The Semantic Web & the Evolution of Search

Jul. 31, 2012 | by Allyson.James

An Introduction…

The internet is a vast expanse of data. Each web page is a document marked up in HTML so that search engines, browsers and other programs can display the data to a human. It is unstructured and linked together to form a huge web of information and most definitely can be defined as ‘Big Data’.

When Tim Berners-Lee invented the World-Wide Web, its aim was to facilitate communication not just between humans but also allow the participation of machines. He first coined the term ‘Semantic Web’ in a paper written in 2001 when looking at whether his goal had been realized.  He stated:

“A major obstacle to this goal is the fact that most information on the Web is designed solely for human consumption. Computers are better at handling carefully structured and well-designed data, yet even where information is derived from a database with well-defined meanings, the implications of those data are not evident to a robot browsing the web.”

And so the semantic web came about. It is a group led by the W3C to add structure to data on the web in the form of semantics, i.e. the meaning and relation of words together. This, in essence, means categorising data in classes. So Meat Loaf would no longer look like two words to machines, but instead could be marked up as a dish, or a pop artist.

The evolution of the web to become semantically linked will have great benefit to search. As machines will be able to understand the context of data, search engines will be able to return more accurate results to humans.

How are semantic principles currently being used in search?

Google Knowledge Graph

Perhaps the highest profile usage of semantic technology was unveiled by Google in May this year. Using databases of information already available to them, such as Wikipedia, Freebase and Google Local, they created a graph containing more than 500 million objects and 3.5 billion facts with links between the objects representing the relationships between them. In effect, they have created their own semantic web of information which they can now query and return semantically correct results.

The below screenshots shows the knowledge graph in action. Performing a search on ‘Meat Loaf’ you are presented with organic results mainly focusing on the dish. However, on the right hand side, the knowledge graph has picked up that ‘Meat Loaf’ is also a singer and offers alternative search results based on this knowledge.

Google Search Results Meatloaf

When you look at the singer’s results, the graph’s relationship features really come into play.

Meatloaf Google Knowledge Graph Results

In the screenshot above you can see the information pulled out of the knowledge graph, featured on the right hand side. Including a short biography snippet from Wikipedia, song titles, upcoming shows and other similar artists who have been linked to Meat Loaf in the graph, this feature gives searchers looking for quick information on the singer what they need without having to click through the organic search results.

Schema.org

Schema.org is an extremely interesting joint venture between all the major search engines. Whilst not following the Semantic Web group’s usage of RDF technology to the letter, the specification provided by Schema.org of HTML used to nest semantics within existing content of web pages has proved extremely popular. The different schemas include Book, Movie, Recipe, Review and CreativeWork which leads to markups like ‘Author’.

Google Author Rich Snippet

 

 

 

 

Once content is marked up using Schema.org, Google and other participating search engines can use this semantically linked data in features like rich snippets (as shown above) to provide more prominent positioning within search engine results, leading to a very good reason for taking the time to mark a site’s content up semantically.

Companies adopting semantic technology

Whilst a lot of the Semantic Web’s development has been in academic and research fields, some companies, such as Volkswagen UK, have adopted semantic technology to enhance their own websites.

Volkswagen have used Ontologies (sets of objects and relationships between objects) defined by the Semantic Web group and also defined some of their own, more specific ontologies to describe their cars and specifications.

By marking up all their product pages using these vocabularies, their on-site search is semantic and is able to return more precise results to consumers.

How will the advent of the semantic web effect search engine optimisation?

A large part of SEO has always been based around keyword searches and optimising a web page to be as relevant to that keyword as possible in order that a search engine deems it relevant enough to display in its search results.

With the shift towards semantic search and search engines using natural language processing to understand the meaning of a user’s search, SEO practitioners can no longer simply rely on using keywords to gain rankings within their desired search results. The context of these keywords is now significant and content must be relevant.

But this is nothing new. Content has become a huge focus for search engines not just because of semantic search and cannot be bypassed. The advent of the ‘Knowledge Graph’ however, could be troublesome for those in the optimisation industry. With their vast information databases linked together semantically, Google is moving towards not becoming a gateway to the internet and a user’s search, but a destination. This means that a new competitor for search engine clicks will be Google themselves. Even if they continue to include links to other websites returned from a search query, they will of course hold the prestigious top space on a results page, and many searchers may never click on a link to answer their query.

One problem to this approach which has been highlighted is the difficulty in keeping such a huge array of information up to date. This may lead to information trust issues. But this could be solved by the release of the ontology and markup that Google are defining and using and the expansion of Google’s databases via other websites marking up their own content for inclusion within the Knowledge Graph. This is much like how schema.org, on a very small scale, is allowing for rich snippets to be displayed within the current organic results.

Whatever the plan for the future, this will be a long time coming. Search engine optimisation may shift focus away from keywords to semantic markup. But as an ever-changing industry, adapting to this evolution will not be something the SEO world will be afraid of.

Be Sociable, Share!

    Comments (3)

    • ebinx

      The big worry is what you mentioned in your article- i.e. Google is shifting form a middle man to the destination for more and more searches.Oct 13, 2012 05:59 pm

    • Introduction to SW | Semantic Web & Databases in Information Technology

      [...] The Semantic Web & the Evolution of Search [...]Sep 5, 2012 07:43 am

    • SIGIR’12: State of Semantic Search | ReStreaming

      [...] The Semantic Web & the Evolution of Search (connect.icrossing.co.uk) Share this:Like this:LikeBe the first to like this. This entry was posted in Uncategorized by Dušan Omerčević. Bookmark the permalink. [...]Aug 17, 2012 08:03 am

     
    Please note: the opinions expressed in this post represent the views of the individual, not necessarily those of iCrossing.

    Post a comment