Best of 2013: No 13 – Search in the Knowledge Graph era #bestof2013stod
Estimated reading time: 7 minutes, 40 seconds
During the holiday weeks we will be showing you the 15 best read posts of 2013. Except for on Christmas and New Years day, each day you can read the best articles again, going from number 15 back to number 1.
Now it’s time for number 13, originally posted on January 10 2013, a post by Gianluca Fiorelli.
One of the most interesting news around Google Search last year was the introduction of Knowledge Graph, in May for English queries and in December 2012 for many other languages. Now the Knowledge Graph “covers 570 million entities, 18 billion facts and connections, and about three times as many queries globally as when we first launched it”, as Aaron Brown proudly wrote. To refresh how Knowledge Graph works, nothing is better than (again) watching the video a which a few Googlers present it:
Objects connected, not simple keywords/pages connections. Those objects are defined as “entities” in the post I linked up above. And entities are something Google has loved for a long time already and, if we are SEOs who pay attention to the evolution of Google as a Search Service and not just to how its algorithm works so to game it, we should know how Google is now more an Entity Search Engine, and not just an information retrieval search engine based mostly on the relative importance given to links. Google is working on the concept of semantic search ever since the acquisition of Applied Semantics (aka: Adsense) in 2003. If we read how Oingo (the previous name of Applied Semantics) was defining its technology, we can see this (kudos to Enrico Altavilla for making me noticing it):
Oingo Meaning-Based Search is powered by the Oingo Ontology, a highly detailed database of over 1,000,000 words and meanings, linked by millions of relationships in a semantic network that is constantly changing to reflect the currency of everyday language.
This is a terminology that reminds us about the one used by Google when presenting the Knowledge Graph, with the difference that Freebase and Metaweb, acquired by Google in 2010 and on which Knowledge Graph is based, are much more complex and evolved.
Google does not simply have all the information it is crawling everyday and that it takes from Freebase, Wikipedia (via DBPedia), CIA The World Factbook and many other open and linked data sources, but also the information it can retrieve from the volume of queries done by the users and the links those users have clicked on the SERPs presented for those queries. If you paid attention to the video, you have heard Johanna Wright saying this:
We are in the early phases of moving from being an information engine to becoming a knowledge engine.
Somehow it is a declaration of intent that reminds us about what Vic Gundotra said at the last SMX Social Media Marketing Expo:
I think people are just now starting to understand that Google+ is Google. At some point, we’re gonna have a billion users. And eventually, you’ll just say: ‘Google has a billion users.
It is not my intention to discuss if Knowledge Graph is really applying semantics or if it is more an über complex database, which crosses all the relevant data that Google already has so to offer those answers, which may be statistically more relevant to its users; people like Bill Slawski or (again) Enrico Altavilla (check this post in italian and here in its gTranslated version in English) are more qualified than me to offer some possible answers. But as a web marketer it is my duty to investigate and see if Knowledge Graph can represent a competitive advantage for my clients, and my best tool – once studied posts of people better than me – is observing the SERPs.
When doing a search on Google for “The Hobbit”, you will see the Knowledge Graph box on the right. It presents information from Wikipedia but also – and this is something relatively new – the “Recent Posts” voice, which links to the Google Plus Brand Page of the movie. This connection between Knowledge Graph and Google Plus is surely interesting and may lead to experiments and speculations, because we will see how “The Hobbit” case is not the only one. Then we see an alternative Knowledge Graph result, “The Hobbit” book. So, let’s click on the book link.
In this case, and as usually it is in Knowledge Graph, we can see how all the links are sent to a sort of pre-packaged SERPs. Those SERPs, hence – and if we are in the publishing niche – are the ones we should aim to see our site ranking in. But the most interesting thing is that main image of the box is taken from Google Books and links to “The Hobbit” page on Google Books. If we click on that image we arrive here:
The page is personalized – I was logged in when doing the search – which is why the links to the eCommerce selling “The Hobbit” book are Spanish. But that Google Books page offers also the opportunity to find a bookshop close to where I live and reviews. Question: how many of you use Google Books to extend the opportunity of reaching your target? Ah!… Google Books authors (even if they publish just eBooks) are into the Authorrank if they use the rel=”author”…
Knowledge Graph, though, is even more useful for brainstorming content marketing ideas. Remember, it relies – a lot – over the volume of queries done by the users about certain topics, so it can be considered a great complement to Google Suggest.
In the case of “The Dark Knight Rises”, Google also offers us a good amount of images, which all link to Google Images SERPs. I clicked on one of those images and found this:
That post from Moustache Magazine is not even ranking in the Top 50 (neither in the Images vertical), but thanks to Knowledge Graph it is in the first page and with some possibility of being reached by a user like me.
Before I was saying that Knowledge Graph and Google Plus seems very well-connected, which is somewhat logical if you think about it: brand pages in Google Plus are an entity sign, and so are the Google Plus profiles.
Searching for Madonna, Google presents the Knowledge Graph of the pop star (also in google.it, where, as a curiosity, should be presented the option of Madonna as the Virgin Mary).
Also in her case we see the “Recent Posts”, which are nothing but the Madonna’s Google Plus updates.
But how extended is the Knowledge Graph when it comes to people? Are only very well known artists, writers, politics and historical figures considered right now?
No. Also people quite normal (even if they are stars for us) like Danny Sullivan and Rand Fishkin are nodes in the Knowledge Graph:
Danny Sullivan’s bio is taken from Wikipedia, but what is especially visible is, again, his Google Plus latest update. More over, you can see how Google reminds me that I have Danny in three circles on Google Plus. In fact, if I am not logged in, the Knowledge Graph box of Danny is not presented to me (while the Madonna one still it is). This seems indicating how personalization is playing a role in the Knowledge Graph as well.
Quite similar is the case of Rand Fishkin:
I said “quite similar”, because in his case Google is using a snippet from the SEOmoz page on Wikipedia where Rand is cited as the co-founder of SEOmoz, as he has no page on Wikipedia. I know that Rand does not pay much interest in that page, but its use by Knowledge Graph should make him change his opinion.
What I have learnt is that Knowledge Graph should be taken into consideration by web marketers and SEOs, even though it doesn’t present a direct linking to other sites than Google, Wikipedia and the various sources used by Google. Why? Because it presents a great opportunity for Branding (Brands too are present as nodes in the Knowledge Graph), and because we can “influence” some of the sources Knowledge Graph use: Wikipedia and Freebase above all. And we can try to influence also the same query volumes, so important in the Knowledge Graph, creating actions of predictive marketing, which relies over the correct perception of the incoming trends in search. I’ve learnt that image SEO is not just an option, and the same should be said about video, in Knowledge Graph. That structured data is used by Google (event, for instance concerts’ dates are presented in KG), that using Google products like Google Books is something I should consider more seriously and not just as an “enemy” for my eCommerce. I’ve learnt the personalization is the real big challenge we as marketers have to deal with more and more now and in the future. And that Knowledge Graph and Google Plus are strongly connected, and both, possibly, are to Authorrank. These are just few things I have observed playing with Knowledge Graph searches during an hour. Surely many others can be “discovered” with more time and you surely can suggest some of them in the comments.