Show Mobile Navigation

March 18, 2014

, , ,

Agent Rank is here! The Google+ Verified Entities Authority Blueprint

Joshua Berg - 11:16 AM

Google+ was built on Agent Rank's link based reputation score template - AgentRank+ Part ①

1. Agent Rank the Patent Vs. Author Rank the shifting baseline.
2. Separate Google Author Credibility Score based on authenticity.
3. Back to square one for answers in the Agent Rank patent.
4. AgentRank+ is the Plus in Google+! The blueprint for G+ authority.
5. All ⑭ Claims of the Agent Rank patent already active as described.
6. The Million $ question: Does AR already affect ranking in Search

Agent Rank Plus - Part 1 - The Google+ Verified Entities Authority Blueprint Exactly one year ago I published, Is Google+ Page Rank the culprit behind the current suspicions of active Author Rank? The first of my PageRank in Google+ series titled. The story came about during a period of much analysis & discussion on the topic of what could be giving the noticeably strong ranking authority in Google Search to certain reputable Google+ Profiles, which tended to easily outrank all others, even outranking profiles with reshares of their own posts.

These reputable authors did not have nearly as high numbers of followers as some others, nor even as many +1's & comments overall, yet their uncanny ability to receive high exposure both in the Google+ platform itself & high ranking in personalized & non-personalized search were noticed by many.

“Within search results, information tied to verified online profiles will be ranked higher than content without such verification, which will result in most users naturally clicking on the top (verified) results. The true cost of remaining anonymous, then, might be irrelevance.” – Google Chairman Eric Schmidt (Feb, 2013)

For quite some time I had noticed that the Google+ Profiles (which are Authors) and the Google+ Pages (which are Publishers) all have their own PageRank score (viewable under certain conditions), which I referred to as Google+ PageRank or G+PR and that the entities with the highest G+PR score always had the strongest ranking authority.


Embedded here is that Google+ post in which I first explained this phenomena, and then continued it with a 6 part series on the power & authority of this G+PR. Explaining how this authority is built & how like PageRank it is a link based reputation score, with an authors score being a function of incoming links through sharing of his content, citation of his content, +mentions of the author & even the authors own rel=author tags.

Unbeknownst to me at the time was that all of these functions, their operational structure, authority to rank & a detailed description of how this was designed by Google as a link based algorithmic reputation score for authors & publishers, were exactly what is described in Google's Agent Rank Patent written years before Google+ was ever created.








For follow up on this topic here is the rest of this series last year Google+ PageRank...



The idea of building a system of verified entities with reputation scores towards improving semantic search has been around a long time and I believe the Google+ system is already Google's functional fulfillment of that desire. However it was previously not confirmed to me until my recent scrutinous study that Google's Agent Rank Patent is the template with which that system was built.

Google's Agent Rank Patent As Studied - by Joshua Berg

United States Patent: 7,565,358
Granted: July 21, 2009
Inventors: Minogue; David (Palo Alto, CA), Tucker; Paul A. (Mountain View, CA)
Assignee: Google Inc. (Mountain View, CA)
Filed: August 8, 2005



The following paragraph buried deep inside the document, was particularly enlightening as to the type of algorithm we're talking about here:

In one implementation, an agent's reputation can be derived using a relative ranking algorithm, e.g., Google's PageRank as set forth in U.S. Pat. No. 6,285,999, based on the content bearing the agent's signature. Using such an algorithm, an agent's reputation can be determined from the extrinsic relationships between agents as well as content. Intuitively, an agent should have a higher reputational score, regardless of the content signed by the agent, if the content signed by the agent is frequently referenced by other agents or content. Not all references, however, are necessarily of equal significance. For example, a reference by another agent with a high reputational score is of greater significance than a reference by another agent with a low reputational score. Thus, the reputation of a particular agent, and therefore the reputational score assigned to the particular agent, should depend not just on the number of references to the content signed by the particular agent, but on the importance of the referring documents and other agents. This implies a recursive definition: the reputation of a particular agent is a function of the reputation of the content and agents which refer to it.





┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅

Here I'd like to mention a special thanks to +David Amerland for advisement on the document study part of this project & adding useful commentary to my *Agent Rank Patent As Studied - by Joshua Berg .
*Note: David's comments should be viewable in the Google Documents margin comment bubbles.

Josh, I have placed my comments in the document. I have gone over it several times now. That there is an Agent Rank in effect is without a doubt. That it is demonstrable for G+ content is also self-evident.

I am not 100% convinced on its ability to rank content, particularly when it comes to improving the ranking of a website or individual web pages. An experiment I took part in, carried out by Eric Enge demonstrated that this could be done but it was not permanent, the rankings dropped off a while later. Personally I think the mechanism is there but not in effect in that Google deprecates it when it lacks corroboration of the ranking signals from outside the G+ environment. I mean that if a profile with a high AR promotes content, you would expect to find corroborative signals of that content, independently, across the web (citations and social sharing on other social networks from profiles not associated with the original profile) - lacking those the initial jump in ranking in search is deprecated. Conversely when it does find those then it becomes more permanent. Google has a patent allocating Trust Rank for content in search that acts as a filtering system for the veracity of contenthttp://goo.gl/1yLcey.

...I think you've pretty much nailed it here regarding AR.  
┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅┅




1. Agent Rank the Patent Vs. Author Rank the shifting baseline.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔



For purposes of this discussion I specifically use the term Agent Rank, rather than referring to what has become an almost mythical term, Author Rank.

The first refers to a very specific term used by Google to define what it calls "Agent Rank" in its patents. Agent Rank covers a broader spectrum of entities than just authors, also describing publishers, editors & reviewers. There are numerous references to this in multiple Google patents, so to be clear on exactly what I am discussing we'll use that term & be referring directly to these patents.

The second term "Author Rank" has become a kind of shifting baseline (in my opinion), by which ongoing articles on the topic consistently refer back to other writings on the topic & their interpretations, instead of back to Google's original documents. While the specific term Author Rank is not used by Google in their related documents, it has still become popularly used &arguably taken on a life of its own. This is not to say we shouldn't use the term in reference to reputation scoring of authors, but if we're talking about the broader context of Google's patents that distinction matters.

One reason for the confusion about what "Author Rank" really is, may be that the Agent Rank patent the concept was inspired from was published, then theorized & written extensively about years before Google+ was built.



A shifting baseline is a type of change to how a system is measured, usually against previous reference points (baselines), which themselves may represent significant changes from an even earlier state of the system. - Wikipedia



A side benefit is we may avoid stepping in Author Rank Paper where ideological debates never cease. ;-)

That said many highly insightful articles have been written referencing the term "Author Rank," and from certain writers many of the practical principles of improving the quality of your author authority are still just as true as they ever were. A favorite author of mine I highly recommend on this topic has been +Mark Traphagen & his extensive analysis of authorship in search results, as well as how to improve the quality & authority of a Profile's Authorship.



For quite sometime now we've been receiving a slew of vague answers & cryptic messages, about when "Author Rank" is coming from Google reps in the know. Some of these "answers" then being interpreted in a variety of confusing ways. Believe me I've looked up & been back through all of the answers as part of this study specifically in light of my current understanding of Agent Rank, to insure that my theories were not contradicted in them.

From this review it appeared quite clear to me, that the reason we kept getting the wrong answers is because we were asking the wrong questions...
  • Is Agent Rank / Author Rank already here?
  • When is Author Rank coming?
  • How will it affect ranking in the Google Search results?
  • How will types of content apply to an author's rank?
  • Must all applicable content be topically related?


Whereas I feel that the answer to most of these questions lies in only one & that we already had the pieces of the puzzle to, but just needed to fit it together:

  • What is Agent Rank?




2. Separate Google Author Credibility Score based on authenticity.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔



Another reason to be cautious about using the term Author Rank, is that Google has a much newer unrelated patent called Reputation Scoring Of An Author, which describes an entirely different type of system that is used for authentication & determining a credibility score for authors which can be included in the author's reputation score for purposes of determining the author's rank.

This Credibility Score may also be closer a type of Trust Rank, than more specifically an author rank. The scoring is primarily based on reviews & the quality of authentication, this was previously used in Goolge's Knol, and is not a link based algorithm at all as is our current Agent Rank.

That being closer to a tiered professional human review system, with increased focus on verification of academic and professional credentials & known association than an algorithmic system that could be easily applied across a wide number of entities. And I'll be happy to further detail this at a later time.



Reputation scoring of an author
United States Patent: 8,645,396
McNally , et al. February 4, 2014
Filed: June 21, 2012



...An authentication score is determined for a contributor of the multiple contributors. The contributor's name and a representation of the contributor's authentication score is published online for display on one or more second computers in association with the online content received from the contributor.

...determining by the computer server system, a reputation score for an author of the online content item, wherein the reputation score is based at least in part on reviews of the online content item that have been provided by one or more authors other than the author;

The method of claim 1, further comprising: determining, by the computer server system, a credibility score for the author of the online content item by obtaining personal information about the author that relates to education or employment of the author and verifying that the received personal information about the author is accurate; and adjusting the ranking of the author based on the determined credibility score.




3. Back to square one for answers in the Agent Rank patent.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔




When I finished writing the Google+ PageRank series a year ago, there were still significant questions left unanswered.


If Google created this entire entity authority distribution system, why haven't we heard of numerous patents & papers referencing it?

How is it possible that something as seemingly significant as this has gone unspoken of & unseen when we have so much detail on most everything else?

Was the authority distribution system we were seeing planned all along, or is it just a happenstance of the PageRank algorithm already active in indexed content?

Matt Cutts frequently states, "PageRank is our opinion of the reputation of a page", so wouldn't the G+ PR on a Profile be Google's reputation score of an author?

What on earth is this Agent Rank patent for, if ten years after its designing & continuing updates thereon they are still not using it?




In part 4 of my series PR in G+ - PageRank is Still The Guts of Google Search, 3/22/13, I wrote...

In a follow up to this, I want to explain more about why I believe PageRank will continue to be the backbone of future algorithms, including AuthorRank & why I "speculate" that ranking Authorship (possibly beginning to be) expressed through G+ PageRanking could be the next evolution of this. If it looks like I've been leading up to this debate it's not an accident, because I believe they are inseparably combined.



I have long suspected that the phenomena of a PageRank type authority in Google+ was intrinsically related to any potential Author Rank, if they were not already one & the same. Unfortunately I never continued writing about what seemed like a strongly opinion saturated subject of debates, instead moving onto writing about more of the analytical details of Google+ SEO.

A few months ago while analyzing a number of anomalies regarding Google+ Author's content in Google Search, I came up with some new ideas on a potential model for ranking authorship content & decided to research the topic further for potential testing.

My search for original corroborating research led me not long ago to read Google's Agent Rank patent for the first time, and what do you suppose I saw there? Well if the goal was to write something as confusing as possible, but still to include all of the important details, I think they did an excellent job.






4. AgentRank+ is the Plus in Google+! The blueprint for G+ authority

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔



The following is what I have learned from Google's Agent Rank Patent As Studied - by Joshua Berg and I would especially like to encourage you to go to the this link where the entire patent is dissected with my detailed notes of my opinions on each of the different functions. It's really not as complicated as first impressions might lead one to believe. With my notes & highlighting there I think it should help you get a good sense of what it is talking about. The main excerpts I have included directly in this article, are still only a fraction of the entire patent, on which I have written many more notes.

In summary though, here are the main opinions I would like to explain about this patent...



① The Agent Rank patent describes in detail a link based algorithm for determining a reputation score of verified agents. This link based reputation scoring algorithm is very much like Google's original PageRank & appears to work together like a partner of PageRank, both corroborating authority of PageRank'ed content in several ways & receiving authority back from PageRank'ed content.


② The phenomena I have previously referred to as Google+ PageRank, in almost every way fits perfectly with the Google's Agent Rank description. So these two as I have thus far described them over the last year, are more or less one & the same thing, with the biggest difference being that I actually now know it's called Agent Rank.

Aside from that, there's a tremendous amount of new information now able to be learned from our much increased understanding of the patent. A lot of questions & theories that we've had about how it works, can now either be confirmed or debunked from the descriptive details in the patent. Also very interesting is to be able to see how it was planned & the potential directions for where it is going.

Here is what I have found to be one of the best descriptions buried several pages within the patent that explains the PageRank similarities & how as a link based algorithm it is applied.




In one implementation, an agent's reputation can be derived using a relative ranking algorithm, e.g., Google's PageRank as set forth in U.S. Pat. No. 6,285,999, based on the content bearing the agent's signature. Using such an algorithm, an agent's reputation can be determined from the extrinsic relationships between agents as well as content. Intuitively, an agent should have a higher reputational score, regardless of the content signed by the agent, if the content signed by the agent is frequently referenced by other agents or content. Not all references, however, are necessarily of equal significance. For example, a reference by another agent with a high reputational score is of greater significance than a reference by another agent with a low reputational score. Thus, the reputation of a particular agent, and therefore the reputational score assigned to the particular agent, should depend not just on the number of references to the content signed by the particular agent, but on the importance of the referring documents and other agents. This implies a recursive definition: the reputation of a particular agent is a function of the reputation of the content and agents which refer to it.

In this manner, the reputation of a particular agent can be calculated by an iterative procedure on a linked database. A linked database (i.e. any database of documents containing mutual citations, such as the world wide web or other hypermedia archive, a dictionary or thesaurus, and a database of academic articles, patents, or court cases) can be represented as a directed graph of N nodes, where each node corresponds to an agent along with all of the content pieces associated with that agent, and where the directed connections between nodes correspond to links from a content piece of one agent to a content piece of another agent. A given node has a set of forward links that connect it to children nodes, and a set of backward links that connect it to parent nodes.


FIG. 3 illustrates a linked database...






③ The Agent Rank patent describes far more potential implementations than will ever be used, and some of them really make me wonder if they didn't just throw as many possible ideas in there they could possibly think of to make the document & hence understanding of its actual implementation that much more mysterious to discover.

A key word in the document though is "implementation," which when running a word cloud on the document, I noticed is close to within the top 10 most common words used. Mostly it is used as a kind of function separator, referring often to "in another implementation," or "in one implementation."

The main point to understand from that, is that it would appear there are a good number of described functions that are not used & will likely never even get used.



④ One of the most common misunderstood aspects that I have seen about Agent Rank, is that it absolutely has to rank authors topically, whereas in Google's patent that is specifically described as optional. To highlight this point, the Agent Rank patent begins with the 14 specific claims (detailed below) described, topical ranking is not even listed as one of them.

This particular implementation may not necessarily currently be functional, at least not in a big way. That said, I have some suggestions on that & here are my notes out of the document.


Particular embodiments implement techniques for computing agent ranks on the basis of a corpus of content signed by those agents, where the corpus optionally contains explicit links among documents and signed content. The agent ranks can optionally [important to note] also be calculated relative to search terms [for this a Topical PageRank could easily be used (TPR is a big topic with lots of available research)] or categories of search terms. [This topical categorization of rank while only optional to Agent rank, is not particularly known to be functional] For example, search terms (or structured collections of search terms, i.e., queries) can be classified into topics, e.g., sports or medical specialties, and an agent can have a different rank with respect to each topic.
[That said, I propose that within Google+, this may already be used to rank topic related Google Plussers for following, such as more personalized suggested users under Find People. And could be expanded to broader Suggested User Lists (SUL) such as the plus.google.com/people/follow .]



⑤ Currently functional all Google+ authors & publishers receive reputation scores that can be seen as the Toolbar PageRank (TBPR) of a profile or page. This is not however viewable until an agent receives sufficient rank to move their PR above 1 and until after which time a TBPR update also must occur. In my opinion it is safe to assume that this visible G+ Profile rank can be considered your visible AgentRank. Except that, we do NOT know this specifically from the patent.

Although as Matt Cutts has repeatedly expressed, "We have very fine-grained notions of PageRank within Google. Outside of Google, PageRank is truncated to 10 levels that are visible in the Google Toolbar." And you can be sure that the actual Agent Rank also has to be very fine-grained, from which we will only see the example.


There are actually a lot more details I could continue on here, but I'll let you review the notes in my
Google's Agent Rank Patent As Studied - by Joshua Berg for more in-depth details.




5. All ⑭ Claims of the Agent Rank patent already active as described.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔



The Agent Rank Patent begins by specifying 14 specific Claims, followed by the Description & then the Summary, both containing far more details. From my thorough dissection of this document, it is my opinion that all 14 Claims of the Agent Rank patent are already functioning & observable as described and have been for quite some time. There are however several optional implementations that probably are not & a few that perceivably never will be.

So I'll explain each of these 14 claims one by one & describe how I believe they currently function:


[Disclaimer: The following are my current opinions from personal analysis & observation. Several of the claims have multiple applications & certain of my specific interpretations can be wrong, or applied differently. In fact I have changed my opinion of several features applying to certain claims several times already as I've studied along & am open to other suggestions. So although I have no doubt about my overall premise of this patent, if there is a devil here, he could be in some of the details.]



① The Google+ system itself, with its built in ability to verify digital signatures with authors, publishers and their content within the system, is the primary & most important claim in the Agent Rank patent.



Patent Claim 1:

A computer-implemented method of determining a reputation score for an agent [in other words the reputation scoring is algorithmic], the method comprising: receiving multiple digital content items in a data processing system comprising one or more computers, [the Google Plus system] each of the received digital content items being associated with a digital signature that uniquely associates the digital content item with a respective agent and an assertion made by the respective agent concerning the digital content item, the respective agent being one of multiple agents, each of the multiple agents being associated with at least one of the received digital content items; validating in the data processing system the associated digital signature for each of the received digital content items, and identifying, based upon the associated digital signature, the agent responsible for making the assertion concerning the received digital content item;...




② When an agent shares a content item from another agent (a link to a page, a share, a reshare, or a mention), that link becomes a positive assertion of that item, because functionally Agent Rank is a link based reputation scoring algorithm. Agent Rank determines and uses the current reputation score of the agent sharing the content to decide its value.Thus authority can be given from agent to agent by sharing their linked content items in at least any one of the following:

  • Your Page is a different agent from your Profile so has its own authority.
  • A Page can share, or link to and receive links from other Profiles or Pages.
  • Another Profile sharing a content item of the first agent.
  • Another Page sharing a content item of the first agent.
  • A Profile or Page sharing a link (+mention) of the first agent.

Patent Claim 2:

The method of claim 1, wherein the digital content item associated by digital signature with the another of the identified agents comprises a link to one of the digital content items associated by digital signature with the first agent, and wherein determining the reputation score of the first agent further comprises identifying the link.




③ The link to the first agent's webpages, or even the agent's Google+ content (digital content items) from the share of another agent, is considered an assertion of reputation for that first agent. Therefore all the links from other agents to the first agent & their verified content, are considered towards the reputation scoring of that agent, based on the reputation score of the referring agent.

Patent Claim 3:

The method of claim 2, wherein the link [described in the previous paragraph] further comprises an assertion by the another agent about the content item associated with the first agent, and wherein the reputation score is a function of the association by the another agent.




④ Web pages may include content pieces by other agents different from the author, for example it may have digitally signed comments by other agents. The publisher of the web page itself with its own digital signature and verification link, is also a separate agent from the first.


Patent Claim 4:

The method of claim 1, wherein the web page includes at least another content piece associated by digital signature with an agent different from the first agent.




⑤ Within Google+ itself, digital signatures are automatically appended to content items. Google also has & is progressing on methods of automatically assigning authorship to some external website pages.

Patent Claim 5:

The method of claim 1, wherein the digital signature that uniquely associates the digital content item with an agent is appended to the digital content item.




⑥ The repository that stores the agent's verified digital signature for verification with all other content items is built into Google+ itself.


Patent Claim 6:

The method of claim 1, wherein the digital signature that uniquely associates the digital content item with an agent is stored in a repository separate from the digital content item.




⑦ Through Authorship external web pages (digital content items) wishing to claim association with an agent, must include a rel=author/publisher (Authorship) link to to the digital signature stored in the repository (the Google+ agent).


Patent Claim 7:

The method of claim 6, wherein the digital content item includes a link to the digital signature.




⑧ The web pages must then have a link pointing to them from within the Google+ Profile Contributor-to authorship system, which verifies that agent's claim to the content. The Google+ Page must also contain a link pointing to the website they wish to be associated with to be verified as the Publisher.


Patent Claim 8:
The method of claim 6, wherein the digital content item is a target of a link associated with the digital signature and stored in the repository separate from the digital content item. [Where G+ profile Contributor-to section associates agent with his content]




⑨ Agent reputation score (Agent Rank) is portable across the wider web through the use of Authorship tags. An agent's verified content therefore, wherever it is published, will be considered towards the reputation score of that agent.


Patent Claim 9:
The method of claim 1, wherein metadata [rel=author tag] associated with at least one of the digital signatures indicates that a reputation score associated with the respective agent is portable across electronic storage locations for the associated digital content item in a network of electronic storage locations. [Google Authorship associated with agent’s rank (G+PR)]




⑩ The Agent Rank (reputation score) of an agent, is already used to determine priority ranking of content within Google+ itself. For example, in search relevance within G+ and exposure of content & agent throughout the platform. Agent Rank has also long been used in determining ranking order of all Google+ content items & agents in Google Search.


Patent Claim 10:

The method of claim 1, further comprising using the reputation score of the first agent to determine an ordering among resources containing the digital content items. [Using the agent’s rank to order its content, possibly within the G+ platform itself.]




⑪ Aside from rel= tags, agent reputation score can also be attributed to an agent through the websites that they claim verified ownership to. Google is also actively improving its ability to algorithmically understand the relationships between authors/publishers & their content besides the use of rel= meta tags. This paints the matter of reputation scoring potentially wider than exclusively link based.


Patent Claim 11:
The method of claim 1, wherein the reputation score of the first agent is also a function of an unsigned digital content item that the first agent is associated with as an owner of an electronic storage location where the unsigned content item is found. [Ownership of a website passes reputation to the agent (possibly rel=publisher passing authority to its verified page.]




⑫ The Google+ system itself (or the built-in functions of) is/are also a specific fulfillment of the Agent Rank patent. Then we go into further details of how an Agent Rank is a function of not only the links from his own author verified content, but from citations in Authorship verified content written exclusively by other agents, acting as a corroborating trust signal from other trusted agents


Patent Claim 12:

A machine-readable memory device [the computer system] comprising instructions operable to cause a data processing apparatus to: receive multiple digital content items, each of the received digital content items being associated with a digital signature that uniquely associates the digital content item with a respective agent and an assertion made by the respective agent concerning the digital content item, the respective agent being one of multiple agents, each of the multiple agents being associated with at least one of the received digital content items; [Within Google+ all the content items from multiple agents (or links from) are received, then associated & validated to each individual agent. ] validate the associated digital signature for each of the received digital content items, and identify, based upon the associated digital signature, the agent responsible for making the assertion concerning the received digital content item; and determine a reputation score for a first agent of the identified agents, [Then a reputation score is determined for each agent (Authors & Publishers).] wherein the reputation score for the first agent is a function of: a) for each of the digital content items associated by digital signature with the first agent and referenced by a digital content item associated with another of the identified agents, a reputation score of the another of the identified agents, [Part of the reputation score for each agent comes from the references to their content by other agents, based on that is based on the referencing agent's reputation score. This includes the citation and sharing of agents content within the Google+ platform.] and b) a reputation score of a second agent associated with a second digital content item that references one of the digital content items associated by digital signature with the first agent, the reputation score of the second agent based on a second assertion made by the second agent that specifies a role of the second agent with respect to the second digital content item; and wherein at least one of a the digital content items associated by digital signature with the first agent is a first subset of a web page, the web page including at least a second subset of the web page that is not associated with the first agent, and the reputation score is not a function of the second subset of the web page. [It would appear that authorship verified webpages of other reputable agents, whom cite the first agent, or their content, provide significant corroborating trust factors applied to the cited agent who is not an author of the content. Of course this includes citation & sharing the content of other agents in the Google+ platform as well. Note there are additional notions that could also be applied here.]




⑬ There are currently several kinds of pages that match the description of signed content from multiple agents.
  1. Every G+ post is an authored page, with signed content assertions by other agents. These assertions can be in the form of a comment, or a plus 1, which are all verified to respective agents.
  2. Every G+ reshare is an authored page with content signed by multiple agents.
  3. A webpage on the wider web with Google+ comments embedded into it, such as Google Blogger pages, all have such functionality.
  4. Google+ Communities are websites with their own reputation score & signed content from multiple agents, where each community sub-topic is a webpage with signed content from multiple agents. Communities are described further.
  5. Lastly the patent specifically describes ad space being content not assigned to agent.


Patent Claim 13:
The machine-readable memory device of claim 12, wherein the web page includes at least another content piece associated by digital signature with an agent different from the first agent.




⑭ The hardware system, also containing the software that algorithmically processes the data to use the reputation scores of the agents to determine order of content resources. While this can refer to the ranking of content & prioritizing resources within Google+ itself, further details in the patent also explain its use in the reputation score ranking of content within Google Search.


Patent Claim 14:
The machine-readable memory device of claim 12, further comprising instructions operable to cause the data processing apparatus to use the reputation score of the first agent to determine an ordering among resources containing the digital content items. [The ordering of content (at least) within Google+ itself will be based on the agent’s reputation score.]



Thus all 14 claims in the Agent Rank patent are already observably functioning as described.




6. The Million $ question: Does AR already affect ranking in Search?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔



⓵ Authors (Google+ Profiles themselves) with high reputation scores already rank high in Google Search this is already a strong effect of Agent Rank.

⓶ Publishers (Google+ Pages) with high reputation scores also rank in Search results.

⓷ Google+ Posts by Authors & Publishers with high reputation scores also rank higher in Search than those with very low scores.

⓸ Google+ Reshares by Authors & Publishers with high reputation scores also rank higher in Search, frequently outranking the original post, at least earlier on (sometimes balanced with unengaged posts by what I refer to as the Google+ Stream Effect).

⓹ Google+ Communities all have their own reputation scores which affect their ranking in Search & their scores are also a function of the reputation of the agents which they are associated with. There are more details in the Agent Rank patent related to this.

⓺ Google+ Community Subtopics inherit their reputation scores from the main community, but have separate (usually lower) scores based on the activity types in that subtopic. Each community subtopic is ranked individually in Search that is affected by the reputation they retain.

⓻ Google+ Local Pages have their own reputation scores, which I'll strongly suggest affects their local ranking more so than any review, or +1 type factors, but in my opinion it is probable that a Localized PageRank algorithm is used (which is another topic altogether).

⓼ YouTube Channels & Videos have reputation scores & you can be pretty sure these all have an affect on their rankings in Search. Recently YouTube comments became integrated with Google+, and the reasons for this are quite clear, as well as the direction this will go. Comments by high reputation agents already rank higher here & of course their association with the content will have its functional affect.

⓽ In-depth Articles is sort of a bonus idea we might include, courtesy of a Tweet from +Matt Cutts to our friend +Mark Traphagen, but for want of space in this episode I am not going to get into my specific theories on how that might work.




Webpages⁈  In my opinion, I have already seen an observable model for how Agent Rank can be used to affect the ranking of webpages with reputable Authorship content. The concept is quite simple and perceivably not difficult to test, but you'll have to look forward to the next part of this AgentRank+ series for answers on that.



0 comments:

Post a Comment