[/vc_raw_html][text_output]How can a machine understand that the phrase “hot dog” typically relates to a popular sausage-based snack, and not to a poached pooch?
It was this puzzle – and millions of others like it – that led to a shift in the functioning of Google’s algorithms at the turn of the millennium. Their engineers understood that to create the frictionless frontier they sought to build between a user’s thoughts and the information the user desired, their machines would need to go beyond keyword matching to comprehend the nuances of language in use.
Fuelled by Ludwig Wittgenstein’s assertion in his Philosophical Investigations that “the meaning of a word is its use in the language”, this began the march towards increasingly sophisticated ‘semantic search’ that continues to gather pace today.
So what is Semantic Search?
In short, the purpose of semantic search is to go beyond the ‘static’ dictionary meaning of a word or phrase to understand the intent of a searcher’s query within a specific context.
By learning from past results and creating links between entities, a search engine might then be able to deduce the answer to a searcher’s query, rather than provide ten blue links that may or may not provide the correct answer.
The move towards this approach accelerated when Google purchased Metaweb in 2010 and expanded their database of entities and properties significantly from 12 million to well over 1 billion within the Google Knowledge Vault. As a consequence, searches for people, cultural artefacts and places (to name just a few) increasingly show results like the below:
After this initial upgrade, there have been significant advances in Predictive Search, Local Search and Universal Results (Video, Images, etc.), particularly through the Hummingbird algorithm overhaul. As a result of these developments, a typical Desktop results page has started to look as follows:
It is worth noting the irony at this stage that in aiming to provide the one true answer to a query, Google’s results pages have in fact pullulated in various directions to comprise a panoply of potential answers across various media formats. This is, however, reflective of the inherent ambiguity of our queries and it is a complication that Google Now on Tap seeks to disentangle where appropriate.
Factor in the recent announcements on App Indexing and the inclusion of app content on devices that do not have the relevant app installed, and the picture of a universe of entities tied together by behavioural, location and historical data becomes ever more complete. The more data taken in by semantic search technologies, the smarter and more accurate they become.
Is Semantic Search a Google-only phenomenon?
Although we have focused on Google here, it would be a huge oversight not to reference the efforts of all of the big tech players in this space. Artificial intelligence is integral to the product strategies of both Microsoft and Apple, to name just two of many, and as we have discussed elsewhere, Facebook’s recent movements are arguably the most interesting out of this competitor set.
The social media giant has a stranglehold on the messenger app market that will undoubtedly stand it in good stead as it integrates its incredibly rich data from Facebook user profiles and histories. As a result, the battle between Google and Facebook to dominate this Mobile-first world looks certain to increase in volatility over the coming years.
What does Semantic Search mean for SEO?
We need to understand the best way to respond to a user’s desire or need, beyond merely matching some content to the keyword we expect them to use. Content should therefore be created with the aim of providing the most comprehensive answers possible to the likeliest questions among our audience segments. Measurement frameworks should also be created with these aims in mind, as we should not always expect immediate commercial returns from informational resources.
Behavioural data can be assessed through analytics platforms and this approach should be at the centre of any SEO strategy. These insights tell us quite directly how effective our content is in satiating user demand, so they should shape every aspect of the on-site experience and can influence audience profiles for off-site targeting.
We can also help search engines by using structured data to highlight entities and the relationships between them. This can even influence the increasingly conspicuous ‘Quick Answers’ that appear on Google results, so it is an approach well worth pursuing:
By integrating this content with our social media and PR strategies, we can increase the popularity of our brands’ web properties and create the vital relevancy and authority signals that can help pages appear more prominently in organic search.
In return for this careful planning, we can also glean valuable knowledge about the behaviour of our consumers and potentially predict their future requirements. This provides a better user experience and increases conversions, so these developments in semantic search create mutually beneficial scenarios for consumers and marketers.
Semantic search has been the raison d’être of many search engineers for a decade or more; and as such, we can start to predict where this market is shifting. It was Google’s Amit Singhal that summarised this enterprise best:
“The holy grail of search is to understand what the user wants. Then you are not matching words; you are actually trying to match meaning.”
Read more on how semantic search shapes our SEO offering.[/text_output][/vc_column][vc_column width=”1/4″][visibility type=”hidden-desktop”][gap size=”20px”][/visibility][vc_widget_sidebar sidebar_id=”ups-sidebar-2″][/vc_column][/vc_row]