Since Chat GPT rolled out last December, everybody has been worried about robots taking over the world. What SEO people really should be worried about is Google’s new Search Generative Experience (SGE).
Google has been using machine learning and AI for many years now, but SGE is a development that really moves things forward. In this article, I’ll delve into what it is, how it works, and how it will change SEO forever.
What Is Search Generative Experience?
The Search Generative Experience, also known as SGE, is Google’s new search experience utilizing generative AI to answer search queries. It’s basically Chat GPT at the top of the Google search results.
It looks like this:
You’ll note the massive answer at the top of the page. They’re like featured snippets on steroids.
(FYI – To get access, you have to be signed up for Google Labs in the US. Highly worth it if you want to see the future of SEO.)
Each section breaks down as follows:
There’s the top section, which directly answers the query. This is pretty much a featured snippet that we’re used to. The difference is that it’s not directly quoting an existing website. Instead, it’s coming up with its own formulated answer.
The right-hand side features websites that it relied heavily on to derive answers from. These are clickable links.
The middle section expands on the original answer, and the bottom section encourages you to click on or come up with follow-up questions.
Here’s what an SGE SERP looks like with a commercial/ecom query:
You’ll notice that the Google Shopping ads immediately proceed the SGE tab, followed by the organic search listings. No search ads appear.
The SGE output is different from the informational query:
It makes ten different product recommendations! It draws from many different brands, websites, and products.
Here’s a “best” keyword search in the software space:
You’ll note that the search ads precede the SGE output here. The rest is all normal stuff.
How Does SGE Work?
At a quick glance, there are a few things that we can learn about SGE:
- It relies heavily on Google’s current Knowledge Graph database
- Google is still giving prominent placement to its advertisers
- The SGE output takes up a MASSIVE amount of space on the page
Each one of these items has its own implications that I’ll delve into below.
1. SGE Relies on Google’s Current Knowledge Graph
I like to think of Google’s Knowledge Graph as its brain. It’s how it maps out relationships between words and things. Things, in a technical sense, are “entities”.
In the examples above, you can see that SGE features websites and brands that it is already showing on page 1 for its organic search results. This is good for the SEO industry because it means that SGE outputs are using the inputs that we give it.
In other words, just like current SEO – it’s still gameable.
SGE affects branded search too. Here is a query for the mega finance website NerdWallet:
Notice that it draws on the same data sources as the Knowledge Graph does.
If we click on any of the down arrows, this is confirmed:
Note that Google utilizes the same web sources for both to understand the brand. This includes Wikipedia, the App Store, and PC Mag.
No surprises there: Google has been using trusted “seed source” sites to inform their ranking algorithm for decades. But the interesting one to me is that Google uses NerdWallet’s “Terms of Use” as a source page.
This isn’t a one-off thing. Let’s see the other sources for down arrows:
Their “App” and “Best Budget Apps” pages. (Note that they list their own app on the best budget apps pages.)
Terms of Use.
And the “Security” page.
The takeaway: Google is using on-site trust signals to understand what your website is about – and rank you accordingly.
Maybe this is nothing new to you. But I think it’s revelatory because this is undeniable proof of what Google understands. (And also about what it doesn’t!)
Establish Your Brand
In my last article, I discussed why Google doesn’t trust blogs and affiliate websites, how to establish that trust, and how to reverse your traffic decline if you got hit by the Helpful Content Update.
In other words, it’s all about establishing EEAT with Google.
SGE is giving us a massive insight into what Google considers trustworthy.
Clearly, it read NerdWallet’s:
- Terms of Use page
- App/Product page
- Security page
Other pages that might be worth having include:
- Privacy Policy
- Editorial Guidelines (if you’re an online publisher)
- Staff or Team
- Customer Reviews
- Press & Awards
If you had told me five years ago that having a T&C page was essential to SEO and that posting a “Reviews” page built trust, I would have considered that nonsense.
I often see “Reviews” pages from the brand itself ranking #1 for brand + “reviews” keywords. (Eg, “hubspot reviews”.) It makes no sense to me that Google ranks “Reviews” pages from the companies themselves. As if any company would publish bad reviews about itself!
But I think it’s an easy way for the algorithm to separate real companies from fake sites – so they use it.
All of your reviews could be totally fake, your press awards could be made up, and your T&Cs could be 100% boilerplate – but all of it would still engender trust with the Google machine.
(For the record, I don’t advocate doing any of that! Honesty is the best policy.)
Let’s continue with what we’ve learned from SGE.
2. Google Still Gives Prominent Placement to Advertisers
Of course it does! To this day at least 90% of Google’s revenue comes from ads. To be clear, that’s 90% of their $100+ Billion annual revenues.
It would be massively against Google’s financial interests to create a search experience that didn’t fill the page with as many advertisers as possible!
As long as the overwhelming majority of Google’s revenues come from advertising, I don’t think Google will ever create a user experience that doesn’t offer many different organic search results.
Why does this matter?
For years, many people in the search community have seen that Google has been transitioning from a “search engine” to an “answer engine”.
20 years ago, Google used to look like this:
But now it looks like this:
For a commercial query, it looks like this:
I know that this is an obnoxiously long screenshot. But look how long you have to scroll until you reach a traditional organic search result!
Of course, ads receive priority placement.
Over the past 7 years, there has been a huge rise in “featured snippet” search results that show this:
So many people think, “Google is only going to show one answer! SEO is dead!”
But that’s not the direction that Google is going in. Current knowledge panel search results draw from multiple websites and sources. Those results look like this:
SGE is the most recent iteration of this type of answer. IE, a search experience that just gives you an answer.
And it looks like this:
SGE is layered on top of knowledge graph results. Expanded, it looks like this:
As of right now, Google’s “answer engine” search results still provide ample opportunity for publishers to receive referral traffic.
Rest assured that both ads and long organic search result pages aren’t going ANYWHERE.
But how will that affect SEO?
3. The SGE Output Takes Up a MASSIVE Amount of Space on the Page
In case you didn’t notice, SGE output takes up a massive chunk of prime search real estate!
At this point, we can only speculate about how that will affect referral traffic. But if I was a betting man, I’d say it would result in fewer clicks to the search results.
The most recent data that I could find shows that almost two-thirds of Google searches result in zero-clicks:
Source: SparkToro based on SimilarWeb data
Their earlier findings, which I cited in my digital marketing statistics research, found that this was a trend over time:
I don’t expect this trend to go away with SGE.
External referring clicks go down 5%, 10%, or a whopping 20%; we don’t know at this point.
You don’t give away prime, above-fold placement without seeing a massive increase in accompanying clicks. That’s just common sense user experience.
But given Google’s love of ad revenue, Google isn’t turning into a one-answer “answer engine” any time soon.
How to Optimize for SGE
I explained how to find trusted seed source sites that you need to get reviewed by in my last article. Go read that if you haven’t.
In short, you need to get more trusted mentions in your company’s “About the Source” queue:
Your Google Business Profile, Google reviews, third-party reviews, editorial reviews, forum chatter, and premium site business listings all combine a series of signals that Google reads.
You want to get as many positive third-party reviews as possible to signal to Google what you are. This is showing Google that you are “trustworthy” – the fourth letter in the acronym EEAT.
That same process of optimizing your site’s reputation is how you optimize your entity for SGE.
Consider the additional components that we learned earlier: on-site signals. These on-site signals include:
- An “About” page that tells Google what your site is about
- Organization schema markup that tells Google what you already told it in your About page
- A real address in your contact page and sitewide footer
- Human faces and names on your About, Staff, and Blog content pages. (Bonus points for using author and person schema markup with appropriate sameAs links!)
- Boring legal pages, including “Terms & Conditions”, “Privacy Policy”, and “Security”
- Pages that show excellence and legitimacy: Press, Awards, Careers
Pro tip: Make sure that all of those new EEAT pages you create end up in your website sitewide footer.
All of this is for optimizing for your entity.
But about optimizing for specific keywords?
I started with your entity on purpose. Google needs to know what you are before it can recommend you.
For example, say you’re a credit monitoring company called Credit Sesame. It’s a big company with a big following, but not as famous as category leaders Credit Karma and FICO. Let’s see what the traditional, current Google algorithm thinks about when you search “credit monitoring services”:
Three of the four first organic search results are the three credit bureaus, which now have their own credit monitoring services: Experian, TransUnion, and Equifax.
Who were they recommended by? The top publications recommending them: CNBC, Forbes Advisor, US News & World Report, and Business Insider.
Many searches like this include a “Perspectives” carousel. It highlights internet chatter from Quora, Reddit, and YouTube.
The social media articles and internet sites all contain “best” recommendations. Google reads those recommendations and uses them to inform their search results.
This is what the current algorithm thinks are the top brands:
Related searches:
Notice that the “Related searches” include brands along with keywords. Experian, TransUnion, and FICO Scores all appear.
Remember: Google ranks things, not strings. So your entity optimization statements on your About page, schema markup, Google Business Profile, and third-party “reviews” all need to say the same thing about you. They need to say what you want Google to say about you.
If you want Google to say, “XYZ Credit Company is the best in the market”, then you need to feed it that data in the aforementioned sources.
If all of the third-party reviews say, “Experian is the best credit monitoring company”, that gives Experian a massive boost in the rankings!
Here is what SGE says:
Notice which brands come up: the same ones on the first page of Google! The best articles get highlighted, and their brand recommendations are used.
If you’re a brand, you need to get mentioned as the “best” in all of those articles! (And believe me, this is much easier than it seems. A little budget and a conversation with the right in-house person go a long way…)
If you’re a recommendation site, you need to look like a real company so that Google trusts you.
Remember those “Related searches”? Look into “autosuggest manipulation” and see if you can’t get your brand mentioned in keyword searches. (This is a trick that few people in SEO even know about!)
What about your linking strategy? Try sprinkling in your brand name with the keyword in external backlinks now and then. For example, “credit sesame credit monitoring service”.
(Credit Sesame, if you’re reading this, you know who to call! 😉 )
Data in, data out.
Feed the machine!
Is there anything else that you think helps with SGE optimization? Leave a comment below.
