Google’s AI Overviews have long been planned by the search engine giant. Now it’s completely out of the testing phase and into production, the feature promises to change the whole search landscape for many users and websites hoping to rank highly.
Websites such as iPullRank, Peak Ace, Authoritas and SERanking have all joined the foray of AI Overviews/Search Generative Experience (SGE) research to give us some early insights and inklings as to how this new AI tool works. We’ve summarised some of the main findings here.
Quick disclaimer
Before we get started, it’s worth acknowledging the fact that this research, at least for now, is still developing. AI Overviews and Google Search are constantly changing. While these results, statistics, and insights might be relevant right now, this could certainly change over the coming weeks and months so keep checking back for updates.
Everything we know about AI overviews so far
AI Overview triggers
Most reports conclude that the majority of search keywords will trigger AI Overviews to the present. Half of those keywords will happen automatically, while a quarter of those keywords will require the user to trigger the overview manually.
AI Overviews and featured snippets
Featured snippets have become a common sight on Google search results. They take a portion of content from a page’s body copy and serve it directly to the user without them ever having to leave the search results. Ranking in featured snippets is a way of knowing that Google is responding well to your website’s content, but it seems that AI Overviews may be trying to steal their dinner.
According to Onely/ZipTie, e-commerce keywords return an AI overview roughly 85% of the time. 0.17% of them return a featured snippet, while 1.16% of them return both a featured snippet and an AI overview.
Sources used in answers
While it can be safe to assume that AI overviews are informed by top-performing search content, this isn’t always the case. A lot of AI overviews are informed by results outside of the top 10 organically ranking pages for a given query.
ZipTie reports that 53% of sources used by AI Overviews rank outside of the top 10 results – they go even further by stating that 43% of results aren’t even ranking in the top 60 results. Authoritas makes similar claims, specifically that 62% of sources linked in AI overviews come from outside of the top 10 organic search results.
Keyword length
We’ve long thought that the best kinds of keywords to optimise for in order to rank in AI Overviews are longtail keywords, which are keywords that are longer than five words in length. Research by SERanking seems to confirm this theory, with there being a 17% gap between longer and shorter keywords when it comes to how often they return an AI Overview.
For example, one keyword query would trigger an overview 12% of the time; nine keywords, however, would return one 32% of the time – a 20% difference.
Ranking and linking
Mike King reports that 91% of AI overviews feature at least one link to a source ranking in the top 10 organic ranking positions for a search. Comparing the first three links with the first 3 organic rankings, Peak Ace reports that in the vast majority of AI Overview responses, there is mostly either zero or a partial overlap.
What does this tell us?
So, that data gives us some initial glimpses into the mechanics of how AI overviews work. Now, let’s discuss what this tells us.
Ranking highly doesn’t guarantee success
One of the main findings from this tells us that ranking highly in search results isn’t the end all be all that it once was. AI Overviews consider a vast range of sources when determining what information to provide to Google users – especially those outside of the top 10 organic search results.
Now, this isn’t to say that websites should spend loads of time de-optimising pages, but it does tell us that supplemental, high-quality content will become increasingly important in getting websites featured. Remember, a site can feature in an AI Overview more than once, so it’s possible to have a landing page being linked to, as well as an in-depth blog post that’s related to said landing page.
Content quality matters
Of the pages that do get linked to within AI Overviews, a common recurring theme revolves around content quality.
Content is already under a lot of scrutiny from the eyes of search quality raters and EEAT (Experience, Expertise, Authority, Trust). Introducing AI Overviews adds another dimension of scrutiny that will take some time to truly tease out.
While things like keyword stuffing are well and truly out of the picture with content, there still exists a viable strategy around optimising content for longtail keywords. This last bastion of keyword optimisation may too fade out in favour of more originality within website content.
Get in touch
As AI Overviews continue to evolve and reshape the search landscape, staying informed is critical for navigating this new era of digital visibility. At Balance, we specialise in helping financial services brands adapt to these changing dynamics with strategies that prioritise quality, relevance, and adaptability.
Looking to optimise your content strategy for AI-driven search? Let’s explore how we can elevate your digital presence together. Get in touch with us today.
