What is Google’s Search Generative Experience? (SGE)

What is Google’s Search Generative Experience (SGE)?

Google is never one to stop innovating and their new strides in AI integration into their search engine represent some of the biggest changes we’ve seen yet. 

As business owners and agencies come to terms with what these incoming changes mean for the future of Search - let’s take a look at what they’ve actually shown us so far, as well as what that might mean for our everyday lives.

What is SGE and what does it mean?

SGE stands for ‘Search Generative Experience’ and is Google’s acronymic name for their new prototype search engine results page (SERP) layout that integrates AI more prominently than ever before. The ‘Generative’ aspect is key here, as it describes the most fundamental shift from the experience we already know and (mostly) love. 

Google’s plan is to use its in-house generative AI models such as Gemini to summarise, condense and signpost the web pages and content in its index. This comes in the form of new widgets and snippets on the SERPs that can understand and digest the content on the websites that appear in the organic results, answering user questions without ever having to click on any of those pages.

Examples of SGE changes

Below is an example of a search query with SGE enabled - in particular for a ‘question’ query:

(Image Credit: Google)

One of the biggest factors that determines the clickthrough rate on SERPs at the moment is how high up the page an organic link appears. This peaks sharply for the first result and rapidly falls away, with almost 70% of clicks going to the top 3 organic positions (source: firstpagesage.com).

Another hugely impactful change as part of SGE then, is likely to be the visual space it occupies - detracting from organic links. In the example above we can see the initial view a user has in the current form of the experiment includes no organic links at all. You have to scroll past not just the generative AI widget that summarises the answer - but also the already implemented ‘People also ask’ widget, to actually get to an organically ranking page.

It’s not just limited to question-based content, however. We’re also seeing Google experiment with shopping and purchase-intent terms. The same visual impact occurs here too, but now with the added concern of direct links to the places to buy these goods. 

This could obviously have huge ramifications for the businesses operating those web pages - if the SERP itself can answer your question, there may not be much of an incentive to click through to a ranking page. At the moment, this new experience is an opt-in prototype in their Search Labs process.

When is SGE happening?

After a fair amount of pushback from site owners and SEOs during the test phase, Google slowed down it’s plan to roll out the new SGE. Initially it had set an expiration date for the experiment towards the latter part of 2023 but without much fanfare this date has been extended.

Instead of rolling out the whole experiment, Google now appears to be breaking out elements of the test to trial it in live search results. This means you may start to see parts of SGE appearing in SERPs already, or if not you should expect at least some level of generative AI presence in the next few months.

Reading between the lines and given the scale of concern among the industry - we’d be surprised to see a full rollout in 2024.

What does SGE mean for SEO?

It’s difficult to quantify the impact of the introduction of generative AI into SERPs without straying into speculation. User behaviour around featured snippets, ads and other widgets on the SERP indicates there’s still an audience that likes to click on organic links, even if it grows smaller with each new addition - but will that audience need to click on any links if Gemini can answer their query without leaving the SERP?

The search terms most likely to be impacted are ones that can be easily answered, summarised or otherwise satisfied by a roundup widget in the SERP.

This means that sites relying on generic terms or high search volume question terms are going to see the biggest shift in their performance.

Site owners should look to make their content more valuable and in-depth than can be easily provided by Google through gen-AI. This means longer and more insightful copy, the integration of more visuals and high-production-value content that can’t be automated and a user experience that trumps what the SERP can provide.

A key element of understanding and mitigating the risks from these changes is in the format of the answers that Google gives. In early experiments there was very little in the way of credit or sourcing in the generated responses, leading to backlash from the wider SEO community around Google essentially ‘stealing’ the content and using it for its own purposes with no credit. Since then Google has introduced more clear and prominent sourcing, although still not as much as many would like. What this means though, is that there is now at least some value in your site content being seen as a good source for the generative AI to use. Especially in the shopping example above, we can see that it’s possible to still drive value through the new SGE experience if you can work out how to best optimise your site to appear in the sources listed.

In many ways, embracing the tenets of Experience, Expertise, Authority and Trust (E-E-A-T) that Google has encouraged, could be an effective bastion against generative AI. Building sites that can become a reputable source for information, that provide clear evidence of why users can and should trust them and that have deeper knowledge than simply scraping the internet, might be one of the best ways to future-proof the value SEO brings to any business. Especially in a world that seems determined to move away from those classic ten blue links.

What do we at Reflect think?

As we’ve seen with the introduction of BERT, E-E-A-T and other Google attempts to understand and better ‘qualify’ content for search results, the process is slow and steady. Rather than a single announcement and rollout, change in the search world happens consistently and regularly, morphing over time from one dynamic to another. From a business perspective this kind of change is deceptive. Rather than being able to pivot and adjust when presented with immediate change, the most successful companies deal with the issue by understanding what Google wants and where it’s headed. For example, they have talked for years about their goal of making rankings more trustworthy in a digital world of misinformation and deception. Sites that demonstrate their credentials, clearly state signals of trust and treat users respectfully have seen gains from repeated algorithm updates in the last few years. We’ve come to understand these changes as E-E-A-T, where Expertise, Experience, Authority and Trust are actively rewarded - but companies that understood the trajectory being indicated were already attempting to better position their sites to succeed in those ways.

We believe businesses that stay ahead of the curve and understand that in a world of generative AI and automation, humans still come first - are best placed to succeed. Simply knowing that this is coming is an advantage, so we need to make the most of that. We need to understand the value of our content in a world where direct clicks become harder than ever to achieve. How do we show users that our content is more rich than can be summarised in a few sentences? Do we demonstrate clearly why our site deserves to appear in these generated sections in the first place? Most importantly of all, no matter what changes we see in this new framework, we will never replace the human element of search - the searcher. Put them first.