a guy who stated that I should read this post since our website is using nofollow… first of all: i read the article already, secondly: our website is doing fine at this moment… so why should I change it?

    PR calculation by the way is getting much less important anyway. So I really do not get why everyone is still brabbing about it :S

  97. @Matt: Enabling the Thesis theme option “Add Noindex to Archive Pages” does actually result in an added “noindex, nofollow” meta tag –which, I agree with Andy Beard here, is sub-optimal to say the least.

    In order to have a “noindex, follow” (or just “noindex”) meta tag on archive pages instead, you have to provide your own hack. That’s exactly what I did on my blog: I noticed that meta tag, didn’t like it, and modified the theme. Here’s my quick’n’dirty fix:

    In file libclasseshead.php, lines 35 and 38, find:
    $meta['robots'] = '';

    replace with:
    $meta['robots'] = '';

    The above applies to Thesis 1.5.1 (the file path is relative to the theme directory). There’s probably a better way to do it that does not require editing the theme files (e.g., via WP hooks).

  98. Uh, looks like the HTML code was stripped from my comment above. Anyway, just find and replace the two occurrences of “nofollow” with “follow” in libclasseshead.php, and that’s it.

  99. Hey Matt, why are all of our comments NoFollowed here?

  100. >>>Matt, as you know, I was kind of annoyed when you suggested sculpting to a room full of SEOs back in 2007

    Danny, I was on the panel where Matt suggested that and I point blank asked on stage what happened when folks starting abusing the tactic and Google changed their mind if you recall (at the time, I’d seen some of the things being done I knew Google would clarify as abuse and was still a nofollow unenthusiast s a result at that time). And Matt dismissed it. So, I think you can take home two important things from that – 1. SEO tactics can always change regardless of who first endorses them and 2. Not everything Matt says is etched in stone. <3 ya Matt.

  101. Matt, thanks for explaining the whole thing in detail, I’m sure from what’s being talked about, there’s quite a lot of confusion already, and I just hope this post clears it all.

    Simple question – Lets say I have a blog/site with lot of outgoing links (avg 10 links per page). All the outgoing links (in the editorial content and user generated ones) are nofollowed, while all the internal links are “open”. I might have manually “opened up” some links in the editorial content because I’m so sure of their authority (ex:-google faq pages).

    Now, is this a good situation that I have here ?

    Is it right to assume that – I should not be linking at all to websites that I don’t trust rather than nofollowing them ?

    Of course, I agree that it all shouldn’t be this complicated. Stick to quality links (to authority sites), and if at all there are areas where an outgoing link occurs, nofollow it if it makes sense to you.

    Thanks much for your time, and please excuse my ignorance, honestly, its all a bit confusing now.

  102. I think everyone is over analysing this and looking for some advantage that isn’t there for SEO purposes.

    Write good content, link out if its worth linking out and your happy to associate your name against it and if adds depth or explanation, don’t link out if its to bad material neighbourhoods.

    Stop trying to find ways of making search engines push you to the top of SERPs, if your content deserves to be there… it will be there and the PR will follow the page not the other way around. Works for me anyway.

  103. Is using nofollow worthwhile for sculpting anchor text i.e. on “read more” or “click here” links, if there’s keyword rich alternative such as on news stories (title + more both link)

  104. OK so in short adding the nofollow to a link will stop Google from passing juice through that link but that nofollow link can still greatly change the weight and ranking abilities of the page it resides on.


  105. Great article Matt, and it answers alot of questions we have…

    Though I suspect that page rank *is* distributed through nofollow links, or at least nofollow links can still effect search results.

    The reason why I think this is because certain sites come up in google search results for keywords that are not on that page. Upon checking the cache, google kindly informs us that the search term appears in text linking to this page.

    In some cases (on some of my own sites) I noticed that the *only* websites linking to those pages use the nofollow attribute. So, regardless of whether pagerank is distributed or not, it appears nofollow links can still effect search results.

    Would this be accurate?


  106. Hi Matt

    Do run of site links like navigation links, footer links get lesser importance than a normal link? Will the target pages accumulate enough importance (read PR) even if they got lesser each time because there are so many links to them?

  107. bing made me do it

    Why tell us now? Come on…..

    The reason you are telling us now is because that Bing has been grabbing headlines and Google search authority is being questioned.

    I’m not a Bing lover by any means, but the general Google reaction about Bing over the last couple of weeks is consistent with this post.

    If the Google formula wasn’t being attacked with the relative success of Bing, you would have kept people in the dark … Now under pressure, you come out with this post.

    Sad it took competition for Google to start waking up a little and disclose more information about how you index web pages.

  108. It seems that this blog post of Matt is a waste of time since everyone still seems to have the same questions 😀

  109. Is it good practice to nofollow links to areas on a site that are blocked by robots.txt?

  110. Matt, this is an excellent summary. I finally got around to reading “The Search” by John Battelle and it was very enlightening to understand much of the academia behind what led to the creation of Backrub.. er Google.Looking at how many times the project was almost shutdown due to bandwidth consumption (> 50% of what the university could offer at times) as well as webmasters being concerned that their pages would be stolen and recreated. It’s so interesting to see that issues we see today are some of the same ones that Larry and Sergey were dealing with back then. As always, thanks for the great read Matt!


  111. Sounds very clear about Nofollow.Thanks matt for such a clear picture.Previously i saw some back links shown for my site which are with nofollow attribute ( a link from wetpaint page).But now i am clear as u said those are BUGS

    How good is bing towards nofollow ?

  112. Henry Phillips

    Let’s say that I want to link to some popular search results on my catalog or directory site – you know, to give a new user an alternative way of sampling the site. Of course, following Google’s advice, I have to “avoid allowing search result-like pages to be crawled”. Now, I happen to think that these pages are great for the new user, but I accept Google’s advice and block them using robots.txt.

    Under the “old” regime, it’s easy. Nofollow the links to those “unwanted” pages, and the juice flows to the pages that Google wants to index, and that I want to be indexed.

    New regime:

    Option 1: allow those links to be followed, keep excluding the destination pages with robots.txt (this is the one that seems to be in line with G’s recommendations).
    Problem: I’m flowing valuable PageRank to pages that neither I nor Google want indexed. What’s more, the destination pages won’t pass PageRank (because Google isn’t crawling them to discover links). I’m passing PR to what are effectively dangling pages.

    Option 2: keep nofollow on the links, keep excluding destination pages with robots.txt.
    Problem: every nofollow link that I put on my page reduces the PageRank passed to the pages that I want indexed – I have an incentive to remove these links entirely – but I wasn’t gaming Google with these links, I put them there for the users’ benefit!

    Option 3: follow the links, allow Google to index the destination pages
    Problem: in breach of Google’s guidelines re search-result like pages. Don’t want to be flagged as a spammer.

    Option 4: remove the links from the home page and allow the page to flow maximum PR to the pages I want indexed.
    Problem: probably the best SEO solution, but my users have been deprived of what (IMHO) is useful functionality. You may not agree that it is useful, but if it’s my site surely I’m entitled to my opinion.

    It seems that I am left with a trade-off between optimal site structure for SEO and providing (my idea of) useful functionality for my users.

    This seems to run counter to the Google mantra of “don’t worry overmuch about SEO, just create great content for your users”.

    Is it possible to clarify whether the “new” nofollow interacts with robots.txt-blocked pages in the way that I have set out?

  113. An interesting post Matt; I was always a bit unsure as to how Google’s Page Rank system truly functioned. I still have a few questions, but a very insightful post none the less.

  114. This is a great write-up Matt. Thanks for sharing your thoughts.

  115. If I understand this correctly —
    If I was able to write a blog post that was popular and it got lots of comments, then any links that I would have put in the body text would be devalued with each additional comment – even with ‘no follow’ being on the commenter’s links. So it would seem that in some sort of perverse way, the more popular (by comments) a page is, the less page rank it will be passing. I would have to hope that the number of inbound links it gets would grow faster than the comments it receives, a situation that is unlikely to occur.

    Even this post devalues the page a little more 🙂

  116. The big polemic of SMX Advanced!!!

  117. Henry Phillips

    I see that Nick @7.08am basically managed to put my query into a single sentence!

  118. Thanks Matt, I still need the “Geezy” button! LOL

  119. Reading this article on Google’s ( is quite clear. Reading your post is somehow confusing.

    The documentation and you do not say the same thing. So, either the documentation is one year obsolete, either some of us here don’t get it. In short: will nofollow links be followed from now on? Is PageRank sculpting dead?

  120. Sweet, I just updated my blog and removed anything related to nofollow. Plus, all valid too!
    Thanks for the PR Sculpting info..

  121. Concerning your footnote though, It’s interesting that Google doesn’t pass anchor text, but does pass on keywords in the URL itself. One could still obtain a benefit by merely appending additional variables to the URL.

  122. Thanks for the clarification and additional information!

    I think 2 key points summarize everything nicely, I will be sure to recommend this (as I always have for clients that want to Pagerank sculpt)…

    “1) making great content that will attract links in the first place, and 2) choosing a site architecture that makes your site usable/crawlable for humans and search engines alike.”

  123. Matt,

    Thank you for providing such a straight-forward explanation of the nofollow changes and page sculpting issue. It helps to have such a clear picture after all of the back-and-forth and analysis on other (some of them very good) blogs.

    So, at the end of the day, implementing nofollow simply reduces the amount of authority a page will pass on to subsequent pages. The ‘points’ that had once been redistributed to non-nofollow links now simply fall into a HTML Twilight Zone. You can no longer hoard and pass PageRank on a few links.

    What does this do to what many may believe are good uses of nofollow, such as footers, log-in and other template like links that are necessary to the consistency of a site but really provide no contextual value to the content of that page? Are there some exceptions here, or is it all or nothing?

    I think something else I’m hearing here is that Google is making this change to ensure that sites are built for optimal human consumption. Personally, I’m a proponent of The Paradox of Choice, so I am always looking to put less on the page to ensure users get exactly what they expect and have a clear path of where to go next.

    The nofollow attribute made it easier for sites to disregard this principle and instead put more on a page and then simply redact what seemed superfluous. So, would it be accurate to say that the nofollow change was implemented because it prevented the algorithm from emulating human analysis of a page/site?

    Could the nofollow change could be interpreted as a form of usability guidance? For instance, I’ve recently removed drop-down menus from a handful of sites because of internal link and keyword density issues. This wasn’t done randomly. Tests were done to measure usage and value of this form of navigation that made it easy to make the change – allowing usability and SEO to dovetail nicely.

    Essentially, the 100 link per page guidance is being maintained.

  124. The article helped, but I think I’ll learn more about it at WordPress Camp Dallas. Thanks for the article!


  125. thanks matt. i still consider you has more than just a valuable source of information regarding seo… 🙂

    btw, you might want to check this out

  126. Very interesting. Really useful article.

  127. Very interesting and informative post – I will link back to this sometime this week.


  128. Thanks for the update and the clarification on several points. I’ve never been a fan of PageRank Sculpting and I’m now glad that I didn’t go down that route.

    I’m still amazed that the certain changes that Google makes to the algorithm go unnoticed while other factors get a lot more attention.

  129. I am not worried by this; I do agree with Danny Sullivan (Great comment Danny, best comment I have read in a long time). I will not be changing much on my site re: linking but it is interesting too see that Google took over a year to tell us regarding the change, but was really happy to tell us about rel=”nofollow” in the first place and advised us all to use it.

    SEOmoz has an article that offers a great amount of information too: VERY INTERESTING FOR ALL webmasters.

  130. I’d like to echo Danny Sullivan with the unintended (?) consequence that Matt is rather glossing over so far:

    1. Apparently, external linking of any kind bleeds PR from the page. Following or nofollowing becomes a function of whether you want that lost PR to benefit the other site. Since nofollow has ceased to provide the benefit of retaining pagerank, the only reason to use it at all is Google Might Think This Link Is Paid. Conclusion: Google is disincentivizing external links of any kind.

    Muratos – I’ve never nofollowed Amazon affiliate links on the theory that search engines probably recognize them for what they are anyway. I have a blog, though, that gets organic traffic from those Amazon products simply because people are looking for “Copenhagen ring DVD” and I hard-code the product names, musicians’ names, etc. on the page rather than use Amazon’s sexier links in iframes, etc.

    But this leads to a question — if my husband wants to do a roundup of every Wagner Ring Cycle on DVD, that’s about 8 Amazon links on the page, all bleeding PR away from his substantive insights. If he, instead, wants to do a roundup of every Ring Cycle on CD, that’s about two dozen items worth discussing. The page would be very handy for users, and would involve considerably more effort on his part… but no good deed goes unpunished, and in the eyes of Google the page would be devalued by more than two thirds.

    Result: Knowing what I know now, I would put the Amazon items on another page and offer one single link to that page. In other words, actively inconvenience the user because of a decision made by Google. Was this really the intent?

    2. Does a nofollowed INTERNAL link also bleed PageRank? Doesn’t that actively punish webmasters who use nofollow in completely . I think Danny makes the case that nofollow at the link level isn’t a cure for duplicate content, but many hosted sites and blogs don’t have the full range of technical options at their disposal. I myself use a hosted service that’s excellent for SEO in many ways but doesn’t give me a per-page HTML header in which to put a canonical link tag. Conclusion: Google would rather we NOT hide duplicate content if nofollow is the most straightforward way to do it.

  131. As amit agarwal mentioned, i would not want to have my privacy policy, contact us pages to show up in search results.

    so would you recommend to keep the nofollow on these or take them off?


  132. Matt, thanks, I revise my explanation to:

    “Google itself solely decides how much PageRank will flow to each and every link on a particular page. In general, the more links on a page, the less PageRank each link gets. Google might decide some links don’t deserve credit and give them no PageRank. The use of nofollow doesn’t ‘conserve’ PageRank for other links; it simply prevents those links from getting any PageRank that Google otherwise might have given them.”

    Getting closer?

    I agree. PageRank has been a black box. If you’re wanting to open up more about it to help people form mental models, then we now seem to have these two major factors:

    1) Number of links on a page matter. The more links you have, the less PageRank any link is likely to get.

    2) Perceived link importance matters. If Google decides a link is somehow important on a page, it will allow more PageRank to flow to that link than some others.

    Are there other major factors you can share? For instance….

    A) Do internal links get less PageRank than external ones (IE, links within the same domain versus link leading outside a domain)?

    B) Do repeated links within a domain get less PageRank or discounted automatically? (IE, if you see the same link for “Privacy” on all the pages in a web site over and over, do you simply stop assigning PageRank to it?

    I’m sure there are plenty of other questions out there. And knowing answers to things like (B) could cause some people to stop worrying about whether they should nofollow navigational links and so on.

    In terms of the nightmare that I’m worried about — and that you say has been running quite well for over a year — we’re talking different things.

    I agree, there’s no nightmare in people who decided to do PageRank sculpting needing to suddenly remove nofollow. If they’ve been doing well, then don’t fix what ain’t broke. That’s the basic rule of SEO. Don’t mess with a site that’s performing OK.

    I still think you’re going to cause a new form of sculpting, where people will remove links from their pages other than using nofollow, in hopes flowing PageRank to links they think are important. You’ve said number of links matter — and that nofollow doesn’t reduce those links — so some will keep chasing after whatever extra oomph may be out there.

    Philipp, you’d asked:

    “If the 5 links that were previously nofollowed are not meaningful, than they shouldn’t be on the site in the first place, or should they? Googlebot should not be treated substantially different from visitors. If you’re exposing humans to the links, they should be meaningful, and why should meaningful links not be evaluated by a search bot?”

    PageRank sculpting came out of the idea that virtually any page will have links that are important for users but not necessarily that meaningful to receive any PageRank that a page can flow. Navigational links are a primary example of this. Go to a place like the LA Times, and you’ve got tons of navigational links on every page. Nofollow those, and you (supposedly in the past) ensure that the remaining links (say your major stories) get more of a boost.

    This tactic wasn’t new and existed in ways before nofollow. But it shot up in popularity after Matt pointed out how YouTube was using it as a way to reward important pages. That’s what bothered me about the advice. It was contrary to the regular advice of don’t go to extremes for search engines.

    By the way, YouTube currently is all over the place. It nofollows links in the Spotlight and Featured areas, where you assume there’s some editorial oversight. But since some of these show on the basis of a commercial relationship, maybe YouTube is being safe. Meanwhile, Videos Being Watched now which is kind of random isn’t blocked — pretty much the entire page is no longer blocked.

    Seth, you’d said:

    “Correct me if I am wrong but what Matt is actually saying is that the nofollow thingy is already going on for at least a year… if nothing shocking happened to your website within that year what are you guys all worrying about?”

    A couple of things are worrying some people….

    1) Google suggested a tactic that was picked up by some advanced SEOs, and now those SEOs worry they’ve inadvertently flagged their sites for greater attention. If Google sees a lot of nofollow, then it might decide it’s found a site that went too heavy on the SEO and apply penalties, etc. Personally, I wouldn’t worry about this.

    2) Some people sculpted understanding that Google worked in a particular way, and now learning it has changed, assume they must change things or somehow be hurt in rankings. That’s not the case. Google may no longer be rewarding sculpting as it had, but that’s not the same as punishing it. If things are working fine, leave them alone.

    3) Some people don’t believe things have changed. In fact, if things really changed substantially a year ago, you’d think a few of the advanced SEOs out there would have noticed this and talked about it. But nada. There are lots of reasons why the change could have happened and not been spotted. Sculpting might really have been a second or third order factor, as Matt calls it — not helping things as much as some have assumed. SEOs that spotted it might have stayed quiet. Or, it didn’t change — and still hasn’t changed — and sculpting does work even better than Google thought, so it wants to put out a message that it doesn’t, in hopes of putting the genie back in the bottle. That’s probably the major conspiracy theory out there.

    Me, I didn’t like the sculpting idea from the start. I linked to what I thought should get links and figured that was pretty natural, to have navigational links, external links and so on — and natural has long been the think Google’s rewarded the most. So I didn’t sculpt, even after Matt helped put it out there, because it just made no long term sense to me.

  133. Matt I am so confused. That’s why I just build my websites and try to earn links. I will get old pretty fast if I try to stay on top of what Google is doing. I’m done reading SEO blogs, they are all speculators. Pagerank Sculpting has been the hot topic for a while now and for you to say it hasn’t worked for about a year.

  134. Hi Matt,

    Great post. I’m posting a link back to this article from our blog along with some comments. I do have a question. In your article, you post “The only place I deliberately add a nofollow is on the link to my feed, because it’s not super-helpful to have RSS/Atom feeds in web search results.” Yet when I look at this article, I noticed that the comment links are “external, nofollow”. Is there a reason for that?


  135. Very useful article!

    I’ve been having trouble upgrading the pagerank of my 3d animation webcomic website.

  136. Great article. But for me, it came too late, I prefer using Bing now rather than Google search.

    Just Bing.

  137. Matt:

    First of all, let me say thank you for bringing this topic to light. Would you suggest in all fairness that simply being more selective about which links (leak) where within a site and consolidate them, or is there another significant method we could implement to avoid such a broad link graph within a website.

    Let’s face it, leaky pages with off topic links within a site should be avoided to achieve optimal relevance, so what method is considered ethical and will not be frowned on from Google? (javascript on click commands, the concept of using Iframes or url shorteners with nofollow?, etc.).

    Any ideas you can pass on?

  138. I’m looking forward to the new WP comment plugin that displays all comment URLs as plain text for us to copy and paste into the browser, as well as the ensuing Greasemonkey script that converts them back again!

  139. Hi Matt,
    I agree wholeheartedly on what Danny said. Nofollow will turn out to be a link-killer which ain’t good for the web. Sad thing.

  140. If anything it suggests that a) you pay attention not to ways to ‘cheat’ the SE’s but regular things like internal page navigation.

    b) it also suggests that people will comment and you will enable them to comment in blogs regardless of link, but you’ll be more stringent in spam detection. Dumb comments deleted, good ones stay….. wasn’t that the point of the net/community?

  141. One thing I think everyone is forgetting (particularly Michelle and Danny Sullivan) is the first part of Matt’s post. The idea that every link gets $1 (or that Michelle has flushed 25 links worth of PR down the tubes) is based on the random surfer model and no longer holds true.

    If Google’s already decided to devalue links in comments, as they have with blogroll links, then what’s the point in keeping them, leaving them, or making them all purple instead of blue?

    There’s a need for a skilled SEO to assess the link structure of a site with an eye to crawling and page rank flow, but I think it’s also important to look at where people are actually surfing. The University of Indiana did a great paper called Ranking Web Sites with Real User Traffic (PDF). If you take the classic Page Rank formula and blend it with real traffic you come out with some interesting ideas……

  142. Advanced link analysis includes differentiating between sections of pages and treating links differently. What makes you think G or other engines treat links in the editorial section and comments section of the webpages the same as each other. Especially for those content management systems that are widely in use like wordpress, joomla, etc. The advice here is helpful and has nothing to do with creating a nightmare. All those who are asking questions here and envision a nightmare would agree that links in the footer section are treated differently. How is that possible if sections on a page are not classified and treated differently.

    I still do not understand the obsession behind manipulating something that you cannot measure. Define, measure, analyze, improve and control. With the information available or lack of information in this case, you only define the problem. What you cannot measure, you cannot analyze.

    With the amount of data G is working with, decisions have to be taken to benefit the majority of users and improve the quality of the search experience. Not, a few site owners that try to change the importance of their pages via manipulating links and limiting bot access.

  143. Regarding:

    “I wouldn’t recommend it, because it isn’t the most effective way to utilize your PageRank. In general, I would let PageRank flow freely within your site.”

    Maybe the best thing for a little larger fact site would be to choose “10” articles and news stories which is most relevant for the reader? But if so is it anything else to think about that? Should one for example use “rel=tag” for pages of for the “keyword” topical top relevance?

    And looking at say references would it be a problem to link both the actual adress of a study and the DOI (read DOI as anything similar)? Even if they terminate at the same location or contain the same information? The is that it feels better to have the actual adress since the reader should be able to tell which site they reach. But also the DOI have a function.

  144. Matt, could you tell me in which month the change rolled out please? I’d like to go back through my stats and see if I can line it up.


  145. What about wikipedia?They are no following all their external links but the internal links are all do follow…Plus they get all the backlinks from tons of other pages…

    This way the dominate the serps using the no follow attribute to circulate the pagerank juice amng their pages…

    Why don’t they get slapped?


  146. Wow that was a lot of reading from your stuff (Matt) and all the comments.
    I have just one quick question for you Matt—

    If it’s been 1 year for the change in “no-follow” – why didn’t you remove the “no-follow” on all the comments on this new blog design? Seems to me if you really believe that the “no-follow” has been changed and offer no additional value then wouldn’t you have just removed it from your own site?

  147. It is also hard to know who is using non-follow on their blog comments as well. I guess we should all open up our blogs to comments and try to delete the ones that are obviously spam. That could be more work for us.

  148. An interesting blog post, and even more interesting debate in the comments. Danny raises some interesting points.

    The fact this change happened a year ago and hasn’t been picked up by the SEO community suggests that it hasn’t made a big difference. I think if Google worked anything like the simplified example then it would have been immediately apparent.

    I too would be interested to get a bit more information about how the PR flows. In particular how does Google treat links in the master template of a website (i.e. main navigation, footer links – privacy, term & conditions etc).

  149. My page rank on went from a 7 to 4 last month. It was akin to waking up in the morning and finding that the family jewles shrank 45% in size. I am on a quest to build my site right. Any sugestions Matt?

    My guess is a static home page, link to the best stuff to help visitors and bots, remove any dup contnet, take all outgoing links off the home page to other sites??? Are these my issues?

    Help Please!



  150. Why does this have to be so confusing for everyone Matt???????????

    You got one person (Cutts) saying that “Nofollow links definitely don’t pass PageRank.” and then you got other people (including Cutts sometimes) describing links out leak juice from a page with the bucket theory. Ok so if I no follow all those comments links in the comments I am OK right? Nope not according to Matt today as he said:

    “So what happens when you have a page with “ten PageRank points” and ten outgoing links, and five of those links are nofollowed? Let’s leave aside the decay factor to focus on the core part of the question. Originally, the five links without nofollow would have flowed two points of PageRank each (in essence, the nofollowed links didn’t count toward the denominator when dividing PageRank by the outdegree of the page). More than a year ago, Google changed how the PageRank flows so that the five links without nofollow would flow one point of PageRank each.”

    With that statement about it makes it sound like links with nofollow do flow (or leak, but not pass?) PageRank, LOL

    Can we make this any more confusing than it has to be?

  151. LOL @ the majority of the comments…. Paranoid hyper-SEO at it’s best

  152. Matt,
    Thanks for taking the time to address this issue.

    I think it would go a long way if you addressed the concern of a publisher like me who has several successful pieces of content which have hundreds of comments on them with comment name links.

    How would you recommend we proceed? It seems that based on what you say comment name links dilute the ability of a page to pass Rank. Should we remove all the comment name links? Should we move comments to a separate page? What would you do?

    Knowing how Google handles comments in general (both from a link perspective and content perspective) would ease a lot of the concerns we see addressed here.

  153. One question:

    From a customer experience perspective, we currently have three duplicate links to the same URL i.e. i.e. ????.com/abcde These links are helpful for the visitor to locate relevant pages on our website. However, my question is; does Google count all three of these links and pass all the value, or does Google only transfer the weight from one of these links. If it only transfers value from one of these links, does the link juice disappear from the two other links to the same page, or have these links never been given any value?

    From a customer perspective, these links are very valuable as it gives them different options on a page (from above and below the line), but is the website losing value by having these links? I’m hoping not- as it could become a weigh-up between customer value and SEO.

    Any advice would be appreciated.

  154. I’m growing tired of this game between Google and the rest of the online community about how to “manipulate” my content and code to better rank in your system. It seems that you guys have completely over complicated the game. If I add a nofollow tag, why on earth would any page rank be added to that link. I just told you to NOT FOLLOW it! The fact that it receives any rank at all is absurd.

    Honestly, this I’ve read your blog for about 4 or 5 years now and the more I read the less I cared about creating new content online because it feels like even following the “Google Rules” still isn’t the way to go because unlike standards, there is no standard. You guys can change your mind whenever you feel like and I can become completely screwed. So screw it. I’m done trying to get Google to find my site. With Twitter and other outlets and 60% of all Google usage is not even finding site but Spell Check, I don’t care anymore.

    I’m done. Done worrying, done “manipulating”, done giving a damn. I spent 10 years learning semantics and reading about how to code and write content properly and it’s never helped. I’ve never seen much improvement, and I’m doing everything you’ve mentioned. Reading your blog like the bible. The most frustrating part is my friends who don’t give a damn about Google and purposely try to bend the rules to gain web-cred do amazing, have started extremely successful companies and the guy following the rules still has a day job.

    So while I enjoy your writings and trying to crack the nut that is Google, I don’t care anymore. It’s become too much work, and it’s hard enough just coming up with original content.


  155. I was really hoping that the nofollow would cut down on comment spam, but it has not.

  156. Hi Matt – what if you have a bunch of domains you own and you link them to each other, does having the same WHOIS info on them help SEO or hurt it.

    I am linking them because they are related services and would prefer to keep one WHOIS. But want to make sure Google does not ding SEO?

  157. Thanks for clearing that up Matt. Good to hear that someone cares about letting us poor SEO’s know.

  158. You cannot prevent people in the SEO industry from buying into nonsense and bad advice. PageRank sculpting has always been a waste of time and resources.

    However, despite numerous attempts by Googlers over the past two years to persuade people to stop engaging in BAD SEO PRACTICES, the SEO community has continued to try to sculpt PageRank.

    It’s Google’s index. Let them manage it anyway they please. A truly good SEO would never develop a site just for Google anyway. There are plenty of other search engines out there and collectively they still draw more monthly searchers than Google.

    But then, the SEO community (which now holds itself to the standard of measuring success by number of conversions) is still focused on the obsolete metric of counting pageviews for measuring search market share rather than actual search conversions.

    No one should be surprised at all the weeping and gnashing of teeth going on right now in the flames of PageRank hell.

  159. If we have the page A that has 3 links (internal) and one of them are nofollowed that points to a page that has been blocked by robots.txt. How many pagerank will the other 2 pages receive? 1/3 or 1/2 of the pagerank of A?

  160. Guys, be sure to comment as much as possible on this post to reduce Matt’s flow of PageRank.


  161. Hi Matt,

    Will these changes make a difference to “high-authority” sites plagiarizing content from small sites. Many high-authority sites often don’t link to the source of the story as this can pass page rank and devalue their page. I can give you specific examples for this.
    I think Google should do something about this for sure.

  162. Thanks for explaining this Matt. I have just disabled commenting and deleted every comment my site had ever got. Let the PR flow, baby!

  163. Matt, Thanks for the super explanation. I have been reading about this topic all over the internet and I actually never understood what just have changed.

    Thanks for the clarification

  164. Yes i did notice this change back when it happened because my forum tanked in Google, i didn’t know what caused it though.

    I use nofollow internally fairly extensively, my forum has Linear, Threaded, Hybrid, Print Friendly viewing modes plus each post has a link to a single showpost version. All these look “horrible” when you land on them from the SERP’s, they are duplicate and/or they are not intuitive or good for the user.

    I don’t give a hoot about Pagerank or sculpting it, i was simply doing Googlebot and the search traffic a favor by directing everything to the one normal version of content using a combination of nofollow/robots.txt and i got pounded for it.

    Gee thanks Google!

    Oh well time to kill user experience and open up approx 20 Million worthless pages to consume crawl budget so the worthwhile stuff takes forever to crawl. *sigh*

  165. But a question Matt, would it not be better for Google to treat external and internal links with the attribute nofollow differently? My concern is with articles I have on my blog that have several hundreds of comments.

    For example this page. My program found almost 400 nofollow links on this page. (Each comment has 3). And then you have almost 60 navigation links. My real question is how much percentage of the PageRank on this page gets distributed to the 9 real links in the article? If it is a division of 469 which some SEO experts now are claiming it is really disturbing. You won’t earn much from the links if you follow what I am saying.

    I hope Danny is correct by saying the Google naturally gives more PageRank to certain links than others.

  166. To answer the guy in the comments who said that the more he reads the less he understands about SEO.

    SEO *should* be simple. But its not because of things like this post about moving goalposts. But consider this, if it was made truly simple, then the SEO industry would not exist, hence SEO professionals generally love this continued confusionas it means mortals hurl cash at them!

  167. I haven’t followed SEO in awhile because when ranking well in the past doing what was suggested by Google at the time, I’ve had my legs cut out from under me a couple times. So, everybody who is chasing it now, will still be chasing it in years or until they give up, stick to the basics and forget about it. Traffic will come and it will go.

  168. Matt and Danny,

    I guess all of the comments on this post just shows how much people care about all of this. In essence this is great for Google (how sad would it be if nobody cared when you make a change) and also good for the ‘good’ part of the SEO community (shows we are taking all of this seriously).

    Spam is a poison that in different ways (and in different names) affects many things. Matt, you and your guys do a great job in trying to keep it at bay. But, as mentioned before, with that role and power, you set the rules for the web in many ways. As I have said before even though the JavaScript link change is not (in Danny’s words) backward compatible, it is understandable. I will maintain that the PageRank sculpting thing is not the same.

    Yep, please change things to stop keyword stuffing. Change them to stop cloaking. Definately change them to stop buying links that try to game Google. But, telling search engines to not give weight (that I control) to pages that are not what my site is about or are not really relevant. No way. This is logical stuff here. Maybe too logical. I think deep down you know this Matt too.

    Just because some people have been turning their page, way to, pink (with the Firefox ‘nofollow’ indicator plug in installed) that is not a reason to devalue something that is OK to do. It would not of been that hard to plug in a change that would pick that up as spam and therefore put a ‘trust’ question mark against sites that have been ‘nofollowing’ everything.

    My final (thank goodness) point on this is not that (white hat) PageRank sculpitng was really anything special. It was just quite logical. It really feels like we are going down a wrong route here. Shall we outlaw cars because some people drive dangerously? Or should we do all we can to make driving safer? Not on the same level in any way, but you can see my point here. This is the first time I have felt that you have made a bad call and that is the only reason I am making a case for the logics of this.

  169. (Hey everybody, just wanted to let people know that today is my crazy meeting day, so I’m tied up right now. I’ll circle back around tonight to respond to at least some of the comments.)

  170. I’m with Reuben Yau, it would really help me analyze some data.

  171. Well Michael M!

  172. Great topic and most of the discussion is interesting as well. Love the feedback from the “Power SEOs”. Matt thanks for the response to Chris_D’s comment:

    Chris_D, great question. If you have a single product page that can have multiple urls with slightly different parameters, that’s a great time to use a rel=canonical meta tag. You can use rel=canonical for pages with session IDs in a similar fashion. What rel=canonical lets you do is say “this page X on my host is kinda of ugly or otherwise isn’t the best version of this page. Use url Y as the preferred version of my page instead.” You can read about rel=canonical at Bear in mind that if you can make your site work without session IDs or make it so that you don’t have multiple “aliases” for the same page, that’s even better because it solves the problem at the root.

    I can also appreciate Mr. Sullivan’s request for additional clarity on:

    A) Do internal links get less PageRank than external ones (IE, links within the same domain versus link leading outside a domain)?

    B) Do repeated links within a domain get less PageRank or discounted automatically? (IE, if you see the same link for “Privacy” on all the pages in a web site over and over, do you simply stop assigning PageRank to it?

    Great discussion.

  173. Does this also mean that the ‘reply’ links in comments in wp2.7 (enabling threaded comments) will ‘evaporate’ pagerank too? I notice you don’t have them …

  174. Hey Matt, Internet tends to copy the experience in the real word, where any body can suggest (by mouth to mouth) any good source, this is like a link, but if we use the “nofollow” attribute, this is like say:

    I like coffee from “the corner`s shop” but don`t listen to me.

    Or we refer to something or no.

    In internet we have the option to refer somebody (good source) but say something like (don`t believe or not follow my vote or not listen to me) … That`s not good, that`s breaking the normal rules.

    I understand in Internet any search engine can place his own rules, but one of the goals of search engines is offer a real (or natural) experience to users, if that is the case, the best way to offer a real experience is “reading what people is writing” in other words, following the links like a real person will do.

    😉 my 2 cents..

  175. Malcolm Coles – It’s a moot point IMHO. (Matt can correct me if I’m wrong). Since Google only credits the “first link to any destination url” whether you had 1 or 1,000 to the same reference from the same reference it would be treated the same (follow or nofollow)

  176. Q: Does this mean “PageRank sculpting” (trying to change how PageRank flows within your site using e.g. nofollow) is a bad idea?
    A: I wouldn’t recommend it, because it isn’t the most effective way to utilize your PageRank. In general, I would let PageRank flow freely within your site.

    In my experience this means (the key words are “not the most effective way”) a page not scored by Google (“e.g. my private link” – password protected, disallowed via robots.txt and/or noindex meta robots) whether using or not using rel=”nofollow” attribute in ‘links to’ is not factored into anything… because it can’t factor in something it isn’t allowed.

    I haven’t tested Javascripts and jumpscripts… so these could be either or.

  177. Thanks for all the efforts to spread a better understanding of these concepts.

    I would like to be sure I am understanding what you are saying. My simple mental picture is that in a first pass all outlinks on a web page are assigned a fraction of the web page’s PageRank. Subsequently for some of these that have been no-followed the PageRank contribution ‘evaporates’.

    I take it that for all the other outlinks from that web page their PageRank contribution stays what it was and is unaffected by whether or not other outlinks have suffered an evaporation. In other words, there is nothing you can do on that web page to increase the PageRank contribution of those outlinks that are not labeled nofollow.

  178. Matt,
    I think it is important you distinguish your advice about no-following INTERNAL links and no-following EXTERNAL links for user-generated content. Most popular UGC-heavy sites have no-followed links as they can’t possibly police them editorially & want to give some indication to the search engines that the links haven’t been editorially approved, but still might provide some user benefit.

    Rather than advising UGC sites to now dissallow all UGC linking to preserve internal pagerank, I’d like to think that no-follow tags to external sites is no different than it was a year ago and that the changes you speak of are mostly related to internal linking.

    I’d love to see you address this as there appears to be a flurry of “what should I do” for UGC-heavy sites and for comments on blogs, eventhough I believe this change is mostly targeting internal linking practices.

    Thanks in advance for a little more clarification!

  179. Let’s say I have a reasonably respected tourism website. I include a restaurant directory
    (spread across a number of pages) which lists something like 1,000 restaurants in a large city with contact details and a web link to each of those restaurant’s home page. Given that the outgoing links are relevant to my content, should I or should I not be using REL=nofollow for each link given the massive quantity of them? How will my ranking for pages containing those links and pages elsewhere on my site be affected if I do or don’t include REL=nofollow for those links? My fear is that if I don’t use REL=nofollow, Google will assume my site is just a generic directory of links (given the large number of them) and will penalize me accordingly.

  180. Ah, ic see what you are trying to accomplish now. It’s all crystal clear.


  181. We now know one might use a rel=canonical tag to clear up duplicate content issues of internal links (I’ve seen it do wonders for large newspaper sites); instead of using the nofollow on every link within a site (less work) for the ecom and large db sites. (Thank you)

    There seems to be no end to the discussion, but taking away granular control of link relevance puts the onus back on Google’s (or Bing et al.) to show us how they can crawl the web with ever increasing accuracy while it keeps growing eventually 100+ billion pages & closer to a trillion links.

    With this in mind, should 3-4 query strings always look more like a brand/paid search farm with .5 million “useless” results or will weighted link decay allow for more narrow results to win out.

    We have to remember that Google’s $ model+bots to scour the web have to tow the same line so they can optimize their own pocketbook, balancing a free and open resource – ie. the www, all while taking money from the natural competition that arises from their market share. On the one side, its all about appearing fair and the other, to drive competitive output.

    If Google was to allow webmasters full control of their own fate it would be like giving up the farm rather than giving up to the forces of human creativity. If you feel today were in a crowded market place even with a Google’s superiority complex, wait until the web is completely machine readable and aggregated on pure laws of information. I don’t think most can comprehend the future of data management as we have yet to see readily available parsing mechanisms that evolve purely based on the principles of information theory and not merely economies of scale. Remember not too long ago when Facebook tried to change their TOS to own your links and profiles? We can see that the tragedy of the commons still shapes the decision of production with that of opting in.

    But at some point when Google halves overall link weight due to something like inflationary value, we will see there is nothing left but the proverbial line in the sand…you either step it up or move out of the way of the forces of interaction-mechanics.

    I say this because as Google is watching its own tailspin we normally see the relative growth the web in a matter of years working like the old web maker (spider+crawl) But a system that is exponential has the potential to become (node+jump). All the copy and wonderful content aside, the real use of the tool that is now called the internet will be discovering along the way, what some might call cybernetic or rather android-like mainframes for eco-stellar exploration, or instant language learning, or even mathematical canon though cloud computing.

    Stay 22nd century for your bread.

  182. I use no-follow links to avoid being suspected of having paid links. I mean when someone pays to advertise on my site, they are paying for the click throughs, not the link value. They have a no-follow attribute on their advertisement so Google won’t think I am hosting paid links.

    Does this make sense, and am I on the right track?

  183. Matt, I have a question about PageRank: if a page has a “noindex” meta tag or is disallowed by robots.txt, will it accrue the same amount of PR as if it were indexable?

    I ‘ve always held that a page’s PageRank depends exclusively on its backlinks’ PageRank, but maybe now things have changed, who knows? 😉

  184. Huh? Oh so confused with so many posts. Absolutely brilliant to read all these opinions but Google really needs to provide some facts as I just think these are still opinions.

    I agree that the more facts that you provide and if you were to provide the complete algorithm, people would abuse it but if it were available to everyone, would it not almost force people to implement better site building and navigation policies and white hat seo simply because everyone would have the same tools to work with and an absolute standard to adhere to.

    Google have set the “bar” with search, i think they now need to set the “standard” too and share that with all of us. Come on Matt, knock on Larry and Sergey’s door and give us some definitive guidance.

  185. I have went through this article and the comments above, what a shame to see some people getting tired of optimising their websites properly.. I have always enjoyed doing extreme SEO as its just needed to be done properly, google is the great search engine and among Indians its known as Google Baba! lol..

    Thnx Matt, for sharing your knowledge like always, I always learn good things at your blog and stay updated..


  186. Thanks for the great post Matt. It does help to better understand the big picture.

  187. The flood of iframe and off-page hacks and plugins for WordPress and various other platforms might not come pouring in but I’m willing to bet the few that come in will begin to get prominence and popularity. It seemed such an easy way to keep control over PR flow offsite to websites you may not be ‘voting for’ and afterall, isn’t that way a link has always represented. It would seem Google should catch up with the times.

  188. Matt, don’t you fear creating an uneven playing field for Google if we have 2 basic Webmaster types?

    1) Those who read your blog, along with other SEO sources and apply everything that will help them rank better.

    2) Those who simply write content and sites for their users and couldn’t give a Rats arse about SEO.

    It would seem to me that MOST searchers would want the latter site type 2 pages over the type 1. Over the Years, Google has seemed to move more and more to relying on Webmasters knowing about SEO to return relevant results.

    IMO, Google (and yourself) is either feeding us bullshit, and/or useless information, or Google is relying more and more on Webmasters knowing SEO.

    For example (and ONLY an “example”), your post last week had the title: Straight from Google: What You Need to Know bolding added by me. Why do we “need” to know these things?

  189. Thanks for such a great post and all the useful comments!!

    Recently being torched for aggressive linking to keep up with competitors and doing things like PR sculpting because they were too. This is very helpful information. We have been undoing as much as we can and removing all the internal no follows was one of the items we have done. We have also gone back to linking to useful sites for our users without the no follows.

    Let the love flow if it is good for your users 🙂

  190. This whole PageRank discussion just keeps coming back. I like the math of PageRank, or actually I should say the basic principles of it: The information that’s contained in links and what can be extracted and then used for ranking purposes.

    As far as I can tell there are 2 types of people interested in PageRank:

    1) Those that don’t understand SEO
    2) Those that are extremely aggressive SEO’s.

    The first group often has small sites and what ever they do it’s got little effect and the only SEO they apply that has any noticeable effect are titles, metas and H1’s.

    The second group often has huge sites (or a lot of sites) in which messing with PageRank may actually have some effect and they probably also have the tools to work on a lot of links in little time.

    In my opinion, both groups are wasting their time, but I can understand the interest because PageRank is cool and it makes you look smart if you know a thing or 2 about PageRank. Also it’s something that is very easy to sell to your customers: “We´re going to work on your PageRank!”

    The extra power of SEO though (besides titles, metas and H1’s) is site architecture, focus and understanding your visitors. Those beat PageRank manipulation every time, but it is a lot harder to get right and worse, not easy to explain to customers. (In fact, you´re better of talking about this after you sold SEO services, if at all.)

    With focus I mean making sure that your pages focus on the same keyword everywhere, and your site focuses on the same high level keywords and sections in your site focusing on their own high level (but not as high as the keywords for which you want your home page to rank) keywords. Focus few people really understand while the interesting thing is that you do this almost automatically right if you do your site architecture and understanding your customers, right.

    However, PageRank is still a cool thing. I was surprised to read before that the average number of links per page is only like 10 per page. There must be tons of pages out there with just a couple of links because I can’t imagine building a site with just 10 links per page. Or did you mean “unique links”, Matt?

  191. Thanks Matt. Thats really helpful.

  192. Damn, its getting more confusing now.
    Matt, I’ve been a firm believer of the thought that webmasters shouldn’t really bother too much about the calculations that Google would do while spotting external links on a site. Leave that to Google. You write the content and if you find relevant resources, link to it. Why worry over PR ? In case you’re so sure about the linked site to be “kinda spammy” then nofollow it. That’s it.

    Also, I’ve never found that page rank scultping worked. It might have for smaller sites that have a simple structure to follow. but in case of CMS’s handling a large number of pages and dynamic websites, its not practical to have an intricate graph of how your page rank flows. I mean, even if you did, wouldn’t it be easy and clever if you just leave some thumb rules (like always nofollowing an external link) and leave it to Google for the rest ? Rather ocus on the content ?

    So, I’d appreciate if you can give us a simple solution to this, may be a few “guidelines” to follow while using the nofollow tag, or linking generally, so that a webmaster can just write it down his white board, and focus on his content rather than fretting over complex calculations.

    I’m so damn sure there are many “Google friendly” websites out there which doens’t really follow any of these rules but have great presence on the serps, purely because of their “awesomeness” in the content.

    Thanks for your time.

    Cheers !
    Mani Karthik

  193. I think this change is for the bad. It’s only going to encourage more spamming on blogs.

  194. It is odd how no one complained before google went public about this. I wonder how many complaints matt got before the announcement even though the change took place over a year ago. I was still using them. Don’t act like you weren’t

    The one thing I like about this is that it refocuses SEO’s on what is really important. Stop using tricks and figure out how to help clients really help their clients.

  195. “It is odd how no one complained before google went public about this. I wonder how many complaints matt got before the announcement even though the change took place over a year ago.”

    david wolf, no one had complained or even noticed, which is another datapoint that helps point out that this isn’t a high-impact issue in general.

    Everfluxx, a page that is blocked by robots.txt can still accrue PageRank. In the old days, blocked Google in robots.txt, but we still wanted to be able to return for the query [ebay], so uncrawled urls can accumulate PageRank and be shown in our search results.

    Bob Dole (interesting name), you’re certainly welcome to use Bing if you prefer, but before you switch, you might check whether they do similar things. I know that Nate Buggia has strongly recommended not to bother with PageRank sculpting in the past, for example, or at least that was my perception from his comments at the last couple SMX Advanced conferences.

  196. In my opinion, both groups are wasting their time, but I can understand the interest because PageRank is cool and it makes you look smart if you know a thing or 2 about PageRank. Also it’s something that is very easy to sell to your customers: “We´re going to work on your PageRank!”

    You lie and feed spin to YOUR customers, nice to know.

    Mani, could not agree more your statements. It’s no wonder the SEO industry has such a bad name and it’s 99.9% Snake Oil. Still, Google, this Blog and other “SEO” sites are partly responsible for the PR hysteria, link spam and email spam for PR. Google should also put an end to Webmasters being screwed by SEO, but placing a BIG prominet statements on their Webmaster pages along the lines of;

    There should be ONLY one thing Webmasters need to know, and that’s the subject matter of their site. Take 2 extremes. Two sites about fishing a specific coast line area.

    1) Webmaster 1 knows everything about SEO, HTML, JavaScript etc etc and LITTLE about the subject matter and relies on research MORE than experience.

    2) Webmaster 2 knows everything about fishing the specific coast line area through Years of experience and LITTLE about SEO, HTML, JavaScript etc etc.

    Who would rather see above who in the SERPs?

    If Google is now relying on even tiny bits of SEO elements, it’s the beginning of the end for Google, IMO.

  197. Perhaps, my first comment on your blog.
    Thanks for informing..

  198. Hi Matt,

    I have a question let’s say I discovered a very cool little site and I linked to it. I was one of the first 10 people to link to this site.
    Fast Forward 10 months.

    The site I linked to became a web monster PR 10 and all the buzz, it now has 1000000 incoming links, does Google pay attention to this?

    If such a thing happens once, I might have been lucky, if it this happens 10 times it means I know what I’m linking to and should get some credit for it.

    Giving early adapters (and spotters) credit will encourage folks to link quality sites even if they are small.

    Is there any sense in my point? 🙂

    Hope I didn’t waste your time with this thought,
    Good day,

  199. Hi Matt,

    I’m not as funny as Dan Thies, but I hope you’ll read my rant

    I really appreciate that you keep us updated as soon as you can, but in some cases, e.g. WRT rel-nofollow, the most appreciated update would be the removal of this very much hated and pretty useless microformat. I mean, when you’ve introduced it because the Google (as well as M$, Yahoo and Ask) algos were flawed at this time, why not take the chance and dump it now when it’s no longer needed?

    “Don’t code for search engines but for users” should mean that webmasters can link out freely and the engines interpret the links either as votes or not, depending on the context.

    Do you’ve a clue how much utterly useless workload you generate when you update a link policy? That’s a gazillion of man years for each change request. So please just dump all those policies.

    Thanks for listening and have a great day!

  200. Thanks for coming out and making a clear statement about this topic. It is very interesting why Google is choosing to still give link juice to the Nofollow tag. I’d have to agree with the sentiment of several other posters here.. but you have to do what you have to do to try to deliver the most accurate results.

  201. Hi Matt,

    Quick question;

    If you do not want any juice to go through to an internal page (for instance; terms, contact, about etc) would using meta noindex, nofollow not keep the “juice” away from that page, or will it just disappear with that page?

  202. The phrase ‘Pagerank Sculpting’ sounds great. May be it gives a (false) feeling that there is something geeky that SEOs can do.
    But it essentially means “hey, my site is not well-organized”
    Content structure is the ideal way for controlling PR flow.

    May be as Toby Mason mentioned above, if you have a content with say 3 links to the same page and one says ‘Read More’ or ‘Click Here’ they may be nofollowed; just for controlling the Anchor Text flow.

    It might be the emergence of blogs that is causing very high interest in ‘Pagerank Sculpting’ because some important content automatically gets archived and thus hidden. Features like Related posts below posts, featured posts links, effective cross-linking etc should solve this issue.

  203. Should I turn off comments on my blog?

    The no-nonsense answer: Yes! Turn off the comments on your blog and you will see an improvement in the SERPs, moreso now that we can’t nofollow, but this is how it’s always been.. even though it makes your blog suck.

    No one said anything about ‘hoarding’ rank, Mr. Cutts – this guy is simply asking if his comments are going to suck pagerank out of his site since he can no longer label them ‘nofollow’ to redistribute juice. The answer to that is yes.

  204. Hi Matt, what if i have these referrer links inside my site?

    For example, my domain is
    I have a link to an inside page on my site, but the original link is

    would it make sense if i add nofollow to to prevent duplicate content penalty right?

    I tried link rel canonical and i still see duplicate urls on the google index.

  205. Matt,
    Just a question which someone sort of covered earlier, but not quite like this ?
    Say you’ve got a content heavy product page in a shop, with say 10 product variations and therefore 10 links to the same file with different url parameters that fetches the product price, bigger image etc. from the database. (Because its just v. low content data for the customer).

    We don’t want all those details / prices etc indexed, so the fetching file is noindex, nofollow.

    How do we stop pagerank disappearing to wherever along those 10 links ? Nofollow would have done it previously yes ? Because we actually want 3 other links to related products on content-heavy pages to get the rank passed.

    I think that’s clear ?

  206. Hi Matt,

    My main concern is the point already raised by Andy Beard and Danny Sullivan.

    What happen to blog comments? You said “I wouldn’t recommend closing comments in an attempt to “hoard” your PageRank.”

    Well, if we consider the changes you mention, closing comments looks like a very good idea to me (under an SEO point of view).

    Here is the logic: the main source of links to a blog are its most popular posts. But those posts also get dozens, if not hundreds of comments. In the end of the day, therefore, you will get only a fraction of the PageRank value of that post flowing back to internal pages and posts.

    If a certain post has 200 backlinks and 300 comments, for example, you can forget about any internal PageRank flow.

    How do we solve that problem?

  207. Something I still do not understand though. Google still provides us with confusing information:

    Google states this: “We don’t follow them. This means that Google does not transfer PageRank or anchor text across these links.” —> this means “sculpting” would still work.

    If the above is true then Google and Matt seem to have very different opinions.
    If the above is false then Google should be updating their info faster to prevent webmasters from getting in a confusing situation like they are in now.

  208. Great Article! I agree that website owners and the SEO community should focus on creating quality content, intuitive navigation and interactivity. Not passing page rank to forums is plain manipulation which should be and is rightly being discouraged.

    The choice to link or not to link should be based on improving usability and not to artificially manipulate page rank. The best way to give more importance to certain aspects of the website is in the sitemap planning.

    I am still not sure what is the best practice for blog comments. Wouldnt it be detrimental to allow a follow through link to a site which we know nothing about?

  209. Like Sebastian, I dislike Pagerank building strategies !

    Everywhere we find “nofollow” / “javascripts” / “absence of anchors”…

    People have to edit more relevant content.

    Thanks Matt for this detailed article 🙂

  210. So What you are saying is that page rank sculpting is not an important issue, page rank is just one of 200 factor that determine serps. Webmaster doesn’t have to worry about this stuff because the impact is miniscule, right?

  211. Matt …Thanks for your time, Took a few reads, but I get it!

    Can you clarify…
    If no-follow is the blocker that allows us to prevent an outgoing link from getting its allocated % of the PageRank after decay. What happens to that link juice. Does it evaporate or does it pass back on to itself?

  212. Stefen Arkinson

    Hi Matt , first upon I want to say you thanks for this blog on Page rank. This is really a very helpful blog for everyone. But do you not think this one a little bit tough for new comers , means how you explain it.

  213. Having introduced No Follow specifically to stop blog comment spam it is ripe for Matt Cutts/Google to now state that actually all those comment links ARE wasting your PageRank after all. That’s done blog owners who have allowed extensive commenting when No Followed an extreme disservice.

    The outcome will be obvious – direct all commenting to a separate forum. Is that really what was intended by all this? I very much doubt it as it absolutely detracts from the continuing value of the original page in terms of added content.

    There’s a lot of frustration being vented in this comments section. It is one thing to be opaque – which Google seems to be masterly at – but quite another to misdirect, which is what No Follow has turned out to be. All of us who produce content always put our readers first, but we also have to be sensible as far as on page SEO is concerned. All Google are doing with this kind of thing is to progressively direct webmasters towards optimizing for other, more reliable and transparent, ways of generating traffic (and no, that doesn’t necessarily mean Adwords, although that may be part of the intent).

  214. @ProFx: javascript to prevent your links to get followed is getting useless: javascript now can be crawled and followed as well 😀

  215. >no one noticed or complained

    If you were using robots.txt and meta index settings that matched your nofollow sculpting you wouldn’t have noticed.

  216. I’ll do my links in flash. It’s best than have my other “good links” penalized by other useless links.

  217. Thanks for sharing Matt,

    I wonder why Google just doesn’t get rid of the attribute. This will prevent anyone from wasting time on sculpting.

  218. Dave (original) June 16, 2009 at 11:21 pm
    “You lie and feed spin to YOUR customers, nice to know.”

    Dave, if you don’t know how to read right, please refrain from commenting. It’s really sad.

  219. Well I read through this whole thing only to determine that I’m not going to do one single thing differently after this post than I was doing before this post.

    Was an interesting read, tho.

  220. It’s great that you’ve finally demystify PageRank sculpting.
    A practice that I’ve always argued as pointless and waste of resources.
    This will hopefully get SEO professionals focusing in things that really matter: site architecture => content => site reputation.

  221. Peter made a very good point in all of this, and Michael Martinez did in a backhanded way as well. Talking about a concept related PageRank sounds cool. It doesn’t actually have to be useful or practical, and it usually isn’t; but as long as the impression of something productive is given off, then that can be all that matters in the eyes of those who lack sufficient knowledge.

    A lot of the problem lies in the name “PageRank” itself. The term “PageRank” implies that a higher value automatically equates to better search engine ranking. It’s not necessarily the case, it hasn’t been the case for some time, but it sounds like it is. As stupid as it sounds, a semantic name change may solve a lot of this all by itself. Some of the old-school crowd will still interpret it as PageRank, but most of the new-school crowd will have a better understanding of what it actually is, why the present SEO crowd blows its importance way too far out of proportion and how silly the industry gets when something like this is posted.

    I don’t know exactly what that name should be, but something along the lines of “Inbound Link Approximator (ILA)”, or “Rough Link Factor” or (pick a term that implies an external approximation and that doesn’t seem to imply a magic bullet for search engine referral traffic.)

  222. Google states this: “We don’t follow them. This means that Google does not transfer PageRank or anchor text across these links.” —> this means “sculpting” would still work.

    No, what Matt has pointed out is that the ‘pagerank that does not transfer’ does not go to your other links either. The fact your other links do not get the benefit is what changed.

  223. I only read about 25% of the comments, but it seems people are more upset about not being able to “work the system” anymore. Google does sound like a broken record telling everyone to just build quality sites and not sweat the small stuff. I personally have dropped the nofollow strategy from my SEO.

  224. Give me a choice between PageRank flow or having my anchor texted crawled, I have and always will choose the anchor text.

  225. Matt has mentioned that sites that block Google with robots.txt still can accumulate PageRank even though Google does not crawl them, so links pointing to such sites will still pass PageRank and be used in the equation.

    My question (and I believe someone else asked it above and Matt you said you checked) is pages that are blocked with a noindex meta tag or with robots.txt (where robots.txt blocks a file or directory, not a spider). If these pages receive PageRank.

    If you have 10 links on a page, 5 being normal, 5 going to pages with a noindex meta tag, and the page has 10 points. Do the 5 normal links get 2 points each, or 1 point each?

    The issue being, this change makes it a bad idea to nofollow ANY internal link as any internal page is bound to have a menu of internal links on it, thus keeping the PR flowing, (as opposed to nofollow making it evaporate). So no matter how useless the page is to search engines, nofollowing it will hurt you. Many many webmasters either use robots.txt or noindex to block useless pages generated by ecommerce or forum applications, if this change applies to those methods as well it’d be really great to know, so we can stop sending a significant amount of weight into the abyss.

  226. Thanks for the useful information about the “nofollow” attribut. What about this PR Tool?
    PR Tool Does it fit in a way?
    Greetings from Germany

  227. “david wolf, no one had complained or even noticed, which is another datapoint that helps point out that this isn’t a high-impact issue in general.”

    It wasn’t a high-impact issue while SEOs were in the dark about it. The effect it will have from now on is another matter entirely.

  228. wow this is interesting stuff…
    seems like everyone is taking it differently though.

    thanks for the info matt

  229. What a read – and the article was good too! It’s amazing that it has been changed for so long and wasn’t realized, but I guess the mysteries that surround rankings make it tough to really know what is going on.

    @seth r – Matt’s article is probably more up-to-date than that page, but I think it just goes to show you there isn’t one clear answer when it comes to search engine ranking… 🙂

  230. Well since reading this articles I’ve implemented the following:

    Since doing so I haven’t seen much of a difference (except perhaps a small climb)?

  231. Matt,
    I have read a lot on this and have not found an answer to my one questions.
    If you have $5 and 5 outgoing links and 2 are nofollowed, I understand that $3 will go to the followed links, but where do the other $2 go? Does it stay within that page or is it just lost?

  232. Rob,

    Apparently you missed SMX. The Remaining $2 “evaporates” according to Matt.

  233. …whenever you’re linking around within your site: don’t use nofollow.

    I think I did it by nature 🙂 I am just starting to learn all SEO and SERP stuffs. PageRank only entered my life when I quit the job last month. I don’t rush and I believe in a more natural way to get traffics, such as trying hard to create good contents and releasing some GPL works. You give something, and traffic flows in return. It’s Newton III Law. No sculpting, nor illegal way is necessary.

  234. Okay, this is getting weird. Does anyone else see the irony in a worthless number being assigned a dollar value, or is that just me?

  235. lol wow. I did read all the posts. Out of all of them, Peter above has it exactly right. Amazing that we agree, huh? This is not an issue for the total majority of SEO’s and webmasters. It never was.

    Matt; correct me if I’m wrong, but didn’t you say something like this back when this sculpting issue first came up?

    “”You can try to sculpt your pagerank if you wish, but fixing the issues you may have with your site structure and architecture would benefit you more in the long run. Also; unless you are very skilled with pagerank and SEO, you should probably not attempt anything as it could lead to unintended consequences.”

    I know I’m paraphrasing, but that seems to be the gist of what you wrote or said at that time.

    That was a great post Peter!

  236. btw; All those SEO’s out there probably made some monies off clients, selling the sculpting thang to them. I know some are still insisting it worked, etc, but would they say in public that it didn’t work after they already took a site’s money to sculpt? How would anyone judge if it worked or not definitively? The funny thing is, the real issues of that site could have been fixed for the long term instead of applying a band aide. Of course; knowing the state of this industry right now, band aides are the in thing anyway.

  237. Is this a serious bug in the Google’s Algorithm or Just a
    simple SEO mistake on my part?


    A few of my keyword listings for one of my sites have become
    very erratic in Google – bouncing all over the place. In one
    case, I had the number one spot for years and recently it has
    always been in the top 5. However, lately it has been bouncing
    on and off the first page results in Google. U.S. rankings only…

    Correct me if I am wrong but this is what I think is
    happening to cause these yo-yo keywords within Google.

    In my case, Google was listing both my main index homepage
    and my optimized page for the same keyword phrase. This
    appears to be diluting the ranking of both pages. The optimized
    page was at around 50 and my homepage came in at the 95 spot.
    However, this seems to be more my mistake than Google’s, since I
    inadvertently placed the same keyword phrase in links on
    my homepage going to other interior pages and not necessarily
    to my main optimized page which I wanted and which is usually
    ranked in Google for this keyword phrase.

    Let me explain further, I develop sections/guides on my site
    that are keyworded to different markets generally relating
    to the theme of my site or domain. For the keyword phrase in
    question, I have a well established page that’s linked
    from hundreds of related sites around the web – this is the
    page that should be ranked for this keyword phrase since
    it is the most relevant, not my homepage. Likewise, on
    my own site I have constructed my pages so that PR flows
    to this optimized page. For years, this optimized interior
    page held the number one spot in Google.

    However, recently when Google was spidering and updating
    my site in their index, they picked up that keyword phrase
    and linked it to my homepage as well as to my optimized page
    – causing both pages to drop in the rankings. Now when Google
    re-indexes my site and DOESN’T pick up the keyword phrase
    on my homepage – my optimized page pops back into the #2 spot
    in Google, or at least on the first page results.

    Usually, double listings for a keyword on your site
    is a joyous occasion, Google lists your homepage and
    then immediately lists your interior optimized page
    right after it. As long as these links are displayed together
    or right after one another, you get tons more traffic and
    it can even be better than the #1 spot.

    However, in my case, when they are displayed separately
    and at very different ranking levels – it’s a real downer
    in more ways than you can imagine since we’re talking about
    millions in product sales changing hands!

    Now there could be separate filters at play here and the
    above scenario has nothing to do with these bouncing keywords
    within Google. It seems Google’s Algorithm has become so
    sophisticated and complex, it probably has developed a
    mind of its own and nobody can tell it what to do!

    But if this is actually what is happening and this same
    scenario is widespread throughout Google’s rankings… can’t
    believe my site would be unique or different from thousands
    of others on the web or I the only one making this SEO
    mistake. Besides, if Google is picking random keywords and
    listing your homepage instead of your more relevant optimized
    page which has the more appropriate content, isn’t your index
    much less effective at supplying the most relevant content to the

    Doubt if this question will be answered but I had to give
    it a try…

  238. It’s so funny to have spent a lot of time and effort PR sculpting and then to have given up on it to find this blog and see that the advice from Google is essentially “don’t worry about it.” Something, I think, Shoemoney has been saying for years.

  239. But you use nofollow on your blog although its default on wordpress but you can easily change it. But I think you probably should change that about only having nofollow on the feed…

    I don’t think google penalizes you for having no follow or sculpting pagerank, it’s just that you may end up leaving out important info and articles out in the process. If google was to penalize for this it would be rather foolish.

  240. Also it’s something that is very easy to sell to your customers: “We´re going to work on your PageRank!”

    Peter, I read your post and understood it just fine, thanks. Assuming it’s English of course. Don’t worry though, you are not the only SEO that preys on the Webmasters ignorance of PR and makes money from that. After this post by Matt, it will likely get even worse. Now that IS “really sad” for the SEO industry clients.

    As you can see from my post above, I’m of the opinion that Webmasters should not have to know a single thing about SEO and know everything about there subject matter. That would make the whole WWW a much better place when searching via a SE like Google. Don’t you think? Oh, hold on, you have vested interest where-as I don’t 🙂

    I really think Google keeps it’s enemies sweet and close, all the while gathering info on ALL SEO tactics, so they can compare and discount them where warranted. Put yourself in Google’s shoes. It relies on returning the most trustworthy and relevant pages in the SERPs for any given search term. That IS the all important foundations of the Google empire. IF that can be artificially manufactured by SEO and money, Google has lost, not only the battle, but the war.

  241. “you are not the only SEO that preys on the Webmasters ignorance of PR and makes money from that.”

    Dave, he was saying that’s what others do to make a quick buck.

  242. I still don’t get it. When I have 10 page rank points, and 5 go out with do-follow – where do the other 5 go? I thought there is no PR lost?!?!?!

  243. Great post. I read everything and sort of think its better that page rank sculpting is off now. Thanks for confirming this Matt.

  244. Something a lot of people seem to have overlooked was hinted at in Greg Boser’s comment above. Greg identified that there is a major (and unfair) disparity with how authority sites such as Wikipedia disrupt the linkscape by run-of-site nofollows. Once Wikipedia implemented the no-follows, previously high-value links from Wikipedia were rendered worthless making the site less of a target for spammers. Increasingly large sites are following suit in order to cleanse their own pages of spam.

    Google can’t like this. Although its great for them to have spammers out of the Wikipedias, they’re also losing a lot of very authorative input for their PR algorithm. Think about it – if every site in the world put nofollow on every link Google’s algorithm would be worthless overnight. There has been ongoing speculation as to whether or not Google ignores nofollows from certain sites like Wikipedia, something Mr Cutts has outrightly denied (but also admitted that it would be very useful to have more granular control over nofollow so that it was not an all-or-nothing situation.)

    Adjusting how Google treats nofollows is clearly a major shift (as the frenzy in the SEO community has demonstrated). So, if Google were to adjust how they treat nofollows they would need to phase it in gradually. I believe this latest (whether in 2008 or 2009) change is simply a move in the direction of greater changes to come regarding nofollow. It is the logical first step.

  245. HI MATT,

    I think it’s not good for site owners because it will help small sites to grow up fast by spaming on our sites like hide links and texts on comments !!

    thank u any way ,

  246. Thank you Matt for your enlightenment.

    I like the nofollow tag it gives webmasters control!

  247. Matt,

    Just a heads up, someone is using your content from this post on the forum below –


  248. Matt, why do you gotta whip out terms that I can’t find authoritative references for like “reset vector”? 😀 (please throw me a link if there is one)

    So ordinarily there’s a 10-15% chance that a surfer gets bored and leaves a page to another random page. Given a page with 10 links, that leaves 85% chance that a random surfer will click on one of those links (though in reality, probably the chance that a surfer will click on a link that’s near the “hot spot” – page upper left – will be higher).

    Say 4 of those links are nofollow (or 34% chance). By “reset vector” are you saying that the 34% now gets added on to the 15%, so that from Google’s POV there’s a 49% chance that a surfer will leave the page to wander off to a random page?

    I’m asking because the word on the street is that the nofollowed PageRank basically gets redistributed to the entire web and I just want something a little more detailed.

  249. Halfdeck; Don’t you think the big problem is that Google is giving too much information to the industry? I stated a long time ago this fact, wondering why they wish to constantly hand out more information when they should have known the industry would try their best to exploit anyway. Not only that, but wanting more and more no matter how much Google hands out is something that is very clear as well. You just stated you want “more detail”. Why? I’m thinking too much detail handed out over the years is Google’s biggest problem right now. Considering the total majority of websites on the internet don’t know what a nofollow attribute is anyway, what exactly is Google gaining by giving up parts of their algo to the SEO industry? Big mistake. They should actually just shut up.

  250. Thanks for clearing the nofollow part for me, although i still have some serious doubts on what en when to do it made the myth clearer for me.


  251. The topic has been going wider and wider. Around the web everyone writes about PageRank sculpting and gives their own examples and confusions 🙂

    I just don’t care about it unless it is an untrusted content or for a page which is not open to all 🙂

    Be simple!

    All time good explanation from you Matt

  252. Very nice article but basically nothing really red, hot & new. If you are trying to build links to help your website’s PR then do not link to websites with nofollow.

  253. This is my first comment on your blog. Thanks for clarifying this issue.

  254. Hi Matt,

    This question is only slightly tangential to the discussion. I have always heard that a higher amount of pagerank was passed on to links higher on the page (or in the code, as the bot reads). You imply that this is also inaccurate. Can you confirm that?
    Cheers and thanks for a great post!

    P.S. Thanks for the very clear diagrams for us simple-minded writer types!

  255. Thanks for this great article matt, it’s very helpful for all of us 😉

    But i just don’t understand why would Google do such a think, PR flow within my website should be my own job as a webmaster… what if i want to pass more power to the most important categories within my site ???

    However, thanks for the info.

  256. PR sculpting is always a hot topic of discussion. Matt thanks for providing in depth information. The best thing is to manage the site architecture that will eventually help to boost PR.

  257. How does this all relate to disallows in the robots.txt? My ecom site has 12,661 pages disallowed because we got nailed for duplicate content. We sale batteries so revisons to each battery where coming up as duplicate content. Is PageRank being sent (and ignored) to these internal disallowed links as well? One of our category levels has hundreds of links to different series found under models, the majority of these series are disallowed. If PageRank acts the same with disallows as it does with nofollows, are these disallowed links are hurting our
    PageRank flow right? I am not looking to “sculpt” per say just fix a previous “fix” if it needs it. We walk a fine line between SEO and usability, of course leaning towards the customer.

  258. I used to work for a small agency and was in charge of the SEO so am familiar with the previous notion of what nofollow did. I now work for a College and we outsource our SEO. I just finished 2 days of recommended nofollow work on our 5000+ page site and then read this article.

    I think I’m going to go home and drink now.

  259. Firstly, we should notice the new nofollow rules is fair. If you want to modelize surfers behaviour (Google’s goal), you have to ignore nofollow links but have to count it.

    Now there’s the wikipedia case. Okay Wikipedia provides a great unique content but wikipedia’s SEO success comes from a systematic nofollow use (AKA SEO Black hole).
    The new nofollow rules will reduce the link juice capacity of each wikipedia page. Of course deeper wikipedia page will suffer the most and should drop in the SERP.

    If wikipedia drops in the Google index, this is perhaps a good news for another knowledge product…. knol (by Google)! Surprised? 😉

    My 0,02€

  260. it’s good to remember that not all the links pass page rank and that many links have the no follow attribute. And, in reply to Dough Heil, it’s good that Google sometimes allow us to know more about its technology.

  261. Don’t you think the big problem is that Google is giving too much information to the industry?

    Yes, but Matt as well. I’m of the opinion, though, that the information is either false, worthless or already in the Google Webmaster guidelines. They also likely plant seeds and watch what happens.

    My main concern though, is Google appears to becoming reliant on sites doing MANY things for SE only. It also appears that Google is lowering the bar for YouTube videos in the organic SERPs and forcing their insertion as the cost of relevant pages. It even seems they are now doing the same for pictures, despite BOTH having their own SEs. I fear Google is attempting to increase profits, for it’s shareholders, in a rather impatient manner.

  262. I think there are real issues over this change:

    1) The blog comment issue

    The more popular a blog entry is….the more comments it gets…the lower its page rank…wtf??
    (This page being an example, due to its 600ish “nofollowed” links in comments must be about the lowest page ranked page on the topic of “page rank sculpting” on the whole internet – which seems ironic)

    On a blog the page rank should go to the main article pages. Now it just gets “evaporated” if you use “nofollow” or scattered to all the far flung nooks and crannys which means google will not be able to see the wood for the trees. The vast majority of a site’s overall page rank will now reside in the long tail of useless pages such as commentors profile pages. This can only make it harder for google to serve up the most relevant pages.

    2) The privacy issue

    Nofollow was a neat way to keep pages out of the index such as User Profile pages on a teenage community site. But now webmasters get penalised for “nofollowing” any internal pages so the incentive is to “follow” all pages even those you should really have kept out of the index.

    3) The communication issue

    This change is too significant in terms of impact on peoples business, livelihoods, technical developments plans to not have communicated one year ago.

    How many tens of thousands of hours have been wasted globally on page sculpting code tweaks that now need to be undone.

    How many sites dived in page rank with no clues or advice in google webmaster guidelines during 2008/2009 to point them at this possible cause.

    4) The OMG how much PR has been flushed down the toilet these last 12 months issue

    Even the biggest sites including Google sites seem to be totally unaware of the implications.

    I find it amazing that has been “nofollowing” its featured videos for last 12 months (still doing it as I type) when it now seems that this means “i dont trust this content” and “i want to page rank to flow to this content”. In fact a quick glance at a youtube page tells you that youtube are currently flushing 50% of their page rank (very approx) down the toilet on every page.

  263. These are only directions to SEOs for consideration and there are many more factors behind this and considering the facts these information are not too much 🙂

  264. First, thank Matt for the great analysis and useful article.

    But to me, it’s hard to understand. Should I understand that pages with high PageRank, backlinks to other sites using the “nofollow” attribute linked is not good for the PageRank of destination’s site?

    With PageRank, which distinguish between natural urls, natural HTML structure .. or not? If so, how to calculate PageRank of landing be like?


  265. Matt, above you say: “More than a year ago, Google changed how the PageRank flows so that the five links without nofollow would flow one point of PageRank each.” so doesn’t this mean that a link with rel=nofollow still gets pagerank?

  266. These are only directions to SEOs for consideration and there are many more factors behind this and considering the facts these information are not too much

    Considering you run the risk of being banned or penalized from Google IF you link out to a site that later becomes a “bad neighborhood”, I would say there’s is little choice BUT to nofollow ALL outbound links.

    Besides, blackhats USE any info from Matt and Google and then proceed to charge clients for being banned from the Worlds most popular SE.

  267. Dave makes a good point. I never understood the purpose of the nofollow attribute in the first place. If we look at a site like “Yell” they use nofollow on all their outbound links. As it turns out, this practice isn’t preserving any of their pagerank, and it isn’t passing on any pagerank to sites that are relevant.

    What if every page on the internet used nofollow on ALL their outbound links? Wouldn’t this make pagerank obsolete? The majority of forums and blogs already use this practice, and surely links from these sources can be a good indication to the relevancy of the pages being linked to.

    Though as I mentioned in an earlier post, I believe that even nofollow links can effect the placement of the sites being linked to.

  268. Well it certainly takes a lot of moderation to use dofollow on a blog. I receive a lot of spam comments – but almost all are 100% spam and they probably link to some of the darker parts of the web.

    But if trusted people really contributes to my blog I think it’s ok to give them some link love for the effort.

    I normally never nofollow links to sites i trust – have linked to this blogpost without a nofollow – so I guess this means I trust Matt.

  269. Hi Matt, we are currently about to upload a mayor project that includes applying no-follows to a numerous amount of links. This project was not conceived as a method of sculpting the site but because in out webmasters tools google alerted us that the site had too many links and this could affect crawling. This is mainly because we have a high listing and because we offer the users many filtering options which makes for great user experience as several surveys indicated. The idea of the no-follows was to apply them to these filtering links to tell google not to follow them. Do you think this is the way to go?. After reading your post I am affraid that we might be making a mistake.

    Thanks and continue with the good work.

  270. I don’t want to dig through Matts posts, but I could swear you posted one about sculpting pagerank and how to do it, now you do a 180 degree about face and say don’t!!


    I don’t use Google for searching anymore any way.

    I BING IT!!!!

  271. Thinking of page-rank always makes my head hurt. Can we be 100% sure that the nofollow tag will not leech out PR? We don’t really know what google is doing do we?


  272. I agree that there is no point in trying to over analyze how the PageRank is flowing through your site. Just focus on great content. Link out when it actually helps the reader. This is what Google wants – for you to give good quality content to their users. So if you are doing that, they will reward you in the long run. No need to worry yourself with these types of link strategies.

  273. So if I have 10 points of pagerank on a specific page, 5 internal links to my other pages, 5 outbound “nofollow” links to other people’s pages.

    I will only distribute 1 point each to my 5 internal pages? And no points to “nofollow” external pages. I keep 5 points to the original page? Right?

    You’re saying, if I want to flow more pagerank points to my internal pages, I don’t even want to have any other nofollow external links.

  274. Doug you make a good point about Google giving out too much information – they opened Pandora’s box when they started and now too much of the content has escaped. It was always much more fun when everything was speculation and what you could work out for yourself.

    Also given that the original reasons for implementing the ‘nofollow’ tag was to reduce comment spam (something that it really hasn’t had a great effect in combatting) – the real question I have is why did they ever take any notice of nofollow on internal links in the first place? It seems to me that in this case they made the rod for their own back.

  275. Argh! This means that I cannot remove incentive from spammers to blogspam other than by actually excluding the blog pages themselves with robots.txt!

    Obviously I have to clean up the spam on the site, but *how* much spam I have to clean up is a function of how much incentive the spammers have, and this change removes a major tool in denying them that incentive.

    You guys ….


  276. So what kind of JavaScript links can I use instead of nofollow links so that I don’t loose PR on those?

    I’d assume that Google doesn’t have a full featured JavaScript interpreter when crawling websites, so there should be some way to have JavaScript links that work in browsers but that Google cannot follow. Does somebody have any idea how to do it?

  277. Just a quick question.

    Does the position of a link on a page effect the amount of PageRank that flows through it? ie a link at the top of the page passes 10% more juice than a link at the bottom?

    If so it would be possible to sculpt your PR by putting your more important links towards to top of your page, it would also help users find the important content of your site therefore deserving of some PR recognition. It would also stop all the link spam in the footers of sites.

    Just a thought.

  278. Hi Matt, i had seen word “bad neighborhoods” many a times in your posts and other documents of google. Can you please explain us what do you exactly mean by “bad neighborhoods”??

  279. Matt, It would be great if you can provide detailed explanation on Danny’s this part of the comment

    Say I have an article on a blog with 5 links in the editorial copy — some of those links leading back to other content within the blog that I hope to do well. Then I get 35 comments on the article, with each comment having a link back to the commenters’ sites. That’s 40 links in all. Let’s say this particular page has $20 in PageRank to spend. Each link gets 50 cents.

    With nofollow before the change, I could have (if I were worried about flowing PageRank), kept any of those comments from getting some of my PageRank spend. Nofollow them all, and the 5 remaining links each get $4.

    With this change, I can still get the $4 if I simply don’t allow comments. Or I show comments, but I use an iframe, so that the comment actually reside on a different page. In either case, I’m encouraged to reduce the number of links rather than let them be on the page period, nofollow regardless. If I’m worried my page won’t seem “natural” enough to Google without them, maybe I allow 5 comments through and lock them down after that.

    I guess many bloggers would like to know whether we should use iframe/js for comments or leave them as they are (in case Google is not considering them already for PR calculations)


  280. I think it’s important to remember that the majority of website owners are NOT at all technical and savvy as to how this whole system works when it comes to SEO, but they still get along and do their best to put something worthwhile on the net. Unfortunately this same majority can so often damage their website rankings without ever knowing it, and lead to an under-performing website as a result, regardless of the quality of their content. As I’ve learnt as a new website owner for the first time, there’s a lot more to running a website than just doing the right thing and trying to produce quality, when time and time again the experts in SEO always win out on the SERPs regardless of quality. One keyword phrase I searched on repeatedly over recent years resulted in the same EMPTY site being returned as the number one result, truly, it had NO content.

    I suppose for those people, including myself who just keep trying to our best and succeed, we just need to keep trusting that Google is doing all it can to weed out irrelevant content and produce the quality goods with changes such as this. Meanwhile the “uneducated majority” will just have to keep getting educated or get out of the game I suppose.

  281. It’s nice article about a pagerank.

    However, it remind me that it looks to me that Google is getting biased towards old pages. Old pages with some quality will have much more links comparing to new pages with better quality :). For one query I tried, there are approx. 90% of different results for the same query but to show only pages from the past year (showing first 50 links).

  282. Matt,
    I have the same questions as Muratos but didn’t see the answer here…

    “Maybe, Google should develop a new tag as well something like rel=”commented” to inform spiders about it to give less value and wordpress should be installed default with this attribute :)” Do you think this could happen???

    “And my vital question about Amazon affiliate links….Should I nofollow them all or leave as they are?”

  283. Come on face it… link spamming works in google and google no longer seems to care about taking real action.

    If you want to increase your sites ranking all you need to do is start buying links and spamming forums/blogs/social networking sites.

    For example… has over 21 thousand PAID links that they acquired in a about six months… this site has been reported almost 21 thousand times to google (I hired people to report ALL the paid links) and the site is still rising in the SERP like a rocket… As they continue to purchase links.

    How can you compete against that?

    I can give hundreds and hundreds of more examples… but I think this one says it all.

    Google has dropped the ball.

  284. I won’t blame MC. Google, knows what they does. These are things that webmasters need not worry about. Well, it won’t make much difference as far as I think. I don’t use no follow tags specifically – I use WP for blogging purposes and it does rest of the things for me other than writing content which I do. I think it is the content and the external links that sites point to – which should be considered. I mean, if a computer blog owner posts a really fantastic computer article about something related to computer, and also puts some links to external pages (which are really useful for the readers), then that post, should be ranked high in gooogle – And I think google does this well – So, webmasters, just concentrate on yur website/blogs etc and leave rest of the things to Big G.

    CEO, DigitalPointing

  285. Well,good news for those who have zero comment all the time.

  286. Dave (original),

    later becomes a “bad neighborhood – You need some trust in web and the word “later” don’t fit for some conditions. You think a lot and by adding a nofollow you do not trust the site fully right! And hope reviews would help

    blackhats USE any info – When it comes to blackhats there are many ways that would be researched and done with their method and it is not like information’s are provided so the advantage goes to them. They would do in any cases. Things should change and those should been revealed and things can’t be idle by worrying something would be misused 🙂

  287. @Ronny – At SMX Advanced it was noted by Google that they can, and do follow JavaScript links. They also said that there is a way to provide a nofollow to a JavaScript link but they didn’t go into much detail about it. Vanessa Fox recently wrote a lengthy article about it over on Search Engine Land which will likely address any questions you might have:

  288. I’ve seen so many cases of webmasters nofollowing legitimate external links it is not funny. Any external link on their site is nofollowed, even when quoting text on the other site. IMO, the original purpose of nofollow has long been defeated in specific industries. As more webmasters continue doing everything they can to preserve their pagerank, the effectiveness of nofollow will continue to erode.

    I personally nofollow links to my privacy policy and contact form. Even though these are excluded in robots.txt, I prefer that extra layer of protection so that the pages are not indexed. Anyone that has ever had their contact form blasted continuously by spammers knows what I mean. And yes, one could add the noindex meta tag. But let’s face it, not everyone is a skilled PHP programmer. On dynamic sites its not as simple as adding a meta tag…

  289. Jag, sorry mate, I don’t understand what you are trying to tell me 🙁

  290. Great post, Matt. Actually this is good news for me even it’s not clearing me enough because i still confuse few things. i should read your article and whole comments here over and over again. i’ll follow you on twitter. Thanks matt.

  291. Dave, what makes you so bad about this old update?

  292. In reference to coments on blogs for which the nofollow was introduced to help.

    It is good that Google can decide what is a bad neighbourhood and not count the outgoing link, but is it also not true that my blogs can be punished for linking to a bad neighbourhood?

    In better words, someone leaves a coment and links to a bad neighbourhood, is it not posible that my blogs can suffer, even though I am helping my visitors?

    If we no longer get penalised for linking out to bad neighbourhoods please lets us know.

  293. Is this advice still valid: “If your site has an infinite calendar, add a nofollow attribute to links to dynamically created future calendar pages.”

  294. Great post about Page Rank, I am a newbie to understand but I am really want to know that.
    Beside, I am confuse between nofollow and dofollow, what’s exactly benefit each tag?

    Thanks for your post.

  295. Matt I wanted to say this piece of information is causing huge waves in the SEO industry. I have one question for you guys, why don’t you release this information before?

    I think if there was more transparency and openness between search engines, and the SEO industry, both would be better off… after all this is an industry dependent on your search engine, why would you not try to have better relations and connections with SEO’s?

    I don’t really get it, doesn’t seem very Google like.

  296. Dave, what makes you so bad about this old update?

    Scroll up.

    Beside, I am confuse between nofollow and dofollow

    No such attribute as “dofollow”. You should nofollow ALL links to pages you have no control over. Google might ban you, if a site you link to, becomes part of a linking scheme.

    The default for SE’s is follow links without “nofollow”.

  297. If you’re Matt Cutts and a billion people link to you because you’re the Spam guy at Google, writing great content is enough. For the rest of us in hypercompetitive markets, good content alone is not enough. There was nothing wrong with sculpting page rank to pages on your site that make you money as a means of boosting traffic to those pages. It’s not manipulating Google, there’s more than enough of that going on in the first page of results for most competitive keywords. Geez Matt, give the little guy a break!

  298. Dear Matt

    nofollow always has been, and remains, nothing more and nothing less than a tool for manipulating Pagerank. How can it be characterised in any other way that does not boil down to that?

    I’ve never been particularly enamoured with nofollow, mainly because it breaks the “do it for humans” rule in a way that other robots standards do not. With other standards (e.g. robots.txt, robots meta tag), the emphasis has been on crawling and indexing; not ranking. And those other standards also strike a balance between what’s good for the publisher and what’s good for the search engine; whereas with nofollow, the effort has been placed on the publisher with most of the benefit enjoyed by the search engine.

    nofollow is beyond a joke now. There is so much confusion (especially when other engines’ treatment is factored in), I don’t know how you expect a regular publisher to keep up. The expectation seems to have shifted from “Do it for humans and all else will follow” to “Hang on our every word, do what we say, if we change our minds then change everything” and nofollow lead the way. I could give other examples of this attitude (e.g. “We don’t follow JavaScript links so it’s ‘safe’ to use those for paid links”), but nofollow is surely the worst.

    Given that “only a tiny percentage of links on the Web use nofollow”, why don’t we just get back to focusing on humans and drop nofollow? It has failed, and given that all it ever was was a tool to manipulate Pagerank, it was bound to do so. Has Google done any tests on its search quality taking nofollow into account vs. not taking it into account, I wonder?

    Yours more in hope than expectation


  299. @Matt ..

    So instead of using page rank scupting would using sitemap priority do the same / a similar job ?

  300. Interesting, that you post this four days after I posted on the same topic on my own blog:
    If I didn’t know better, I’d think you were reading me! 😉

  301. Fine info. However I am still confused what is better for me as a site owner to add nofollow or not to add.
    If I do no one will wish to have link exchange with me but it may reward me in google rating
    And If I don’t……
    I have worked very hard to have pr4 on one of my sites and I do not want to undo it.

  302. Hey Matt,

    I have come across some sites. They use dofollow tags only when GoogleBot visits their pages. If you visit the site using a default UA like Mozilla firefox or IE, the source code has nofollow.

    Can you please explain?


  303. @ Gerry White – Great question Gerry.

    I work on a site that allows users to find what they are looking for by clicking links that take them deeper and deeper into the site hierarchy. Content can be categorised in lots of different ways. After about three steps the difference between the results pages shown is of significance to a user but not to a search engine. I was about to add nofollow to links that took the browser deeper than 3 levels but after this announcement I won’t be…

    I believe this is a common problem but no one seems to be able to give a definitive answer to how to deal with it. I (now) see 2 options:

    1) Use meta noindex on the deeper results pages and also block them in robots.txt. On the assumption that search engines won’t pass pagerank to them.
    2) Use Sitemap Priority. Give lower priority to pages deeper in the hierachy.

    Surely Google can give some guidance on what best practice is for this scenario?


  304. Nothing to scroll up Dave 🙂

    Hope you understand the post and there are factors behind many else tons of sites would be banned

  305. Hey Matt, It’s time to have a post summary of all the comment 🙂

  306. Wow, amazing post. I’ll be reading this again and again. It’s just completely undone a lot of my own “theory” about how exactly page rank works and is transferred from page to page. Bookmarked bookmarked and bookmarked again!

  307. Hey Matth, thanks for sharing. I’ll have to mention this on my blog.

  308. nofollow is beyond a joke now. There is so much confusion (especially when other engines’ treatment is factored in), I don’t know how you expect a regular publisher to keep up. The expectation seems to have shifted from “Do it for humans and all else will follow” to “Hang on our every word, do what we say, if we change our minds then change everything” and nofollow lead the way. I could give other examples of this attitude (e.g. “We don’t follow JavaScript links so it’s ’safe’ to use those for paid links”), but nofollow is surely the worst.

    Given that “only a tiny percentage of links on the Web use nofollow”, why don’t we just get back to focusing on humans and drop nofollow? It has failed, and given that all it ever was was a tool to manipulate Pagerank, it was bound to do so. Has Google done any tests on its search quality taking nofollow into account vs. not taking it into account, I wonder?

    Yours more in hope than expectation

    Agree. It would seems the days of writing unique quality content pages are no longer enough. I.e, now you need to know SEO tricks, or use AdWords.

    The day Google floated on the Stock Market was the beginning of the end. NOW, Google MUST please it’s shareholders insatiable greed.

  309. so why the heck are you no-following the comments? o_O

  310. Expect to see major blogs and media sites allowing user’s comments to limit the display of those to the last 10-15 and put the rest in an archive.

  311. Hmm very controversial post. At least for a webmaster point of view, why would someone post useful comments on my blog or anyone’s blog if he gets nothing in return?

    There is a theory that even with the noffolow tag, a link pass visitors to that link. But if you think about it for a sec:

    “If everybody and every link would have a noffolow tag in it, how would some pages have PR7 or more?”

    That would happen ONLY if Google is unethical and help some of the websites out … well THEN I understand. But wait… Google did this from its first day of presence right? Oh so now is not much as a surprise anymore for me.

    BTW: Question : “Okay, but doesn’t this encourage me to link out less?”
    And your answer: “I wouldn’t recommend closing comments in an attempt to “hoard” your PageRank”. Nice work in avoiding the answer. That answer has really nothing of value in it because people were looking for an answer RELATED to the question. How helpful.

  312. Great Post,
    I think that many are overly concerned with outbound links and if they spent more time on internal linking and sculpting, they would find it requires far less external sources to acquire a high ranking for competitive terms.

    We have terms that rank based on in all inclusive relevance in the hundreds of thousands, just from a blog post. However, without pulling off the internal links one could not expect that type of link juice from just a post alone.

    I have yet to truly play around with no follow, I am more of the philosophy or leveraging older content by either re-writing it or 301 it to preserve the authority, but to each their own.

  313. This is where I see the importance of the comment rating plugin if such a WP plugin exist. There are a lot of insightful good comments here, but I wish the best ones based on user rating would be listed first. This is a loooooongg list of comments.

    Back to the topic, I am not a fan of the PR sculpting by nofollow, but I believed in it in the past, I was just too lazy to implement. So I guess I was doing the right thing. Be lazy in sculpting by nofollow and do other thigns more worthwhile. :p

  314. Don`t mean to be snidy, but how come you still put a no follow tag on these links? I`ll tell you a good reason to keep no follow on the blog comment, It will reduce the number of spam that you will get. But I guess that might be a little off the subject, or maybe not.

  315. Allow me to do a little Pagerank sculpting here …

    In Bot Obedience: Herding Googlebot back in 2006, Matt, you made a good post about how to sculpt Pagerank – one option being to use nofollow. Note there are many other options that still exist.

    Oops, sorry, I think I’ve made a nofollow link on your site to another page on your site … you don’t mind I’ve just sculpted your Pagerank, do you Matt? 😀

    While I’m at it, let’s link to it again for emphasis, but this time let’s bounce the link off a friendly redirect.

    That’ll teach you to change the rules while the game is in progress! 😉

  316. I don`t understand. But I did get a big piece of understanding from off the comment response you gave. Such as google does throw out hints of what does or does not work. Let`s face it. Google leaves us guessing. But I think that I did grab a nugget from this post.
    Page rank is good but without relevance it still could be pretty dead. Good post. I learned a lot.

  317. so why the heck are you no-following the comments? o_O

  318. If you wouldn’t recommend nofollows on comments then what is your reasoning for doing such?

  319. webtechnepal, links in comments usually lead off-site so nofollowing them isn’t Pagerank sculpting per se, although it is of course Pagerank manipulation. 🙂

    However, my comment above linked to a page within this site and this link is still nofollowed – so this is Pagerank sculpting on Matt’s own site. WordPress does not differentiate between on-site and off-site links when nofollowing links in comments.

  320. Thanks for the informative post!! But won’t the nofollow tag in a way deter the flow of quality comments to my site? Also I was reading an article that said about 50 outbound links which are do-follow is not really too bad for a website to have (does not in a way affect the site’s PR in a negative way). Is this true?

  321. Hey Matt,
    I just have one question regarding how rel=”nofollow” is being treated by Google. At SMX you said that rel=”nofollow” used to be a way to control the flow of page rank within in your site and now wouldn’t be as effective.
    In a follow up interview with Danny Sullivan you made a comment that Google ultimately determines which links pass rank and how much.
    Would it be fair to say the stated change of how Google treats rel=”nofollow” effects primarily links to other internal pages?
    I noticed quite a few comments to this blog post that including links with the assumption that these leak link juice, however since these external links are all rel=”nofollowed” Google would treat these links differently, right?

  322. And what should we do at all about links on blog comments?

    I like to receive comments on my blogs, and I definetely don’t mind letting commentators add links on it (I moderate all comments and any spam is deleted before even appearing to visitors, I’m talking about real ppl talking about real stuff with real links).

    But I also don’t wanna lose PageRank on every comment with a link… If I can give PageRank and lose none, I wanna let the comment there, even without nofollow. But if I lose PageRank on every link, even inside original post, EVEN MORE if nofollow also takes PageRank out of me, I may just start using JavaScript or simple text without anchor for links… I definetely don’t like this idea, but I dislike even more losing PageRank on each outlink on my site. I’d just link top quality sites that I actively wanna vote for Search Engines.

    So, my question is: Do we lose PageRank points when we link below-medium, non-spam, sites? Do we lose PageRank points on every comment our blog receives? Even if we use nofollow on them?

  323. If nofollow links are invisible to the algorithm, then why do my clients end up ranking well for the blog commenting name I use for them which is almost always nofollow’ed?

    I think nofollow links are just handled differently. This is how I think we are safest handling this issue: Use nofollow links for every link someone else gets to add to your site. And then, of course, to your “Print this page” and shopping carts that are useless to search results. It’s pretty simple.

  324. ”You can try to sculpt your pagerank if you wish, but fixing the issues you may have with your site structure and architecture would benefit you more in the long run. Also; unless you are very skilled with pagerank and SEO, you should probably not attempt anything as it could lead to unintended consequences.”

    That sort of solidifies my thoughts that Google has always liked and still likes sites that are most natural the best – so to me it seems like it’s best not to stress over nofollow and dofollow – regarding on-site and off-site links – and just link to sites you really think are cool and likewise comment on blogs you really like )and leave something useful)… if nothing else, if things change will nofollow again, you’ll have all those comments floating around out there so it can’t hurt. And besides, you may get some visitors from them if the comments are half-decent.

  325. Hi Matt

    i asked about the no follow attribute, as my colleagues have them and they still got more page rank than me and links even though most of theirs is no follow.

  326. @James
    Absolutely right. Matt’s comment “but rather other parts of our system would encourage/reward those [out]links” does indeed say a great deal, as my understanding is that outbound links haven’t historically contributed positively to any form of rank for the donor site. However I’m prepared to be told otherwise.

    I also agree that SEO’ers can suffer from BO (backlink obsession) yet shy away from being equally generous with their own links.

    Therefore seems like this could be a time for change.

    My feeling is that Google dangled the PageRank sculpting carrot and for obvious reasons we grabbed it eagerly as it became another SEO tool that handed us some more control. It seems like it has been taken away which I guess is the natural evolution of search algorithms.

    @Matt Cutts
    I’m sure I read somewhere that Google considers all SEO to be black hat. Aside from the obvious requirement for quality content a la linkbaiting, would you agree with that statement?

  327. ok i need a lil insight on something . my site is an art community and the data on homepage changes all the time i mean users constantly posts new artworks . so thumbnails and links on homepage updates every min. thats why i place a nofollow in the structure so they are automatically nofollowed. is that bad ?

  328. Matt, thank you for this very interesting indeed.

    I am very careful about sites I link to, and also whilst I allow comments on my blog, I moderate them all first. If I feel someone is linking to a bad neighbourhood I always decline the posts.

    Thanks again.

  329. Google promoted hardcore the nofollow attribute, and now you guys tend to back down.
    Of course the nofollow is not beneficial to the Web. It went too far into sterilizing links. At first, goal was to combat spam, but greedy websites used it to sculpt PageRank. Can we say it nofollow totally backfired on Google ?

  330. I am not sure this no follow tag works well, because i have a website where even applying the tag Page Rank leakage does happen.

  331. Why bother linking at all if your going to add no follow. Or is that too easy for you guys. Surely content is still king.

  332. Thanks for sharing this, Matt. I’m happy that you took the time to do so considering that you don’t have to. What I mean is, in an ideal world, there should be no such thing as SEO. It is the SE’s job to bring the right users to the right sites and it is the job of webmasters to cater to the needs of the users brought into their sites by SEs. Webmasters should not be concerned of bringing the users in themselves. (aside from offsite or sponsored marketing campaigns) The moment they do, things start to get ugly because SEs would now have to implement counter-measures. (To most SEO tactics) This becomes an unending spiral. If people only stick to their part of the equation, SEs will have more time to develop algorithms for making sure webmasters get relevant users rather than to develop algorithms for combating SEOs to ensure search users get relevant results. Just do your best in providing valuable content and Google will try their best in matching you with your users. Don’t waste time trying to second guess how Google does it so that you can present yourself to Google as having a better value than you really have. They have great engineers and they have the code—you only have a guess. At most, the SEO anyone should be doing is to follow the webmasters guidelines. It will benefit all.

  333. Thanks very useful post. However I’m not using “nofollow” in my blogs but they are not loosing any pr.I think if your outgoing links concerned about your site subject , it’s not sculpting anf pr of your site.

  334. I don’t do any page sculpting because I want the whole site to be regularly indexed by the search engines.

    Nor do I add ‘no-follow’ to external links at the moment.

    Maybe I will suffer for this in the future.

    My pagerank is 3 and many of my incoming links are no-follow so they must surely be adding to the PR which makes me question the usefulness of the no-follow tag in the first place.

  335. Steve, sometimes good information to users is a consolidation of very high quality links. We have over 3000 links to small business sites within the SBA as well as links to the Harvard and Yale library, academic journals, etc. But because we have the understanding that there should be no more than a hundred links in a website (more now from what Matt said) we have used nofollow on all of them out of fear that Google will penalize our site because of the amount of links.

    Yes the links we have are found elsewhere but our focus is saving our users and clients time so we consolidated the links because it takes hours and hours and hours of searching to find them and some searchers are not very savvy when it comes to looking for, and finding, good quality information. I look at the links like a library, my library has these books, so do a bunch of other libraries. I think it is a shame that I have to hide my books from Google because I have to many really good ones because it is seen as a BAD thing in Google’s eyes. Darned if you dont create a good site, and darned if you do.

    Matt, my biggest complaint with Google and this “page Rank” nofollow nightmare is it seems we need to have a certain type of site to get ranked well or to make your crawler happy, you say you want a quality site, but what my users deem as quality (3000 links to the best academic information on the planet for business development) is actually looked at by Google as a bad thing and I do not get any rank because of it, makes it hard for my site to be found, and people that can really use the information can not find it when you yourself would look at the info and think it was fantastic to find it all in one place.

    You just do not give us any clarity on how to really produce quality.

    I compare the latest Google search results to this: Mcdonalds is the most popular and is #1 in hamburgers… they dont taste that great but people still go there. BUT I bet you know a good burger joint down the road from Google that makes awesome burgers, 10X better than Mcdonalds, but “we” can not find that place because he does not have the resources or budget to market his burgers effectively.

    The internet was the little guy savior, simple sites could rank well locally. Sadly your company is in the process of destroying that. In this economy small business with zero page rank that are listed on page 22 of results, need to be found in order to survive. My customers are really suffering because of the work that is coming out of Google, it keeps getting worse. Their conversions are still good coming out of Yahoo and MSN and now Bing. They do not have the resources to produce blogs, forums, or $5,000 websites let alone pay for Adwords when they are just trying to pay rent and not a lot of people can do their own web production.

    Too much useless content is being created ONLY to make your search engine happy, I saw a lawn mowing site that had a forum and a blog.. This Google appetite for content is getting out of control.

  336. This must be one of the most controversial attributes ever. I participate in photographic communities. The textual content there is quite sparse, as it is a visual medium, with only basic descriptions. However, the community is very active and the participants leave a lot of meaningful comments. Now, with the “nofollow” used everywhere the photographic community is punishing itself for being active and interactive without knowing it. WordPress and Pixelpost now have “nofollow” built in almost on any list of links (blog-roll, comments etc). The plug-in and theme developers for these platforms followed suit and yes, you’ve guessed it – added “nofollow” almost on every link. So, every time I leave a comment without being an anonymous coward or if some one likes my blog and links to it in their blog-roll than I’m or they are diluting the rank of my blog? Does it mean for my own good I should stop participating in the community? Should I visit hundreds of blogs I visited in last three years and ask the owners to remove my comments and remove my site from their blog-roll to stop my PageRank from free falling?

    What happened with creating content for the audience rather than “robotic” recipients.

    What happened with Web 2.0 and organic communities of interests.

    There was a time when anchor tag created the Internet, now it is becoming its nemesis because of one poorly chosen, misunderstood due to undocumented and part secretive nature non-standard attribute. I’m looking forward to HTML version X, where anchor tag will become deprecated.

  337. This pagerank theme is getting understood in simplistic ways, people are still concerning about pagerank all the time (talking about SEOs). I just use common sense, if I were the designer of a search engine, besides of using the regular structure of analysis, I would use artificial intelligence to determine many factors of the analysis. I think this is not just a matter of dividing by 10, is far more complex. I might be wrong, but I believe the use of the nofollow attribute is not a final decision of the website owner any more is more like an option given to the bot, either to accept or reject the link as valid vote. Perhaps regular links are not final decision of the webmaster too. I think Google is seeing websites human would do, the pages are not analyzed like a parser will do, I believe is more like a neural network, bit more complex. I believe this change make a little difference. People should stop worrying about pagerank and start building good content, the algorithm is far more complex to determine what is next step to reach top ten at Google. However nothing is impossible.

  338. “I use nofollow if I link from one of my sites to another one of my sites, or to a paid sponsor of one of my events.” purposeinc, that’s still true.

    So would it be a good idea to put nofollow on links to a versions of a website in a different languages hosted on the appropriate ccTLD?


  339. I liked your starting sentence, very nice, “random surfer”:

    Even with NOFOLLOW, having link on Google homepage improves traffic to one’s weakest site millions (if not billions) times. Obviously.

    The probability that the random surfer visits a page is its PageRank. And, the d damping factor is the probability at each page the “random surfer” will get bored and request another random page. One important variation is to only add the damping factor d to a single page, or a group of pages. This allows for personalization and can make it nearly impossible to deliberately mislead the system in order to get a higher ranking. We have several other extensions to PageRank…

    So, Google tries to be honest; powerful companies do not care and use other marketing tools, not necessarily AdWords.

    Guess, who is on top of “Made for AdSense” list?

  340. From an SEO point of view I just get more and more confused with this stuff! Just keep it simple – don’t tell anyone how google works and then we can all just get on with writing content that matters.

  341. Hi Matt,

    Thanks for the information.
    Again great to know what Google is up to an how they operate. I had a read of the link to “principal eigenvector of that normalized link matrix.” and had a head spin. Thank God I don’t have to know I(P) = blal bla bla to rank.

    I believe I’m just going to stick to what I do – on page SEO. Google just seems to be using Nitro in their search engine vehicle and the average man is using a standard gasoline 1 litre car and cannot keep up. What’s next after PageRank, Paragraph rank?

  342. Hey Matt,

    Question: Why are you using nofollow on your links TO your comments on your site, then?

  343. Is over use of the nofollow considered to be abuse by google? I see a lot of talk about what is and is not good practice and Matt has kindly taken the time to post quite thoroughly on the subject. But I can’t help but feel it will come down to some sort of percentage, like KEI, whereby overstepping will result in a possible penalty.

  344. “I would let PageRank flow freely within your site”
    Thats exactly what i am doing in all my projects. No nofollow. But i have just some small webprojects, maybe for realy big sites this will work diffrent?

  345. Matt statements and the arguments here are really confusing everything. The more I read, the more I am lost. What’s the relevance of the no follow tags now?

    Now I guess, I will just have to wait for Google to make a simple FAQ about it.

  346. I think that knowing this makes life easier. No more time wasted sculpting PR.

  347. @matt: I notice a bit of WordPress-related talk early in the comments (sorry, Dont have time to read all of them right now..), I was wondering if you’d like to comment on Trac ticket( – Related to the use of nofollow on non-js-fallback comment links which WordPress uses – Its linking to the current page with a changed form.. the content and comments should remain the same, just a different form.. I think the original reason nofollow was added there was to prevent search engines thinking the site was advertising multiple pages with the same content..

    If you’d like to throw a comment in there related to all your WordPress-y thoughts of the use of no-follow that’d be great, I’m no SEO-expert, and to be honest, I generally shy away from what a lot [of SEO guys] will claim, so input from someone who knows what they’re talking about (and i reconise) would be great.

  348. Thanks for this article, I wanted to understand better how Google pagerank works and I appreciate the diagrams. I know that this isn’t precise, but it helped me to understand a little better what pagerank means and what is going on.

  349. Excellent article.

    I have read more that 25 articles about nofollow and its usage and effect.

    But to good to header from you (official google person)

  350. Nofollow attribute does no effect your pagerank anymore. Whenever looking at your backlinks they are out there. their effects can be seen on new websites pagerank with only building nofollow backlinks. I’m totally agree with pagerank sculpting.

  351. Thanks for the info on nofollow and pagerank. It makes sense that this will always be a moving target less everyone will eventually game the system until it’s worthless but at the same time it’s worth it to know a few tricks. I still have open concerns on how freshness of content factor in, the only time i’m ever annoyed by search results these days is when the only links available (on the first page at least) are articles from 4 years ago.

  352. Thanks for this post, I am however still slightly confused by no follow. It seems the debate still rages on SEO forums. I don’t think I am the only one who is still confused. Could a small amount of awareness of no follow and how it works cause people to structure their sites and links in an artificial way (to try and preserve their own PR) ??

  353. I normally don’t link to any other websites, i feel the PageRank will ”flow” outside, and build related content around my niche. However i found many low PageRank sites ranked higher than me. I am confused.

  354. I think the fact that you removed no-follow from the links on your site is very noble. Lots of sites have commentors contributing valuable information to the topic at hand in a ‘forum’ sort of way if you will and these comments make for a very interesting article.

    As for people striving to get a very high pagerank, even though I think it may be achievable eventually, persons need to understand that great content will get your site up there instead of spending all your time trying to increase pagerank.

    My 2c

  355. What an article… thank you so much for the priceless information, we will be changing our pages around to make sure we get the highest page rank available to us, we are trying to get high page rank sites to link to us, hopefully there is more information out there to gather as we want to compete within our market to gain as much market-share as possible.

    Again thanks for the thoughts.

  356. Gotta love Google. They turn the entire SEO/webmaster world on its head with an announcement of a new attribute in 2005. We all go out and make changes to our sites to take advantage of this new algorithm change that is said to benefit out sites. And then 2 years later, they change their mind and rewrite the code – and dont bother to tell anyone. And then a YEAR LATER, they make an announcement about it and defend the change by saying “the change has been in effect for over a year, so if you haven’t noticed obviously it isnt that big a deal”

    Talk about arrogance. That is like stealing a penny per transaction from a bank account for a year – and then saying, well since it took you a year to notice the theft, it cant be that big of a deal.

    I dont know if Google gets its kicks out of keeping Search Engine Marketers and Webmasters jumping through hoops – or if they are in cahoots with the big SEM firms – so that they get this news and these updates before the average guy on the street. Either way, they are seriously getting a bit too big and powerful and the time is RIPE for a new search engine to step in and level the playing field.

  357. thank you so much
    im going to let my website link even if its with “external nofollow” attribute may be google give me some points form you 7 PR 🙂

  358. The nofollow tag is being used for page rank sculpting and to stop blog spamming. In my mind this is tant amount to manipulating page rank and thus possibly ranking position in certain cases. I do post to regularly blogs and forums regarding web design and this improved my search ranking as a side effect. Whats wrong with making an active contribution to the industry blogs and being passed some Pagerank. Google needs to determine whether the post entry is relevant then decide to pass pagerank after the analysis or just decide that blog should not pass PR in any event. Whats gone wrong with the Internet when legitimate content pages do not pass PR?

  359. This means that no matter what every comment on your blog must be moderated because I also believe that Google penalizes you for linking to bad sites. This why many people choose no follow attributes so that people can post and you not worry about whether your site will get penalized for the link.

    Unless there is another way to get around this then every post must be moderated.

  360. Matt Cutts, it’s Shawn Hill from Longview, Texas and I’ve got to say, “you’re a semseo guru”. That’s obviously why Google retained you as they did. Very informative post! As head of Google’s Webspam team how to you intend to combat Social Networking Spam (SNS)? It’s becoming an increasingly obvious problem in SERPs. I’m thinking Blogspam should be the least of Google’s worries. What’s your take?

  361. I’ve been a web developer since 2002 and this is the first time I’ve heard of the “nofollow” attribute for the link tag. I feel a bit like an idiot but this is huge for me right now. I’m nearly embarrased to say but I’m doing a lot of SEO work atm and this is like a magical tag. Thanks big time, this an absolute gem!

  362. My thoughts on this topic. Matt, I’m not really believing you… 😉

  363. I like the regular do-follow. I think the net is getting too strict. Remember in the 1990s when the Internet came out? I miss those net days.

  364. It is good to get more clarity on nofollow and the concesquences from someone at google.

  365. Bad Idea, many, many websites nowadays use nofollow in every single external link to sustain the value of their PageRank, not only to prevent comments spam, which is in my opinion defeating the original purpose of nofollow. Google should’ve “marketed” nofollow as nospam without going into much details of how that links to PageRank to keep webmasters away from getting into the idea that nofollow means better PageRank

  366. I still would like to know what you think about linking to other sites from within the posts a blogger publishes.

  367. I’m a bit confused now. If it is better not do place a nofollow, so why does it happen on this site? On the other hand, many sites have a place, where visitors can place links (like this one right here). And so the owner of the page does not guarantee for the links placed (makes sense, since they are not reauthored). So I don’t want do be associated with bad sites and do a no follow. Seems like a workaround, lets see whats going to happen in future.

  368. Very interesting article, I’ve removed the nofollow tags from my site, though i noticed that the nofollow pages did have a lower pr than other pages… interesting to see what effects it has… pity it takes so long for updates to happen as it’s hard to know what tweaks work.

  369. All you need to know about getting results is all told to us publicly via here and the YouTube video series, all you gotta do is listen.

  370. Surprising you find partial indexing with at least a link. Craigslist listing show all the time so I guess the no follow tag is a joke. You decided I don’t know.

  371. Hi Matt (and everyone),

    Two weeks ago I changed a few internal anchor text links for a HTML SELECT Label in order to save some space in the menu bar. Today, when I saw in Google the Cache (text-version) page of my site I realized that all the links in the HTML SELECT Label cannot be followed. So I understand that Googlebot doesn’t follow this links and obviously there’s no inbound ‘link juice’. Is that so?

    Thank you and excuse my english, it’s no my mother tongue.

  372. I’ve been concerned about nofollow attributes for quite a while. It was hard to convince my boss about what was being said about this, but thanks to your post now realizes how it really works. Thanks Matt.

  373. Everyday you see evidence as this might just be the wrong use of tags considering no follow. Look at the tagging they set up for google place pages.

  374. The web was created with the intent of sharing information, not hoarding it … While nofollow attribute is a great way to not vouch for information on certain sites, it’s definitely not a useful tool to increase your site’s pagerank …
    A second reading of the post reveals that the post is more about pagerank flowing out of your site than being contained within. The only ones who need to worry about who they are linking to are the ones who are linked to very often. So unless your site has a high pagerank, you may or may not use nofollow and it will not make much difference.
    Having great posts is a great way to have links flow towards your site and have a high pagerank.
    As the engine techies say “There’s no replacement for displacement”.

  375. Hello Matt,

    I read your post carefully. I’ve always been concerned on making high page rank. This posts help me. Thank you so much Matt.

  376. “1) making great content that will attract links in the first place, and 2) choosing a site architecture that makes your site usable/crawlable for humans and search engines alike.”

    This is so funny. Google stifled the notion of linking to “great content” the minute they let on to how important linking was to passing pagerank. In effect, the importance of links has indeed led to pagerank hoarding and link commoditization which in turn leads to all of the things google doesn’t like such as spammy links, link farms, link selling, link buying, etc. What you end up with is a system, much like our economic system, where the rich get richer and poor get poorer. Nobody has a problem linking to CNN, as if they really needed the links. On the flip side who wants make a dofollow link to a site that’s 2 days old, great content or not when you can provide your visitors a nofollow link which is just as valuable to them. The whole notion of benefiting from a quality outbound link is a joke, the outbound linker receives 0 benefit when you factor the outflow of pagerank.

  377. Does PR flow to iframes, in particular, external ones?

    if yes, does rel=nofollow work for iframes, to prevent PR flow to it?

  378. I might look at iframes as that PR should flow to the source and nofollow might make no sense as because why would you iframe it in your site if it has no content to follow? and that spam is usually not a case here because, for example, it is not common to give visitors who place comments the ability to use iframes.

    On the other hand, one might say that he is placing an iframe for an external source to show something, but not recommending the content and might actually be contradicting with his. And so, nofollow.

    But what I am really asking about, is what is technically going on in terms of Google PR algorithm? PR Flows with iframes? nofollow works with iframes?

  379. I’ve been trying to tell directory owners not to use “Nofollow” on their links. They aren’t doing the right thing by themselves, the submitters or Google

  380. If I understand correctly, no incremental link juice will flow to the “followed” links, but on the other hand, using nofollow will still stop link juice from bleeding through links that are followed.

  381. Really great post. I understand the no follow attribute before. It’s too bad so many bloggers throw no-follow on their comments, it discourages people from extending the conversation. I for one do not plan to use no-follow for my comments, but will remove spammy messages.


  382. @Dan in Canada:
    Therefore it isn´t that useful to use nofollow-links. Furthermore I think it isn´t grateful for people who put a great effort in writing comments that the nofollow-tag makes their links useless.

  383. I want to know one thing matt, My twitter pr was 4 last week, but this week is showing n/a. may i know that reason of this, I did not give any link there, and I did not do any cheating. So what haapend there,My twitter account name is seoexpertinda,I followed you.

  384. This is an excellent posts and really clarifies the subject. I actually quoted on (bottom of the page):

    Or, as Matt Cutts from Google states it: “The essential thing you need to know is that nofollow links don’t help sites rank higher in Google’s search results.”

    Even thought this blog is not an official blog from Google, I hope that the wording on my site is fine.

    HP Jeschke

  385. There is still a way to sculpt your pagerank using JavaScript. Not only it has the same effect as old-nofollow but it’s really simple.

  386. (spread across a number of pages) which lists something like 1,000 restaurants in a large city with contact details and a web link to each of those restaurant’s home page. Given that the outgoing links are relevant to my content, should I or should I not be using REL=nofollow for each link given the massive quantity of them? How will my ranking for pages containing those links and pages elsewhere on my site be affected if I do or don’t include REL=nofollow for those links? My fear is that if I don’t use REL=nofollow, Google will assume my site is just a generic directory of links (given the large number of them) and will penalize me accordingly.

  387. thank you wery mach 😉

  388. If you have a question about your site specifically or a general question about search, your best bet is to post in our Webmaster Help Forum linked from

  389. Someone know when google will update pagerank again?


  390. Will link velocity be a new factor in determining Pagerank when a company builds it’s link structure abnormally fast in 2010?
    Just curious to get an opinion from the source itself.

  391. It seems as if PageRank Sculpting works and this is all the big 52 fakeout. The algorithm and specific penalties handed out are still based on the technical engineering. I would go as far as to say good site architecture should include some sculpting in order to be effective. Otherwise we build it and it bleeds PR.

  392. Great article, I’m currently working on a few political blogs trying to get some content up and running along with figuring the entire google/seo thing out. Very confusing especially for a beginner. I’m also going to try the custom search you outlined.

    Thanks again!

  393. “Print this page” and shopping carts are useless to search results.

    Thank you for article.

  394. I really like this site. You guys seem to know what your talking about. Can anyone tell me why i got rejected from Google adsense???

  395. Thank you very much Matt for all your work to make more transparent the rules of Google. You opened here a subject who arouses heated discussion – 394 comments approved at this time in a year and a half show enough about it…

    There will be always an unfinished story: the webmasters or SEO experts who will want to know the complete algorithm of the search engines for their needs and the necessity of those SE do not show all the truth, to avoid attempts to manipulations of the results in searches.

    …And a commercial purpose of course… people who can`t find enough knowledge to reach the top positions in search engines will spend money in advertising –> AdWords in this case… Things not so hard to understand.

    Good luck to everyone,

  396. I did not share this idea about pagerank.
    Because one million of BL with pagerank 1 cannot equal to a BL with pagerank 5.

    I cannot believe..

  397. I am a great fan of Google’s way of doing things(search engine and others). But often find it testing webmasters behavior against a little change Google suggest. After sometime things go back to their old place… 🙂

  398. Why doesn’t google publish a document explaining everything about the (one of the most discussed) page rank.

  399. Uf da.

    The net effect of just having read all 398 comments on this thread is that I would rather give up on SEO and buy PPC ads for my clients instead. Isn’t that the goal after all? : )

    I really hope that folks don’t take the idea of disabling comments to heart… first that isn’t much fun for you the blog owner or your visitors. Second… I just did a cursory glance at the SERPS for ‘pagerank sculpting’ (how I found this post). Interestingly enough, the number of comments almost has a direct correlation with the ranking of the URL. I’m not so certain that there is a causal relationship there. But I would certainly consider that Google probably has figured out how to count comments on a WP blog and probably factors that into ranking. I know that I would.

  400. What about links to an affiliate site? Are they considered “bad neighborhood” sites, even if they do give value?

  401. So if I understand it correctly, the PR points are divided by the total number of links, not by the number of links without “no follow” as it used to be earlier.

    Then the only solution is linking less and not using “no follow”.

  402. Am I safe in assuming that any site that has PR can be considered a good neighbor?

    e.g. when checking blog comments if the sites that are posted all have PR, can I assume Google will view them as good neighbors and the links won’t harm my site?


    Jamie Dolan

  403. Greetings Matt,

    Great info about PageRank and linking. I have only one contention. You recommend that we allow PageRank to flow freely (even in blog comments), and yet this appears in all your outbound links in your blog comments…

    rel=”external nofollow”

    I don’t get it…did I miss something?

  404. Quite interesting. If we doing the page rank sculpting with or without no follow links, will it affects our page rank or not ? I agree with Alex , yes the google can give an elaborated information for understanding the page rank deeply.

  405. No matter what, I’m never going to understand PR. For some reason I was a PR4 for quite a while, dropped to PR3 then PR2 and eventually PR0 for quite a while… still don’t know why.


  406. Re the “Print this page,”…it is indeed useless. I have gotten rid of that phrase in my two websites. Also…re comments on blog, about 99% of my comments are garbled useless phrases from foreign countries. I have no idea why. Of course, I just delete them.

  407. If you could just link me in static HTML to the top place of all Google searches then I wont take up anymore of your time.


  408. I have a few websites powered by wordpress and all of them are dofollow. Simply, I think that people who spend some time to write a quality comment deserve to get a backlink.


  409. Great and insightful information as usual M. Cutts.

    I wish you could talk a little more on how link to good neighbourds brings more value to the site and how it affects search engine analysis.


  410. Page Rank is very confusing, i dont think that anyone will completely understand this ever growing algorithm.

  411. I don’t follow the PR thing at all. I have a site with 2 In links and 2 outlinks with a PR3, I have another site with 30+ In links and 2 outlinks PRO. They are almost the same age and are on the same domain. How can this be? There is more than just adding inbound and outbound links.
    btw the cummulative PR for the PRO links is 10 times that of the PR3 site.

  412. Many peopel say that high PR means high ranking.Is that right?

  413. So if I understand it correctly, the PR points are divided by the total number of links, not by the number of links without “no follow” as it used to be earlier.

  414. Re the “Print this page,”…it is indeed useless. I have gotten rid of that phrase in my two websites. Also…re comments on blog, about 99% of my comments are garbled useless phrases from foreign countries. I have no idea why. Of course, I just delete them.

  415. Ok, everyone been infering from Matt’s comments that all of these nofollow comments would kill the page rank of this post. Which means this page should have shown up on page 1 for the phrase I searched which was “does google follow nofollow”. In spite of all these nofollow comment links it still was presented as the most relevant page, which it probably is.

    case closed?

    I personally don’t believe that nofollow links pointing to your site do nothing for your page rank. I have pushed sites to page one using nothing but nofollow links on other’s sites.

    I concede they were not high competition keywords (probably less than 20,000 seriously competing sites, in search for phrase) but I still outranked them with just nofollow links.

  416. I jeft employment with a big e-commerce company that I’d promoted to the tops of search engines for some huge big money terms, number ones for bar stools, chairs, patio furniture and many others. We were under attack from big organisations with big sites and budgets so anything to get the top spot seemed worth considering.

    At the time I was strongly advocating page rank sculting by inclusion of no follow links on “related product” links. It’s interesting to note that my proposed technique would have perhaps worked for a little while then would have lost its effectiveness. Eventualy I reached the point where my efforts delivered diminishing returns which was perhaps unavoidable.

    In the end the site owner entirely re-structured the sites with a usability payoff and modest declines in SE positioning for the big terms.

    It is interesting to read the article which however as others have commented leaves questions unanswered. The ideal internal page rank flow is clearly down the hierarchy and back up again evidently and no page to page links down the line. If this effect is big enough to have provoked an algorithm change then it must be substantial. Removing those related product links altogether would improve ranking and degrade the user experience of the site which surely is undesirable. I suspect the lesser of two evils was chosen.

    It’s certainly a fascinating subject.

  417. Thank you for the detailed information. I was confused about follow/nofollow and the consequences for my shop and blog. I guess now I got it. Greetings from Germany.

  418. I am glad as a SEO newbie for the Page Rank clarification, so that I can best focus my efforts on building good content for my window cleaning business.

  419. Nofollow links discourage comments. On the other hand, perhaps people just trolling for PR links are not the best contributors to the conversation.

    And in the end – write good content (and this post was)

  420. when checking blog comments if the sites that are posted all have PR, can I assume Google will view them as good neighbors and the links won’t harm my site?

  421. I have not at all seen the results I would expect in terms of page rank throughout my site. I have almost everything pointing at my home page, with a variety of anchor text, but my rank is 1. There is a page on my site with 3, though, and a couple with 2, so it certainly is not all about links; I do try to have somewhat unique and interesting content, but some of my strong pages are default page content. I will explore the help forum. (I guess these comments are nofollow :P) I would not mind a piece of this page rank …

  422. I would recommend that guys like you should make your blog dofollow to encourage the online marketing community… what say Matt? 😉

  423. Thank you so much Matt! I’ve been researching about this pagerank sculpting thing and the nofollow attribute. Everytime I find an article or blogpost about it I find a different opinion. Until I found this post coming from Matt himself:) I’m relieved, now I don’t have to confuse myself further with just opinions elsewhere. Thanks again!

  424. So it looks like PR points are divided by the total number of links, not by the number of links without “no follow” as it used to be earlier.

    Then the only solution is linking less and not using “no follow”.

  425. I found this site from a *deep breath* Google search. I was trying to find out if having many links on a page lowered page rank. After reading the article and about 200 coments I am now more confused than ever. (Yes, I admit I gave up reading them part way through…so much for the user experience! )

    Thanks for trying to explain, Matt. It is not my intention to attack you.

    I am now going to go away, try to ignore Google and write a few posts! 🙂

  426. If the algorithm really works as Matt suggests, no one should use nofollow links internally. I’ll use the example that Matt gave. Suppose you have a home page with ten PR “points.” You have links to five “searchable” pages that people would like to find (and you’d like to get found!), and links to five dull pages with disclaimers, warranty info, log-in information, etc. But, typically, all of the pages will have links in headers and footers back to the home page and other “searchable” pages. So, by using “nofollow” you lose some of the reflected PR points that you’d get if you didn’t use “nofollow.” I understand that there’s a decay factor, but it still seems that you could be leaking points internally by using “nofollow.”

    Of course, it’s possible that the algorithm has some method of discounting internally reflected (and/or directly reciprocal) links (particularly those in identical headers or footers) to such an extent that this isn’t important. Evidence to support this the fact that many boring pages that are linked to by every page in a good site can have very low PR.

    All of this leaves me with more questions than answers. Which, I suppose, supports the don’t bother using “nofollow” because it’s a waste of typing time theory.

  427. Well, it seems that what this article says, is that the purpose of the no-follow link is to take the motivation away from spammers to post spam comments for the purpose of the link and the associated page rank flow; that the purpose of no-follow was never to provide a means to control where a page’s pagerank flow is directed. It doesn’t seem that shocking to me folks.

    Yes, the more links on a page the smaller the amount of page rank it can pass on to each, but that was as it was before. With regard to what happens to the ‘missing’ page rank, it seems that if this is the case all over the Internet, and it will be, the total amount of page rank flow is reduced the same all over so you don’t need as much page rank flow to your good links to maintain relative position.

    It also seems that the underlying message is that google is constantly trying to find ways to identify the value of a page to it’s users and as it does so it will promote those pages more strongly in it’s search results and demote those that offer less real value, and it does not care how much you invest in trying to game the system by following ‘the rules’. As a small web site operator with no SEO budget and little time to apply the tricks and best practice, I think this is probably a good thing.

    One final note is that if the links are not directly related to the subject, or you have no control over them, such as commentors’ website links, maybe you should consider putting them on another page, which links to your main content. That way you don’t leak page rank, and still gain hits from search results from the content of the comments. I may be missing something but this seems to mean that you can have your cake and eat it, and I don’t even think it is gaming the system or against the spirit of it. You might even gain a small sprinkling of page rank if the comment page accumulates any of it’s own.

    (I wish I knew how Matt C prevents this log getting spammed, now that would be a useful post…)

  428. So what’s Google’s take on websites no-following every single link they have, like wikipedia?

  429. Thanks for this great article! Reading your blog has been far more helpful to me than taking any other SEO courses… Keep up the good work! 🙂

  430. Hi Matt, can you please answer this question for me. Its been bugging me for awhile. You said, “I wouldn’t place nofollow on my blog category or archive pages”, but wouldn’t this create duplicate content issues?

    Since my category and archive page lists all my post’s content, would google be understanding of the blog structure and not raise any warning flags? (by the way, this question also pertains to tag urls)

    And if google can recognize and doesn’t pentalize me as having dupicate content, does it matter if I have a plugin which automatically places a canonicalization tag for all pages? I wasn’t sure if this would mess up anything.

    I look forward to your response. Thanks and great post!

  431. Nice article and you are trying to keep it simple for people. I think most of the techies are always trying to out-think everything. Keep it simple and build a nice looking site with quality.

  432. Great article Matt – thanks. I’m a bit of a newbie with SEO stuff and just came across this after I trawled through one of my sites trying to add nofollow tags to everything that I didn’t think mattered! LOL.. I shall subscribe to your feed in the future 🙂

  433. Thanks for the information. Being new to SEO and managing my site had some great insight on Pagerank and learned quite a bit on the Nofollow. Thanks!

  434. PS- Blog is great. Spent quite a bit of time sifting through over the past few days. Great insight.

  435. Great article, but I think I need a drink after reading it and all the comments…

  436. Thanks for article Matt that cleared things up. For internal linking I don’t use nofollow except for pages I really don’t want search engines to go and I also have those in robot.txt.

    But on a few of my blogs and websites, I do have banner advertisings that I do have nofollow on and I want to make sure that is not confused with “PageRank sculpting” by google.

    Anyway thanks

  437. Thanks for saving me hours of work adding “nofollow” I havent touched SEO for about a year, when we added nofollow on pages such as “about us”, “contact us” etc to prevent link juice leakage. Now I see that in the current model it is not a good idea.

  438. I didn’t consider the idea of hoarding PR in the way you’ve suggested. I’ve always found it to be much healthier to outlink about 1/5th of the time throughout your pages. I don’t really have any great amount of evidence for this, but still, an interesting (if old) look at SEO. I learned something today, thank you.

  439. I am so confused now.
    I understand that if you put a link on a site that is nofollow will not pass a vote for pagerank but those it still count as a backlink and a improvement in your serp?

  440. Is there anyway “no follow” links can positively affect your search rankings? Most bloggers have to assign this attribute as they can’t trust every site they link to nor do they want to decrease their own PR. Is there any potential for search engine to reward a page that accumulates “no follow” links even if they accrue 1/10 the weight they previously did? This would reward those who take the time to actively blog.

    In regards to link sculpting I think the pro’s of having the “no follow” attribute outweigh the few who might use it to link sculpt. Those crafty enough to link sculpt don’t actually need this attribute but it does make life easier and is a benefit. Without this attribute I would simply change the hierarchy of the internal linking structure of my site and yield the same results I would if the “no follow” attribute didn’t exist.

    Either way I strongly feel that the “no follow” attribute is a win win solution and has enhanced search engine performance.

  441. Hi Matt, can you please answer a question for me. You said, “I wouldn’t place nofollow on my blog category or archive pages”, but wouldn’t this create duplicate content issues? An article can then have more than one link.

  442. If I understand correctly, no incremental link juice will flow to the “followed” links, but on the other hand, using nofollow will still stop link juice from bleeding through links that are followed.

  443. Hi,
    could you please answer me about should we place nofollow attribute to the login page?

  444. Perhaps, This is my first comment on your blog. You providing nice information here.
    Thanks for informing.

  445. This is more helpful then you’ll ever know. We’ve been working hard on our site ( for an industry we didn’t was very competitive which is day spa in Perth. However, it seems that due to Pagerank a lot of our competitors are ranking much better than we are. I’m wondering if there are visual aides like videos (youtube etc..) that you would recommend for us to watch that would give us a better understanding of this? Thanks as Always

  446. Thanks for that matt. I see there are advantages and disadvantages to this story. I guess we have to decode the advantages for our selfs – Keep your updates and methods coming matt very true and helpful:)

  447. Great article, thanks for posting. Really appreciate the transparent information on Google’s PageRank algorithm. I will be avoiding nofollow links on internal pages from now on.

  448. Every mechanism or algorithm is good untill someone brake it. In my opinion as people tend to scam the search results, google is getting more and more consevative upon indexing and ranking search results. When I search a word or a phrase I see more wikipedia, amazon, google, youtube, etc. links returning my search, even the page name or headline does not cover the keywords in the phrase. I’m getting afraid that this may lead to an elitist web nature in the future.

  449. Hey Matt,
    Great article, thanks.

    I do have a question though- I have a site with Google sites and apparently Google sites makes all of my outgoing links “nofollow” automatically. (Oddly enough except the outgoing link to Google) 🙂

    My question is- Does having all outgoing links as “nofollow” hurt my site, or should I have a few “regular” outgoing links? And if so, how do I know to whom I should link to and how many “regular” links I should have?

    Thanks again.


  450. Anyway, PR from what i’ve heard will not be a criteria anymore. It’s no more present on google webmasters tools since months. PR is just a “simple” criteria but absolutely not as important as webmasters think.

  451. Interesting, just after 4 months I explained the Google nofollow paradox, all of sudden Google decided nofollow is not that important anymore.
    You shoukld recongize me a dofollow link to my website then. 🙂

    BTW: this blog (and this page) is still packed with nofollow links even if most comments do not seems to be spam at all.
    So if you are worried about the PR of your blog and you still use nofollow, why should we live the PR flow as you suggest and don’t be worried about using nofollow?
    This sounds like another paradox, but this time is just a question. 🙂

  452. @aker
    No criteria is not right. The PR is no criteria for ranking in serp’s. But a page with a higher PR has often good backlinks.

  453. Matt – if what you posted here is true, then why are the comments on your own website nofollow?

  454. So if I understand it correctly, the PR points are divided by the total number of links, not by the number of links without “no follow” as it used to be earlier.

    Then the only solution is linking less and not using “no follow”.

  455. Thank you for providing such a straight-forward explanation of the nofollow changes and page sculpting issue. It helps to have such a clear picture after all of the back-and-forth and analysis on other (some of them very good) blogs.

    So, at the end of the day, implementing nofollow simply reduces the amount of authority a page will pass on to subsequent pages. The ‘points’ that had once been redistributed to non-nofollow links now simply fall into a HTML Twilight Zone. You can no longer hoard and pass PageRank on a few links

  456. So if I understand it correctly, the PR points are divided by the total number of links, not by the number of links without “no follow” as it used to be earlier.

  457. Hi Matt,

    This is an old post so I hope you’re still reading your comments! 🙂

    How does Googlebot interpret the situation where there are two links on a page going to, say, page “Y”, but one link has a nofollow tag and the other doesn’t?

    Example in our nav bar:

    [a href=”/path/index.epx” title=”News & Events” rel=”nofollow”][img src=”/navigation_image.gif” width=”200″ height=”40″ border=”0″ alt=”Text” /]

    [p class=”NavBarItem”][a href=”/path/index.epx” title=”Text”]Text[/a]

    Paths are the same, but we’re referencing the same destination twice – once/first with a nofollow, then a 2nd time without. I think that this has the potential to be confusing to a bot.

    Your thoughts on this?

  458. It’s a lot easier to explain to my bosses when there are pictures 🙂

    Thanks Matt

  459. Hey there Matt,

    What a powerful post you put together here. It’s always interesting to hear your perspective on PR versus all the fluff you hear out there in forums and other nonsense. Thanks for the advice man.

    Btw, would to recommend using Disqus comment system on your blog?

  460. Here’s my take on the whole pagerank sculpting situation. As I understand it, the basic idea is that you can increase your rankings in Google by channeling the page rank of your pages to the pages you want ranked. This used be done with the use of the ‘no folow’ tag. That said, things have changed, and Google has come out and said that the way ‘no follow’ use to work has changed. In short, using ‘no follow’ to channel that page rank juice is no longer as effective as it once was.

    These days I don’t really worry about ‘no follow’, the way I do page rank sculpting is to simply link to the pages I want ranked – that is I use my internal linking structure and link to those pages, I want to rank, more often.

    The only use I have for ‘no follow’ now is with regards to page reputation. For example, as part of my site navigation if the link to my homepage said Home, I’d put a ‘no follow’ on that, because you don’t want Google to think I’m trying to rank for the keyword home. I’ll ‘no follow’ those that are completely off topic!

    To summarize, quit worrying and using ‘no follow’ for page sculpting purpose! The times have changed and you need to get with the program. Keep up to date with what’s new in search by following us on twitter, friending us on facebook and signing up to our RSS feed.

    Your SEO Coach,
    David Jenyns

    Ps. I also made an audio for you guys to answer in more detail:

  461. Matt,

    This is good but it still doesn’t explain everything properly in my opinion. All I hear is “create good content” which is clearly true but it’s not all there is to it.
    The content is the product, even if you have a brilliant product if you don’t put the effort in to letting people know about it, it means nothing.

  462. Thanks for the information Matt. It does take a lot of work and patience to get this going. I will be referring back to your site.

  463. I’m in the wedding industry and recently a Wedding SEO Company began touting PageRank sculpting as the missing link for SEO. So naturally I got intrigued and searched for your response to PageRank sculpting and your answer for anything SEO-related is always the same. “Create new, fresh, and exciting content, and organically the links and your audience will grow.”

  464. Your explanations are very straight forward and make a lot of sense. Page Rank Sculpting has definitely been a conversation of so many SEO pros and others in the industry alike. It’s good to discuss what works versus the myths. 😉 Thanks for this info.

  465. @Danny I agree with you. It also encourages the so called “power-SEOs” to use older techniques to sculpt their pagerank…

  466. Do you think the whole concept on Page Rank is similar to the laws on abundance and sharing Matt? I’ve found the more I give away the more just keeps coming back. I’ve left my site as dofollow so I’m hoping that it’s just flowing throughout as you mentioned. Sounds like a really nice thing to do anyway!
    NOT an SEO expert as you can see 🙂
    Thanks for the post

  467. I pretty much let PageRank flow freely throughout my site, and I’d recommend that you do the same. I don’t add nofollow on my category or my archive pages. “The only place I deliberately add a nofollow is on the link to my feed…”

    Matt, contrary to you say also you use the attribute: rel=’external nofollow’ here, in your “comments” not only in your “feeds”

  468. PageRank license of google expires, license is owned by Stanford University. Will be a renewed license, ie: is there a future google pagerank?

  469. is this for real thats the Stanford University the PageRank license is owned? Read about..but I always thought that Larry Page invented it, therefore the name…if it is true it ends in January or December 2011 ?

  470. May be google will be again “Sponsore” of Stanford University for this year with some 50-100 million dollars 🙂 I really think that Matt will dont tell us this.

  471. Hi Matt,
    Thanks for the article (and lead-off links as they were good info too) but I did not quite get – if there was a penalisation by Google for sculpting – from the article or whether it was just bad practice? And also to echo what someone else asked ‘is it WORTH actually undoing this type of work on websites SEO’s have worked on’ or simply change the way we work with new sites?

  472. Thanks Matt for the informative post. However I do have some questions regarding blog comments. Let say a blog post of mine have PR 10, the page has 10 links, 3 of them are my internal link to my other related post, the other 7 links are external links from blog comment. Based on your explanation, even the 7 external links are nofollow, my 3 internal link will only get 1 PR each which is still the same if the 7 external link is dofollow. Therefore there is no point of adding nofollow for the sake of keeping the PR flow within your own links. Is this correct?

  473. I just did a consult and opinion letter for an extremely large 200,000+ page corporate website that had been forced to temporarily remove their html sitemap due to some compromised code that overloaded their server and crashed the site. A number of individuals at the company were concerned at the potential, negative SEO implications of removing this page, loss of page rank equity transfer to sitemap targets and a feeling that this page was providing the robots with important pathways to many of the orphan pages unavailable through the menu system. This article was helpful in debunking the feeling that a page with 200,000 links off of it was passing any link juice to the targets. PS. XML sitemap in place.

  474. Interesting….so when are you planning to remove the nofollow attribute from your blog Matt? 🙂

  475. I dont understand how internal links have any effect since the Googlebot finds all pages eventually. or is it just the case that you have the site interlinked which gives a bonus i.e is it the navigation factor which adds value?

  476. Sudeep Chakravarty

    Dear Matt, I have two question, can you help me out?
    What happens when:
    1. A page casting more than one vote for a single page?
    2. When a page cast vote for itself?

  477. Hi Matt,

    I don’t understand why is not good to put nofollow on pages like about us, contact us,…
    You said that you like free flow of pagerank and let google to decide but those kind of pages are first pages from a site and they will have some pagerank for sure if we not put nofollow

  478. So in a nutshell, don’t bother using internal nofollows, and only use nofollows to sites that you don’t “vouch” for.

    I must say though however, that internal pagerank sculpting has always worked very well for me. Thanks for the suggestion Matt.

  479. Thank you Matt for casting some light on the page rank issue. I now have a much better understanding of it and will help me alot with my web development.

  480. I like that you said you let PageRank flow freely throughout your site. I think that’s good and I’ve steered many friends and clients to using WordPress for their website for this very reason. With WordPress, it seems obvious that each piece of content has an actual home (perma links) and so it would seem logical that Google and other search engines will figure out that structure pretty easily.

  481. So this is “Pagerank Sculpting”… hmm. Well, I really don’t see any reason for bots following my links in the sidebar on every page. These are shortcuts for users. For bots links are placed within website at more relative context areas for better understanding. Thanks, for this post. Now I know I don’t have to worry about my implemented aproach.

  482. In my example, if I am passing PR to a local eatery by having a do-follow link but there are 9 nofollow links on that page and I only had 10 points to begin with then that lowers the value I can give from my local foodie blog to that site. In that case would it actually be better to either disallow comments on that page or to disallow links associated with the comments on that page? I mean if my client is a food blogger (and some are) and they tell the restaurateur “when I write about you it will be good for your Google juice because I will place a link to you with my post” then they would really be diminishing the value they could give by having an increased number of links. Kinds of sucks for the blogger who wants a lot of comments, no?

  483. Thanks so much for this information. I haven’t been using ‘no follow’ tags as a general rule and now I feel much better for this. This has been very helpful for me as I’m a new blogger, just learning the ropes.

  484. This is an old article, so my comment is really late to the coffee table. Still, I think it’s important.

    I run a political discussion website, and it features commentary by writers who represent views that are largely found outside of the most popular realm of political opinion. Thus, the tendency will be for us to link to other sites more than other sites link to us, simply because our ideas are not popular.

    In an effort to be as relevant and as factual as possible, we require our columnists to document every factual claim they make (or nearly so) to their sources. As a result, we end up with a LOT of outbound links. Tons, in fact. We do get back links, but not nearly as many as outbound links given.

    We added the no follow tag to help us reduce the amount of PageRank flowing out of the site, but what you’re telling us is that a significant amount is still flowing out anyway.

    So here’s my question: isn’t it fair to say that PageRank is essentially biased in favor of the political status quo and against the political cutting edge folks who come up with the new ideas that would shake up that status quo?

    The same question could be applied outside of the realm of political discussion: doesn’t reliance on PageRank actually discourage the spread of ideas that don’t capture the imagination of the masses due to factors like complexity, “un-sexiness”, and being “different”, while encouraging the spread of fads, gossip, and cliche-thinking?

    Let me clarify that the web, with Google’s help, has aided the spread of unorthodox ideas. My question is specifically about the role that PageRank plays in the process.

  485. Very Nice and informative post and thanks for sharing. I really like it.
    Good work and keep it up!

  486. Everybody seems to be making a big deal about no-follow. It seems that these days a no-follow link still has some sort of value

  487. PageRank sculpting has been a selling point for some seo’s for years. It’s good to know that Google has come out and said that it is a fruitless endevour.

  488. Thank you very much for this and all the other info you divulge. Your blog seems to be the only place to get reliable information like this. There’s too many people out there with too many speculative “opinions” on Google and SEO in general.

  489. My most linked page, according to Webmaster Tools is my html sitemap. I provided the sitemap as a quick way for users to see the structure of the website and navigate easier.
    Initially I nofollowed the sitemap, but after “careful consideration”, note the irony :), and reviewing some of the information on this blog I decided to remove the link from all of the website and leave it only on the home page.
    I think that removing the link to the sitemap shouldn’t be a big problem for the navigation, but I wonder what happens with the disclaimer and the contact page? If nofollow doesn’t sink the linked page, how can we tell the search engine that these are not content pages. For some websites these are some of the most linked pages. And yes for some the contact page is worth gaining rank, but for my website is not.

  490. This was a very good explanation, personally I removed the nofollow attribute from my wordpress comments but instead I try to moderate the comments more and make sure the links in them only go out to quality sites.

  491. OI like WordPress for this exact reason. Let the page rank flow.

  492. I feel it is best to just link to content within your site that is useful and helpful to your site visitors, ie. to place links *naturally* within your site pages. As you come to keywords in your body content such as SEO, that refer to content on another page about SEO, then link to it. This way the PR just flows naturally.

  493. So in a nutshell, don’t bother using internal nofollows, and only use nofollows to sites

  494. Hi Matt,

    Thank you for the article. It was just mentioned in an SEOMoz Whiteboard Friday comment by Justin Briggs so I checked it out. It was about linking out in moderation and with creative anchor text to good sites, even competitors, as it helps the end user and may help the the websites SERPs.
    -Ken Jansen

  495. There is still a way to direct your page rank using Java Script. Not only it has the same effect as old-nofollow but it’s seriously simple.

  496. Thanks for the article. I’ve left my site as do-follow so I’m hoping that its just flowing throughout as you mentioned. Therefore there is no point of adding no follow for the sake of keeping the PR flow within your own links. Is this correct?

  497. Hi, thanks for this post I have been totally confused about the value of nofollows and you have saved me a great deal of hassle, please keep the posts coming its great to be able to come somewhere that provides up todate info.

  498. Id love to see the answer on this question. I think it would be great if this were the case, because the current structure really does encourage link farms and spammy nonsense. It may not be good in-links, but it does count, and people really ride that wagon all the way – and unfortunately, it works for them.

    Michael D June 15, 2009 at 10:10 pm

    Correct me if I’m wrong, but “parts of our system encourage links to good sites” gives me the idea that a site can be “rewarded” for not only creating good content and good site architecture, but for also providing relevant outbound links to other (perhaps more authoritative) sites. Do I have that correctly?

  499. Why do so many people spend so much time researching SEO and page rank? Its really not that hard to figure out, (I am speaking in a nice tone by the way =) – all you should need to be focusing on is advertising and building your website in a manner that is ethical, operational and practical for the content and industry that your website is in/about. If you are not up-to-something, then google will know it, and they will rank you accordingly. If you spend so much time trying to figure out how to get to the top, I bet you google spends triple that time figuring out how to figure out how your trying to get to the top. So and and so forth…and your not going to win. Have good content not copied, stay away from to many out bound links especially affiliates, post your backlinks at places that have something to do with your site, etc etc… Is it an American thing, I don’t seem to see it as bad in other places of the world, that is “always trying to figure out an easy way, a quick fix, a way to not have to put in the effort…” anyway… Thanks for letting me vent. Please not nasty replies. Keep it to your self = )

  500. I find it strange that by creating ‘no follow’ links on a page, the link juice of that page is still diluted, and the juice that would have otherwise gone to the linked sites is just evaporated.

    People who are concerned about using ‘no follow’ on internal links are fools, as your overall link ratio will remain the same regardless. Its like choosing whether to have 10 links on your homepage pointing to itself, or none – both of which make no difference to your page flow.

    On another note, I would like to express my contempt for Google and its so called terms of service regarding the legitimate acquisition of links. why should it care if links are paid for or not? Thanks to the invention of pagerank, it is Google itself that has cancelled out reciprocal linking and has stopped people giving out links due to fear of them losing pagerank, and blogs and forums are worthless thanks to the nofollow trick. so it is now impossible to get decent links organically, without having to pay for them, and those who do give out free links are considered fools. Google has brought this dilemma on itself, and yet it seems like punishing us for trying to get links other than freely! Face facts, no one is going to link to someone without getting a link in return! google has invented pagerank which is like a currency, and so people expect to be paid for links, as giving out links devalues their pagerank and so compensation is now required. It is forcing people to use underhand methods to get links, mostly the ‘paid’ variety.

  501. Considering this article is older now I wonder what has been changed if anything with passing PageRank and such. No-follow links are still effective because you can still get traffic from people where you leave your link even though it is no-follow and you can still add value to the overall discussion.

  502. Great post Matt. At the end of the day, it’s all about great original content.

  503. hang on I’m totally confused now! Let googlebot run freely on your enter website you say? With all due respect why in the world would I want to remove the nofollow from duplicate links for example (eg: home link & logo link that create loops)?
    Why would I want to waste the dedicated time or kilowatts that googlebot has spare for my website (depending on age and relevance score) by letting him waste half his time in my “shipping policy”
    I way rather google finally finds a page that was never indexed yet with a nofollow tunnel but Google doesn’t want us to decide what’s right and wrong.

  504. Things are constantly changing, there is even evidence that nofollow links do count on some occasions. Its really a very complex subject as there is a formula behind the algorithm that takes many factors into consideration trying to guess what factors come into play is very difficult. I always focus on making the site as useful as possible to as many people as possible this is the end goal for search engines as well as webmasters. Webmasters who do this whilst observing the search engine’s guidelines should not have problems in reaching the top.

  505. Thanks for explaining rel=nofollow in detail. I have noticed that the page rank passed to sitemap and privacy page as well and that’s why i though to add nofollow on privacy and sitemap page from the home page link. What do you suggest?

  506. It seems to me the web is a better place when the best thing to do is let it work naturally without manipulating specific factors like this. You should get credit for being open enough to give links as well as receive them without encouraging “hoarding” behaviour.

  507. I guess I have a ton to learn. I always figured a no-follow link was better than no link at all. Thanks for the info!

  508. Hi Matt..

    The Big G always changing the way to rank a site. To me, just do just keep on doing as long we stick to the T.O.S

  509. As for one of the points I the article, I always switch off comments on my blog. Just me though, I don’t have time to moderate them!

  510. I would like to know how Google is handling relevancy with so many websites now jumping on the “no follow” wagon? Seems like just about every major website has no follow links, so with the Panda updates this year what’s happening to all that lost link power? Seem’s like this tactic will stagnate the growth of up-and-coming websites on the internet to me. Am I right here?

  511. Hey Matt,

    I am convinced that Pagerank deploys traffic. I am still fairly new to internet marketing and I read all of your blogs. I always learn so much. You always have useful information and I like how your blogs are perfect from an SEO standpoint. Keep up the good work

  512. For most parts the sophistication in this system is simplified here. I still have trouble understanding the difference between letting link flow withing my pages without thinking about a loop. For example, page A, B and C link to each other from all angles therefore the link points should be shared. But in this loop formula, page B does not link to A. It just goes to C and loops. How does this affect navigation bars? As you know they are meant to link stay on top and link to all pages. I’m lost.

  513. Great post Matt. It can be challenging trying to keep up with google.

  514. Has anything changed in relation to links with Google nofolow Panda

  515. Why is allowed to get away with link sculpting? In every article they publish, there’s a section called “References”, which is ostensibly the articles that were research to create the content.

    But…eHow tags every single “Reference” link with a rel=nofollow tag.

    That doesn’t pass the “I can’t or don’t want to vouch for this link.” test, as they are clearly saying that the links are relevant to the posted article.

    That can’t be for any other purpose than sculpting, and keeping, PageRank on

    Why is it that they can get away with this stuff?

    An example is here:
    Do a “view source” and look at the links in the “Reference” section.

  516. Cheers Matt for discussing this. While important for the authority of your site… after much testing over the past 2 years we’ve seen more and more that links trump pagerank when it comes to ranking power.

    We’re even seeing that links on a page with low page rank, but many back links to the page are more powerful than high PR links.

    Just thought I’d throw in some of the results we’re seeing.

  517. It’s funny, because after all of this, guess what. These comment outgoing links are.. you guessed it.. no-follow.. lol

  518. Good explanation but I must agree with the Angel Demirev’s post. Why is this blog no follow.
    If Google has really implemented guidelines that discourage and quite possibly restrict the spread of ideas then aren’t they essentially doing some major evil?

    Here’s a link to an interesting article on no-follow links and bloggers – not looking for any link love, just wanting to share ideas:

  519. Matt, HyperArts is a Web dev company started in 1997. We have thousands of sites we’ve developed over the years with a “Website by HyperArts” and variations that links back either to our home page or, if done with a specific CMS, to our page about that CMS.

    Should I add a “rel” tag to these links? Perhaps “nofollow” or “author”?

    Semantically, what’s the best way to do these links, some of which appear in the footer of all the pages in a CMS-driven website?


  520. Sorry. I should clarify: The “Website by HyperArts” always appears in the footer of the page.

  521. I like this post….
    This will help me a lot..
    Thanks for posting Matt…..

    Rahul Nigam

  522. One of the reason I love Google as a company is how helpful and transparent they are (to a point that means that knowledge isn’t exploitable by SEO ;).

    Thanks for another great post Matt

  523. Thanks Matt. This post cleared up some uncertainty we had about navigation planning for a new directory site. We’ll just let page-rank flow naturally and trust that if Google sees our content is serving a community, the site will be crawled sufficiently.

  524. Re: Cameron’s Comment. Google transparent? Maybe. Great products for users – yes… but they operate from lofty towers. Can’t get a hold of them. Can’t contact them. They are the ONLY company in the world with zero customer support for their millions of users. Who really knows what they are doing from one month to the month in regards to ranking sites… etc.

    As for no-follow links, it is my understanding that even a no follow link will ad link juice to a sites page rank. But I agree with Edward that adding to the discussion can also bring traffic from interested partakers.

    This is a great site you have here Matt. Lots of traffic. No Google ads? I am curious…

  525. There are soooo many opinions on “follow” and “nofollow” and how the various engines use this information. I recommend going to the source (aka Matt) for an answer right out of the horse’s mouth! Thanks for the info!

  526. What i have learnt with comments only allow them if they give value to your blog i have used this for one of my main blogs bpd and me and it worked i have let comments threw witch were spamee and it just got a google page rank of 2 after a year learning by mistakes google page rank is always going to be a mystery and people will try to beat it they might for a short period after that they get caught out but the people who write good quality content will be the winners and keep writing quality content a question might be does google count how many no follows there are i wounder

  527. Hi Matt, I have a question about PR: N/A. With the recent update I found many sites including mine went from PR: 3 to PR: N/A. I Googled for to find it its banned, but I found its not banned, I posted this question on Google Webmaster forum and couple of other places but I didn’t get any help to fix it. I don’t know whom to ask, or how to figure this out. Could you please help me out please?

  528. Your position about nofollow link have been changed since this post ?

  529. Thanks Matt. thats the first time I have seen page rank explained in any kind of details.

  530. I note that the links in comments here are ‘nofollow’???

  531. Should have added in my previous comment that our site has been established since 2000 and all our links have always been followable – including comment links (but all are manually edited to weed out spambots). We have never artificially cultivated backlinks but I have noticed that longstanding backlinks from established sites like government and trade organisations are changing to ‘nofollow’ (and our homepage PR has declined from 7 to 4 over the past 5 years). If webmasters of the established sites are converting to systems which automatically change links to ‘nofollow’ then soon the only followable links will be those that are paid for – and the blackhats win again.

    my site page rank flow is 89% is good or bad?

  533. mmmhhhh
    I see my nofollow links in other sites when I check for SEO. I mean, I comment here, my link is a nofollow but so far it is recognized as a link from here to my site…
    I would like to make sure I understand. is the nofollow just to avoid rank leakage but google still crawls this?

  534. Can I just remind Google that not all “great content” is going to “attract links”, this is something I think they forget. I have great content on my site about plumbers in Birmingham and accountants in London, very valuable, detailed, non-spammy, hand-crafted copy on these businesses, highly valuable to anyone looking for their services. But no-one is ever going to want to link to it; it’s not topical or quirky, is very locally-focussed, and has no video of cats playing pianos.

    The point is “great content” and “attract links” do NOT necessarily go hand in hand. In fact I’d say any local business page that attracts lots of inbound links is MORE likely to be spam than one that doesn’t.

  535. frankly i read but very confused about page rank about links. Can anybody explain me more details please

    For example ,
    1. i have 100 links on my home page . then is this good or bad?
    2. Should i keep good quality page main directory pages as direct link on home page and rest links should make with nofollow on home page?
    3. If i put law quality pages nofollow link from home page then this is good idea to have good page rank and SEO for home page?
    4. My site had PR 5 before now i notice its down to 3, why this happend?

    Anybody can answer me please