In May, Google announced that it was testing a new, experimental AI-powered search experience. Aside from understanding more complex queries, Search Generative Experience (SGE) uses AI to generate answers in the SERPs. This means fewer users need to visit websites and, to make matters worse, SGE generates its answers from content in the search results users no longer need to click. Website owners are understandably concerned about losing traffic to SGE, but how much of a threat does it pose?
Search Generative Experience is an experimental new search experience that incorporates generative AI into Search. So, if a user types in a query like “what is search generative experience,” they’ll get a generated response at the top of the results page.
Google uses its own GPT-like generative AI technology to generate these responses, using content from around the web. For this particular query, Google returned a single sentence answer with a longer response accessible by clicking the Show more button.
In the full SGE response, Google includes two short explanation paragraphs, two bullet point lists, three sources and an image. The image is credited to the original publishers and each text element is also cited, sometimes with multiple sources. Users can access the original sources of information by clicking the v icon.
At the bottom of the SGE section, users can ask follow up questions, to explore the topic in more detail, and Google also provided some generated prompts.
Google is currently testing Search Generative Experience with selected users in the United States, India, Japan and other selected areas. Google hasn’t officially announced a rollout date for the UK but we expect this to happen sometime between January and March 2024. Once it’s available to people in the UK, users will be able to opt-in to the experiment via Search Labs.
Keep in mind that SGE is still in the experimental stages of testing. So, it’s too early to assume that the version participants are currently testing will make it into the public search experience. Anything and everything could change before that happens. Even still, we’re gradually getting more clues about what the future of search could look like from the SGE experiment.
Back in May, Google teased a preview of Search Generative Experience. When users enter a query, the results page renders and SGE generates an AI snapshot, pulling information from relevant websites.
We know Google is testing citations, including links to the pages it sources information from. It’s also testing follow-up questions and topic exploration to save users from typing multiple queries.
SGE’s generative AI snapshots are the big concern when it comes to losing traffic. Early analysis by Michael King reveals some interesting insights from the SGE experiment so far:
One of the biggest takeaways here is how much faster SGE is generating AI snapshots now. When it first launched, SGE took anywhere between 11 and 30 seconds to generate AI snapshots, but it now clocks an average of 6.08 seconds – a significant reduction.
As Michael King states in his analysis, it takes users an average of 14.66 seconds to click on a search result. This means any AI snapshot that takes longer than this to generate is unlikely to threaten organic traffic.
Marketers need to pay close attention to AI snapshot generation times and how this affects CTRs. Everything is speculation at this point, but let’s say SGE needs more time to source information and generate responses for complex queries. This could reveal opportunities to optimise for long-tail keywords that get more visibility for those blue organic links.
One (very small) study of 23 websites found that the aggregate drop in organic traffic due to SGE was 18-64%. Obviously, we can’t take much away from such a limited study but nobody has had access to SGE for long enough to draw meaningful conclusions.
Aside from a serious lack of data, we simply don’t know how the final version of SGE will function. A lot of variables will determine how much traffic websites lose and the list will only grow as Google refines the experience further:
As we saw when featured snippets first arrived, most discussions revolve around how much organic traffic AI snapshots will take. However, the more important question will be what kind of traffic will they take away from websites and how much damage will this cause to business performance?
While featured snippets have taken a lot of traffic away from websites, they mostly show for informational queries. Of course, this is a problem if you rely on informational queries to drive and monetise traffic – for example, you’re a publisher relying on ad impressions.
For most companies, though, the queries that drive real revenue are more commercial by nature, though. So, between the searches that show paid ads and those triggering featured snippets, the loss of commercial traffic is relatively modest compared to their prevalence and impact on zero-click searches. The big question with SGE is how much commercial traffic it will take away from business.
The presence of AI snapshots for shopping queries in SGE suggests they could cause more damage than featured snippets. However, they’re still predominantly showing for the more informational kind of shopping queries. After all, you can’t buy a TV from generative AI responses, no matter how good SGE’s snapshot may get.
Another parallel with featured snippets will be whether we can optimise for SGE AI snapshots. Over time, we found ways to optimise for featured snippets and, as competitive as they are, the rewards are big.
If it’s too early to calculate the threat of Search Generative Experience, it’s definitely too soon to start looking for opportunities. That being said, when the time comes, marketers and brands will need to look for every opportunity they can in the SGE landscape.
Now, this is a purely hypothetical example, but Google has always struggled to interpret and answer long-tail keywords – something it is constantly addressing with its AI technology.
Back in 2021, Google unveiled its MUM update, which specifically targeted complex and long-tail search queries. This was a key step towards the interpretational power of SGE and who’s to say businesses won’t benefit from Google’s advances with long-tail queries?
Let’s remember that commercial long-tail queries are often the most valuable because users have a strong, specific idea of what they need. If Google can leverage its AI gains to drive this kind of traffic to commercial businesses while snapshots handle the more informational stuff, the future of search looks more positive.
Again, this is all hypothetical, but Google needs publishers and businesses to keep producing content. If it sucks all of the value out of content creation, it will have nothing to feed its algorithms other than outdated, low-quality AI content. The output of SGE will plummet but – more importantly – the organic search experience will disintegrate.
In all honesty, marketers can’t do much about Google’s Search Generative Experience at this point. The programme is still firmly in the experimental stage and nobody has enough meaningful data while it’s only open to select users – other than Google itself.
Until then, it’s time to start putting a plan of action together:
Keep in mind that early data could be skewed by users themselves testing out the new experience. Google will probably get inflated engagement metrics early on and websites may also see a disproportionate loss in traffic as SGE rolls out to their target audiences.
Remember, it will take time to understand the real impact of Search Generative Experience and there’ll be plenty of speculation in the meantime. Don’t make any knee-jerk reactions or rush into anything until you have the right data to inform decisions.
If organic traffic from Google is essential to your business, the only sensible action you could take is to diversify your traffic sources. Regardless of SGE, you never want to be overly reliant on traffic from Google Search or any single channel.
Measuring the impact of SGE will be the first step in knowing how to deal with it. If you lack the analytics system or personnel to track the real impact of Search Generative Experience, our data science team can help. Call us on 023 9283 0281 to speak to our search analytics experts or submit your details here and we’ll get back to you.
Dave is head of SEO at Vertical Leap. He joined in 2010 as an SEO specialist and prior to that worked with international companies delivering successful search marketing campaigns. Dave works with many of our largest customers spanning many household names and global brands such as P&O Cruises and Harvester. Outside of work, Dave previously spent many years providing charity work as a Sergeant under the Royal Air Force Reserves in the Air Cadets sharing his passion for aviation with young minds. He can often be found in the skies above the south coast enjoying his private pilot licence.
Digital marketing under-performing and not sure why? We'll do a free SWOT analysis and reveal priority focus areas and quick wins
Categories: Events, SEO
Categories: Machine Learning, Martech