What is semantic search? How do I do that?!?

Posted 25-07-2013 by Alexandra

Have you heard lots of chatter about how search engine optimisation is changing? That search engines are upping the stakes, moving the goal posts or just generally trying to defeat SEOs? Don’t listen! There is one thing that Google and search engines have always and will always aim to do: Give the searcher the best possible experience through the websites that they show in their results. What has changed as new technologies emerge is how search engines determine what the best possible experience is and how they can identify the relevant characteristics in web pages in order to deliver on their primary function – enabling you to find what you are looking for. This is where semantic search comes in…

First things first; the Semantic Web

Before you can understand semantic search you need to have an understanding of the semantic web. Manu Sporny provides an introduction to the semantic web in this video. To paraphrase his explanation the semantic web gives computers context behind the content on a web page and the relationships that content has to other information on the web. This is through additional code being added to a webpage that will help computers to better understand exactly what the page is about and the context in which the words on the page should be understood. This code is known as microdata or RDFa. The best example I can give of how this is more useful than basic keyword search was used by Aaron Bradley in the 3rd section of his blog post on semantic search. If you were to search for “Paris Hilton” should you be returned results about the heiress or hotels in Paris? Based purely on traditional keyword search search engine computers would only be able to deliver results which include the words “Paris Hilton”. Unable to differentiate between the person and the hotel, you would see a mixture of the two results, and the order would be based on other ranking factors. Using microdata or RDFa web developers can indicate to search engines that the content on the page is about a person or about a hotel. This will enable search engines to better group and deliver the information you are looking for.

So, Semantic Search

Semantic search is where, through the semantic web, search engines can understand the context of and intent behind a search term. For example, if you search for “red soled shoes” your intent could be to look at pictures of red soled shoes, find information on Christian Louboutin (the famous designer of red soled shoes), see the shoes a celebrity was wearing in a magazine yesterday, or buy red soled/Christian Louboutin shoes.

How can computers understand my intent behind a search term?

Crystal brian graphic — Crystal Brain by Nevit Dilmen http://commons.wikimedia.org/wiki/File%3ACrystal_mind.jpg

We know that through the semantic web and microdata search engine computers can better understand the context behind the content that it has indexed and may return in the search results. But that does not explain how search engines can determine what it is that you, the searcher, wants to know, i.e. your intent. According to Tony John in his blog post ‘What is Semantic Search?’ search engines can go one or two steps further than the semantic web to guess the intent of your search in order to give you the best possible results. This includes:

Current trends or news

Was a celebrity spotted wearing Christian Louboutin red soled shoes recently?

The searcher’s location

Am I near a Christian Louboutin store? Am I looking for a UK website?

Word variations/plurals

As shoes are generally bought in pairs and are referred to as plural “red soled shoes” will generate more purchasing oriented results than “red soled shoe” which suggests you are not looking to buy

Appropriate synonyms

“Red bottom shoes” generates the same results as “red soled shoes”

Generalisation and specificity

As well as listing the specific Christian Louboutin website the ebay results for red sole shoes is listed

Natural language queries

“buy red soled shoes” gives the same results as “purchase red soled shoes”

The combination of words and how that changes the overall meaning of the search

“Real red soled shoes” gives different results to “designer red soled shoes” Are you with me? Or are you day dreaming about a gorgeous pair of Christian Louboutins? Here, do both:

Picture of shoes — Image From Christian Louboutin Website: http://eu.christianlouboutin.com/uk_en/shop/women/lady-peep-sling-patent-8.html

Obviously semantic search is not foolproof, there is no way a computer (or a person for that matter) could determine exactly what you want to know based on a generic 3 word search phrase, but it can start to make a better guess than has previously been possible.

Semantic Search and SEO – What to do.

In his post how semantic search will affect SEO in 2013 Joe Bradley made some excellent recommendations for strictly following SEO best practice. This is undoubtedly the first place to start and I recommend taking a look at Joe’s post to find out more about the fundamentals of best practice SEO for future proofing your website’s performance.

Microdata

The next stage is to join the world of the semantic web. Whilst this is not vital for SEO at the moment the focus on the semantic web is only going to increase in search engines’ priorities so being an early adopter could make a huge difference to your SEO campaign. To find out more about implementing microdata on your website visit http://schema.org/docs/gs.html and http://dev.w3.org/html5/md-LC.

Content Formats

This is great for creative content marketing as through the use of microformats in semantic search the onus on written content is slightly reduced. Properly formatted files, such as images and videos, can be entities that have the potential to rank in the search results, if Google determines that file type to be the possible intent of the searcher. This greatly increases the ways that your brand and website can rank in the Universal Search results in Google. With semantic search if someone searched for your brand name, for example, the results could include:

Your website
Your logo
Pictures of your office/your staff
A brand promotional video
Other websites that you are listed on

This can be seen already with Google Knowledge Graph which aggregates different data formats into a single ‘fact sheet’ about the search term.

What Next?

Whilst being an early adopter is good more speed and less haste is definitely the message here. The first thing to do is make sure that your web developers are up to speed with html5 coding and then allow them to spend time studying http://schema.org/docs/gs.html and http://dev.w3.org/html5/md-LC. Test! Do not put microdata on every page of your site in one fell swoop! Try it out on different pages on your site and give a reasonable amount of time to test the impact before extending implementation across the site. Good luck entering the realms of the semantic web. Let us know how you get on with a comment below! Follow my content and social media contributions to the blog to find out more about the importance of website content strategy, or sign up to the ThoughtShift Guest List, our monthly email, to keep up-to-date on all our blogposts, guides and events.