Machine learning has truly entered the search market this year, with Google revealing that it has introduced a new facet to its integral algorithms called RankBrain. The implications for search engine optimisation are potentially huge, especially if this pseudo-AI (artificial intelligence) becomes more influential.
Firstly, it will be better equipped to deal with search queries which are rather more ambiguous than the concise phrases that Google is already so good at handling. And since around 15 per cent of the queries which Google receives each day have never been entered before, there is a significant need to improve its ability to handle unique, vaguely worded questions that its global audience generates. The new algorithm should be able to assimilate new, ambiguous queries, make a guess at what the user is really trying to find and assimilate this knowledge to make sure that it is better prepared to deal with similar queries in the future.
Another area in which it should enhance the user experience relates specifically to voice search. It is built to tackle conversational, long form search queries that are increasingly common thanks to the popularity of mobile devices and the personal assistant solutions they offer. This is interesting information in its own right, but Google remains tight lipped about the specific aspects of how this algorithm operates, which complicates the process of developing an SEO response to its rollout.
In fact, the power of this platform to impact a site’s ranking performance is already significant, since Google has said that it is now the third most important ranking signal used to determine the order in which search results should be presented to users. But why has its introduction caused so much fuss and how will it change SEO?
What is RankBrain?
Google’s Greg Corrado told Bloomberg quite a bit about RankBrain in a recent interview, providing details about when it was implemented, how it works and just how many queries fielded by Google it is influencing globally. As with Google’s other search algorithm updates, it was built to make the search results more relevant based upon users’ queries whilst in parallel ensuring that higher quality pages rise to the top. But while previous rollouts such as Panda, Penguin and Hummingbird have all been designed and tweaked by human engineers, RankBrain is capable of thinking for itself and should improve over time without the need for direct intervention. In essence, Google’s past updates have been reactive; the result of engineers creating and adjusting algorithms to meet the needs of users. This one will be proactive, automatically learning to get better at determining how to serve queries which have traditionally outfoxed Google.What are its parameters?
The mention of AI can make this algorithm update sound as though it will be an all-powerful part of Google’s search arsenal and thus a key consideration for SEO, content marketing and CRO. But in reality it is just one of the hundreds of ranking signals that are used to rank pages. And when you dig deeper, it seems to have some fairly specific applications within the search engine ecosystem, mostly relating to semantics.
