5 Things Google’s Latest Patent Tells Me About Links

Anyone with an interest in linking, link building and algorithms who hasn’t yet read the latest Google patent might want to have a look. It’s located on the USPTO database at an impossibly long URL, so here’s a shortcut: http://bit.ly/9pbQfr.

Note that this patent, while only granted a couple months ago, was filed June 17, 2004. That’s over six years ago.  It’s reasonable to assume that much of what it contains has already been in use, but it’s equally reasonable to assume that some aspects of it are already out of date.

The patent contains 19 specific claims, and clocks in at over 9,400 words. As with most search related patent documents, it reads like a combination of legal and technical documents, and is extremely difficult to understand.  I’ve read through some parts of it four and five times and I’m still processing it.

I understand if you don’t read all of it, but there were a few things that jumped out at me. Let me also state that I don’t read patent documents looking for loopholes, and I don’t recommend you do either. Your time is better spent elsewhere. I read them for validation of what I believe to be true, and to see if the advice I give needs to be adjusted, due to my misunderstanding or changes to something I believed previously.

Here are five things that I spotted that I’ve tried to reword or distill into actionable strategies.

1. The earlier in the content the link appears, the better. No surprise, but… a second link to the same site from the same document is not always devalued, as some in the SEO biz say.  (claim 12)

2. In claim 17, where I read “the topical cluster with which the source document is associated, or the degree to which a topical cluster associated with the source document matches a topical cluster associated with a link”, I interpret this to mean keyword based anchor text does not have to be present for Google to do its thing, nor does the presence of keywords within the anchor mean the link is more valuable.

3. In Claim 18, I read that the color of your links matters. If this is to help with identifying hidden links, that’s a no brainer and makes sense.  But, if it means something else…hmmm.

4. User behavior and interaction with links on a page may be used to determine importance of the page being linked to, but this is not treated the same way for every page on which links exist.

5. The value of a link is independent of the type of document or file within which it is found. As stated, “A ‘document,’ as the term is used herein, is to be broadly interpreted to include any machine-readable and machine-storable work product”.  I interpret this to mean a link from a document other than an HTML file has as much potential to impact the algorithm as a link from a plain old web page. To put it another way, Google is filetype agnostic. If you earn links from a document that’s produced as a PDF, or even in MS word, if it’s linked to and accessible to users on the web, it’s a link like any other link.

I’d be very interested to read your thoughts and comments on this patent. There are many more implications for link building, and I just chose five to get the ball rolling.

Opinions expressed in the article are those of the guest author and not necessarily Search Engine Land.

Related Topics: Channel: SEO | Link Building | Link Week Column


About The Author: has been creating linking strategies for clients since 1994. Eric publishes the strategic linking advice newsletter LinkMoses Private, and provides linking services, training and consulting via EricWard.com.

Connect with the author via: Email | Twitter | Google+ | LinkedIn


Get all the top search stories emailed daily!  


Other ways to share:

Read before commenting! We welcome constructive comments and allow any that meet our common sense criteria. This means being respectful and polite to others. It means providing helpful information that contributes to a story or discussion. It means leaving links only that substantially add further to a discussion. Comments using foul language, being disrespectful to others or otherwise violating what we believe are common sense standards of discussion will be deleted. Comments may also be removed if they are posted from anonymous accounts. You can read more about our comments policy here.
  • jmellon

    Interesting, thanks for the heads up; I look forward to reading the whole thing. Curious however that this patent number did not show up a the Google Patent Search!

  • http://www.seobythesea.com/ Bill Slawski

    Hi Eric,

    Google is just pushing some patents out these days. Actually, there have been 51 patents granted to Google at the USPTO since this patent, Ranking documents based on user behavior and/or feature data , was granted on May 11th. There have also been 34 pending patent applications published since then that have been assigned to Google, too. So it’s really just not quite Google’s latest patent.

    A very strong caution I would raise to anyone attempting to take away something of value from this patent is that the various link and page features that it describes should not be viewed as isolated from each other. The link, source document, and target document features and user-behavior considerations listed or hinted at are intended to be weighted based upon a vector score which considers some or all of those features together.

    For a simple example, if the earliest link in the content used some anchor text that might be considered completely unrelated to the rest of the text on the same page, it might not pass along as much PageRank as the next link in the content that used anchor text very related to the content of the page.

    I can provide a fairly simple takeaway, which is that the link most likely to be clicked upon by a visitor to the page is the one that may pass along the most PageRank. But, if anyone wants more actionable guidelines than that, I suggest that they struggle through the legal and technical language, dig deeply into the patent and the assumptions behind it, consider how things might have changed in the five years since it was published, and then test, test, test.

  • http://www.ericward.com Eric Ward

    Thank you Bill! Your excellent work on patent examination and explanation has helped me for years, and I welcome your comments and feedback.

  • http://www.seobythesea.com/ Bill Slawski

    Thank you, Eric.

    I definitely agree with you that patents aren’t places to go to in an effort to try to find some kind of loophole. What I’ve found them very useful for though, is the chance to view something from a very different perspective, from people who are actively building and working on search engines and algorithms.

    Chances are really good that the PageRank that we’re all familiar with from the early patents and papers about it likely started changing rapidly even in Google’s beta phase, and I think this patent does have a lot of implications from how best to link internally within your own site to topics such as valuing linking opportunities in other places.

  • http://www.sefati.net asefati

    Nice thanks for the article. On the link color note I think that is very accurate observation. Google has said it many time “write content for users” so obviously it is a good practice for the link to have a different color than the text and background color so users can see it or even bold and stand out.

    Also I think websites traffic where your link is really important. Many SEOs go by PR but I don’t trust PR one bit. As far as I know Google PR could not be updated or they could be playing with us (SEOs) to detect natural link building process.

    If I have to go through trouble o get a good link, I rather it be from a high traffic website with good number of backlinks and low alexa ranking that is relevant to my website and doesn’t sell links or if it does, it has good link neighborhood.

    Again thanks for the article and keep it up.

  • http://www.benjarriola.com BenjArriola

    1. The earlier in the content the link appears, the better. No surprise, but… a second link to the same site from the same document is not always devalued, as some in the SEO biz say. (claim 12)

    - I believe it really makes no sense to devalue the 2nd link.
    - I also Google has no intention of devaluing the 2nd link.
    - But after several test I have done, I guess others have too like Rand Fishkin and Michael Van De Mar, only the link text of the first link is read.

    Even if experiments are consistent on Google, I think Google does not want to it to work tis way.

    The patent is how they want it to work, but somehow… it seems they are still fixing that issue.

  • http://www.search-usability.com/ Shari Thurow

    Excellent article as usual Eric. I, too, like to look at search engine patents to get a perspective I might not have considered.

    Regarding link colors and background colors, it is referred to as “click affordance” or “visual affordance” or “clickability” (personally, I like the term “click scent”) in website usability. It has been around in software development for a very long time. Long before Google came into existence. :-)

    We SEOs have got to stop thinking we shouldn’t emphasize the usability and usefulness of footer links or a 2nd, 3rd, or even 4th link on a page. When you only give access to a web document with 1 link only, depending on where you place it and how you format it, you make that link (and content) less findable. If you put too many links to a document on a page, you also make that single link less findable. There is a certain redundancy, formatting, and placement that is acceptable and even PREFERRED by users.

    But I don’t let Google, Bing, or Yahoo dictate that information to me. Users give me better answers and guidelines than a search engine does…or an SEO who knows jack about website usability, information architecture, or UXD (user experience design) for that matter.

    My 2 cents.

  • http://learningseobasics.com/ Kimberly

    I think it’s important to note that there was information provided in this article which was set straight in commentary. As much as I hate to say it, if it weren’t for Bills comment, I would still be confused as to the cluster reference and anchor text being singled out. The article was an attempt at deciphering a patent which may or may not be relevant, and 51 patents later. Nice attempt, but in my opinion, didn’t belong as a tutorial of sorts. Not on SEL, a place people go to because they think God is here.

    The title of the article should’ve been “I’m not certain, but..5 Things Google’s 2004 Patent Tells Me About Links”

  • Duane Forrester

    Just like Sahri says – awa lways, excellent work Eric. :) Thanks for taking the time to review and offer your point of view.

  • Duane Forrester

    Google = filetype agnostic
    Duane = typing agnostic

    Apologizes to Shari for butchering her name above… :(

  • http://www.seobythesea.com/ Bill Slawski

    Hi Eric,

    After seeing and hearing some feedback on your post, I think it might be worth looking at your 5 takeaways more closely.

    1. “The earlier the link in the content the better” – no, that’s just one factor amongst many that might be considered. ” a second link to the same site from the same document is not always devalued” – the patent is silent on that issue, and when the claims discuss “first links” and “second links,” it isn’t referring to that situation.

    2. While this might describe the relationship between the source document linked from and the target document linked to, other features listed involve the anchor text used within the link, the topic of that text, and their relationship with the source and the targeted documents. Again, these features shouldn’t be considered solely in isolation from each other.

    3. The patent specifically mentions the possibility that a link might be devalued because it is the same color as its background, but that’s just one example of a broader category – chances are that differences in color between text and links could have other implications as well.

    4. Not quite sure what you’re saying here about user-behavior data.

    5. The language defining documents in the patent is a standard stock definition which tends to appear frequently in Google patents, and helps potentially broaden the scope of the patent itself. Not sure if it should be ascribed too much weight when talking about how Google might handle links within PDFs or Word Docs.

    @ Shari – First Rule of Usability? Don’t Listen to Users

  • http://www.search-usability.com/ Shari Thurow

    Hi again-

    @Bill…I hope your last comment was tongue-in-cheek, because any qualified website usability professional knows to objectively observe user/searcher behaviors and actions. Because people say they will do something on a website (that includes search engines) and then do something completely different. I observe this all of the time.

    In other words, Bill, I usability test. You don’t. I’m not an attorney. You are. ’nuff said.

    I don’t think Eric was stating that these patent items were facts. I believe (and please correct me Eric if I am wrong) he was saying that these items in this patent provide good guidelines for site and search optimization. In this sense, I agree with Eric’s observations wholeheartedly and look forward to reading more.

    P.S. Nielsen’s article is a good one to read.

  • http://learningseobasics.com/ Kimberly

    1- “Users give me better answers and guidelines”

    2- ‘Objectively’ is a good word. Difficult to find, but a good word.

    3- How does one know what another does at a private home & desk concerning usability?

    4 – I realize that this issue left the realm of educational.This is a respected website, I guess, for various reasons. I only need the education piece and oftentimes, it comes not from what I hear but from what I see is missing, avoided, contained. You don’t need to be an attorney to have eyes.

    5- My point abut Eric’s post was not as supportive as it should’ve been and I am working on my ‘social approach’..I have much to work on in that arena. I guess I will never quite understand the importance of *cough* ‘presentation’ as well as others do. It seems to be a hassle and creates facades that I just can’t draw an attraction to, no matter how marketable it may be.I discuss what I see and oftentimes, to my downfall because face is most adored.

    6- Eric, if I have offended with my approach to what I noted, I apologize. I do it with a speak truth frame of mind and for that element in particular, I am not sorry.

  • http://www.michael-martinez.com/ Michael Martinez

    I think this statement from the patent agrees with a point Bill raised: “The weight of a link may be a function of the rules applicable to the feature data associated with the link. A link’s weight may reflect the probability that the link will be selected. ”

    The rules may be general or specific to the document, so how your document “behaves” very much may affect how its links are weighted.

    Eric, I think we can now attribute the concept of a “cluster link” to you — a link from one cluster (of documents) to another (cluster of documents) may be evaluated in the Google algorithm (sometime in the past, presently, sometime in the future, or never).

  • http://www.search-usability.com/ Shari Thurow

    Hi again-

    If this whole “link probability” is to be believed, then the search engines really have it wrong. Because the presence and clickability of a link can be extremely valuable to users whether or not they click on the link.

    I observe that all of the time during usability testing. In fact, people “see” items on web pages with peripheral vision that are important to them, and they focus on items without clicking.

    Kimberly/Bill- With field work in usability testing, you can and do observe users in the home as well as the work environment. I have observed, and tested, users in a home environment for many years. IMHO, maybe you should learn more about usability before making comments and poor/inaccurate assumptions about it.

  • http://www.ericward.com Eric Ward

    I have learned my lesson and will leave the patent reporting to others. I also should have been up front about my motives for writing this article. I like exploding myths. In the link-building-as-SEO world, people clamor for absolutes. First link best, all others suck,
    18% anchor text good, 28%, bad. Recips=FAIL. That kind of thing. If you have heard me speak or read my stuff, you know I am fond of saying there are no absolutes with this. You can’t examine any particular signal or factor or element in a vacuum. What would appear as perfectly normal in one instance can be total spam in another. Give me any scenario, I want to explode it by example. Because too much time and money is wasted by rank chasers with agendas rather than by people trying to help content find it’s natural place on the web. One thing that is very close to being an absolute: I need to stay away from patents. Shari and Bill, stop me if I try this again.

  • http://learningseobasics.com/ Kimberly

    One of the names mentioned was busy self promoting & blind praising to give actual constructive criticism.

    Yea, Danny, I know..I’m kicked out again for this comment. Sheesh, can’t win for winning. ;-) (yup, that’s really the word I intended to use)


Get Our News, Everywhere!

Daily Email:

Follow Search Engine Land on Twitter @sengineland Like Search Engine Land on Facebook Follow Search Engine Land on Google+ Get the Search Engine Land Feed Connect with Search Engine Land on LinkedIn Check out our Tumblr! See us on Pinterest


Click to watch SMX conference video

Join us at one of our SMX or MarTech events:

United States


Australia & China

Learn more about: SMX | MarTech

Free Daily Search News Recap!

SearchCap is a once-per-day newsletter update - sign up below and get the news delivered to you!



Search Engine Land Periodic Table of SEO Success Factors

Get Your Copy
Read The Full SEO Guide