With Google Search’s Amit Singhal ready to move on, enabling John Giannandrea, who has been incharge, machine learning and artificial intelligence in Google, the search giant is sure upto some gradual but massive change in its processes. Taking into account the official announcement by Google about the launch of Rank Brain in a video interview given to Bloomberg, it sure seems to take the SEO world by storm.
In plain computational linguistics, semantic search seeks to improve search result accuracy by understanding the searcher’s intent and the context of the terms in the order of their appearance in the search dataspace, to generate more relevant results. Put into a layman’s terms, it basically throws in relevant key terms as soon as something is typed into the search bar of the search engine based on the expected context of the words.
Though it would be too early to predict the exact nature of Rankbrain, what can be said with some strong conviction is that it is a type of new ranking signal most probably working in conjunction with the current overall search algorithm used by Google Search, named Hummingbird that basically uses semantic search techniques.
Rankbrain incorporates the concepts of artificial intelligence and machine learning for near perfect result to the query submitted in the search bar. It embeds vast amounts of written data into mathematical entities, called ‘vectors’ – that the computer understands. This is sure going to be more effective at handling queries of the sort never-heard-of-before and presenting a better success rate.
As told in the video interview given to Bloomberg, Rankbrain is the third most important ranking signal out of over 100 such signals. As for these signals, they are nothing but signs to convey instructions or information for a wide variety of attributes relevant from the point of view of streamlining online search results. For example if some words are in bold, it generates a particular signal that might be completely different from the one generated when some mobile-friendly query shows up. In a nutshell, all these signals together figure out the best possible search results for any query in the search bar.
Statistically, Google search entertains close to 3 billion search queries every passing day. In the year 2007, about 20-25% of those queries are never seen before . By 2013, this figure came down to 15%. It still accounts for about 450 million unique query per day, that may contain complex terms, multi-word queries, long-tail queries, etc.
What Rankbrain does is, it provides help in interpreting searches that people submit to find pages that may not have the exact words that were searched for. In a nutshell, all it does is streamline the search queries and align them according to the expected context and relevant information to give the best result to the user, free of ambiguities and as closer to the actual meaning as can be.
Though Rankbrain had a gradual roll-out in the early 2015 but it went live globally only for a few months now. Since then it has been processing a really large fraction of data. It also dictates to a large extent how any page ranks in the search results. However, it shouldn’t be confused as a new version to Hummingbird algorithm; it rather coexists and optimizes the Hummingbird.
The final word, what should this mean for SEO at present? Not much for daily activities, though optimizing web content using semantic seo practices would definitely help in the long run. This would in a way offer us a glimpse of what future SERPs would probably function based on artificial intelligence. Can it be the next cornerstone in search algorithm space? Let’s say… only time can tell.