Every day it seems a part of SEO is dying or a new method is the next sliced bread. It’s not often something comes around that actually seems to ring true but semantic web search seems like it could be just that. Imagine a search engine that will provide answers to search queries by verifying data which links through to trusted documents and sources. In this world keyword and synonym placement would be obsolete thanks to semantic search technology.
But how would you optimize for something so complex?! Simply put, in the future SEOs will have to create on page entities that answer specific and targeted queries that can then be picked up and indexed by search engines, or ‘answer engines.’ Plus, when we say ‘the future’ it’s not as far ahead as you may think, as proven by Hummingbird.
But What Really Makes Up An Entity?
You know the Google Knowledge graph that we wrote about not too long ago, that perfectly summarises entity search. Each person or place that appears in the Knowledge Graph is an entity. Albert Einstein is an entity as is the Eiffel Tower, they both have verified values and information attached to them, because of this Google knows the exact information to return when you want to find out how high the Eiffel Tower is. These entities are created through the use of structured data. Not too long ago Google released their structured data highlighter that allowed any webmaster, even with little to no knowledge of coding, to mark up their onsite data. This helps search engines workout what a product looks like on your site as opposed to an article and therefore helps them categorise it accordingly.
However, semantic web search goes one step further than just differentiating between products and articles. Using something called triplestores, semantic search utilises a vocabulary similar to that of Facebook graph search to read structured data to answer searcher intent. This is the step that bridges the gap between search engines simply reading and indexing information to match it to key terms to them actually understanding that data and searcher intent.
What On Earth Is A Triplestore?! And How Does It Help Semantic Search Technology?
A triplestore is a massive database that stores triples. But what are triples? Well, you and I use triples all the time. It’s a basic sentence structure that forms the backbone of most speech, the subject, predicate and object triple. For example, John is chasing the ball is a prime example of a triple. John and a ball are both entities and John’s actions link the two and form the relationship. So in search terms you could think of a triple as ‘red high tops near Slough’ or ‘Chinese restaurant in London.’
You will already have fairly accurate results returned for these terms thanks to mobile and voice search but they are only really effective in location based queries. In the future ‘answer engines’ will have access to triplestores that will store billions upon billions of possible triples which will contain very specific data. When you search they will access these triplestores to internally verify data and relationships using semantic search technology before returning the most relevant results.
How To Prepare For The Future Of Semantic Web Search
You know that structured data we mentioned earlier? You’re going to want to start marking everything up, and we really do mean EVERYTHING. The Google data highlighter is a good place to start by all means but it isn’t as all encompassing as Schema markup. Checkout the Schema.org site for more information on the insanely detailed levels you can go into using it. Using structured data/semantic markup will not just future proof your site but make it even more attractive to search engines.
We all know that search engines love it when you make their job easier and that’s exactly what structured data does. When indexing a page marked up with structured data search engines know exactly what that information means. An added bonus is the fact that the structured data’s appearance in SERPs is proven to improve click through rate, simply because it is more attractive.
Entity Driven Image Attribution: june29
Neuron Attribution: MikeBlogs
Written by Ryan Hill