Would Google Build PageRank Sculpting into linkbuilding?
Google filed for a patent in 2005 that could have transformed how we think about and use links, such as letting webmasters decide how much PageRank a link might pass along, or applying machine-readable labels to links, indicating that some links might lead to “offensive” content (“offensive=very”) or “funny” information (“funny=somewhat”), or where on a page the destination of a link might appear, such as in a footer or main content area. This patent would also include a method to encrypt the content of some links, so that only certain people might be able to access the information that those links lead to. The patent was granted this week.
When Tim Berners-Lee wrote Links and Law back in 1997, as a commentary on the architecture of the Web, one of the statements that he included was that “The intention in the design of the web was that normal links should simply be references, with no implied meaning.” Before 2005, if you surveyed the links you came across on the Web, you’d often see a combination of anchor text describing the destination of those links and the actual URL of the links in question, but not much in terms of “opinion” about the destinations of those links. At least not something within links that a computer program or a search engine could easily pick up upon.
Starting in 2005, we’ve been seeing additions to the way that links can be written that do express some opinions that search engines can act upon. In an effort to help stop comment spam on blogs, Google, Yahoo, and Microsoft all agreed to not pass along PageRank or link value to sites being linked to when those links included a rel=”nofollow” within them, like in the example below:
<a href=”http://www.example.com/” rel=”nofollow”>Link text</a>
Popular blogging software such as WordPress started automatically including rel=”nofollow” attributes in links from blog comments, and Web encyclopedia Wikipedia started using rel=”nofollow” on all links external to the site. Google suggests in their webmaster help pages on nofollow that a rel=nofollow attribute should be used on links that lead to untrusted content, or which might be part of paid advertisements, or links to some pages on your site that aren’t necessary to index, such as login pages that only registered users of a site might find value in.
Another addition to links recently recommended by Google involves something called authorship markup, where you can use both rel=” author” and rel=” me” in certain links to author profile pages and pages written by those authors to help the search engine make associations between authors and the content they’ve created on the Web.
PageRank Sculpting by Siteowners and SEOs
Many site owners and SEOs started using rel=”nofollow” links to try to control where PageRank might flow through their websites. For example, many would include one or more links on every page to a “contact” page on their sites. Thinking that their contact page probably wasn’t one that they would want to spend a lot of PageRank to, they began to add a rel=”nofollow” in the links to pages like that one. Google’s head of Webspam, Matt Cutts, wrote a blog post in 2009 on this practice of PageRank sculpting with a rel=”nofollow,” noting that it wasn’t really a very helpful practice, and probably didn’t work the way that most webmasters might think it does. Matt is listed as one of the people who came up with the nofollow concept on the Microformats.org website.
Interestingly, one of the named inventors of the newly granted Google patent is the very same Matthew Daniel Cutts. The patent focuses on adding a way of passing along “opinion” with a link, such as enabling webmasters to decide a percentage of the PageRank that might normally pass through a link, with an attribute that might look something like: “linkweight=0.5”. To a degree, it looks like a way of enabling a site owner to do some PageRank sculpting. Other numbers could be used as well, including no link weight at all, so that a link wouldn’t pass along PageRank.
The patent is:
Embedded communication of link information
Invented by Krishna Bharat, Matthew Daniel Cutts, Paul G. Haahr, Radhika A. Malpani, Vibhu Mittal, Marcin Kaszkiel
Assigned to Google
United States Patent 7,979,417
Granted July 12, 2011
Filed: June 30, 2005
A method of processing documents is described. The method includes the operation of receiving a document in a search engine crawler. The document includes an embedded first link tag. The first link tag includes one or more information pairs. A respective information pair includes a respective parameter and a corresponding value.
The parameters in the one or more information pairs may correspond to content at one or more content locations or one or more document locations. The method also includes selecting a method of processing content associated with the first link tag in accordance with one or more of the information pairs.
Interestingly, Google explored the idea of allowing site owners to make some decisions about a percentage of PageRank that might pass through a link, especially since Google had earlier come up with a reasonable surfer approach to determine how much PageRank might pass through links.
I’m not sure what impact the granting of this patent might have on “opinions” that are passed along with links such as those that might include an “offensive” attribute or “funny” attribute shown in the patent as examples. I don’t think that it would apply to a rel=” author” attribute in a link, which is an HTML5 standard, or a rel=”me,” which comes from XFN.
I suspect the future will bring some other ways to include opinions with links, making them more than the “reference” that Tim Berners-Lee described in 1997.
Google’s Reasonable Surfer Patent told us that some links, such as “terms of service” links, likely aren’t given much PageRank, to begin with, so PageRank Sculpting doesn’t have much value being used internally on a site.
I’ve written a few posts about links. These were ones that I found interesting:
5/30/2006 – Web Decay and Broken Links Can be Bad for Your Site
12/11/2007 – Google Patent on Anchor Text Indexing and Crawl Rates
1/10/2009 – What is a Reciprocal Link?
5/11/2010 – Google’s Reasonable Surfer: How the Value of a Link May Differ Based upon Link and Document Features and User Data
8/24/2010 – Google’s Affiliated Page Link Patent
7/13/2011 – Google Patent Granted on PageRank Sculpting and Opinion Passing Links
11/12/2013 – How Google Might Use the Context of Links to Identify Link Spam
12-10-2014 – A Replacement for PageRank?
4/24/2018 – PageRank Update
Last Updated July 1, 2019.
56 thoughts on “Google Patent Granted on PageRank Sculpting and Opinion Passing Links”
That’s interesting and really possible point of view. But Google talks about PageRank, not about positions in SERPs, i don’t care about PR so much :). Thanks for this discovery, very useful.
Even if Google goes ahead with this mainstream I wonder what the chances are of this becoming a widespread practice. Probably minimal unless it gets embedded within link tag interface of main CMS platforms such as WordPress.
This reflects Google’s thinking in that period. Now with Google+ and other signals beyond links coming into play, this may not necessitate adding opinion directly in the link tag. Also, would the mess of nofollow have moved Google away from this direction? Interesting to see the concept though.
Fascinating insight into the way in which Google works and the importance of linking.
On a different note – if this was implemented I imagine the take up of this would be pretty slow. The rel=”nofollow” tag is a logical decision – either the PageRank is passed or it isn’t. As such the options for WordPress were – either do pass PageRank or don’t. They chose not to.
If there were 3 options or more – then the decision making process becomes more complicated: probably requiring human involvement.
Invariably the person wanting a link would want the maximum amount possible passed across, while the person offering the link would probably want to prevent spam by passing as little as possible.
I just can’t see this one working – unless the PageRank value can be tweaked by a rating system – but I’m not sure how this would pass back into the mark up language or not be abused.
I’m not 100% sure how this works, but if this were something Google was actually going to implement, wouldn’t it have been implemented already when the patent is still pending?
It looks to me that would be a great way for Google to help determine websites that are abusing the power of outbound links by selling them. If most of the outbound links in the content area are linkweight=10, then we probably have something either fishy, or very high quality.
It would be interesting to see the mashups Google would put together with this variable approach.
PageRank still plays a role in how a page is ranked by Google. It might not be as prominent as it was when Google first launched, and the search engine definitely considers other signals as well, but actual PageRank does make a difference.
I’m not sure that Google will go ahead with things like letting webmasters determine a linkweight for a link or add opinion related value pairs like the offensive and funny attributes that they describe in the patent, or even the encrypted links the patent discusses. But, the patent does mention that those types of things could be built into a web authoring tool, like rel=”me” is built into WordPress.
It seems like Google has decided to try to help influence standards to have these types of opinion values added to links rather than use their own propriety values. That may be a better approach towards having them accepted by a wider audience. For example, the rel=”author” is an HTML5 standard.
When I read this, I imagined seeing a website filled with links with randomly chosen different weights for each link spread throughout the pages of that site. It was a nightmare. I’d rather have Google use something like the reasonable surfer approach, which I think follows some common sense approaches to determining which links people are more likely to click upon on a page.
There’s probably still some value in using some “rel” values with links, like the authorship ones, and I expect we may see some more as well. I don’t think that will change too much with the addition of Google Plus, and the combination of the author tags and any reputation or credential scores from participation in social networks like Google Plus may actually be used together by Google to create some type of author rank.
Thanks. This patent did somewhat resemble a time capsule to me. I’d guess that the authors might even be surprised at some of the ways that the web has changed in the last 5 years.
I think the ways that Google now values links, with different weights passed along based upon a wide range of features, including the location on a page (header, footer, sidebar, main content area), the type and size and color of font used, etc., may make giving siteowners the opportunity to assign a linkweight rather pointless. There are still some reasonable uses for nofollow, as described in the link to the Google Webmaster help pages that I included in the post, but sculpting pagerank isn’t really one of them.
If Google were to use something like those linkweight values (and I don’t think that we will see them use those at this point in time), I could see a lot of webmasters passing along a minimum amount of value to external links, not because of whether or not they might be selling links, but rather because they might try to keep as much PageRank as possible to themselves.
The patent was only granted this Tuesday, so it’s well within the time for Google to implement it if they decided to do so. When a patent is pending, that means that it’s in the process of being granted or possibly even rejected.
I find it interesting about the ‘opinion’ aspect of it because on YouTube, they recently (I have no idea when but I noticed it a few months ago) added the buttons under the ‘Like’ button of “LOL” or “Cute” or “WOW” types of things. How would you do that on links? I have no idea 🙂
Assigning specific PR to different links on the page would be a nightmare. 🙂
Hey Bill – a very informative read with lots to think about, if implemented then yes this would create the next big debate since PageRank I would guess!
I follow all your posts and have just a quick question… How on earth is it you who brings all these massive notifications to light? – I do NOT mean that in any funny way I am just wondering how you manage to get the scoop before everyone else! 🙂
Fair play, I don’t know why but reading this post just made me think of that!
Re: your response to Donnie-
This would make link building really interesting for SEOs…
If this attribute ever saw the light of day, it might lead to uncluttering some home pages. Cramming links to all the directory pages on a large site is not a very elegant method of signalling to Google that they are important pages.
Thanks for the post Bill! Very informative as always. In theory, this is a great idea but practically speaking, I have a tough time believing this will ever catch on. I know how hard it is to get clients to implement semantic mark-up. A CMS would have to make it very easy to assign link weights or any other meta information about the link for this to be successful.
I hadn’t noticed those additions to YouTube. I wonder how many people are using those.
I really do hope that Google doesn’t even consider the option of letting site owners decide linkweights for links. It definitely would be a nightmare.
I could see a lot of discussion taking place if Google does implement the linkweight aspect of this patent, and I suspect a lot of it wouldn’t be very calm.
How do I get the scoop first? 🙂 I stay up late, get up early, and multi-task when appropriate as much as possible. I consider the research I do as continuing education, due diligence for my clients, and fuel for twitter and facebook and Google Plus and my blog.
One of the problems that I’ve always had with PageRank, at least as it was originally conceived, is that you don’t know the motivation or intent behind a link. A page might be linked to not as an “editorial vote” for the page being pointed at, but instead as a warning or a dislike or even just as a reference.
When someone writes one an “SEO is Dead” post, and I decide to write about it, I’d love to be able to link to it but not pass along any PageRank at all to the post. I can do that now with a “rel=”nofollow”, and that’s one way of passing along some kind of opinion with my link. Instead of a full vote for or against a post I might link to, I could possibly even see giving a 50% vote, or adding some other kind of value, like “funny.”
A lot of scenerios are running through my head when it comes to things like the linkweight value that this patent suggests could be added to links by the people linking. I could see it potentially causing some significant changes on the Web, and in how people link to each other.
Good to see you. I agree that the theory behind this is interesting, that giving webmasters the ability to do more with a link than choose anchor text for it, while pointing to a page. I’m not sure how much value there is in having webmasters attempt to determine a linkweight for links, though.
A CMS would have to make it very easy to use these types of semantic additions to pages, but I could see free open source content management systems like WordPress and Drupal making the effort before some of the very large companies that sell content management systems costing tens or hundreds of thousands of dollars.
WordPress has had XFN built into it since its early days, but I suspect that some of it has been used much more since Google announced that they would possibly show people’s profiles next to search results from them if they used rel=”me” in some links, and HTML5’s rel=”author” in certain other links. I suspect that the percentage of people who might make the effort are small, but I have started seeing some profile pictures next to search results only a few weeks after Google announced their use of authorship markup.
I suspect that we’ll see some open source developers tackling opportunities like this if Google decides to move forward with it. I’m not sure that they adopt the linkweight value though, given the fact that it’s almost impossible to get any idea of which links are passing along how much pagerank on any one page.
this would make it easy to understand link building which can be confusing
First of all, great post. This brings up a topic that needs to be talked about – when will Google start providing personal analytics on links and page rank for individual sites? Yahoo has (soon to be had) Yahoo Site Explorer, Bing has Webmaster Tools with a backlink checker, and Google has still not let it’s webmaster community get access to their personal website’s link data. Personally I think Google has a plethora of data on links, so why not make some of it public?
I think pagerank dying – that’s my opinion! The reason – weird results
One of the purposes of Google is to keep everyone confused with link building and Page Rank. One of the facts remain, Google is looking for good reputable content.
I like the fact that this gives webmasters more control over the type of link they want to give, rather than “do follow” or “no follow.” BUT, overall, I think it will be even more confusing for those that want to make their links efficient for SEO.
Google’s Panda probably came from those that are putting up nonsense site with a bunch of adsense. Google is trying to get rid of those sites. The would like to take them down entirely if they could.
Hi Bill. Very interesting. But apart from SEO’s experimenting with PR sculpting I wonder if webmasters in general will really have that much interest. When they link…they generally do so to say “here it is…go have a look”, rather than to assign a value on the link to the destination site. nofollow was adopted by many as a way to discourage comment spam, so there is motivation there for webmaters to use it. I’m not sure there would be the same motivation for opinion passing links. What’s the benefit for the webmaster giving the link?
Interesting patent. I was pretty frustrated (like many others) with the introduction of nofollow as thousands of links went bye bye. I actually welcome an alternative that will let webmasters make more decisions.
Google can be pretty mysterious when it comes to how much value each link on a page might pass along, but at this point I think they like that and probably wouldn’t change it.
Google has provided some link information since their early days with the special “link” search operator. I believe in the earliest days they would show you links to your pages from sites that had a PageRank 3 or higher. At some point in time, they started showing a random sampling of links instead when you did a link search [link:www.example.com].
Google also shows information about links in their Webmaster Tools, and it provides a lot more information about the links Google knows about there, but I’ve heard that those are also a random sampling, and that Google doesn’t show all of those links.
I believe that there’s still some value to PageRank as a ranking signal, but its importance in that area has been diminishing over the years as Google looks at more signals. PageRank may also be used to help Google determine which links it might crawl, and decide whether a page is included in Google’s main index or a supplemental index.
Google definitely wants searchers to have great experiences when they search. There are many aspects about how they rank pages that they hold closely held secrets, and that may not be a bad idea. There are many people who will take any shortcuts that they can in ranking well in search results. In addition to that, if they make some of the aspects of how they rank pages more transparent, they also risk losing potential competitive advantages against other search engines as well.
I’m not sure how much control it actually provides, and it is confusing.
First of all,we don’t necessarily know at any one point how much PageRank a page might pass along.
Chances are that the “linkweight” may also involve relevance for anchor text as well.
The default follow and nofollow are pretty much all or nothing propositions, and Google has let us know that if we use a nofollow in a link that the pagerank that would have gone to that link is lost, rather than spread out amongst other links. If we were to use a “linkweight:50%” would the other 50% of the PageRank that would go to that link lost rather than sent to the other links on the page? The patent doesn’t even introduce that concept, so we don’t know for certain one way or the other.
Hi Jeff Turner,
There are a lot sites out there that have fairly thin or shallow content running AdSense advertisements, or some other advertisements – including affiliate ads. The aim of Panda seems to be to limit or eliminate lower quality pages from ranking well in search results, regardless of whether they show ads or what kind of advertising they might be showing.
It’s more likely that a system like this would work if it were built into content management systems that would make it easier for webmasters to use it. For example, it was recently reported that there are more than 50 million WordPress sites out one the Web. If WordPress added the ability to pass along opinions in links, or determine a link weight, would could see something like this used widely. Google could also introduce it to Blogger (soon to be “Google Blogs”) as well.
Would there be a benefit to a website owner to labeling links as offensive or funny? Or to only pass along a certain amount of PageRank to a page on another site, or on their own site? Maybe. I expect that some site owners might see some benefit in the ability to do those things.
Let’s say that you want to recommend a number of local restaurants, and link to their sites. There are a couple that you like very much. Some that you think are adequate, and some that you don’t like very much. In your list of links, you might give the two restaurants you like full link weight, for the two that you think are just ok you might pass along 50%, and for the two you don’t like much you pass along no linkweight. I could see webmasters using a system like this in a manner like that.
I’d love to see some studies come out on how effective the nofollow might be in places like Wikipedia, which added that value to all external links. Did it stop or diminish the amount of people attempting to add links to the encyclopedia? Has it stopped the amount of comment spam that gets sent to blogs that are nofollow?
I wonder how much of an impact this would have if it were introduced as an option for people using systems like WordPress, and if it were possible for web publishers to easily use it in the main bodies of their posts as well as in comment areas.
How about this, forget about ranking and concentrate on the quality of our work. I think that is the most important.
Personally I think pagerank sculpting exists in other forms. Nofollow links was just an easy way to do it, you can achieve the same effect by structuring your site in a way that gives certain pages advantages from internal linking and so on.
Do you think perhaps with Google+ we could start seeing something linked into comment sections and actual authority given? This is the nice thing about peerindex, klout ete?
I guess we need to have quality links and not just think about the ranking. But hey that is the goal but then we need to do quality things as well not those black hat work that others do online.
So, with Wikipedia using nofollow on all of the outbound links, does this mean that their content/sources/resources cannot be trusted? If we can’t trust Wikipedia, why are they always #1 in the rankings?
Nofollow usage is heavily flawed.
Improving the quality of our content can have a positive impact on the rankings that we receive in search results. It’s possible that if we spend too much time micromanaging things like linkweights that we may start to ignore things like that, so good point.
I’m not sure that anything beats an intelligently structured site, where you understand who your different audiences are, provide pages and content for all of them, and use internal links wisely to benefit from well crafted anchor text. Using something like nofollow links to try to sculpt PageRank is a little like using a shortcut that might keep you from designing a site with a great information and site architecture.
Is it possible from what you’re saying then, that Google might use these ‘opinions’ instead of anchor text as part of their ranking metrics? Do Google file for many patents annually?
Interesting article. Iâ€™m surprised no one has mentioned the parallel with the sitemap PRIORITY tag. With that it is left up to the webmaster to assign some relative importance to individual pages. I dontâ€™t see if being such a big step doing the same for links.
Very Interesting article.Itâ€™s more likely that a system like this would work if it were built into content management systems that would make it easier for webmasters to use it.
This patent provides a framework that could be added to a content management system that would allow web publishers to add additional information to their links in a machine readable format. I don’t think the intention is to replace anchor text as a ranking signal, but rather to investigate things that we are now seeing come out like the authorship markup that Google said recently that they would follow. The schema.org microdata that Google is backing these days provides additional informaiton with links and other content in a machine readable way as well. Not sure that it supplants what this patent contains, but some of the ideas are similar – more semantic information associated with content found on a web page in a computer readable format.
The linkweight information potentially would have the power to impact search results, but it’s possible that with “nofollow” as an option for many webmasters that Google won’t push something like that.
There are usually a few new patent applications published by Google each week – at least one or two, and some weeks up to ten or twelve.
There is some similarity with the XML sitemap priority, but there’s a difference that might make a difference.
XML sitemaps may help the search engines discover new content to index faster than if the search engines relied solely upon crawling programs, but they don’t have much to do with the ranking of those pages. It does help to have all of the pages you want indexed by the search engines included in your XML sitemap, but if you don’t provide text-based links to them, and no one else links to them either, they are going to have troubles ranking well for anything.
I agree with you. If these types of things were incorporated into content management systems then they would have a much greater chance of being adopted by webmasters and used by the search engines. I suspect that if Google announced that they were going to use these that people might do things like create plugins for systems like wordpress and drupal, or build the functionality into SEO plugins.
Wow. As an SEO, that’s definitely fun to think about. But, if they did start letting us judge how much page rank flowed in each link, they’d be telling us exactly how much PR is on a certain page, and if they told us that, we could make new conclusions on how Page Rank is flowed from one site to the next. This is great to think about because it opens up a realm of new opportunity.
Some of these patents do present some pretty interesting what-if situations. Things would be pretty interesting if Google started letting us do that.
Hello Bill, thanks for clarifying that even though the patent is from 2005, it has only been granted this July 2011, so implementation is still theoretically a possibility. Still, in the six years it took to get the patent granted, the way people and software work today does make me think that these new possibilities would be impossible to manage and make every newbie webmaster struggle with the variety of options.
I don’t know if we will see the kinds of additions to links described here, but Google has been successful in having large amounts of people add things like the rel=”nofollow”, some of the rich snippet additions like reviews and recipe markup, and I’ve been seeing people add authorship markup to their pages as well. There may be a number of benefits to adding those kinds of things, and it’s possible that people might see some value in these as well if Google decides to implement any of them.
Comments are closed.