Subscribe
Share
Share
Embed
The Factory Floor is hosted by the three co-founders of Conversion Factory, the marketing agency at the forefront of SaaS growth, marketing, and tech trends. Episodes are released on Twitter one day early, @coreyhainesco.
Every other week Corey, Zach, and Nick break down what’s working right now in SaaS marketing, share real-world lessons from the field, and give you the strategies you need to outpace the competition.
You can also find us on YouTube, X, and everywhere you listen to podcasts!
Don't fall behind. Subscribe. Like. Drop a comment. Or not. The ball is in your court.
Nick Loudon (00:00)
All right, let's do it.
Hello everyone, welcome to What's Up.
Zach Stevens (00:06)
Sorry. Somebody, it's either Remmy or something, is scratching out. Okay.
Nick Loudon (00:08)
Corey's dog. Clicking his heels.
Corey Haines (00:12)
He's just walking. Let me
put him in the bed.
Nick Loudon (00:13)
Like, where's he going? You know? It's a small place.
Zach Stevens (00:17)
Remmington
Corey Haines (00:18)
can't trim his nails, so he just perpetually makes noise with every single step that he takes.
Nick Loudon (00:27)
All right, let's do it Hey everybody, welcome to the factory floor I am Nick and I'm joined by my beautiful lovely assistants to my Right and left in my current video format. We have Zach Stevens and Corey Haynes
And we're here to talk about something that I don't know anything about. Okay, well, kind of. ⁓ Corey suggested and would like us to discuss LLM SEO, okay? So this is like the melding of SEO and AI in a way. ⁓ Forewarning to both of you. I've been out of the social media zeitgeist, which also means that I haven't really been looking.
at what people are talking about on Twitter for a little bit. So I'm completely in the dark on this thing. I've just kind of been nose to the grind. So Corey, why don't you just give us like what, what are you talking about?
Zach Stevens (01:28)
You
Corey Haines (01:29)
Sure, let me frame the conversation. Have either of you in the last two months, let's just say, asked an AI search tool like a ChatTPT, Claude, even Google, let's not put Google in there. Let's say ChatTPT or Claude or Perplexity for a product recommendation in the last two months.
Nick Loudon (01:51)
Mm-mm.
Zach Stevens (01:52)
actually
done. Yes, I have done a lot of this. I had it. Well, I had a research companies for us to reach out to for our project manager role, which was super cool. And I gave it advice or suggestions that said, we want to have this kind of hour overlap with Pacific standard time. We're looking to fit within this budget. We're looking to. I gave it a long list of things to do and three researched.
Corey Haines (01:56)
Okay, what about?
Zach Stevens (02:20)
a lot and pulled back a bunch of results, gave it to me in a table of pros and cons and with all of the amounts and either alphabetical order or ⁓ descending, ascending. was really cool. I think I could just pull up my chat. Actually, I could see what are the things have I looked at.
Corey Haines (02:40)
That's a good one though.
I just need one example. Pull that up though, just so you have it in front of you. So ⁓ for the both of you, where do you think that it got the information to be able to make a recommendation about what types of companies, products or services to use based on your question?
Zach Stevens (02:56)
pulled from the webpages. Clearly.
Nick Loudon (02:58)
My
guess is from either their website or some kind of like cross-reference with like review platforms like TrustPilot or something like that where it takes maybe some of the top search results and crosses them with reviews or positive feedback from customers. That's my guess.
Corey Haines (03:17)
Yeah. Yeah. No, you're both spot on the, top two there. There's a third one in there, but the top two are definitely going to be like the direct sources of the websites in question of the company's, you know, brands, product services that you're looking to investigate. Number two is going to be review sites. And that can also include a social media forums like Reddit or Quora, for example, which LLMs historically have pulled really, really heavily from. And number three is going to be like blogs,
and or ⁓ podcast content, YouTube content, but like usually blogs and people who are making lists of like best X for Y or best, you know, whatever, it's going to pull from that as well. And it's just trying to subjective experiences. So one of the big differences between Google and/or Bing, just like traditional search engines and LLMs and the AI chat tools that we use today,
is that Google is looking for kind of like objective facts, right? It's like, Hey, you're looking for, ⁓ a remote talent recruiting company. Let me just give you like a big list of them and you can go and like find the differences yourself and kind of do some research, book a call, so on and so forth. LLMs and ⁓
Zach Stevens (04:26)
Mm-hmm.
But GPT,
I was going to say, but this is actually reading the sentiment and, you know, assigning a lot more value and taking a judgment on its own, so to speak, but in a far more unbiased way.
Corey Haines (04:47)
Yeah.
Yeah. LLMs and AI chats are essentially going to try to replicate the user behavior of what you would do with all the Google search results if you were doing the research just by yourself. And so they're going to look through all the website content themselves. They're going to look at review sites and then they're going to try to find subjective stories and experiences of bloggers, people on social media sites.
mentions and podcasts, YouTube videos, so on and so forth to try to get like kind of fill in the gaps of maybe what those websites are not covering or what the reviews aren't covering. ⁓ so this is actually really good news because previously when you're searching for products, it was up to the user, the searcher to go and like fill in all the blanks and find all the research and hold it in your head.
A lot of people are basically kind of doing spreadsheet work in their brains where they're like, all right, let me open up 10 tabs. Let me like read through as much as I can try to sort out the differences and then kind of figure out like what's most important to me along the way. The AI tools do a lot of that work for you. The only scary part is that now it's up to the websites and it's up to the review sites.
Zach Stevens (05:41)
Yep.
Corey Haines (06:04)
to give the information that AI LLMs need in order to do that thinking for you. And so let's just say, for example, we're talking about the remote talent recruiting platforms or services. ⁓ their positioning, their website content is going to bias and weigh the most heavily because that's what chat GPT and LLMs are going to see as like the primary source material. It's going to
Nick Loudon (06:09)
Hmm.
Corey Haines (06:30)
augment that with review sites, with bloggers, with external sources, but it's always going to look at like the website itself and the raw content from directly from the source as the primary source and try to weight that the most heavily. And so this really got me thinking because I think previously in the last 10 to 20 years, most websites historically have had pretty bad copy because it's very diluted. It's agreed upon by consensus.
people are kind of trying to, ⁓ summarize and provide kind of fluffy copy that sounds good, but maybe it doesn't have a lot of substance. And that's kind of been okay because people end up booking a call, starting a free trial anyways, and then they're going to figure out that information for themselves. But that doesn't work with this new paradigm of LLMs and AI chats, because if the LLM can't figure out what you do,
based off of the content that you have in your website, then it cannot recommend you to anyone looking for a recommendation. And so your positioning and the content that you have in your website actually becomes one of the most important things that you can put out there about yourself because that is what an LLM and AI chat tool is going to be using to decide if it should recommend you to someone like Zach.
and why and actually have like a tangible reason. Zach, you're looking for someone in Latin America or South Africa in this price range who can, you know, in these job titles and maybe for part-time contract basis, not full-time. Cool. There's only one option and it's this option because we scraped all the content from all the different services and we found that this is the only one that fits your criteria. If they didn't have that information on there,
they are not getting a booked call. ⁓ and so it's just a really new paradigm that now people are realizing, ⁓ shoot, I need to have my positioning in order. So that like my website content can actually be used by an LLM to get me customers from AI chat tools.
Nick Loudon (08:34)
I feel like ⁓ this it feels like it's kind of forcing the hand of marketers. Does that make sense? Like you like there's no more room for like smoke and mirrors or like, we can use this tactic to like kind of get them in and like we don't have to show our pricing because we're to go do this over here. But like, if I'm like, hey, I'm pulling a list and there's 10 options and there's one of them that
like says like, they don't list their pricing. I'm like, well, just okay, just throw that one out. Like I want to see the pricing. Like I want to cut through all the BS and get straight to like how the service is provided, how much it costs, or even if I'm looking at a product and a service, the same thing, like if their pricing is listed or something, I'm going to be more focused on the ones where I get that information. It kind of like cuts forces marketers to cut through some of the stuff that they might try to do. ⁓
but maybe that's ⁓ assuming that people use this in a majority of cases where they're shopping for a product versus like just, I heard from a friend, I should try this out. Like what percentage of the discoverability of the products is this specific ⁓ method? You know what saying?
Zach Stevens (09:44)
Good question. I don't know, but I feel like it's picking up really quick. I spent like once I discovered ⁓ three and all the things that it could do as far as research and looking at stuff on my behalf. was a game over Google. I'm not looking at these search engine results pages anymore. I am just going to go straight to GPT. And the cool thing as well is that it will also prompt you with additional questions.
of, have you considered these things as well alongside your criteria? Or if you're really smart, which people are actively becoming better and better and better at prompting GPT to get it exactly what you want. You can have it tell you some of the things that you should be looking for. Um, when searching for X thing, uh, another aspect of this, I think is really interesting is how it requires you to drop the.
gobbledygook headlines from your site, like revenue orchestration platform, which means nothing. how the AI can't read that and make a clear assessment of whether to include you in a category that this person is looking for.
Corey Haines (10:56)
Yeah. Yeah. Cause someone's in a, in a search tool in a, in AI chat tool is going to be asking chat GPT, Hey, I'm looking for a new, ⁓ cold email sending platform. What would you recommend? And if you're calling yourself a revenue orchestration tool, you're just like, not even in the conversation. Like you would be lucky if it picked you up and had to in that conversation, but it's probably going to be so muddied up that even
chat GPT is going to have a hard time recommending you because it's going to be like, well, they kind of do this and do this other thing too. And if this is the only thing that you're looking for, and this is like, I'm trying to answer the question, then you know, that's not going to be at the top of the list, right? Which is really interesting because again, forever for the past like 20 years on the internet, it's kind of been true, especially for software tools that a lot of times the best product doesn't win.
a company with the best marketing wins, but not necessarily the best product. And I think that that's ripe for change because now the AI can do so much of the thinking and evaluating in the buying process that again, it skips through all the gobbledygook. It's, it skips through all the smoke and mirrors of kind of the, the marketing ness of hype and brand and, you know, making yourself seem
Kind making these lofty promises that might not be backed up and it just cuts to the facts. It's like, this is what you're looking for. This is the best tool for your case for whatever criteria that it is that you're looking for. And I think that's actually good news across the board because again, for forever, everyone's had really terrible copy on their site, I think, which has made it hard for people to make buying decisions. And so for one, they're going to have better copy, which is just going to make it.
easier for people to make buying decisions in the first place. But for two, then that copy is going to be used in LLMs, which is then going to get used in the AI chats. And that way they can actually get customers. And I've seen many, many, many screenshots of people showing their traffic and being like, dude, six months ago, Chad's GPT wasn't referring any traffic. And now it's my second highest referral source of traffic to my website.
Nick Loudon (13:10)
Holy cow.
Corey Haines (13:10)
it's growing exponentially
and people are realizing like this is kind of a superior experience than Googling something.
Zach Stevens (13:16)
there things that those people who are seeing that seeing that kind of shift, are there things that they are doing that are different to accommodate GPT or any of the other AI tools?
Corey Haines (13:28)
people are starting to,
yeah, people are starting to optimize now cause they're realizing, okay, there's, there's a race now to try to gain an edge. And so this kind of breaches into like, okay, what is LLM SEO really look like? And I'm calling it LLM SEO, but there's actually kind of a fracture right now in the marketing community about what to call this. ⁓ I think, right, right. It's not a search engine.
Zach Stevens (13:50)
because it's not a search engine. I think it would just...
Corey Haines (13:54)
It's kind of a misnomer being like, it's a large language model search engine optimization tactic. It's like, no, that's, those are two different things. So people are trying to call it GEO, which is generative engine optimization. Some people are saying it's LEO like LLM optimization, engine optimization, something like that. ⁓ no one's really decided on what it is. I think that everyone understands the concept of SEO. So it's sort of like it's LLM SEO.
Most people know kind of what that implies. I don't really care to be honest what it's called, but ⁓ we're all.
Zach Stevens (14:23)
Yeah.
Nick Loudon (14:26)
Yeah, dude, yo, you, dude,
let's pick a brand, dude. This sounds cool. Let's take this on. Yeah.
Zach Stevens (14:29)
AI search. AI search, that's what it should be called.
Corey Haines (14:33)
Well, yeah, there's also like AIO, yeah, AI optimization, but it's more specific like AI chat. And there kind of is ⁓ an argument to be made that it's still like a search. So it could be like AI search optimization, which would be AISO. ⁓
Nick Loudon (14:37)
search optimization.
Zach Stevens (14:53)
ISO.
Nick Loudon (14:55)
So dope,
I like that. It doesn't matter.
Corey Haines (14:57)
I forget which company I think
it's, I think it's Claude or yeah, it's either Anthropic or one of the other big AI companies is trying to brand GEO right now. So that's kind of what people are starting to roll with. ⁓ but to answer your earlier question, like, okay, well what are companies doing about this? for one, I think that any smart company right now is looking through their website content right now and being like, what do I give LLMs to index?
Nick Loudon (15:11)
Mm-mm.
Corey Haines (15:26)
So they're looking through all their copy. They're looking through the headlines. They're looking through their positioning. They're probably adding more pages because the more information, the better actually. And they're actually also adding just like a, there's a robots.txt file that you can add to let Google know that you're, you want to be indexed. People are starting to add LLMs.txt files that essentially generates the equivalent of a site map for
LLMs to be able to scour through all your web pages and understand like these are the pages that I want to be indexed in an LLM specifically. So that's number one. Number two is, they are trying to kind of SEOify the content that they produce for LLMs. And so, ⁓ they might be creating content that they,
normally wouldn't because there's no search volume in Google for it, but because they know the train of thought and the follow up questions that AI chats normally ask, then they're creating content around that. So like you said, with the remote recruiting platforms, for example, if it, if they see, ⁓ every time someone does a deep research option for something, it asks three follow up questions and the follow up questions are usually around
specific parameters or filters, then let's create blog content or other FAQ sections on our pages to try to address those filters and those qualifiers.
Zach Stevens (16:57)
So it's getting an insight into what additional information could we use to better make our case. Like, cause you can check, you can check all this stuff yourself and put yourself in the shoes of somebody who is searching for something that would fall on the lines of what your company does. And then you can play out the entire scenario of you went here, you went here, you went here, and you went here. And we've got all those bases covered. Gosh, that's freaking smart.
Corey Haines (17:20)
Yeah, yeah, yeah, exactly.
Zach Stevens (17:22)
It is, it is so straightforward when you think about it, optimizing yourself for this, because I feel like the, curtain's been lifted in a far more easy way. feels like dune where they're like, yeah, we don't even have robots anymore. It's just straight to the point. Like, you know, we've, we've cut out all of the, ⁓ the craziness in technology. That's what it feels like to me.
Corey Haines (17:41)
Hmm.
Yeah.
Yeah. Cause I mean, a lot of people didn't know how Google made their rankings. And so the entire SEO industry was built off of speculating on what factors contributed to ranking higher in Google. Now, I think everyone has a really good understanding of how LLMs work. And it's also actually not so dissimilar from how Google works in the sense that one of the primary sources of one LLM trains
itself on all the data that's fed to it is off of Google search indexing. And so yes, if you rank in Google number one for a specific keyword, then they're going to wait that content higher in the, in the LLM as well, because they're kind of using it as a proxy until they start generating their own indexing of people making native searches within the AI chats. And then like logging all of that.
and trying to create their own ranking system. They're like, why reinvent the wheel? Google has already done a pretty good job of like people trying to make sure that they're showing up in the search results, providing information, building an authoritative website, ⁓ so on and so forth, having legitimacy and authority and expertise. It's the whole like eats E E E A T, framework that Google recommends people to follow and
Nick Loudon (18:49)
you
Corey Haines (19:10)
They're like, this applies to LLMs too. So why not? So if you want to get better at showing up in AI chat results, be better at SEO for Google. Like just do, keep doing that. Yeah, yeah, just keep being better at that. Keep creating more content, optimize for keywords, keep building a higher domain authority website, and keep making your website fast, performant, easy to scan, and technically proficient.
Nick Loudon (19:21)
Two birds with one stone.
Okay, can I ask a semi adjacent question? Like there's two paths, like there's, okay, obviously Google right now is like, it's not AI exactly. There's AI like results that sometimes pop up, especially like when you're asking a question versus just searching for something. And then you have chat GPT, which is like purely AI, right? Answers and results. I'll say results. Does one end up winning this like search engine?
battle where like people just stop using Google together and they just use chat GPT or something like that? Or does Google say, okay, now everything they're all their results are kind of AI generated to ⁓ is, is that the shift that happens? Or is it like, no, there's two types of search.
Corey Haines (20:21)
Mm-hmm.
No, they're, they're converging for sure. I mean, Google is pooping our pants right now. And, ⁓ yeah, I mean, and everyone sees it and everyone's seen it coming now for about two years where they're like, this could change everything because the way that most people use the internet today is they start with a Google search. And even if they're not on google.com, they are still searching through Google, you know, Apple and Google have the deal where Safari is powered by Google.
Nick Loudon (20:30)
Yeah, probably.
Corey Haines (20:53)
if you just type into a search bar in Chrome or in Firefox, like it's all powered by Google. And so they have to figure out what they're going to do because they don't want to lose that advantage and that real estate and, that, that user experience of essentially being the default way that people search for things on the internet. So Google is definitely going to be incorporating a lot more AI. And we see that with like the AI summary, generative results. ⁓ I think that they're,
you know, they're trying to partner with companies. They're trying to incorporate things. they're working on their own LLMs. They're working on their own AI products. They're definitely going to have to basically disrupt themselves, particularly with the Google search engine in order to stay relevant because right now at this very moment, like chat GPT is eating their lunch on the AI chat search experience.
Zach Stevens (21:44)
It definitely does not feel the same as far as searching on Google versus searching within, excuse me, GPT. a hundred percent.
Corey Haines (21:47)
no, yeah.
Nick Loudon (21:51)
I so much better in ChachiPT. So much better.
Corey Haines (21:54)
Yeah, it's completely different, very
superior.
Nick Loudon (21:56)
Let's go, okay, so.
Corey Haines (21:57)
And it's just a matter
time before that catches up. ⁓ the, think that like, can see a world too, where we see with chat GPT, they're also coming up with more traditional search experience type of results with like deep research. For example, they make citations and, ⁓ they're going to ask follow up questions. They might give you a list of things instead just one thing. Right? So you say, Hey, I'm looking for
the best talent recruiting agency for Latin America, they're not just going to give you one, they're going to give you five. So that's kind of the same as like the first page of Google, right? One of the other big paradigms in here that people haven't really figured out what to do with is advertising because again, Google makes a lot of money from serving ads in search results. Currently there aren't really any ads in chat GPT. I think that there will be,
And we're also seeing opportunities, for example, where ⁓ Shopify is building a direct integration with open AI and chat GPT. So that if people search for, you know, best men's razor, that it's actually going to have a list and like a link to purchase directly within chat GPT. You don't have to like leave chat GPT to go buy a razor off of, you know, beard brands, or I'm sure that Amazon is trying to be in that conversation. Yeah.
Nick Loudon (23:08)
my goodness.
the drop shippers are gonna cry
Corey Haines (23:17)
So anything that you would do again within Google, we're gonna see that same paradigm in chat GPT and maybe even more because they can move faster and they're a little bit more nimble as a startup.
Nick Loudon (23:29)
Can I make a prediction, because there's people like us who are on our computers working and doing things all day, like the internet folk and younger people and some older people. And then there's this whole giant swath of people that like, they're not sitting behind a desk all day or they're not on a computer a lot. And when they need to find something, they go to Google and they're not wondering about chat GPT. They've heard about it. They don't care, whatever.
Corey Haines (23:55)
Mm-hmm.
Nick Loudon (23:55)
But they're also very concerned about data privacy and Google tracking their information. When and if Google can figure out how to make their search function AI-ified and much closer to what ChatGPT does.
These people are gonna lose their minds and either be so stoked or freak out that the AI Google overlords are eating their brains alive. That's my prediction is that anyone you know who's so scared of Google is going, but still uses it because that's really the only option that they see is gonna lose their minds and be so afraid of new Google.
Corey Haines (24:13)
Yeah.
Mm-hmm. yeah.
Yeah, it's like, if you were scared
of Google or Facebook, wait till you hear about chaggpt. It's the final boss of things that boomers are scared of.
Nick Loudon (24:36)
Dude, yeah. Yeah, don't use
Zach Stevens (24:42)
Mm-hmm.
Nick Loudon (24:42)
chat GPT for like a year and then be like, what do know about me? Don't do it. Don't ask. man, you'll be, you'll be scared.
Corey Haines (24:46)
Yeah, yeah, don't ask it.
Don't ask questions you don't have to answer too. ⁓
Zach Stevens (24:52)
I answered that
question yet.
Nick Loudon (24:54)
Yeah, it knows.
Corey Haines (24:55)
It's
it's crazy. Yeah, especially now also it knows it can like search through all of your past chats as well So like it has a really really good understanding of who I am as a person I've basically divulged every single personal detail I ever could to chat to BT and So I'd like my friends that I was just on this trip with on my little workation They were like how often do you use AI and I was like When am I not using AI?
Nick Loudon (25:02)
Mm-hmm.
Yeah, every day.
Corey Haines (25:24)
Like I
was logging all my expenses through like a chat to BT chat. I was doing translations live where I would just take a picture of me. Like, you know, translate this menu from German to English. I was asking it for like things to do around the city. I was asking it about like recommendations for what I should eat when I had like an esophagus thing going on because I antibiotics, whatever. Like I was using it multiple times a day, every day.
Zach Stevens (25:48)
yeah.
It gets its hooks in you pretty quick. It gets it, it gets its hooks in you really quick. I have, I haven't been able to do it yet because I think that I, it just wasn't functioning the way that I wanted it to, but the AI chat during a drive is actually something that I thought was really useful. Well, especially for someone like me, because I, I process externally, you guys know where
Nick Loudon (25:50)
And they're just like... yeah.
Interesting.
We do.
Zach Stevens (26:14)
My ideas aren't fully fleshed out until I've verbalized them. And that's actually when they become the most concrete, but I could do that with somebody else and, or, you know, without there need to be a body nearby. I can instead talk to Vale, who is my chat GPT voice. No, she's not my AI girlfriend. She's, she's my AI, my AI assistant. And her name is Vale.
Corey Haines (26:31)
your AI girlfriend.
Nick Loudon (26:34)
No, we're just friends.
Corey Haines (26:36)
This is work wife.
Nick Loudon (26:41)
Dude.
Zach Stevens (26:42)
So, you know, obviously I'm going to talk with her. That's a great assistant.
Nick Loudon (26:44)
I, what I would love is
Corey Haines (26:45)
That's funny.
Nick Loudon (26:47)
my AI version of, know, the, my veil would have my voice and my name. And I could like add, yeah, a hundred percent. I would much rather have that than talk to someone I, a different name. Anyway. ⁓
Corey Haines (26:54)
Can talk to yourself?
I wouldn't want to talk
to myself. I'd want to talk to like David Attenborough, whatever his name is, someone with like an awesome voice.
Nick Loudon (27:05)
No, no.
No, because then it's like cheesy to me. It's like cheesy to talk to someone else's voice. Like, just talk to me. Like, this is my thing. Talking to me. I was going to say, it's funny that you mentioned that the people you were on this trip with were like, how often do you use AI? I think once you get a positive result or
Zach Stevens (27:07)
or David Goggins.
Corey Haines (27:10)
Yeah, yeah, yeah, different cocks.
Nick Loudon (27:28)
something for the first time, like maybe you have a big problem you're trying to fix, technical problem, something like that. The first time you fix something and AI did 99 % of the lifting, you will never escape. Like you'll never go back, which is crazy. But yeah, I'm really curious to see how the search engine and AI stuff plays out. Any final thoughts?
before we stop talking about this, because we've been talking for a minute.
Zach Stevens (27:55)
fascinating.
Nick Loudon (27:56)
Awesome.
Corey Haines (27:57)
Yeah, do you guys feel like you have a better understanding now like LLMSEO, GEO, whatever we want to call it, kind of what goes into it and why it's important?
Zach Stevens (28:06)
I thought that I wasn't sure if we were going to talk more about the use of it from a programmatic front, especially with things like MCPs or ⁓ anything along those ends, like using LLMs to try and increase your SEO, but instead using them as the search engine. But you know, those two things might play hand in hand really well, having the AI write things that it is going to find beneficial to then search for itself. And then you will have completely outsourced your job to a robot.
Corey Haines (28:19)
I see.
Okay, the AI
Zach Stevens (28:35)
you
Corey Haines (28:35)
circle jerk.
Nick Loudon (28:36)
So great. Super interesting topic. Thank you, Corey, for bringing it in. ⁓ Yeah, I'm curious to see like the next iteration. We'll probably have to do like a follow up to this, I'm sure, in like six months or sooner. Geez. Okay, thanks everybody for listening to the factory floor. See you on the next one.
Corey Haines (28:49)
We'll check in in six months. Six weeks.