How Google is using AI in 2016

SEO Executive Rory Long looks at Google's increasing use of artificial intelligence and the implications this has for how Google interprets 'useful' content.

There have been many significant developments in SEO practices in a relatively small number of years and with each new algorithm update, marketers have looked at ways to adapt to the complex set of rules that Google uses to rank search results.

However, in October last year, Google announced that machine learning had been part of its Hummingbird algorithm for months. This move to incorporate artificial intelligence – nicknamed ‘Rankbrain’ – into search results has been something of a game-changer in Google’s ultimate quest to deliver a brutally efficient user experience. With artificial intelligence, Google is now able to understand a wide range of search queries better than ever before and importantly, interpret semantic meaning. In this blog post, we will examine the implications this has for SEO, specifically in terms of how we should approach creating content.

The end of keyword targeting?

For many years, the main focus for SEO content was selecting the keywords you wanted to target and placing them strategically throughout your site. Keyword research was vitally important – after all, if you didn’t have your keywords throughout your site, how would Google end up finding it?

Although this is still the case to some extent, with the help of Rankbrain, Google is much better at interpreting meaning when it comes to keywords. For example, if you wanted to rank for ‘Glasgow hotels’ you would need to mention this keyword throughout the copy, and also within your title tag and H1 tag. Google is increasingly able to interpret semantics so if you had ‘luxury hotel experience in Glasgow' in your content, it’s more likely this would also rank for the term ‘Glasgow hotels’. What does this mean for content? It means that content writers can now focus on creating genuinely descriptive page copy and still rank for generic industry terms without worrying how many times they have mentioned the word hotel or the location for example.

It’s also important to consider that longer tail search terms are becoming the norm. According to Martin Laetsch, director of online marketing at Beaverton, 75% of search queries are three to five words long. Speaking to the American Marketing Association, Laetsch described how “the search engines are figuring out that if people search for the word ‘marketing’, or any one or two word query, they don’t get the results they want. To get quality results that are most likely to answer their question, they have to go to three, four or five word queries. As content creators, when you’re thinking about optimization, you have to think about that”.

The increasing importance of long-tail keywords can be seen by the fact that Moz, the well-known marketing analytics and software specialists, has created a new feature aimed at helping content creators understand the way that search engines interpret different topics and phrases. This development will also give prevalence to longer content.

It used to be the norm that most shared content was around 300 words long, but articles between 1,200 and 1,500 are actually performing better in search. The changes Google is implementing are ranking longer, more informative pieces highly with the aim of answering the longer tail search queries of users.

Analysis of user interaction

The focus for SEO used to be about getting the initial click from the search results. However, Google is now increasingly able to examine post-click user interaction. How long does the user spend on the site? Are they leaving your site to look back through other results? Google analyses this data to determine whether users are able to find the information they need on a website and using it to assess which sites should be returned for certain queries. Google has always placed importance on user-friendliness in their search rankings, with a focus on site structure and site speed in particular. With Rankbrain, Google is able to place more of a focus on user interaction and with machine learning they are able to gain a better understanding of human interaction and behaviour which allows them to analyse whether they are satisfying user needs in their search. With this in mind, it’s going to be increasingly important to consider how users will interact with your content and to make sure that the content will satisfy the query of the user.

Increasing role of voice search

Voice search is taking on an increasingly prominent role in search and according to a study from Northstar Research, 55% of teenagers use voice search every day, with 56% of adults using it because it helps them feel more "tech-savvy". This has meant that search queries have begun to take on a far more conversational tone. When conducting a voice search, people ask questions such as “where is the nearest cash machine?”, whereas if people were carrying out a text search, they would be more likely to type “cash machines Glasgow city centre”. This allows Google to focus on the who, what, where, and why that takes place in a normal conversation.

SEO Expert Michael Peggs recently stated his top three tips to get ‘voice search ready’. The first tip was to implement long-tail keywords with voice search in mind. The second was to use an FAQ strategy to think of what questions your customers are asking and addressing those answers across your online presence. Finally, Peggs emphasised the importance of writing natural sounding content to reflect the language that customers use in voice search.

Voice search is signalling a shift in the language of search terms towards a far more conversational and human sounding tone that will require content creators to provide answers to increasingly inquisitive search queries. The implications of artificial intelligence and voice search for keyword targeting and user interaction brings up an interesting question – what key areas should businesses focus on in order to adapt to these changes?

Understanding your audience

With the increasing prominence of long tail keywords, Google’s better understanding of semantics through Rankbrain and voice search becoming more prevalent all point to one thing – it’s now more important than ever to understand your audience. For example, what kind of questions are they asking and how do they expect them to be answered? And which type of content performs best? Cyrus Shepard of Moz cites that the most important SEO tip for 2016 is to focus on your audience. Shepard says, “In the past, it was about marketers trying to promote what they wanted people to see. Today it’s about delivering what people actually want to see that will give you an SEO ranking boost”. In terms of content, this will require a much greater level of analysis of what your audience wants and how to answer their questions.