Episode Transcript
[00:00:00] Speaker A: Welcome to 10 Speed Sessions. I'm Nate Turner, co founder and CEO of 10 Speed, and today I'm joined again with my co founder, Kevin King. In this session, we're going to be discussing the overlap between traditional SEO and what's now, I think, sort of settling into being called generative Engine Optimization, but AI search, LLM optimization, still a lot of different terms floating around, but just chatting about like, what's different, what still works and then just kind of generally through the lens and framework of why people with SEO expertise are well positioned to lead this shift. So very specifically going to be going through just seven areas where we see both overlap between kind of traditional SEO and geo, as well as some of the ways in which it's needing to iterate or it's changing for the better in most cases. So, yeah, with that I think we can go ahead and jump in. So the first one is audience and prompt behavior research. So, you know, kind of going from keyword volume and intent to really prompt analysis and conversational query tracking.
[00:01:12] Speaker B: Yeah, it's a big one. I feel like the list is actually in a really nice linear order for how people should be approaching or how teams should be approaching what they should, how they should optimize for LLMs or AI search.
So yeah, I guess like the gist of this is that traditionally SEOs or content marketers have done keyword research using tools like Ahrefs or, you know, whatever your SEMrush. SEMrush keyword planner back in the day to get keyword volume right, understand search volume around the things that you want to rank for in traditional search, obviously break down intent along with that and understand exactly the type of content you should be creating to map to that intent to be successful.
But now with the shifting world of AI prompt analysis and conversational query tracking, which is the way we put it is not necessarily the full shift, but like the additive thing that you should be doing.
Because the one thing too I want to call out in this is that like search isn't gone and there is keyword data that does map to that that you should still look at. So it's pairing these two things together.
But I think the, the that's the big shift is that or expansion of what we're already doing is, is that, hey, now the skill set that we have for keyword research really well maps to this new like conversational query tracking and prompt analysis, which the tools that have been around like ahrefs have some functionality. I'll stop there because I could probably continue to Ramble.
[00:02:45] Speaker A: Yeah.
[00:02:46] Speaker B: Without some gut.
[00:02:47] Speaker A: I think that it's a good point that it is a continuation because it was before very one to one. I mean you would literally, you know, create pages or you know, content that was mapped to a very specific keyword. And then I think you had that progression into, you know, understanding the semantic aspects of that to where it was shifting a bit more to the topic and intent era.
But now to your point, it is shifting much more into that full question query and less of trying to assume or intuit the intent behind a keyword phrase. And really there's a lot of different questions and queries that are coming out of that.
[00:03:36] Speaker B: Yeah. And I think the, the bigger picture challenge or.
Yeah, challenge to address that. I think most people are wandering around, especially since we put this in number one, we get the question a lot around how do we understand what we're being surfaced for within ChatGPT or perplexity? I don't want to discount also AI overviews within Google Search because that's a whole other thing that doesn't get reported on within like Google Search Console or anything like that. And has like a different way, there's a different way to like look at that keyword data or prompt data.
So there are tools like this is the practical application of like how you can understand that. Like AHREFS is one we're getting behind pretty strongly. They have a great brand radar tool.
We're evolving tool because it's really fresh out of like beta and everything, but where you can, you know, get really clear visibility based on what they have access to from the, through the APIs around what you're being surfaced for or what your domain is being surfaced for. And they will break out individual queries or long tail queries or prompts that seem to be surfacing your. Your brand. Yeah. So I'll stop there again.
[00:04:52] Speaker A: Yeah. And I, I think the other point you made earlier that I wanted to, to just highlight real quick is that there is, I think you maybe didn't use the exact word, but you said like it's additive. And I think that that is really critical because that is a lot of what we're seeing is that there is, there are still a lot of, there's still a ton of search on, on Google and other search engines.
So obviously it has been shifting and there's sort of cultural shifts and all that. But like there is still a ton of search volume, there is still data around that.
But now in addition to that there is all these other queries. And so I Think that's a big part of, like we said, that skill set that SEOs have very much maps to them. It's the same process that you're really trying to understand the audience and the research and know the queries and all that to then be able to work towards it. So I think there's a ton of overlap there. But a really good point that it is in many cases additive and not just purely replacing traditional search.
[00:06:03] Speaker B: Yep. And I think the, to put a really, like, good point on it is that it's, it's possible to see where you're appearing in these things.
I think that that's been for at least I'd say the past six months to a year. Like, felt like a black box. It's not as much anymore. The one caveat I'd put on it is that with conversational queries or prompts, they are the new like definition of long tail. They can be. There's so many different variations that like, the data you do get will be somewhat directional in the sense. Like, they're not going to have every single variation possible for your brand just yet. Like, it's still going to be an evolution, but you can get a really good sense of like, what it is. And the one thing I would call out too is when you do that analysis, which we've been doing a lot of, and you like, do some competitive gap analysis or just look at what the conversational queries are that are servicing your brand in, in those platforms, you will see like nine times out of ten, it is very well mapped to your top traffic posts or on page.
But there seems to be a correlation there, which is kind of reassuring to some extent, but something to look out for. You're not going to get any. I don't think you're going to see like major surprises when you go and see like what people are finding you for in those platforms. Yeah, yeah, yeah.
[00:07:26] Speaker A: I would say the, the last piece there real quick is, I do think that there's still a decent amount of that, like you mentioned the last few months it felt like a black box. I think it is still for a lot of people because the majority of these tools, including that brand visibility radar, they're fairly cost prohibitive for a lot.
[00:07:48] Speaker B: Of people, for sure. Good point.
[00:07:50] Speaker A: You know, smaller companies, I don't think I've really seen anything sub 300amonth.
[00:07:55] Speaker B: Yeah.
[00:07:55] Speaker A: You know, so I think that's.
There's going to be a lot of companies that, that's just not important enough to them.
Certainly something we're able to to surface for clients and whatnot. But do think that's probably somewhat of a factor in that currently as well.
[00:08:12] Speaker B: Yeah, which also is a good point. Cause that is like there is cost to that, it is expensive to get that data. I'm not only for you to get access to that for the companies that are building these new tools, but I want to give some at least practical advice like still, manual prompt analysis is very, very, very effective similar to how you would do. I mean we still to this day do manual searching in every search engine when we're doing competitive analysis to validate something we saw in a tool. So definitely start there. And a good way to do that is to look at the things that you have tried to prioritize from a non brand visibility perspective and play around with some conversational prompts and see where you fall, if you fall in there and if you're getting sighted, if you're not.
[00:09:00] Speaker A: And then you can potentially follow on after the manual query response to just literally ask why.
Yeah, why you know your brand or the brand you're looking for wasn't included and yep, you know, inquire a level deeper as well. So totally, yeah. Number two, crawlability, indexing and bot access. So you know, shifting from ensuring googlebot and know search engine bots have access to really opening it up to you know, LLM crawlers and AI specific agents.
[00:09:37] Speaker B: Yeah, I think this one is pretty foundational for any, you know, technical SEO or technical marketer I guess you want to call it is just it's always been important to make sure that your house is in order and so to speak, that site structure is good. Right. In all of the fundamental things that we had been given as a tool set from back in the day, like robots, Txt and all that stuff are, are set up properly.
And we have learned especially over the past, I'd say month or two there's been a lot more people talking about analysis that they've seen. Like if you pull your crawl logs you can see similar to how you used to see chat or Googlebot or if you wanted to go that level deeper you can see Googlebot and what pages it's hitting. If you do that now you can start to see some of these AI chat tools in their bots in your server logs. Just a way to validate that that's how they're accessing their site, which is very similar and probably actually almost exactly the same as a search engine bot. Which just means all the things that you would hope to or prioritize from a technical SEO perspective, you would be prioritizing for LLM visibility, which again just reinforces the point that this is just an expansion or additive aspect of what you should be doing to make sure that you are set up for success. And it's not some other thing that you need to be doing. It's just accounting for these new tools.
[00:11:05] Speaker A: Yes. And similarly, and kind of what I mentioned at the beginning, it's not, oh, this is this whole totally different thing. I mean it is, but it doesn't require.
Now I need to go figure out from square one who can help me with this, how do I do this? It's like you should just be doing this anyway from anything you were doing from an SEO standpoint. But there are some nuances. There is the LLMs Txt, there's some different schema things and just on page formatting that are helping and making a difference. But ultimately it is the basically the same process, same skill set that's, that's really leading a lot of that and like you said, kind of having that house in order to, to make sure that's being indexed correctly and favorably.
[00:11:59] Speaker B: Yeah. Again, no index on index. All the things that we've done user. The one thing I want to call out on this one too because this is one of the more, I'd say probably more straightforward technical ones. I don't think we actually have any other that like bucket in that is just also UX and page speed and all the things that we've been like really focused on for the past probably even 10 plus years is like super important. More important than ever now is making sure everything is in good working order there.
Server side rendering, clean HTML formatting, all that.
[00:12:31] Speaker A: Yeah. And I think that it's the same thing. It's like we've never in the last, you know, past few years with clients or you know, past in our careers. Technical SEO was never viewed as like this is a huge growth lever. It, it was let's not have anything that's impeding performance.
And so I think it again just sort of follows that same rule. Like this is not like, oh let's go add some schema or add this txt file. And all of a sudden we're going to be in every, you know, every LLM response. That's not, that's not what anyone's saying. But right there are things that can be detrimental to, you know, actually showing up and performing. So yep, number three, reformat and refresh for AI friendly structure. So a little bit of what I was just saying about the on page, but certainly more than that. So you know, from publishing and then forgetting about it to you know, actively refreshing, consolidating and structuring.
You know, in this case now also specifically for chunked retrieval.
[00:13:38] Speaker B: Yep. Yeah, I, this is one of my favorites because it goes back to, I think the time that I ended up getting into SEO was at the time when like this, this thing specific, this, this actual like tactical application, however you want to put it, was the like starting to become more and more important. People just put out content. It wasn't like structure mattered or anything.
And it's even more important than ever. So what that means is LMS love to have really, really concise, as you said, chunkable content which we can explain a little bit more of. But that means clear H2s, clear H3, bulleted lists, you know, numbered lists, call up boxes, those types of things, anything that like helps the user from a scannability perspective. But also those bots, because they want to, they can't extract that information very easily to be incorporated into an LLM or generative. Generative response.
[00:14:37] Speaker A: Yep.
[00:14:38] Speaker B: So it's super important and it should be super important for you to think about that. Anyway, it's like think about it more as how can I make this piece of content as user friendly and accessible as possible. Yeah, that's really what you should be thinking of even if you weren't considering an LLM.
[00:14:55] Speaker A: Yeah, absolutely. Yeah. In the chunk retrieval aspect is interesting and I won't try to speak beyond my technical abilities but you know, as it is, I think sometimes there can be this notion that, you know, any sort of bot from search engine or AI is indexing.
You have a 2,000 word blog post or everything on your homepage, it's indexing it all in the way you created it as one unit. And that's very much not the case and really wasn't the case already with search, but in this case breaks that out into chunks and then in the vector database can look at that differently and kind of pull and parse how it wants to have all of the right pieces and say, well this one piece of this page or this one little sentence in this blog post does apply to the question being asked here.
And it can pull just that to surface that. And so obviously we talked a little bit before, we've seen that already with the way it could handle AI overviews.
Going back further, it used to be that the meta description was what showed the blue links and then it started to understand actually deeper into Your content. This is what speaks to specifically what's being asked and we're going to show that as the meta description.
So there are definitely pieces of that that I think are familiar for people.
This is just happening at a much larger and more technical level that I think, to your point, requires a lot of refreshing and reformatting content to still work for a search engine, but now also AI friendly.
[00:16:44] Speaker B: Yeah. And I think to that, that point as well, which if you pair that with the audience and prompt behavior research, you get into a nice, especially if you have like a nice cadence for how you guys are doing it, from whether it be a quarterly perspective, a monthly perspective, depending on team structure, who you're working with, you get into a nice rhythm of actually like trying to test and see how to get your content cited in those things. So if you're doing regular audience and prompt research, either manually or with a tool, and then taking that research and like looking, hey, okay, this got cited this way by this competitor or whatever, what have you.
Now let's go and maybe restructure our piece, maybe to beat that piece, to be more better positioned, improve the language, elevate it, make it more direct, make it more concise.
All those things just test around. That's going to be really impactful in getting you improved citations and visibility, but also to connect it back to what we've done. That's what we've been doing on the organic search side of things forever.
Yeah. Competitive analysis, looking like, hey, how did this brand win on a term that we want to win on?
Go look at the content they created and you try to always out beat them. You want to be unique. The thing I can't emphasize enough with this because I don't want it to sound like you're just going and copying what people are doing is like people need to be surfaced for certain things, so it makes sense. But just try to obviously make it as competitive as possible and differentiated, I guess is what I'm getting at.
[00:18:20] Speaker A: Yes. That is 100% the process, feedback, loop, flywheel, that's been core to SEO for, for a long time. And, and I think to your point, like, one thing that's been, that I've thought about is, you know, the overwhelming majority of companies and people in general are just trying to rank.
[00:18:44] Speaker B: Right.
[00:18:45] Speaker A: And not many people get the experience that we had at Sprout and have had with clients as well, where you're literally, you know, flipping with a competitor for the number one spot.
[00:19:00] Speaker B: Right.
[00:19:01] Speaker A: You know, back and forth where you're like, Wait, okay, what did they change and why did that Let them overtake it and then you change something to take it back. And truly seeing that very direct.
For one, URL competitiveness of it is a bit of a more unique use case, but is generally the concept overall the lack of progress. If you're not trying to get replaced by someone in the top, you're just trying to get up to the top. The lack of moving up is also a signal in any case. Or for an AI, the lack of being mentioned or cited is also signal of like, okay, well whatever we tried didn't work. And moving on to the next thing.
[00:19:54] Speaker B: Yeah, agreed. And one last point on that, because I think it's important is the nurturing of existing content has always been important. It's even more important than now now. So I would say because I think we still, you know, encounter questions of like, why are we spending time revisiting this piece or that piece?
It's for these reasons specifically. It's because if we've chosen what we're going after strategically, we've built with intent, it maps to ICP and all that. Like, we should spend time making sure that it's always the best piece that it can be, and especially if it's further down the funnel, capturing high quality traffic like it's doing something for the business, then it makes sense for you to continually reinvest time into making sure that that's good because all of these things will play out successfully for you.
[00:20:42] Speaker A: Yes. Yeah. And your, your point is, right? Like, that's been the case. Yeah, yeah. Whether or not people have done it. Because I think, you know, the. We started with saying, you know, publishing and forgetting, which I think even in the last year has still been highly prevalent for a lot of people that they never do go back and refresh or tweak it to make it perform better.
So certainly a lot there.
And yeah, I mean, I think so much of what was already being done from an SEO skill set standpoint, especially once like featured snippets started, that really put a huge emphasis on like bullet points and tables and like clear structure and formatting and sometimes even like formatting a direct answer to the query early, you know, that would drive that. Like a lot of that is a direct shift here. Like the stuff you were doing already from an SEO standpoint is now applying directly here, just with a different bot and retrieval system.
Number four, we have topical depth with clustering, and I think this is a big one.
So moving from pillar content and topical authority to supporting AI's need for entity rich multi angle content structure.
[00:22:18] Speaker B: So yeah it's.
[00:22:19] Speaker A: Let's jump into that.
[00:22:20] Speaker B: Yeah so if you're listening to this or watching this then you have some existing knowledge of topic clusters historically also called hub and spoke probably made Most infamous by HubSpot but also like adopted by the entire industry as a way to make sure that you are getting as much visibility which would be creating a big epic pillar piece. Some people call them skyscrapers for a variety of reasons that connected to going really in depth on a topic exhaustively so and then creating a bunch of supporting pieces that go deeper on individual sections within that piece making sure everything's internally linked in like creating this kind of like really nice hub and spoke cluster of content right now that is actually at least the general approach to that is still is now more important than ever. We've been told by Google and there's been some good articles written about it when they released AI overviews that they use a thing they specifically call query fan out which when broken down and we can link to some resources but it is almost.
It's very similar to what we've been doing for topic clusters and if anything it's coming now directly from Google which is actually very reassuring. I'd say the slight difference in the way that you should probably be thinking about it is like really making sure that individual pieces of content are pretty concise.
I think we had tended to go over. We'd over index on like being super exhaustive even on the supporting pieces of like here's your pillar that's super exhaustive. Here's your individual individual piece hits super exhaustive. This is that like for everything we've talked about before it should be chunkable scannable really get to the point.
So like your approach to the content is.
Is probably has got to shift a little bit but like the general direction of how you're going about it is very similar. Yeah yeah the.
[00:24:26] Speaker A: The pendulum is definitely swung a bit there. You know where before it was hey your competitor has a piece that's 4,000 words so you need to write one that's six or 8,000 words.
It's all about just outdoing them to hey, a lot of this stuff that's overly verbose at this point or unnecessarily exhaustive does need to actually become a bit more concise. And I think that this is interesting and this does also kind of the understanding chunk retrieval and some of that stuff certainly plays in here.
But I think it's probably worth talking A bit more about the concept of query fan out and what that means.
[00:25:16] Speaker B: Yeah. So basically you have a topic that you want to cover, which is the way you should be thinking about it, not necessarily just a, A keyword, something holistically that represents like the, the, the product that you're selling and the themes and, and value props that it offers or problems that it solves for your icp, and then creating individual pieces of content that break out from that topic that support specific subtopic subtopics, different angles and, and have different use cases. So I feel like you'll probably come up with a better example on the fly than I will.
But basically being somewhat exhaustive in breaking out all the different ways that someone could possibly use your solution would be a way to think about it in very concise manners.
I'm blanking on a good vertical use case example, but.
But yeah, it would be basically breaking down things from multiple angles to make sure that you're supporting a broader topic and then obviously many topics because you can approach that for different ways. So, like, whatever the solutions are, that your product is still speaking very generically about it.
[00:26:34] Speaker A: Yeah, but I think it's fair. Like, again, a lot of this had already shifted from we're directly going keyword by keyword to try to rank, to understanding that if we want to rank for a topic, that topic may have dozens, hundreds, maybe even thousands of queries.
[00:26:56] Speaker B: Right.
[00:26:57] Speaker A: And iterations that could map to that. And so that concept is not new.
But then applying that now to the topic can and does also then fan out into a lot of different, longer queries, which can have even more variation. And again, still understanding intent behind it and trying to solve that problem. But ultimately a lot of that is not new. We've kind of already gone through that shift from one to one keyword to topic.
[00:27:31] Speaker B: Yes.
[00:27:32] Speaker A: But now has again, sort of a different application because of how query behavior has shifted.
[00:27:39] Speaker B: Yeah. And I think what's still important is making sure that no matter what you cover, even if it's this, like, concept of fanning out, is that it's still relevant to your icp. This, it could be interpreted as, like, cover it exhaustively in every possible angle. But, like, if you did that, you might hit on topics that, like, are actually not super relevant to your audience, potentially if you go too far down the rabbit hole. So it's really making sure that it's really tight maps to what you guys do or what the website is offering and, and kind of go from there.
[00:28:10] Speaker A: So for anyone who's Making you know, a new strategy or you know, looking at building a roadmap for the next quarter or anything like that.
Anything.
Any specific nuances you'd add there, you know, related to this?
[00:28:28] Speaker B: Yeah, I would actually because it kind of ties to another one later on down.
Down the list which is be thoughtful of like the intent and where things maybe historically would be landing in your funnel. Right. That like classic awareness, classic middle funnel, classic bottom funnel stage.
Probably prioritize things that are a bit further down the funnel mainly because we've seen that like lots of that type of content stuff that has higher intent versus like educational like broad. Broad intent gets surfaced a lot more incited incited in in LLMs.
But always just be mindful of balancing all of those things with what you need to educate your audience on.
So I would say that that's a big one.
That's probably one of the better ones to.
[00:29:21] Speaker A: Yeah.
[00:29:21] Speaker B: Give from building a roadmap.
[00:29:23] Speaker A: Yep.
[00:29:25] Speaker B: Cool.
[00:29:26] Speaker A: Before moving to number five, just wanted to mention we do actually have a number of resources that very much relate to this that will all be linked in the show notes as well. So yes, we'll make sure we're not necessarily like calling them all out directly here but some good stuff generally from Kevin Indig Aleda.
[00:29:50] Speaker B: Yep.
[00:29:51] Speaker A: Kyle Poyar, like quite a few folks there have written up some good stuff and so we'll, we'll make sure to share the links there. Yeah. Number five, make content citation worthy in AI answers. So kind of moving from just being included in the answer which was really the, the initial focus to being attributed or cited as the trusted source.
[00:30:16] Speaker B: Yeah, you'll see it too because that might be a little confusing. Is that in ChatGPT depending on how you're using it. Because you could be using it for just like interacting with the main bot or actually searching for the web. Sometimes you will quite literally be called out in that response or it'll just have what would be similar to a direct link with a source and all that.
And that's going to be one of the more important things to continue to optimize for because if you're getting cited in the answer there's only like a really specific type of query which is likely more bottom funnel that you're getting like cited in which is still important but you need to be optimizing for kind of both these things. But I think the thing that's going to end up becoming even more important is like the trusted sources because when now people have so many, so much more experience using these Tools I do think similarly to like anything that they've used from searching on the web. You want to understand that like the information you're getting is accurate or being attributed to a good source. So that would be the big thing that I think now we want to be optimizing for.
[00:31:29] Speaker A: Yes.
And yeah, I think again this is an area where.
[00:31:36] Speaker B: So much of the.
[00:31:38] Speaker A: Fundamentals are a direct transfer or the exact same of what we've been doing already on the SEO side.
So again this is not something where kind of have to like freak out and figure out how to be cited. It's a lot of the same stuff.
Everything that was happening with eat, which then became eeat.
Right, right.
Was all I think a very good primer on this which was, you know, so what if you're like the 8,000th domain to write about this topic? Like why, why should yours be surfaced?
[00:32:17] Speaker B: Yeah.
[00:32:17] Speaker A: You know, and that's where a lot of like first party data or unique data comes in. But also like who is the author?
Again a lot of the fundamentals on your site like do you actually have author bios and credibility signals and things like that that are going to influence that? It's basically the same but again there's other stuff there and how it's pulled in and how it's cited and not exactly the same but there are many, many shared building blocks between the two for sure.
[00:32:52] Speaker B: Again it goes back to the whole on page or more crawlability, indexing, bot access one at number two which is just being. Making sure that your, let's just say your expertise house is in order. I mean you still go to websites nowadays and you have you know, brand, not brand team but like it'll just be like team brand name or whatever as the author to your point, have good authorship strategically make sure that you're incorporating people who have a footprint. This is like this is the harder thing to optimize for but everyone should have been driving towards this in general is like just making sure you're having authors that have some credibility online beyond your website. If it's just really on the website that might be tougher like social profiles and all that.
[00:33:37] Speaker A: Yep.
[00:33:38] Speaker B: That show that they have a historical, a history of speaking and as experts on these subjects.
Yeah.
The first party data is really big. I think that's going to go a long way and we'd already been seeing that shift as well as something that people should be prioritizing so that meaning like do your own research reports or mine your own data to come up with something compelling to talk about like why they should use your product and how that's impacting an industry or things of that nature or just do good third party research that aggregates that stuff.
[00:34:11] Speaker A: Yeah, totally. Yeah.
So do you think that every company is just going to start putting like 10-15 q a format questions and answers on there page?
[00:34:27] Speaker B: Yes, I do, I do. I've seen and read and had discussions where that that's definitely being prioritized and I think people are going to run with it in a way that like we've run with these types of things in the past.
But you know it's, it's not a bad thing. Like if you look back to like FAQ snippets being on, on the bottom of every blog post that came out of an era when that was something that you knew. I think I believe it was FAQ schema was being used to actually get in incorporated into the people also asking Google search which then like they deprecated and so like hey we're not looking at that anymore. Yeah but we, we work with clients a lot of times now that our clients now where like it's a strategic decision that when we implement it it should provide value. So like if you're going to do it just make sure that like you're ash like don't just go to the people also ask and find all those things like come up with relevant FAQs that like you've heard from your sales team or, or maybe something that maps more accurately or appropriately to that topic and, and be more thoughtful about it. That's the only thing I'd say. But yes, people, businesses and websites are going to start doing that a lot.
[00:35:42] Speaker A: Yeah. Yeah. So maybe some actual frequently asked questions by your customers, you know and not just purely an SEO play.
[00:35:52] Speaker B: Yeah. If anything that's a good way to go at it is like customer FAQs specifically.
[00:35:57] Speaker A: Yep.
All right. Number six multimodal and middle of funnel, bottom of funnel optimization. So you know, kind of moving from SEO for UX and conversion to AI scannable content that supports answer variation and user types.
[00:36:18] Speaker B: Yes.
So this is something we have been very adamant about. I alluded to, I just kind of alluded to it a bit back but mofu bofu content middle funnel bottom funnel content that has high intent. Historically what that type of stuff has looked like has been you know at the middle of funnel we've kind of defined it as versus or sorry not versus but like best X software type listicles and then at the bottom of the funnel more like Brand versus brand type pages or you can even just go like your brand and it's like pricing pages. Right.
It is so important now to have that stuff like not only prioritized and published to your website because especially on the pricing front or anything that breaks down features or anything and be highly accurate and objective because if you don't and someone asks can I chatgpt compare these two products and you haven't published that, it's going to ingest and spit back some summary of other websites that have talked about it which 9 times out of 10 is not accurate because they don't know your product.
Right.
[00:37:28] Speaker A: Yeah.
[00:37:30] Speaker B: So the, the more practical application in that too is like use tables, comparisons, all that type of stuff.
All of the ways that you would str, like improve the structure of like any other type of content is even more important here because you need to lay that out in a way that's like really clear for people to make decisions off of, but also for the LLMs to, to. To to scan and ingest.
So yeah, like I love this one because it is so important. It has been important and I think now more than ever it's even more important because you wanted to rank for the best X software piece or versus Page and you should still. It's kind of table stakes what we've talked about in the past, but now it's even like more elevated table stakes, I guess.
[00:38:12] Speaker A: Yeah, yeah, I absolutely. I mean, I think, I mean I think we're all still kind of learning what like user behavior looks like.
[00:38:23] Speaker B: Yes.
[00:38:23] Speaker A: With LLMs and how much weight or you know, stock is put in to the answers that it gives. And you know, do people just take that as.
As fact? Like oh, this says that whatever X software doesn't integrate with this. And so I'm never even going to go to their website.
Is that what's happening or are people going a bit deeper to your point on making sure that you're properly talked about in a comparison type of setting?
I think some of the user behavior stuff is to be determined. But just like anything else, it's no different than before when we would. You'd say like hey, you, if all of your competitors have, you know, us, us versus you page and you don't have one, then like anyone who's searching for those things, you're just letting that perception be dictated by your competitors.
[00:39:25] Speaker B: Yes.
[00:39:27] Speaker A: So now it's the same thing.
But VLM and we do know that there is a good amount like that type of content does break out well and does inform A lot of how the LLM is surfacing those types of answers.
[00:39:45] Speaker B: Yes.
Yeah. It couldn't be more important now. And anybody who's like on the fence or has been hesitant about it in the past, I would say anecdotally we have seen traffic that like some of our clients are getting from LLMs be of higher intent and higher quality. Smaller sample size so far, but it def. It's definitely driving high quality traffic.
And you can't make these connections because attribution on these types of things are. You can't make them directly, but it is, it is safe to assume that they're coming from those types of queries.
Lower funnel, no one is necessarily going like we've talked about being cited and what gets surfaced in other. On other episodes like it's, you know, a lot of top of funnel traffic might be coming. That's not what's driving traffic to your website from an LLM is that they searched like what is X non brand thing, got an answer and then somehow you got cited. It's from a lower funnel product focus question or query where they were evaluating products or whatever. So like feel good that that is something that you should be investing your resources are tied into.
[00:40:57] Speaker A: Yeah, absolutely. Yeah. And there's been a recent study that was showing a lot of very similar or even verbatim stuff from meta descriptions on a page were kind of what was coming through in an LLM in terms of how they were wording it, describing it, some of that. So just you know, all again kind of fundamentals of like scannable stuff, captions, alt text, meta descriptions, like all those types of things.
[00:41:30] Speaker B: Yep.
[00:41:32] Speaker A: Really kind of making a big part of that.
[00:41:39] Speaker B: Yeah.
[00:41:40] Speaker A: I thought I had one other thing to say, but I can't remember what it was.
[00:41:42] Speaker B: Oh, it'll come back.
[00:41:44] Speaker A: Yeah. So that's fine. We'll move to last one number seven, performance monitoring in AI ecosystems. So moving from SERP tracking and GA source reports, GA being Google Analytics to prompt visibility, citation monitoring and bot crawl audits. So I know this is a big one because this is just a big part of. We've talked about this before like with, you know, companies and marketers reporting to leadership and just kind of like there's a big expectation shift, you know, within companies that that has to happen and understanding how to, to view the performance on this.
[00:42:25] Speaker B: Yeah, yeah. I mean it's kind of closing the loop on all of these things. Right. Like it naturally you start over at 1 on this and you have your research but you do all these things that we've talked about and you should be in monitoring mode and have the system set up to understand if all the things you've been doing are getting visibility in that. And so there's a bunch of different ways that you can do it which call back to the first one. You do have those third party tools but this is more about like the first party stuff that you should be building dashboards.
I should preface that like because you said like from search, from SERP tracking and GA source reports.
Because there aren't anything, there isn't anything that's like super pre baked to give you that visibility so far. Like even if you look at GA4 it's not like breaking out LLM traffic has like its own channel. You have to customize some of this stuff now and maybe it'll catch up but. Or use a third party tool or like we've been building, we've built our own really strong LLM dashboard and looker which gives us really great visibility into individual landing page traffic from each individual LLM.
You know, where we're getting the most from and all that. So that's really important.
And then you know, doing some more manual connecting of the dots to get some attribution on. All right, well if this is the top, this is what I was saying earlier where you can see the top trafficked LLM referred page is you can see that. So you can make some pretty, pretty simple or pretty confident conclusions around like the, the queries from LLMs that are driving that. Because you built this piece about a thing, people are asking about that thing. So it's going to be a variation of the thing that you optimized for.
[00:44:10] Speaker A: Yeah.
[00:44:11] Speaker B: Whether it be super long tail and you can never 100% accurately predict, but that's really big. And then also like I, I kind of alluded to it but you can also look at your crawl logs too to make sure you're understanding like if you're getting, making, making sure that everything's getting access to all that content and stuff.
[00:44:27] Speaker A: Yeah, yeah. And I think the, the important thing is that beyond. So GA can give some of that data and you can see where traffic's coming from clicks from an LLM. However, that's only a small portion of you know, the, the interactions there are people actually clicking through. And so the other piece of that is certainly the mentions, the sentiment, just overall visibility into AI, which is the tool we've built and there are also a number of others that can do that. And so I think that becomes a big part of that too, is kind of shifting to understand not only how you're performing in SERPs, where your position is and click through rate and all that, but then also dimensions and citation rate and sentiment and all that stuff.
[00:45:24] Speaker B: Yep.
Yeah. And, you know, whatever you're looking at, I think the big thing that's going to be the most important is to make sure you're understanding how that stuff is converting. I think right now they're just to see there's a major obsession with the visibility component. So I think a big takeaway from this, which we've touched on, like, you know, having a good understanding of what that visibility is. But if you can't, we. You're naturally going to get asked that question around, like, what is this visibility doing for us? Maybe right now it doesn't matter. And it's all about just being there to some, it shouldn't be, but I think if you have your events and conversion tracking set up properly, you'll have a good understanding of what it is. And you're probably get, if you're not looking at it now, you, you're getting some that are likely of quality. So, like, just figure out the best way to drill in in your CRM and your analytics dashboard that you built.
[00:46:21] Speaker A: Yep.
So we've covered a lot on this, which I think does a good job of, of highlighting where, like we said at the beginning, like, where there's been shift in kind of like the iteration on some of it, but also very, very heavily aligned to skill sets that, you know, any SEO person or agency already has that very much directly apply to optimizing for AI as well.
So I think there's, there's a lot there. However, I do want to just kind of talk about one thing. We, we kind of prepped and like went through a bunch of questions that we've gotten from like, from clients, from prospects, just from, from other people we've talked to.
And so there is just one that I think we should just kind of chat through, because it, I think it is probably a question that a lot of people have, which is what is the point in creating and publishing content if AI is going to summarize it?
And I think that. Yeah, I've heard that. I'm sure you've heard that.
And I do think it's probably maybe a good kind of summarizing question to talk through that covers a lot of this because this is fairly tactical what we've gone through today.
And there is probably a bit of still the conceptual. I mean, we've had People on calls, prospects just be like, oh, I don't care about SEO anymore.
[00:48:02] Speaker B: Yeah.
[00:48:02] Speaker A: Which is like very, very like, Mr.
It's cool. It makes you sound like super, you know, trendy and progressive, that you're ahead of the curve, but like potentially a disservice to your company if there is opportunity there that you want both. It's no different than in the past when we've talked about if you can capture roi, positive opportunities from paid search, but you can also do well in organic search, do both. So I think there's still a lot here and things will continue to evolve over time, but with that, I think there's a lot of folks kind of wondering, do I invest in SEO? And so I think that what we've covered tactically shows a lot of investing in SEO is essentially, it's the same. You're doing a lot of the same things to invest in both.
[00:48:56] Speaker B: Yeah.
[00:48:57] Speaker A: With a lot of nuance. So with that, again, going back to this question, I started with what is the point in creating and publishing content if AI is going to summarize it?
[00:49:07] Speaker B: Well, I feel like part of the answer is actually in the question.
[00:49:13] Speaker A: I agree.
[00:49:14] Speaker B: Which is that it's summarizing something.
You can be that something, which is really important. That's not going to change for I don't know how long. I know AI and LLMs are moving fast, but they need sources to, to cite, to retrain their, their models on. And so that would be like the first thing I would respond to is that, that it has to respond, it has to summarize something. So you should be that thing that it summarizes and then, yeah, I'll leave it there. I'll let you react to that. So for now, agree.
[00:49:54] Speaker A: I do think part of the answer is in the question.
However, I think that at the end of the day, it comes down to authority.
I think is a big part. I think that there is going to be already is a good positive shift away from just cranking out tons of crappy content. Like, yeah, you know, we've talked about before, like that was a thing and people were, you know, outsourcing, writing for, you know, one cent a word and just like pumping out really crappy stuff. Algorithm changes happened, people started to focus on like more quality. Then the eat stuff happened and it's kind of evolved that way. And then all of a sudden you could generate content with AI and people started to go back that direction of let's just pump out a bunch of crappy content.
And so I think that Companies probably will not need to create as much.
[00:50:59] Speaker B: Right.
[00:51:00] Speaker A: And I don't think there's anything wrong with that.
But to your point, creating nothing basically just says we're not going to try to be a cited authority. We're not going to try to have a voice in this in any way. And so I don't think that's what any company is thinking.
We don't want to be the authority. I think that's what everyone wants to be. They want to be the leader of their space, they want to have the most press, they want to have the most accolades across the board. It's not even just an AI thing.
I think that is the, the crux of it. It is really comes down to authority.
[00:51:38] Speaker B: Yeah. And I mean it goes down this existential rabbit hole of even like what a website means to a business, I guess which would always have been the question. Or it has been probably to some extent for some business for a long time. But. But yeah. And there's a spectrum on the like how much content you should publish to. To your point, like maybe you shouldn't but like you know, in Emerge we've seen multiple times now actually ironically within more like AI spaces that like an industry that hasn't like evolved as much or as newer, there's like tons of opportunity to create a lot of content that gets gobbled up by the alums. Right. It's like the new blue ocean, as they say.
[00:52:18] Speaker A: Yeah.
[00:52:19] Speaker B: And then like the other end of the spectrum still exists where you know, if you're a sales led organization and you have more enterprise audiences, you should still be creating content and publishing content. And that's even like, that's acknowledging that LLMs may not be summarizing what you're doing, but you should be creating content to educate your audience in some way. Your sales team needs pieces of content to like send in some emails or I don't know, summarize on outbound calls, whatever you're doing that might be feel like beating your head against the wall. But I think the point is all roads still point back to some level of publishing content. If you're trying to be a business that's making an effort at being not only seen proactive or in an outbound way, but also especially being discovered inbound.
[00:53:10] Speaker A: Yeah, I think you just alluded to the last point that I wanted to make, which is I think for quite a long time now essentially been saying like you, you shouldn't ever be creating content for one singular purpose. So.
[00:53:28] Speaker B: Right.
[00:53:28] Speaker A: Like we're well past people Being like, let's hide this like folder of content from the site. It's just for search engines. Like this whole concept like that's been gone and you know, it really it is that like your content we've said for forever. Like your blog posts should inform. Like that should be sales enablement, it should be customer support material, it should be repurposed into video.
[00:53:53] Speaker B: Yeah, all that.
[00:53:55] Speaker A: And so I think that that is the to your point is is the other crux of it is like it.
If you were ever just doing that purely for SEO1, you still should do it because it's now you know a way to be the authority and be cited and other things. But you never should have been doing it just for. For one thing anyway. And so right. Yes. People are still coming to your website from other ways. Like sales needs it. People self educate. We know that all the time. Like yeah, there's, there's a lot of pieces to it and to bring it.
[00:54:33] Speaker B: Back to the la. The. Not to get the last word or anything, but it's Even if you didn't create any written content, like, even if you were just like, hey, we have really strong products. We're in a specific space where like we just. You need to put out your individual pages that speak to all the products that you have. You need that like it's still a brochure for the things that you do and that will still inform alarms, it'll still inform all these things. So like never hesitate on. On investing in content. It's just like the, the what type and like where you're putting it and all that. So.
[00:55:05] Speaker A: Yep, yep.
Awesome. We'll wrap that up. As I said, partway through we will drop a bunch of links in the show notes.
Be sure to like subscribe. Visit newsletter 10 speed IO we drop episodes there as well as on our website.
And yeah, thanks for tuning in.
[00:55:26] Speaker B: Yep, thank you.
