[Build Backlinks Online] The Panda Patent: Brand Mentions Are the Future of Link Building

Build Backlinks Online has posted a new item, 'The Panda Patent: Brand Mentions
Are the Future of Link Building'

Posted by SimonPenson


There has long been speculation about how Google actually measures "brand
authority." Many times over the past couple of years have those who speak
outside of those fortified Googleplex walls made it clear that brand building is
the way to win the organic visibility war.


That statement however has always sounded wooly in the extreme. How is it
possible to measure an almost intangible thing at scale and via a complex
formula? If you are Google, it seemsthere is ALWAYS a way.


A fairly innocent-looking patent filed last month, which some say could be
the Panda patentmay have gone some way to answering that question.


Within this post we dive into that patent and other supporting evidence in an
attempt tounderstand what the opportunity may be for digital marketers in the
future. As part of that attempt, we offer our interpretation of various pieces
of the patent itself, andalsolook at actual data to see if mentions are already
playing a part in the ranking of sites.


The patent in question, which can be
found here and has been expertly covered by Bill Slawski, may cover the Panda
Algorithm's key workings, but the piece we are really interested in right now is
the information around measuring site authority and relevance using a ratio of
links and mentions, or "implied links."It's this specific area that got both the
team here at Zazzle Media and also at Moz excited. You can see Cyrus Shepard and
Rand Fishkin's reaction right here:

I knew it! RT @CyrusShepard: This Google patent defines non-linking citations as
"implied links" http://t.co/cOxv0irklk Rand Fishkin (@randfish) March 26, 2014

So, what exactly does the patent imply? It is complex, wordy, and difficult to
interpret, but it starts by talking about what Google calls "reference queries:"


"A reference query for a particular group of resources is a previously
submitted search query that has been categorized as referring to a resource in
the particular group of resources."


To most of us, that statement appears as bad English at best, but there are a
couple of ways this could be used. It firstallows Google to look at what terms
people have used previously through search to find and then click on a site, or
group of sites. In doing so, it wouldthen also allow them to "map" semantically
relevant queries (and thusmentions of abrand) to a site in order to further
extend an understanding of the "popularity" or "authority" ofthat entity.


The patent also covers a mechanism for allowing Google to discount some links
and give others greater weighting based on a modification factor:


"The method of claim 1, wherein determining a respective group-specific
modification factor for a particular group of resources comprises: determining
an initial modification factor for the particular group of resources, wherein
the initial modification factor is a ratio of a number of independent links
counted for the particular group to the number of reference queries counted for
the particular group."


This could beespecially important where lots of links from the same company (or
"group") point at a site, as the search engine could discount those from the
true overall picture. It then also gives engineers the ability to look at
"quality" as a measure of the overall relevance to the queriedsubject matter,
which is called out in a separate bit of the patent:



"For example, the initial score can be, e.g., a measure of the relevance of
the resource to the search query, a measure of the quality of the resource, or
both."



Then, the patent specifically mentions that links can either be "express" or
"implied," calling out non-linkingmentions in a rather unmistakable way:


"The system determines a count of independent links for the group (step 302). A
link for a group of resources is an incoming link to a resource in the group,
i.e., a link having a resource in the group as its target. Links for the group
can include express links, implied links, or both. [...]An implied link is a
reference to a target resource, e.g., a citation to the target resource, which
is included in a source resource but is not an express link to the target
resource. Thus, a resource in the group can be the target of an implied link
without a user being able to navigate to the resource by following the implied
link.


What does all this mean? It means that once a connection is made by someone
typing in a brand name or other search query and then clicking on a site it
creates aconnection inGoogle's eyes. The search engine can then store that info
and use itin the context of unlinked mentions around the web in orderto help
weight rankings of particular sites.





If this is
the Panda Patent, as it is part of a wider algorithm, it wouldalso look at the
quality of pages on a site and how "commercial" they are in their targetingthis
wasoriginallydesigned to negatively impact content farms that created content
targeted aggressively at commercial terms those thatthink "search engine-first"
as opposed to "audience-first."


The patent publication was closely followed by a
related Webmaster video by Matt Cutts within which the head of Google's webspam
teamtalked about a "forthcoming update" that would look differently at how the
search engine measures authority:






By arguing that there is a difference between popularity and links, Cutts made
clear that his engineers are looking very closely at how to continue to tweak
the existing algo to make more popular sites rank higher.


That's a big deal.


Those two pieces of new evidence suggest there is a seismic shift underway in
how links are weighted and how relevance is measured the two building blocks of
search.


It's something
Rand here at Moz first touched on back in late 2012 and I covered in detail in
this post a few weeks later.


In it I gave a little background into why Google is hell bent on getting away
from the concept of a link-based economy:



"My view is that Google is really trying to clear up the link graph and with
it valueless links so that it can clearly understand relevance and associations
again.


It's something that web 'creator' Tim Berners Lee first wrote about back in
2006 in this magazine article and Google has been talking about ever since,
ploughing lots of cash into acquisitions to get it there.


So, why invest so much time, effort and resource into such a game-changing
project? Quite simply because its existing keyword based info retrieval model is
in danger of being bettered by semantic search engines."



Let's look into the how and why in a little more detail.




Links




Everybody who reads this article will be more than aware of the importance of
links in building authority.


A lot has changed in this space over recent years, however, as Google has
developed systems to measure where the link is coming from and assign more (or
less) value to it as a result.


The next logical step in that process could be the downgrading of the follow
link within the overall picture of ranking factors. It is something that has
been expected for some time and wasreiterated by Moz's own panel of experts in
the annual
Ranking Factor survey.


We know that follow links have been gamed "to death" in the past, so it would
make sense to make that particular element a little less important in the
overall mix. Nofollows can still tell Google a lot about a site, as can how much
people are talking about it.

Nofollow




By definition a nofollow link does not pass equity, or PageRank, to a site. We
know that for certain. What is less clear is what, if anything, it does pass.


Google certainly knows these links are there, and this latest patent could
suggest they are taking a little more notice of them than they are letting on.


The patent highlights the importance of "reference queries" and "implied
links," and also that Google looks to discount links from the same "group" or
brand and instead wants to concentrate on independent links from unassociated
domains.


Critically, PageRank is not mentioned, which suggests that other factors are
being measured in terms of how much equity is being passed by each independent
link, or mention.

Mentions

It is the "implied link"element that makes most interesting reading, as it is
black-and-white evidence that Google is looking at mentionsas a measure of
authority.


It is logical, after all, that a popular brand would have more people talking
about it online than one that is simply good at manipulating the algorithm and
has invested heavily in link building to that end.


The results from our sample research also support this, with larger,
better-known brands generally attracting greater numbers of mentions than
others.

Testing




Of course, the question remains, is this already in use? To test this properly
would require a monumental amount of measurement across a plethora of verticals
over an extended period of time, and sadly,I did not have the time or resources
to run that project for the sake of this post, but it is still worth sharing
some of the data and an overview of what may be going on.


The caveat here is absolutely that this does NOT constitute any kind of
fact-finding mission, simply an informed commentary on a few anomaliesthat
cannot be explained simply by looking at follow links alone.


To discover if there are any initial signs that this kind of system may already
be in effect, I spent some time analyzing three separate, random SERPs here in
the UK.


They were:

"Car Insurance"
"Mortgage Calculator"
"Mens Clothing"

All three are competitive terms and are "owned" in the main by what we might
know as brandsin the wider world.


Below you can see a simple chart for each of these, showing:

Follow links
Nofollow links
Follow/nofollow ratio
The number of brand mentions in the last four weeks
Ratio between links and mentions

Clearly this isn't ascientific study, but it does serve as a"finger in the
air"analysis from which a few interesting observations can be made.


The first two tables contain data examining the whole domain's link profile,
while the third looks specifically at links to the URL indexed for the
particular term we are analyzing.


The raw data is below and here is an explanation of where that data was drawn
from:

Position Records what position we saw the domain in for the given search term
in google.co.uk.
Follow link for the first two ('Car Insurance' and 'Mens Clothing') this is
the number of follow links across the whole domain. For 'mortgage calculator' it
records just the follow links into the specific URL indexing for that term. The
data is from AHREFS.
As above but for no follow links.
The number of referring domains into the domain ('Car Insurance' and 'Mens
Clothing') and the URL ('Mortgage Calculator').
Mentions How many mentions of the brand there have been in the last four weeks
(as taken from Moz Fresh Web Explorer and using exact match brand term only).
Ratio of follow to no follow links Designed to see if there is a correlation
with this and position.
Ratio of links to mentions A look at the relationship between how many links a
site has and how many mentions in the previous four weeks.

Let's now look at each table in a little more detail but with the understanding
that there are a myriad of other factors that affect the result. After each
piece of analysis I have added general comments:

"Car insurance"


Despite having considerably fewerlinks overall across the domain, Go Compare is
first. Does this suggest they have hit a sweet spot in terms of links versus
mentions and brand metrics?
Money Supermarket proves that more links doesn't win. The site has many, many
more than anyone else in the top five and yet is not first. Link volume
clearlymatters, but it is absolutely not the only factor at play.
Do Compare the Market and Confused "win" and feature in the top five on the
strength of their mention data? A higher ratio than the top three but MUCH lower
link numbers suggests that might be the case.
LV.com is the anomaly as its mention to link ratio is low. This suggests that
sheer numbers of links are potentiallyhelping it rank well as well ason-page
factors that make the brand super-relevant for car insurance.
"Mens clothing"


ASOS and House of Fraser walk away with it here, and interestingly both of these
sites have very similar ratios for both follow and no-follow linksAND forlinks
and mentions.
Next is an anomaly but only because measuring true mentions is very hard due to
the brand having a "generic" word as its "brand." Could this cause issues for
Google going forward in measuring similar brands?
Again,Burton (in particular) and Topman to a degree showthat you can earn your
place with high link-to-mention ratios.
Topman should rank higher. Is it because of link quality? Certainly the ratio
of follow to nofollow is lower than those above them.



"Mortgage calculator"








This batch looks just at the URL ranking for the term, not the whole domain, to
allow us to see both sides of the story.
The BBC clearly runs away with it in every sense and by these figures would be
almost impossible to usurp.
The data is less correlated here; suggesting that domain-wide factors are
definitely at play in factoring which URL should rank where.
NatWest is a really interesting result as it has very few links relative to
those around it. The ratio of no-follow-to-follow is very high, and the brand
gets a lot of mentions. The site is also very relevant for "mortgages" as a
percentage of overall content.

As a further piece of analysis we have also included a "random" site from page
two to support the concept that a combination/ratio of links to mentions affects
rankings.


That site is woolwich.co.uk, a small UK building society (bank). Its own
mortgage calculator page was ranking 16thwhen we ran the analysis, and
interestingly, we can see that its ratio of follow-to-nofollow is low, as is the
number of mentions of the brand in the past four weeks. On pure followed link
numbers the site should rank top five, but instead it languishes on page two.

Is there a perfect ratio?

It's clear that the small sample above is no true reflection of how the ratio
of linkstomentions affects rankings, but it has certainly raised some
interesting points for further discussion and testing.


What certainly can be said is that the measurement of brand mentions is
certainly possible, and Google certainly nowhas the patent to cover it off as a
potential ranking factor.


Mentions alone do not tell the whole story, of course, and links are still very
important in the overall mix of factors that affect rankings, but it is now time
to start thinking about how you can create brand buzz and grow those mentions.


The possible good news for the smaller guys, though, is that it does seem that
Panda is beginning to look at how much of a site is relevant to any one specific
term and giving over extra "authority" to that site in that niche.


A great example arethe building societies in the UK that rely on mortgages to
make up the majority of their business. As a result they do seem to rank better
for mortgage-related terms.


This could really help specialist businesses compete again with the "big guys."

What can you do?

The key point here is how you can utilize this data to improve your own
strategy, and while there is no conclusive proof that mentions and nofollow
links matter, the evidence is starting topile up.


Given what Mr. Cutts said in the video we mention earlier in the post, the
patent for Panda and increasing amounts of data highlighting similar findings,
the argument is certainly there.


My advice would be to begin thinking outside of follow links. Be happy to earn
(and build) nofollowed links and mentions. Think outside the link, because there
IS value in driving mentions and building brand.


How do you do that? The simple answer is to get creative with your
communications strategy and build content that will make people talk about you
and share. At the heart of that is a great ideation process, and I have
previously shared Zazzle Media's own way of creating great ideas consistently
here.


Think particularly around creating content that plays on core emotions and also
give things away. In a detailed report first shared in 2010, titled "
Social Transmission, Emotion, and the Virality of Online Content," the authors
discovered there is a strong relationship between emotion and how likely it is
your content will "go viral."


Amongst the many eye-opening discoveries the publication discusses:

Negative content tends to be less viral than positive content
Awe-inspiring content and content that surprises or is humorous is more likely
to be shared
Content that causes sadness can become viral but is generally less likely to
Content that evokes anger is morelikely to be shared

And finally, make sure you start tracking and reporting mentions and nofollowed
links. Knowing how you are performing can help you iterate your PR and content
strategies to achieve greater traction.


In short, compelling content, created over the long term WILL now win, as it is
being rewarded.
Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten
hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think
of it as your exclusive digest of stuff you don't have time to hunt down but
want to read!

You may view the latest post at
http://feedproxy.google.com/~r/seomoz/~3/AwjQgL6r8oQ/panda-patent-brand-mentions

You received this e-mail because you asked to be notified when new updates are
posted.
Best regards,
Build Backlinks Online
peter.clarke@designed-for-success.com