Search Generative Experience Roundtable Discussion
Webinar recorded on June 27, 2023
Join our weekly live webinars with the marketing industry’s best and brightest. It’s 100% free. Sign up and attend our next webinar.
We hosted a Search Generative Experience (SGE) roundtable with some of the brightest minds in SEO including Aleyda Solis, Bernard Huang, Gaetano Nino Dinardi, Kevin Indig, Dr. Marie Haynes, and Mike King.
And the conversation was top-notch!
We discussed what is SGE, the effects of SGE, future with SGE and AI content, and finally privacy concerns.
Listen to the webinar as audio only on your favorite podcast platform:
Join our weekly live webinars with the marketing industry’s best and brightest. It’s 100% free. Sign up and attend our next webinar.
Watch the full webinar
About Aleyda Solis:
Aleyda is an International SEO Consultant and founder of Orainti, a highly specialized, boutique SEO consultancy.
She’s a popular blogger, speaker, and author of “SEO, Las Claves Esenciales”. Her writings have been published in Search Engine Land, Search Engine Journal, and Moz.
She is also the host of the Youtube video series “Crawling Mondays” where she shares actionable SEO tips.
Follow Aleyda on Twitter: https://twitter.com/aleyda
About Gaetano Nino Dinardi:
Gaetano is a Growth Advisor for many software companies including Cognism, Alyce, and DataGrail. Previously he was the VP of Growth at Aura.
He’s also a music producer and has worked with major artists like Fat Joe & Shaggy.
Follow Gaetano on Twitter: https://twitter.com/gaetano_nyc
About Kevin Indig:
Kevin is a strategic Growth Advisor, creator of the Growth Memo newsletter, and host of the Tech Bound podcast. He ran SEO organizations for companies like Shopify, G2, and Atlassian, consulted for big brands like Ramp, Eventbrite, or Finder, and is an active angel investor.
Follow Kevin on Twitter: https://twitter.com/Kevin_Indig
About Dr. Marie Haynes:
Dr. Marie Haynes has been recognized as a leader in the SEO industry since 2008. Today, Marie is known for understanding how Google’s Search Algorithms work and sharing in layman’s terms how business owners can improve their websites. Marie writes for Search Engine Land and is often featured in authoritative SEO publications.
Follow Dr. Marie Haynes on Twitter: https://twitter.com/Marie_Haynes
About Mike King:
Mike is the founder and CEO of iPullRank, a digital marketing agency trusted by the Fortune 500. They implement a unique blend of market segmentation, content strategy, SEO, and engineering skills to create powerful marketing campaigns to grow businesses.
He’s an international award-winning marketing thought leader and technologist with specific expertise in SEO.
Follow Mike on Twitter: https://twitter.com/iPullRank
About Bernard Huang:
Bernard is the co-founder of Clearscope, the leading SEO optimization software for high-quality content teams. Before Clearscope, Bernard started an SEO consulting agency, was a growth advisor in residence at 500 Startups, and led growth at a YC startup called 42Floors.
Follow Bernard on Twitter:
Read the transcript
Travis: Aleyda is an international SEO consultant and founder of Orainti and the host of the Crawling Monday podcast. Gaetano is a growth advisor for many software companies, including Cognism, Alyce, and DataGrail.
Kevin is a strategic growth advisor, creator of the Growth Memo Newsletter, and Tech Bound and Contrarian Marketing podcast host.
Marie runs Marie Haynes Consulting and hosts the Search News You Can Use podcast.
Mike is the founder and CEO of iPullRank, a digital marketing agency trusted by the Fortune 500.
Bernard is the co-founder of Clear Scope. Awesome.
To kick it off, who wants to define what search generative experience is?
Marie: We're all going to jump in. I don't know that any of us are experts on the SGE. You know, we're all sitting here talking about this thing that we've kind of beta tested, but the SGE, as far as I understand it, is Google showing AI-generated answers first in search for many queries, not all queries.
And I, I'm not fully convinced that what we're seeing today is what we're actually gonna see in production. So, you know, we'll have a lot of theory probably in this talk. But that's my understanding what we're seeing right now is kind of an AI-generated answer, although really it's just kind of stitched together from different websites, and then they show us some websites next to
Aleyda: I agree with Mary that what we're seeing here is like the attempt of Google to try to come up with something to compete with chat G P T that. Like was first was released first. I guess that they felt a little bit like they were in a position that was to be menaced pretty much as the destination to search for information.
And they had to come up with something that had to be integrated, what the current search experience to avoid, let's say, disrupting the usage of what they already have. But at the same time, their attempt, I will say has made, made it not necessarily as useful or as good, even as the bean chatbot.
Right. But, yes, I will say that there are 10 of the integration of this chatbot-like experience within their current interface. To, yeah, to, to stay, to stay in the, in the, in the competition as a search destination.
Mike: Yeah, I agree with everything that's said here. Like, one, we don't know exactly what it's gonna be.
It's also, you know, something really early. But what we're seeing is basically like an alpha product just to appease the market. It's not even like, like El lady just said, it doesn't have the same; it doesn't have parody with the other things that are out there. But you know, it, it isn't like a, a just a direct knee-jerk reaction.
Google was building this stuff in the background for a while. They have this thing called Realm, which is, basically, I forget what it stands for, but it's their version of what's called retrieve augmented generation. And they've had it since, I believe, 2021. 2020. And it is the same framework that's being used for both Bard and S G E.
And basically, what it is, is that you feed results into a Language model, effectively pret tuning it, and then use that to get answers. And so you've got the AI snapshot, which is them being like, okay, here's your answer to this thing. Like a feature snippet on steroids. And then you've got perspectives and, you know, all that sort of stuff.
But to me, it's a colossal waste of time.
Aleyda: Unnecessary feature snippets on steroids when they already feature snippets many times, even below, right? Yeah.
Mike: So, so they spend a lot of money to regenerate the featured snippet. In some cases it's, it's
Kevin: But those are gonna go away. I don't think featured snippets are gonna be there for
Aleyda: And the map packs.
Kevin: The map pack. Yeah. Map packs. Like basically, basically is a replication of the local pack, what they show in SGE. Similar even to e-commerce, right? Like a lot of these e-commerce queries or the results that you see, they're replicated in these organic shopping results at the bottom. So there's a lot of duplication going on right now.
That I don't think will happen when Google rolls this thing out in December. Let's see if they
Aleyda: even roll it out. What, what I do really expect that they actually fix, even, even before I'll say that are links, citations. I mean, how the hell they didn't think of adding over overlay? I mean, the Bing chatbot he was already doing even something better UI-wise, right?
Integrating the overlay with links, external links in a much more, let's, let's say usable way for not, not even for a website, but also for, from a user standpoint, right? But that carousel at the top, right? Like, who will click on that? It's, yeah. Anyway, it's so bad.
Marie: I have a theory as to why Google's not showing links to websites in the answer cuz they're clearly using the information.
Like you can see. It's, it's verbatim that they've taken it from websites and very much like a featured snippet, except that they're not saying which parts come from which websites. My thought, although we don't know this for sure, is that when we actually see something like the Sge Live, those answers aren't going to be stitched together from websites.
They're actually going to be generated with ai. And initially, until today, I thought it was. That they were going to use Bard. I mean, Bard seems to me that it was built to be in search. It's going to connect with Lens with, you know, Google Docs, Google Drive, all those things. But when, if we see Bard in search, that's a different thing than the answers that the Sge are currently providing, which would not be directly from websites.
They're trained on the entirety of the internet but not directly verbatim from websites. And today, I don't know if you guys saw the article in Wired where De Sabi Hass was talking about the new Gemini language model. When I read this, it kinda, I was like, you know what, we're gonna talk about the SGE today, but I.
The Gemini will eclipse everything that we currently know in search. I don't know, I, I feel like I'm, you know, kind of talking crazy talk when you start thinking about everything absolutely changing. But I'd love to hear if any of you guys have thoughts on
Mike: on that. Yeah, I mean, I think we're gonna see dramatic change and, and just think about, well, you know, you've been talking about this a lot as well yourself, but, you know, just the things we've seen in the last seven, eight months have been such a change for everything, right?
For the proceeding of five to six years, a search was really boring. And now it's just like everything is different, right? And I think what Google is trying to figure out is what is the right UX for this. And to your point on language models, I mean, yeah, sure. The language models will be more powerful.
They'll be multimodal by design, and so on and so forth, which is great. But at the end of the day, it really comes down to how users will consume this. And I think the real question is, Know-how, how are ads gonna perform in this environment? And we've seen. Obviously, they've rolled it out to some degree, and there's a lot of companies that are in the beta and so on and so forth.
But if that doesn't pick up, or if that doesn't take in a way that is, you know, just as good, if not better than what we've already got. I think S G E just kind of remains; even with Gemini, it just remains this a novelty that Google has to have because Bing has it. And again, a direct reaction to what the market is doing.
Not necessarily is this the best product.
Aleyda: 100%. I, they, they, they have the capacity, I believe, to do something much better. But I mean, they are too scared to kill their milking cow on one hand and then also alienate the web ecosystem out there, which is already happening anyway. But, yes, the how not to kill the business and milking cow.
And that's pretty much the restriction they have. I mean, they, they should ask Apple. I mean, they have, they're doing an amazing work already with their Visual Pro, like the next thing interface to, you know, browse and interact with like, and as they did with iPod versus iPhone, right? Like they, they, they have been quite successful, like killing the previous product, the previous version or, or, or interface before.
So this is what should happen with Google. They need to sort out how to change and delete the next paradigm of search and kill, let's say a little bit, the former business model that they used to have. While evolving a new one, a new one that will be even more profitable. Right. But I guess that they are not quite yet there.
Kevin: Yeah. But I think that's really part of the problem. Right? It's, and this is, I don't think Bing necessarily was the only reason why Google adopted this model. I think one big reason is because I saw that AI is getting so good that so many other companies can create a ton of content, which really dilutes content as a ranking factor.
And I think that's a that's also in part leading search to just a different user interface than the, than the kind of ranking results that we have today. Right. It is really getting more broadly towards getting an actual answer to your question. And I think Google, I think people at Google have seen that and are seeing that right now.
And that's in part why they say we're probably cutting into our own leg, but what other option do we have? Right. We're not gonna be. Yahoo, where we manually curate results, and then somebody else comes around and gives the answer right away. So, I do agree with all of you that this is a tectonic shift, that the user interface for search changes, and that Google has no other option but to concur, right?
They, they invented a lot of this technology, and they, they always say, oh, we had this, we already had this, and we bring out so much better stuff. But it was not Google, and it was not Bing; it was none of the search engines who redefined the new interface for, for kind of this, this maybe century a bit much, but for this decade.
So, yeah, I think I think there's many reasons for Google try for why they try to do that. And I'm convinced that they will find a way to display ads and maybe even make more ad revenue than they currently make.
Mike: Yeah, I think it's possible. I just think it comes down to whether I do not take it at face value that Google can at this point.
Like they've, they've messed up a lot of things in the last couple of years, you know? Oh, yeah. But to your point on, you know, the new interface. It really begs the question, is chat the best interface for information needs? And I'm not completely sold on that. You know, I, I do spend a lot of time and, you know, this is me speaking as me, like I'm not speaking for the general user, but I do spend a lot of time with like both Google and Chat G P T Open, but I spend more time on Google because it's like, even if I get something outta chat, G P T, and even if you're using a R P R M to get it you know, you still gotta verify that information to make sure that it makes sense.
And so to that end, I don't really trust chat, and I know a lot of users do because, like, oh, it's a computer, it must be right, like we anthropomorphize it to some degree, but I don't believe that chat is the best way for us to meet our information
Aleyda: needs. And, and, and, and there is where I believe that new perspectives.
Filter comes in, and let's say user-curated or creator-driven content comes in that. I think that goal is trying to blend it little by little, but just a filter, just a tap, an extra option. Not even as visible in the main SERPs. But I believe it's, it's there, right? And all of this push of real alternative expert content that they have been talking about.
Now that we have this, pretty much the massification of it, that is AI generated, right? So, so I believe that it goes that way, and I totally believe regarding the lack of trust that the AI-generated content or chatbot answers can have. Because I even, what. A typical user experience that I tend to have with Google, not even the chatbot.
I will go many, many times when it is a nontrivial decision-making process. Go and take a double look at dimensions and visibility and profiles in Instagram or TikTok or social media presence to learn more about what users actually, or real people, actually think about the brand or location or whatever.
Right. So 100% it makes sense to integrate that nowadays, especially with all of this conversation about how Generation Z is using TikTok right. To search for information because it comes from real people supposedly. So there, there's, there's goals that Right. But I think that the critical thing here is how you overlay that in a way that actually makes sense and makes money for Google too.
Bernard: I'm not, I'm not convinced that the, the current model of s g E is the way that. It is the right way to consume this. It's like, I can imagine, you know, there's, there's all the voice assistance, but, and those make a lot of sense to be like, Hey, you know what, what is the highest mountain in the world? Or, you know, those kinds of things.
But, you know, those have already been answering that. It almost feels like, you know, I want to go to like my AI assistant and ask my AI assistant, Hey, you know, do this, what's the answer to that? You know, help me write this like an Excel script. And then search kind of feels like a completely different medium where, you know, I wanna look at all these trustworthy sources, and it does feel like it doesn't quite belong in the search.
Interface. And I commonly find that just the amount of time that it spends generating, you know, I've already moved off of the SERP two, you know that position one or two, right?
Mike: That part. And I think that's a big part of this, this s g E thing. I know they've, they've sped it up in the last, you know, month or so, but I, I did a bunch of scraping cuz that's what I do, and I've scraped like thousands of s g E results, and it was taking as many as like 21 seconds to generate a result.
And a user is not gonna wait that long, even if it's, you know, 10 seconds or five seconds, they're gonna go to the reg regular SERP. And click something rather than waiting for that result. So until they make it faster, until they cash it more, or whatever it's gonna be, to get to the point of having those show up immediately.
It's not realistic that a user's gonna wait for it. And then the ones that do wait for it, and they're like, okay, this isn't that good. They're just gonna go to the regular results anyway. But I think the bigger problem that we haven't really touched on yet, Or to some degree; we have this idea of what it's gonna do to publishers and especially publishers that are in the affiliate space, you know, the wire cutters of the world that are like, here are my reviews on all these different things.
And, you know, I no longer have to go read that article. I can go directly to that product. And so in my example, I, I look for something like, you know, what's the best microphone to record in an apartment in New York City? And it gave me very specific results. And I'm like, oh, cool, I can just read this, and I don't have to, like, go, you know, read three different blog posts and figure it out.
That is gonna have a bigger impact on the internet than any of this, I believe, because that's like the core monetization strategy for so many different publishers. R i p list posts.
Gaetano: Yeah, go ahead. Yeah, I was just gonna say r i p list posts. You know, like even, even, even doing in-house B2B SaaS. Content marketing.
You know, they all wanna rank for modifiers containing tools, software solutions, and platform systems. Very, very rare that you have a ser these days that is not, you know, stacked with g2. What up, Kevin? All the programmatically generated comparison and list pages. And so you, to play that game, you have to do list content.
But I agree with Mike's point, you know, it's, it's gonna be a big shake-up for those affiliate players. You know, I hate, I hate to say it, but, like, I would love to see sites like PC Mag take a dive. You know, I, I hate those guys, right? Like, I,
Kevin: whoa, whoa, whoa, whoa, whoa,
Gaetano: right? Like, me being an in-house Sas, S e o, I've had to pay the middleman as part of my playbook, right?
And, you know, who knows how that's gonna shake things up, and I'm just really curious to see how it all plays out. But I mean, I agree with you guys that it's gonna be a, a, a huge bomb on, on those on those sites.
Aleyda: On the middleman 100%, but not only the middleman. In fact, I have to say that yes, indeed, we can see as the ones that will like not necessarily going to have another option to move forward pretty much rely on those best-off types of queries are the review sites, which are pretty much the reasonable affiliates.
But also, if you think about it, a lot of PLPs of e-commerce websites are pretty much going to list lose their traffic because the PLP experience is being replicated in the, in the s g e in an interface that is highlighting directly product. Pages, product URLs, and coincidentally many of these product URLs, especially for branded queries, Nike sneakers, for example.
If you see the top organic search results right now for Nike sneakers will be the category page of Nike. A couple of other sub-sub PLPs of, of Nikes or whatever. And then merchants offering the same products, whatever, Nike products, whatever. Now, if you take a look at how it is going to affect the Sge, pretty much this PLP experience highlighting products, it's, it's right there.
Goodbye, traffic to help. You click on the product pages, and it doesn't even generally refer, refer to Nike. Website directly, but to the pro's knowledge channel. And, from there, those pages are actually making good use of the merchant center and feed and structured data optimization, whatever.
So welcome to the jungle here. Not only for affiliates indeed but yes, for many merchants and e-commerce websites, especially the very alternative ha, highly established brands that a lot of these branded queries, like assume going to them by default, not anymore, is opening up to many other competitors doing a like a really good job at a much more granular but of the funnel type of actions out
Google is building a marketplace for e-commerce anyway. They're already offering direct connections to checkout for merchants. So that whole affiliate thing is, again, like this middleman situation. It's gonna go away anyway cuz Google is basically becoming a second Amazon try or trying to fight against Amazon.
So, but, but I think the, the kind of core message here, or the probably two core messages here from what we talked about so far, one is that the ecosystem is changing, like as crazy as ever, right? This is part of the biggest change to many of these ecosystems, whether it's publishing or e-commerce, or local search ever.
Number two, when we talk about content or SEO, one of, one of the deep questions that we are all trying to figure out is what is the human contribution of content, right? What are humans actually doing that machines cannot yet replicate? Because a lot of these AI tools, and remember this is only a bit over six months until the like, you know, chat chi PT breakthrough in November.
A lot of these tools are getting really, really good. And even though they still need some more oversight, they're not yet, you know, you can't just let them do the thing and publish the content code. In many cases, we can all see where the journey is going. Right? And I would argue in the next 12 months, when we speak again, a lot of us are probably just gonna publish content with AI tools that need tiny oversight, that we can 10 x the output of good writers.
So then the question becomes like, what's, what are humans actually doing that is still unique, and how can we find our unique skills and double down on them instead of ha being replaced by some ai?
Mike: Well, I think it's more, more Sophisticated than that to some degree. And what I mean by that is what a tool like Sge does is introduces context windows into search.
And effectively, what you're doing is you're creating these very personalized paths because when it's like, here's your first question that you input, and then it's like, well, follow up questions are this, it can learn from what the results were in the pre previous one and do this like very micro personalization for you.
And so what that's really gonna yield is this idea that we have to make even more personal content. So if it's, if you start from a keyword, like the latest, say, you know, Nike, Nike sneakers, and then a follow-up question is like, well, what're Nike sneakers to wear for the New York marathon? And then it's like, okay, it's learned about you.
That is your context. And so it's not just enough to like to have, you know, A page about Nike sneakers in general, like we generally do, and then a page about what sneakers for a marathon, you've gotta have that whole path so that you end up appearing in that pret tuning set that's used for, you know, generating that next result.
And so where we come in is, is doing that, like how do we understand the nuances of the specific audiences that do those specific journeys? Whereas before, we were just so focused on keywords and not so focused on the audience, it really just like turns up the dial on that. Yeah.
Aleyda: I think that puts a good point.
Bernard: Yeah. I think I, I see it more and more, I say it more and more content strategy is becoming more and more the way to think about SEO. And if I'm thinking I. About what's going on in the Google ecosystem. I, I keep, I keep thinking about the manual search evaluator component, right? That's essentially the component where Google employs thousands of, of people to swipe right and swipe left on what looks like a good positive user experience.
And the, the most revised version with the extra e or experience added to the, the P D F, it means that people are looking at things where it, it clearly demonstrates that like somebody has experienced this in their experience, you know, this is what happened, and it demonstrates that. And you're, we're seeing, you know, a lot of manual evaluators swipe right and say yes.
Right? That's, that's a good experience. So I think that we're gonna see that shift play in. And so, yeah, I mean, Us as SEO content creators, we can still leverage AI, and I think that will make us more efficient. But then ultimately, right, it's back to Mike's point, well, what is the, what is the user journey and what is the content strategy that, you know, we are architecting?
Because simply, you know, what are the best shoes for running aren't necessarily exactly the right frame to put it. It's like, you know, I tried seven marathon shoes across, you know, these different marathons, and you know, here's how each of these performed, and the mm-hmm manual search evaluators are, are saying yes.
You know, like those look good. And so we're gonna see, I think, large influxes of that, which then makes purely AI content generated content without human assistant, I think in its current form, just lackluster in performance.
Marie: I think it's a really interesting point about experience. We saw just this week Google released the Perspectives filter.
So that's when you do a search on mobile, you can tap on one of those little bubbles that, you know, it used to just say news, images, whatever. And now Google's figuring out, like, some of those are topics that are really related to your search, and one of those is perspectives, which what, from what I've seen so far, it's showing information from YouTube, from Quora user-generated stuff.
But that's not the only part that Google announced, and we haven't yet seen. They said that they're going to be doing enough updates to the helpful content system, which will reward. Content with experience in search as well, not just for those who tap on that particular filter. And I really think that for all of the faults that AI has and all of the, you know, we're really good at criticizing where language models can go wrong and stuff.
But like it's a brand new technology that it's only gonna get better and better, right? So so, I think, eventually, we will get to the point where language models can very easily answer pretty much any factual question. Even about current events, especially language models that are you know, Google says that Bard is the power of the language model, plus the breadth of the knowledge of the world, the breadth, the breadth of the world's knowledge.
You know there's Bard, maybe not Bard itself, but Google soon will be able to answer pretty much any. Informational query with pretty good accuracy. I, I'm, I'm fairly certain, which means that what's left, if you wanna rank in search, is to actually be like the site or the business that people are seeking out.
And I think, you know, some people are crying the death of Seo. I, I think there are two types of SEOs. There are SEOs who help businesses get found. And that, like, I'm so excited if that's you because the world is changing the, you know, it's gonna be crazy all the stuff that's happening, and those businesses still need to be found.
But there's a whole camp of people that your only existence in SEO is to create content because of your knowledge of how search engines work and then profit from that content. And I think for a lot of businesses, that model is like, it's done because, you know, even if some people still wanna seek out your content, a lot of people are gonna be happy with an AI-generated answer.
So you know, I think, I think we're in for a lot of big changes. And, but yeah, if you,
Aleyda: if your business model where niche websites, affiliates, et cetera, that yeah, indeed, like you, you're up to a challenge there. But otherwise, if you are doing an SEO 100% to a real business with a real product that you provide for a real service, et cetera, et cetera, independently of the interface, that might change, you still need to be found, and there will always be ways to optimize or maximize visibility, and potentially the bar will get a little bit higher.
So the one trick point is that work today will not work very likely in, in those times. So you will need to, well, what is important is to be strategic, as you will mention, to understand how it works, what actually pushes them, or move the needle in the new interface and how you can help those businesses too, to be found for those queries that actually burn business and matter at the end of the day.
Something interesting that you also mentioned, Mary, regarding the perspective filter that I, I think it was Glen Gate that mentioned it over Twitter and I went and double check, is that it seems that the perspective filter is shown on mobile whenever the. The third generative experience is not shown by default when you are testing the interface.
So it's interesting because it seems that Google is, let's say, segmenting the experience and say, okay, this is the type of query where, where it makes more sense to show real users yes. Perspective or, or take. And this is the one that we can easily generate the answer from all the information that, you know, is more, is more factual or, or, yeah.
Information. Let's, right,
Marie: let's, let's go back to this Wired article that I talked about earlier about Gemini that, you know, we don't really know much about because here we are talking about search as we know it. And Gemini, the way that they describe it, is basically a personal assistant that can answer any question and can direct you if you need, you know, if you're looking for a particular website, then it can direct you there probably.
But like, I think. The whole idea of websites, like we're gonna one day talk about this little sliver of history that we lived in with this thing called websites. I, I, I honestly don't think that they're going to be needed for much more. Now I know that's terrifying for those of us who work on websites, but there's, there's such a where, where am I trying to go with this?
That what Dennis was saying is that. What Google is coming out with next. This is not like, Ooh, we had an update, like the helpful content system. This is a whole new product that that Google is coming out with. And he described it as sim, or at least the article described it as similar to chat G p T, but also with the technology that Alpha Go deepminds, alpha AlphaGo used to figure out how to beat like the best go player in the world.
And so they're also making it so that it is really easy to integrate. They talked about how their APIs are gonna be really easy for businesses to use. And so imagine that now, very soon, every business is gonna have the same power that Alpha Go had to figure out, go right from the start to figure out the problems in your industry.
Like that's, that's huge, right? And so I, you know, when I read this, I, I, I feel like how can we still talk about web pages when you know, Google has this, and I had this quote in front of me here from Larry Page in 2000, the year 2000 where he said artificial intelligence would be the ultimate version of Google.
This is the year 2000. So we have the ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. That's obviously artificial intelligence to be able to answer any question, basically, because almost everything's on the web.
We're nowhere near doing that now. However, we can get incrementally closer to that. And that's basically what we work on. And that's tremendously interesting from an intellectual standpoint. So Google's whole goal has been to be an AI-based assistant. And so I, I don't know what point I'm trying to make here other than to say that I think we'll look back at this conversation we had and go, wow, we had no clue what was coming.
Because I think it's gonna be changes beyond what we can comprehend right
Mike: now. Yeah, I think, I think, you know, in the future, we may get to that space, but I think, you know, when we think about things that will tactically matter in the short term, I think we have two key problems that are gonna matter to SEOs that we haven't discussed thus far.
One, if s sg keeps moving in the direction that it is, where there is some sort of, you know, something loads, and then there's something else that loads after the fact rank tracking is over. Like, no one is gonna spend the money that it's gonna take to track millions of keywords and then load something for 20 seconds.
Like, that's just not gonna happen. So content is gonna have a huge problem because if everyone is using these language models that are trained on data prior to 2021, or not even that, like even if they update their models and they start pulling in content from now, like since chat G B T came out, All of that content is effectively polluted to some degree because everyone is using these language models.
And so to that end, we, it, it's gonna be a lot more important that. You're surfacing the content that is actually valuable or new or from experts and things like that. And I think that's gonna be a problem, not just for the language models, but also for search engines. And so I think that information gain is gonna be a really important thing for us all to focus on.
And for anyone that doesn't know what that is, like the, in the audience, I'm assuming everyone up here knows it's this idea that, like, let's say we've got 10 documents about a subject. Okay, which of these documents says something new that the other ones aren't saying? And so being that what all these language models are doing is effectively, you know, rewording what was already being been said, more often than not, or pretty much not at all, they're not gonna introduce new ideas unless they're using a model like you know, any, any model that's using retrieval, augmented generation.
So my whole point here is that We're going to have to focus on, you know, getting clients to be the people that actually say new things when a lot of people are gonna want to go in the direction of like, Hey, let's just fire up this language model and spit out 500 pieces of content. And I, I don't say any of that to discount what you're saying, Marie.
I think it's; it's a really interesting series of, of things that are being presented, ideas and like, you know, it is really us having more of like a futurism discussion when we're talking about all of that. And I'm really interested in what we're all at is going, but I think that there's some tactical things that we're gonna need to solve between now and then, and that's still gonna involve, you know, webpages.
Gaetano: Yeah. To, to that point, you know, I've been trying to get information, gain a part of every single SEO strategy that I do in every client project that I work on, every company that I've been a part of. One example that always comes to mind when I think about this is the query is identity theft protection.
Worth it. If you search that, you're gonna see affiliate players saying, yes, you need it. You need insurance. Yes, you're fucked if you don't have it. Pardon my French. We were the only company that said you might not need it. The truth is, you can do all this crap manually, so you don't necessarily need it.
The only reason that you need it is if you're lazy and you don't feel like changing all your passwords manually. You don't feel like checking your credit report manually and showing them a side-by-side of things that you can do manually without buying the tool. Versus how the tool makes your life easier.
We were the only company that took that approach. Right. So it's easy to just go and copy and try and out machine, learn everybody, you know, out muscle them with links, but you know, when you're up against US news, NerdWallet progressive. Yeah. Good luck. Right. And so to Mike's point on information gain that, you know, that's just one example that comes to mind, but we're all gonna have to do that if we want to compete.
Aleyda: That means that you actually need a, well, potentially a real expert. Right. And have an opinion, a, a unique opinion on, on that 100%. Yeah.
Gaetano: And to Mike's point, you know, clients don't want to do that. It's very hard to convince a client to say, yeah, we saw we offer this, but don't buy us. That's
Travis: Let's keep pulling that thread a little bit.
I know there are a lot of sites out there that kinda wonder, like, what can we do to start future-proofing our sites? Like what, what can they start doing now in preparation for sg?
Marie: I think that anything you can do that demonstrates real-world experience is something to focus on. One of the problems with that is that it's expensive.
You know, I recently reviewed a news website, and that was my top piece of advice was, look, you have almost no original reporting, no original insight. You're very good at aggregating and summarizing. But Google wants to reward original reporting. And what they said was like the cost of that, their, their writers are not trained to be original reporters.
So that, you know, for a lot of businesses, that's gonna be difficult, but if you can. Focus on what it is that you could bring if you took your content and deleted everything that potentially is on other websites or could be repeated very easily by a tool like Chat g, PT, or Bard and looked at what's left.
I, is it something that is worthy of people visiting? And so that would be my main focus right now is to see how we can demonstrate real-world experience to try to be rewarded by Google.
Aleyda: The bar will certainly be higher, and this is something that we need 100% to think of. And also the format that you generate this content, right?
I believe that we have underestimated a lot of video content. One of my main clients right now is, Like it's, it's, it's a player that really leverages, well, YouTube as their main channel. And pretty much what we do in Google is pretty like trying to replicate a little bit of what they have already created over, over YouTube.
Right. And, I have been waiting for this time when there is a much, let's say, natural integration of video results in search. And now that we're talking about information gains and easier ways to find information that can be integrated better. Also, in S G E, I can see video and non-text. Content being that right for visual information.
A lot of how-tos that right now we had a position zero that was one video on that set. Let's, let's see how that plays out with the search generative experience, right? But this is another, let's say, part of the puzzle, right? Like a having real expertise, showing real expertise, and how you show that potentially not creating yet another article that has been covered thousands of times already, but a unique take that if you identify.
That will be much more helpful to a video format to do it in, in, in that format. Right. And that will, will allow you also to allow diversified channels to gain visibility through YouTube, too, right? But yes, I, I, I am a firm believer that, for example, we have underestimated a lot of non-text formats.
With Lens also, we can see how it has been advancing a lot as an input channel, let's say, or platform. But the output will end up being many, many times also image or, or, or video at some point. So I, I, I, I believe that if you ask me, I will be, especially with certain types of clients that have been investing a lot in guides, For very obvious queries how-tos and best off type of, of content to try to sh push them to move forward.
Beyond that, with expert-led content video-related content also for e-commerce clients, a lot of my e-commerce clients have invested a lot of improving PLPs; I'm going to, Hey, let's put a little bit of more work on those PDPs. Now because it's obvious to me that a lot of the traffic and visibility that PLPs used to have will shift on PDPs and that in that expert-led content will fulfill informational intent for many of the brands that they represent.
Marie: If I can add one more thing that I think would be really good to focus on, actually two things is whatever changes are happening with Google ads. I'm not an ads person, but the blog posts that Google recently put out about revolutionizing ads with AI is really worth paying attention to.
I, I can see it where that's actually a big component of search is Google allowing people to. Instead of clicking on an ad, I think you'll be engaging with the event of the website that you're engaged with eventually down the road. And so the next thing is to look at what's available through vertex AI with Google's Vertex ai. Start playing around with it; you have access to several of Google's language models to the Pom a p i.
There are all sorts of things that can be done, and I'm surprised that I don't see more businesses talking about it. Mike, you, you, you're nodding there. Have you played around with the, I mean, obviously, you've played around with the APIs, right? Yeah,
Mike: I've played around with all the things. I mean, you know, all my thoughts are basically from Neil patel.com, so you can just go there.
I'm kidding. I'm kidding. I'm kidding. All right, so as far as like, you know, what we should be doing, I think the first thing is that we should be looking at. The queries that you care about in S G E and then see what sort of formats are starting to emerge there. Like, what are you actually seeing in S G E?
Like is it actually triggering an AI snapshot? Is it, you know, doing the whole big guide thing that it does for e-commerce queries and things like that? You should also be doing that generally in the SERPs, right? Like in regular Google or whatever we're gonna call it. If you're seeing more, like you know, the YouTube shorts starting to pop up and things like that, those are all things that you should start worrying about.
Like, okay, this keyword is threatened. Perhaps I don't want to be here anymore because I don't necessarily, you know, have the content to be competitive here anymore. Information game we'd already talked about. I think structured data is a huge thing. And when I say structured data, I'm not just saying, like, schema markup.
You know, I think that one thing that people don't focus on enough is the idea of semantic triples, where Google is able to, you know, extract information about things from actual sentences. And when they see it across the web over and over, they're like, okay, this is an attribute that hasn't been indicated explicitly in structured data.
So the more that you structure your sentences in that way, the more you're giving them that information. And then, as we discussed before, the idea of the multidimensional search journey. So thinking of your queries kind of in that pattern that we are already using from the people also ask ideas where it's like, okay, well someone searches for this, and they search for this, and then we search for this.
So using that as the mapping for the content that you're gonna create moving forward. And, you know, just thinking about it from a personalized way. And I think a lot of this dovetails into. You know how Google has improved its understanding of content. There's a few, something they, they talked about a couple years ago was like passage ranking or whatever, and it was something that we all were like, oh, passage ranking is so important, and then we all kind of forgot about it.
I think that's still like really important, and it's a function of how embeddings work and so on. And so they're able to understand like the, the, the smaller aspects of the page a lot better. So it's not necessarily you've gotta make 500 different pages to capture these different topics. It's just doing it really well on one page.
And so my last point here is that. The distribution of searches is gonna change dramatically. And I think it's largely a function of, like, we have all been trained to use Google in a certain way. You don't want to give something too specific cuz, oftentimes, you don't get the result that you want, so you start from these head terms.
But with them being better at giving these more specific responses in response to these longer queries, the head terms, the search volume on head terms is gonna get smaller, and then the long tail's gonna get longer. And so, really, we've all gotta rethink our keyword strategies and alignment with that so that we can, you know, go to where users are gonna go as a result of this new user
And part of that is also sourcing how you understand what to optimize for, right? This whole idea of optimizing for, I mean, a single keyword has been that for a long time; maybe the topic is a bit more viable. But what we often forget is that AI goes both ways, right? AI gives answers in search, but we can also use AI to be more effective.
And one part of that is to just source better topics to create content for or pages from our users, right? So there are many tools out there now that allow you to upload tons of PDFs as you could just go out there and use. Transcripts from sales calls or customer support calls, throw them into a tool and then just query and ask what are the common pain points.
In fact, I have done that with a, with a, with a customer, and we've created some content completely outside of any search terms or better-set search terms, and had zero search volume. And lo and behold, when we create content for that kind of stuff, all of a sudden. Those pieces of content get a lot of traffic.
So I think we also, we should like also use the opportunity to throw overboard some of our old kind of methods and workflows and think about what AI AI enables us to do now, right? Another example is lots of these chatbots that everybody has now; those are interactions with customers and not a lot of people log the questions that customers have and what they first ask for.
So we can substantially optimize our experience. I think that websites have become, most websites have become incredibly boring over the last 10 years. It was barely innovation. They all look the same, and not enough people are thinking about how we can provide something truly unique. Some interactions, maybe like you can, like one example here is a Vimeo or script, right, that allows you to.
Edits video or audio, like text. And so they brought that to the landing page where you can try out little features of the product on the landing page in itself. And that is a great way to not only get people bought in before they even sign up and increase their excitement about the product but also to stand out with the unique experience.
And I really think that we need to innovate more about what a good web experience actually looks like. And AI actually gives us some great tools for that.
Aleyda: I mean, we definitely will need to, let's say, reeducate clients, right? It has, it has been quite easy, let's say, especially if we take. Generic search volume as a reference, right?
For say, okay, it, it has this potential of traffic because it has this average search volume per month, this trend, whatever, whatever. But if, if it is going to become so much more segmented and spread through the low, low, low long tail, imagine now having to, let's say, invest in so many different pieces of content rather than, than a single one.
For which many times it was, it already requires a bit, a little bit of like good influence right there. Especially like, for example, if you work with commerce websites to invest in informational content, oh my God, how am I going to invest in content that doesn't bring me money directly, right? I, I believe that we will see a little bit of this need.
Elevate a little bit, our capacity to communicate and, and show value and show the value that is going to be, be a little bit, especially at the beginning, much more complex to show on the potential on, on, be more strategic on, on the developing content that requires expertise and unique takes, for example.
And that is so spread out, and it's not a single long-form guy trying to rank for the very popular top of the fu top of the funeral query that everybody wants to rank with. Right. And you send a bunch of backlinks to improve popularity, and pretty much that's it, right? No, it's, it's not going to be so straightforward anymore.
Gaetano: Yeah, I, I agree so much with that. You know, like, would you rather create demand generation best practices or how to allocate my demand generation spend across paid channels? That's a real problem that demand generation managers have at companies. And a part of this is also gonna be about creating some demand, like SEO has, you know, traditionally been a demand capture channel.
You find volume; you create content; you rank, you get traffic, and you pull it all in businesses, Now gotta think a little bit omnichannel, right? How are we gonna create some of this super long-tail content and get it rolling on some of these other channels? And use paid and organic marketing strategies together to go reach our target customers.
That's the way I see it.
Travis: points. And we have seven minutes left, so I wanna get it to an audience question, which also kind of sums up some of these like privacy concerns. But Chris asked; he works in the biomedical industry, and their content is pulled from their page and reworded and s G E oftentimes incorrectly.
How should they and others combat this?
Marie: I think that Google is. Aware of the issue. And they've been very quiet about it. You know, they've said, well, you know, the websites are on the side and, you know, maybe just like with featured snippets, you know, they took our, our content and then we had a link from it.
And sometimes, that was beneficial, and sometimes those featured snippets stole all of our clicks. I, you know, I think we can continue to complain, but I don't know that there's much more we can do beyond that. But then again, it's possible that Google is doing nothing because this is just an experiment, and what we see go live, we'll have a language model in there rather than like actually have truly AI-generated answers as opposed to stitch-together answers.
So I'm not sure what their recourse is. I, I can't see how Google could go live with just blatantly taking content from everybody's website the way they are and not continue that way. I, I don't know.
Mike: So I, I'm with Marie, like it's not something that we can easily solve overnight, but I think it's an indication or another indication as to why the s e o world needs technical standards.
And it's also the sort of thing that if we were to establish that, we could then probably, you know, enforce that in different places. And what I mean by that is how is it that the biggest search engine is defining what robots do? Text is, that makes no sense. Like why do they get to define it like they are crawling the web?
We should be like, the web should be saying, here's what robots do. Text is, and here's what you need to adhere to as a search engine. And so if we were the ones to define the technical standards, we could say like, oh, robots, sat text now has a rule that means you can't use this copy for a large language model or for generating content.
And being that we don't have that level of control, they Google can do whatever they want.
Marie: Like, can things be removed from the language model? Like if something was whatever, Google trained it on. I mean, we know that Chat g PT was trained up till 2021, right? We don't, I don't know if we know when Lambda or or Palm or anything was, was trained up to, but if something's in there, I don't know that you can just remove
Right? So from the language model itself, no, you can't. But for these results, they're taking, you know, whatever they have in the rankings, and they're saying like, Hey, use these, these pages to then create this result. And so what we, what could be, like, the way I imagine this is like you have a Metabo tag set dot meta robots tag that says like, no generate or something like that.
And then it's the sort of thing where it's like, okay, well, this came up in the rankings. Cool. We've gotta omit this result for the, for the fine-tuning for the answer. So that's what I'm saying, being that we don't have the ability to define that we're just like left to whatever Google decides they want to do.
Kevin: I think it needs to go even further, right? I think there needs to be, and maybe this is something that governments of the world have to regulate, but there needs to be transparency about what data sets these models are trained on. It's not just Google that uses the web and index and rag too to serve AI answers.
It's also chat chip, PT, and all these other companies that are working on models and rights. So you could, for example, technically you could say, okay goo like, you know, in, in Robots Cxt, you disallow Google from calling your site, but then Common Crawl crawls your website, and that's then being used to train a certain model which then uses your, your data.
So I think that's really one of the few areas where I think we need more regulated transparency so that anybody has a choice to say, I want my data to be used for training or not. And, and
Aleyda: at a much, much more granular level, like, so, for example, as it was a good example, the one that Mike gave regarding the feature snippet, right?
For example, for the max snippet or the no snippet we can, we can specify and provide rules to Google to prevent them from showing them in certain CIR circumstances, right? Right. We need totally, and I, and I agree with you, something at a very granular level to be able to select even from a topical standpoint, right?
I'm okay with this type of topic, but not okay with all the types of topics, things like that. Mm-hmm. So it will be, it will be great. I already have a client who's very concerned regarding Google using their data. They realize that if, if they block it, I mean, there's no hundred percent. Let's say straightforward way to do it because they rely on Google search anyway.
So even if they block through the chat, dpt, open ai, common crawl, whatever, it won't necessarily affect this other scenario. And it's, it's tricky, and it just, we don't have means right now, we don't have an answer. And yeah, it would be great that we are always going to go ahead with this new interface, a new way of searching.
At least they provide. Channels and configurations to control that to happen in any case, like the, the, the type of conversation that I have with this client is like, imagine that if you are able and allow and block, this information is Google and all the language models not going to be able to get this information in any other way.
Right? And this is where the expert-based content also comes in handy. If you have unique information, unique content that is actually so valuable that you can control and can damage the experience of people if you're not able or willing to give it. Right? So it's put portal here, but in their case, it was like, no, pretty much they will be able to get this same information elsewhere.
So it doesn't make any sense for them to try to block it in any case. Yeah.
Travis: That's a good place to kind of wrap this hour by really, really fast. Again, thank you so much for your time today. And I wanna give you everybody a last some last words. So a lady, if you kick us off with any kinda last words before we give everyone their day back and then Bernard and kind of round the horn.
Aleyda: I think that we are in this moment; that is, we might be a little bit scared, but it's also very exciting at the same time. My recommendation will be to keep things strategic. Look, take a look at the SERPs. That is the best advice I would see generic, overall applicable to everything, especially now.
Take a look at the SERPs for the most important type of search behavior that you tend to follow and and profit from. And act accordingly to what will keep the visibility and clickability up of your clients or branch results and experience. And let's, let's. The, and also to what we can do in many different ways.
Maybe we become creators all of a sudden, or it can change in so many, so many different ways. And it's certainly exciting. I, I believe that in a year, if we repeat this it will, it will be interesting to see how our takes have changed and how it has everything evolved. Maybe half of us, we don't call SEOs anymore, but something else, I don't know.
AI optimizer, I dunno. Let's see how it goes.
Bernard: Yeah. Aleida, you, you wrapped it up very nicely. I think. The way that I would frame it just for everybody out there is that Google has been cannibalizing its ecosystem since the start of the search. This is just a further cannibalization of what they're already doing.
You know, in the past you would, you would've searched weather, and you go to like weather channel.com. Now Google gives you the widget. Right now people are searching, you know, what are my, like the best shoes for marathons? Google's generating some ai. All of this has been happening, but the search has still stayed strong.
And so I think that you know, we, we need to stay on our toes. We need to understand that the medium is changing and then just be aware that search is still likely to stay around in some way, shape, or form. And we, we just have to evolve to meet where the technology keeps heading.
Mike: Guy, tell me you're going in alphabetic
Gaetano: court. Yeah. Yeah. I mean, great panel. Thanks to everybody for making it happen. I, I learned a lot just listening to all of you guys and me; I kind of agree with everything that's been said. You know, with Bernard's points, the latest points, Mike, Marie, Kevin, you guys all made a lot of compelling points.
A lot to think about. I don't think SCO o is dead. I think, you know, we'll be around, and maybe we'll be called AI optimizers, but at the end of the day, we're still constant strategists and business strategists, and companies need people like us to grow and and to get results. And so we just gotta level up and stay on top of our game.
That's what it comes down to. Yep.
Kevin: I agree. I think, you know, it's, it's so hard to say what happens. We don't even have any traffic data. It's not even life in the public. I mean, so, that is. But I've been thinking a lot about how do you de-risk the future. And one way of de-risking is of building email lists or building ways to reach your customers and potential customers faster without Google being in the way.
So, I would optimize for three things. I would optimize for email signups, making it super simple for people to stay in touch with you and for you to broadcast to an audience. Number two, I would optimize for direct traffic. So how can you just get more people to visit your site directly and get them into a habit of coming to your site directly?
And then number three, how do you optimize your web footprint? How do you optimize, you know, how do you track how often your brand is mentioned, and how do you just get on as many sites as possible instead of relying on one search engine, essentially traffic? Marie over to you.
Marie: Yeah, I; I would like to just encourage everyone to be excited and to try the new stuff that's coming out.
I think whenever there's a change, it's natural for us to be fearful, and it's natural for us to criticize. You know, any new change in, in history caused all sorts of, of criticism. And what we often see is, you know, leaders in our industry or, or, or people, you know, we're really good at pointing out the things that these tools can't do and that they're bad at.
But very soon, businesses are gonna be clamoring not for critics of ai, but for people who actually know what to do with it. And right now, that's none of us. And so, for me, I'm using Chachi PT to, like, make recipes. My husband, who is not technically challenged or technically inclined at all, uses it for gardening.
Like how far apart should the peppers be planted and, and stuff like that. And that might sound like, you know, okay, we're using this silly tool, but in doing that, when I wanna use it to help me understand the search console keywords and figure out which one's actually declined after a Google update, I'm using chat g p T to do that now, you know?
And so we're in the early, early stages of something that's going to be, I think, incredible. Sunder Phai from Google said that AI technology is. Going to have more profound changes than electricity or fire. Like that's big, you know? So so, I don't think we know what's coming. I would say play with the tools, play around with whatever's in Vertex ai.
Like I would say go there now and just explore with it and, and see what, like, jumps out at you. And then just keep paying attention to what's happening. And I, I would encourage people to pay attention to the positive stuff because we're gonna hear a lot of negative because especially the media is hurt by what's happening.
These changes are, are very damaging to a lot of the media. And so we're going to hear the negative news. So I would encourage you all to just be trying and be excited about
Mike: what's coming? Yeah, boy, Marie, definitely stay curious, stay excited. Th these are really, you know, interesting times.
Like I said, after five, six years of SEO being really boring. So it's a lot of really cool stuff to play with and learn more about. Also, demand more from your tools. You know, I, we didn't really get a chance to talk about this, but you know, all this stuff is very much built on, like the semantic analysis, and all of our SEO tools are still doing lexical analysis.
It doesn't really make sense that. Still like in the stone age on that because Google's been in more in the semantic realm for the last like 10 years, right? So demand more from your tools. If you're doing anything with generative ai, you really gotta lean into content strategy because, you know, a lot of people just kind of do things left to their own devices, and then you've got all sorts of content problems all across the board.
So you really gotta get your organization to be like, here's our content strategy for generative ai. Here are the tools we're gonna use, here are the workflows, and here are the governance models. Otherwise, you're just gonna make a big mess. And then the other two things that we talked about, structure data, information gain, and yeah, check out AIPRM.
That's all I got.
Travis: Awesome. What a fantastic conversation.
Director of Marketing, Clearscope
Join thousands of marketers who receive our weekly emails.
We share content marketing best practices and SEO strategies from the brightest minds in marketing. You’ll also be the first to learn about and join our next weekly webinar with the industry’s best.Join today