AI & LLM Visibility: A Practical Guide for Ranking in AI Results

again and David, you're going to be

talking about how to rank in LLMs.

>> Absolutely. And thanks for having me

back. Um, I think there's a lot of

information or uh misinformation about

um how people get visibility in search

engines and and LLM search engines. And

and this is something that that is is

really close to my heart because I a big

believer in SEO and and that

democratization of access to people. And

I think you covered uh a story that was

on our SEO a while back about how um geo

companies were paying SEOs to um spread

disinformation. And there's a lot of

things about like how LLM site things

and how LLMs are looking for things like

schema which is not true. Mark Williams

Cook did an excellent um debunking of

that, right?

>> I made an episode on that as well.

>> Okay. Excellent. Oh, so glad you did

that. Another one I have to go back to.

Um so I what I wanted to share with

people today is like is understanding

the query fan out. And if you can

understand the query fan out, you can

directly link SEO to u LLM visibility.

What's interesting is that Profound

actually came out with a blog post last

week saying it's the query fan out. And

I think for a while they had been in the

camp of oh it's all about citations and

you know brand uh marketing and things

like that. It it's not LLMs are not

search engines. And I think that's a

fundamental um sort of default thinking

that a lot of people have. They assume

that an LLM is trained with all this

information. It's it's sort of like a

cached version of the entire web and

that they can um give results. They

don't. They they cannot store that

amount of information. they couldn't

store the amount of information that

Google handles in an hour. And so it's

very important to know that with page

rank, with Google, even if they're using

Bing or uh Claude in the in the case of

uh sorry, Brave Search in the case of

Claude, um they're actually using um

standard SEO to rank things, but they're

breaking down the query the prompt into

different queries. So, I just pulled up

a random example here, uh, best SAS SEO

case studies, and you'll see Zapier

ranking, Slack, AFS, and, um, I think a

lot of people see the citations, and

what people might see is that they're

putting in a search phrase that, um,

they're not ranking for, right? So, the

first, and the reason I'm using um,

Perplexity is this is just a little bit

easier to do in Perplexity and a little

harder to do in chat chatpt, but you can

do the same. You just need a QFO tool,

and maybe I can share that online. uh to

put in your YouTube uh description

afterwards.

>> So when you look at the um the uh

assistant steps in this case, it's the

same thing, right? So best SAS SEO case

studies, best SAS SEO case studies, but

I wanted to have a look at something

that I rank for because I'm also really

really skeptical about eat in SEO. And I

I I don't think that eat is something

that Google can detect. Google has said

that they can't detect, but it's

everpresent. um it's it's it's very

present in in sort of um a lot of um

marketing information on LinkedIn and

and um X and on Reddit. So let's put in

um E SEO guide and see what comes back.

And so here we can see the the the

change straight away.

That's what I would be looking at if I

was for ranking. So I I try to rank for

these phrases to sort of help um balance

the amount of information out there. And

what uh Perplexi is is is searching

today and this would have been different

a month ago and we call that difference

or that changed the drift is it's broken

up into E-N

SEO guide and 2025. So adding 2025 is

very common in the query fan app and

you'll see it matches a lot of the page

titles uh for some of the content that's

ranking here. So if you go to your

search console and start looking at what

do you rank for? um if you're not r if

you're ranking for edio guide and then

you go to uh chatbt and type in the same

search phrase and you see different

results, it's not because the LLM has a

different search ranking um algorithm.

It's because it's changed the search

query. And if I take this search query

instead of the prompt and I do a search

in Google, I'll see that all of the

results are exactly the same in almost

exact order. And what happens is

Perplexity actually sends its crawlers

to fetch each of these documents in real

time to make that synthesized answer on

the homepage. The the the crawler I

think people think that the crawlers go

out all day every day like Google

crawlers and are building up a copy of

the web. This is a real time uh scenario

and you can literally rank in Google now

and in two minutes repeat the search in

um perplexity and see that you're

actually in the results even if you

weren't there 10 minutes ago.

>> And you're saying so you're saying the

query fanouts are different for for uh

different LLMs.

>> Uh they're different for different LLMs

and they're different at different times

of the day. So, so let's take another

one that I think a lot of your your

audience might be interested in.

>> By the way, can you zoom in a bit more?

>> Yes.

>> It'll be easier.

>> I can. So, let's take an example for um

a a best SAS product uh best SAS CRM for

a company with 200 people. So, one of

the big differences with LLMs is that

people um

let's just do this in the

auto industry. A lot of people are

getting into deeper, more complex um

prompts, right? They they know they're

not dealing with Google anymore. Um and

so sometimes we can actually see much

more complicated fanouts with up to

three queries.

And I'm uh not able to generate one. So,

I'll just I'm just going to stop and try

different. So, um recommend a better CRM

for my sales team at my car parts the

car website

and see what that does.

Oh, and again, it's it just doesn't want

to play play properly for us today. But

again, it's um it's actually broken the

query down into something much more

simple today. Uh before I would have

actually expected maybe three fan outs

>> and then we would have got 30 results.

Here you can see they're just eight

results. And again, if we take the

search, we'll see that basically a lot

of big websites ranked. And I think this

is adding to and I'll I'll zoom in here

again. I think this is adding to the um

frustration and the confusion that the

LLMs are preferring more um you know

much bigger brands right but what

happens is with these drifts is that

they tend to rank sites with a lot more

authority like in solutions doesn't seem

to have 2025 in here and neither does

hedges and company um but slash dot does

and so uh you can you can see that if

you if you're ranking for for example

best CRM or let's say car parts website

without 2025,

you may actually see a different set of

results. And if you're ranking in this,

like for example, here we see Smith.ai,

now we see Reddit appearing. And so that

also brings into play this other idea

that um LLMs prefer Reddit. It's not

that they prefer Reddit, it's just that

Reddit often crops up in Google's

rankings.

>> That's a good point. That's a very

important point. Absolutely.

>> But there was, you know, this method of

marketing is so effective, I had to make

sure it wasn't against Google's rules

before I kept doing it. It's a form of

SEO I call compact keywords. Whereas

most SEO focuses on putting up articles

to answer questions, how, what, when,

compact keywords focuses on putting up

dozens of pages that sell to searchers

who are actually looking to buy. These

pages rank on Google and convert so much

better than normal that when I

discovered this years ago, I couldn't

believe this was allowed. It's less

work, too. The average compact keywords

page is only 415 words. Compact Keywords

is a 13-hour deep course on getting

sales with SEO. A customer recently

said, "Each lesson is dense with

information. You're giving years worth

of experience boiled down into 15 to 30

minute lessons with no filler or fluff.

I feel like I'm gaining a new

superpower. Compact Keywords is about

setting up an SEO funnel that brings you

sales for years and years and years. It

works with AI. It's less work than

traditional SEO and it makes way more

money. You can get it now at

compactkeywords.com.

Back to the podcast. You know, actually

what was very interesting? So, when I

had on uh Jackie Chu on the show, which

was funny cuz that was a day after you

first came on.

>> And so, I was talking about how yeah,

Reddit is always being cited in in

ChachiPT.

>> Um and uh but then he called he's he

actually called me out. He's like, "No,

it was just dropped from ChachiPT

entirely." And and I couldn't I couldn't

get it cited in ChachiPT even though it

is still cited all the time within uh

within Google. So I I but I I suppose

for that moment um it could have been

instructed to to omit Reddit results

within Chacha PT.

>> I wonder if so if you look at the

results sometimes in Google um let's

just do a search for um

credit advice.

there's a certain type of um position

ranking in Google right so here we have

experience at the top we have the people

also ask which obviously every SEO knows

that this is an important part of of

developing topical authority which we

spoke about in our last thing this is

this the discussions and forum um band

and this band is always reserved for

Reddit and other forums here we see

Facebook if you go to Ireland for

example you'll see boards.ie which is

the biggest uh forum board in in

Ireland. Um and I wonder if chat GPT is

just ignoring this whereas I think the

average user just sees Reddit.

>> Yeah.

>> Uh and these are the actual organic

results. Um and then you get perhaps a

YouTube result. I'm sure you're familiar

with seeing YouTube in between the

rankings as well. Oh yeah. And I wonder

if if if Chachbt knows that this result

is really a a boosted result, if that

makes sense.

And that's why they're cutting it out.

>> Interesting.

>> Uh because we I think we we've seen uh I

think Barry Schwarz covered this um that

a lot of um ChateBT's results are coming

from Google. I don't know if that's I I

suspect it's something to do with

Microsoft slowing them down and um

breaking up the was it Windsurf deal um

with their legal analysis taking too

long and Google came in the background

and and bought out the executive team

for $200 million.

um or if Bing is just too slow um or

just has the worst results. Uh that's

what I suspect.

>> Yeah, it could be a few things. There's

been so many tests of people showing,

yeah, look, this is this is indexed on

Google. This is not indexed on Bing and

here Chacht is citing it.

>> Exactly. Exactly. Um and that makes

sense. I think um that's I think they're

the tools are promising. I I know

Perplexity sort of that that's their

strap line. You know, we're getting you

results that are free of SEO. Um, and

and I I I suspect that's why they have

the drift to try and circumvent SEO

tactics. Uh, but but that just means

it's still based on Google, which means

it's susceptible to everything that

Google is susceptible to. Uh, so really

LLM um results are dependent a on Google

getting better and b on how they direct

Google through the drift.

>> Talk about the drift a little bit more.

So the drift is basically where um based

on the results that they get back in the

first search, whether or not they do

more searches and then whether or not

they add different keywords to the the

phrase that they sort of outsource to to

the search engines. Um, and I I suspect

that that's done to get better results

for the synthesization, but also to get

get around people using SEO

um by making it by sort of like

constantly moving the search phrase so

that only the websites that rank for the

most things are always cited, if that

makes sense.

>> Mhm.

Um, and what I wanted to talk about

there next was how to look at um what

kind of queries are coming into your

Google search console and how you can

use Google search console to have a look

at

>> um what's happening. So I have a a test

website called uh money asset lifestyle.

I picked it cuz I wanted as to go as YM

as possible. I can't think of anything

that I could buy for $12 that had more

YM L in it. Um, and I think one of the

interesting phrases that we found is the

word evaluate. Uh, so let me just zoom

in here. Um, and make this more

readable. Um, evaluate in Google can

often be uh synonymous with um with

compare. But if you if you take some of

your pages and I think that's the way to

understand drift is you may need to

publish the same articles under

different um keyphrases to target those

different drift phrases. But you'll

start to see a lot of um deeper

searches, very very long searches um

that use the word evaluate that have no

clicks. And you'll also notice the

average position is in the top 10. In

other words, not one of these phrases is

out is above 10,

>> right?

>> They're all within 10. And that's

because

>> the query fan out only returns the top

10 results.

>> Yet you've got no clicks. We've got 170

queries on a single blog post on if you

if you trust DA measurement, this site

has a DA of like two, right? It's a it's

a very low DA site. But here's all of

the deep queries, right? Evaluate

fintech company ramp. Um, evaluate

fintech company arc on banking and

finance. Uh, evaluate company arc on

banking for consulting firms. All very

deep top three positions. Not a single

click. These are all LLM fan out. So,

you might be thinking at home, "Oh, I

just can't get into um into LLMs."

You're already there.

However, if you want to measure the

clicks, you've got to go somewhere

completely different because the clicks

come from referral traffic, not from

search traffic. And the best way to do

that is through uh Looker Studio. So, I

have a Looker Studio report um that I I

I've built. Very easy to build. Um but

it would take me a long time to show you

how to do it, but we can maybe cover

that another time.

>> Yeah. Uh what I've done is I've uh

brought together the different LLM

engines. So I've got CHPd, Perplexity,

Gemini, Claude. This just a monthly

graph for that traffic. So it tells me

um hey, you know, CatchPD was really

going up. Then everyone noticed CHPD was

sending less traffic. Uh but actually my

perplexity traffic went from 9 to 12 to

13 clicks. And then here I can see what

pages the traffic came from. And I can

also import um any of my uh G4 goals or

or key events. So if somebody made a

telephone call or if they sent a contact

form, I can tie it to the engine and

which uh landing page they came in. So

did it come in through my top SEO

YouTube channels page or my dwelltime

page or my top experts page. Um and I

can

>> You said it took you a long time to set

this up to set this up.

>> When I say took me a long time, I mean

it took it took like 20 minutes. Oh,

that'll be a great episode. Oh, you will

totally do that.

>> You can also get people on Fiverr to do

this for $20.

>> What would someone what would someone uh

search on Fiverr?

>> Oh, uh, looker studio report. Uh, so

what I'll do is I'll jump over to

analytics and I'll show everyone how to

do it in analytics. How's that?

>> That's great.

Um, so basically the the reason for

using um J4 is I just couldn't tie the

engines together in one report. But you

can definitely get all of this uh

straight away from your analytics

engine.

>> Mhm.

>> I'll just zoom in to

>> Yeah, great.

>> A little bit readable. So, under

reports, we're going to look for um

traffic acquisition. This could be under

generate leads or a number of different

reports, but it'll always be called uh

traffic acquisition.

>> I think Harper has shared this a lot as

well on X. He has a bunch of comments on

this.

>> Okay, great. Uh fantastic. Um glad to be

mirroring someone like like I always

say, if if someone tells you this is how

a search engine works or an LLM works,

ask for proof, right? Ask for

screenshot.

>> You should be able to do the same thing,

right? Remember, these engines are

citing data all the time. They're citing

web pages. If if someone else can do it,

they've got to be able to prove it,

right? That's just critical. So, the

first thing you want to do is look for

source. Source is the UTM source. Um,

and it allows you to pick um, you know,

referral. Uh, what we're going to put in

here is chat GPT.

Oh, sorry. I put this in the wrong

place. I need to put source as the

primary. And then I'm going to ask uh G4

to show me the landing page.

It it still says query string even

though Google stopped sending uh query

strings um you know 12 years ago. This

particular site doesn't have a lot but

it's coming into my uh blog post about

financial independent retirement early

books.

>> Mhm. Um, so what I might just do is just

change this over to a

different site that just maybe has more

more pages coming in.

>> Yeah.

>> So, we'll just do this again. We're

going to look for source.

Uh, I can put in perplexity here. I

could put in claude. I'm going to put in

chatbt cuz that's people think is you

can see my chatbtd traffic is actually

going up, right? Which is interesting.

>> Yeah. And then I'm going to turn on

landing page.

And I think this is what people pay a

lot for a lot of visibility tools and

it's great to have that automation. Um,

but I wanted to show you another trick.

So, so this is the data, right? This is

telling me for all of the searches that

people are doing, these are the pages

that ultimately people click through.

And I I reckon there's somewhere around

a 1 to 0.1% click-through rate, which is

horrendous if you're used to SEO and 1

to 40% clickthrough rate, right? Um if I

look at some of the evaluate um search

phrases in Google search console, I look

at the number of clicks I get, it could

be as low as 0.01%.

>> Um but

>> so this is a referral traffic from chat

GPT.

>> Yes, exactly. And while it does also

give a UTM source equals chatg, this is

the way to get all of the traffic. It

only stamps source equals um chatbt to

about maybe 10 15% of the traffic. So

you're better off doing it this way. But

here's a really cool um technique I use

to figure out. What I do is I just do a

a screen grab and I take a screenshot

and I go over to Perplexity

and I'll just copy and paste it in and

I'll say um guess the prompts

people

>> to get to these pages from

and uh it it's a pretty good place to

start. So if you are using a geo tool or

if you just want to get some ideas about

what you want to search for um these are

some great prompts. How does chatbt

compare with seamrush for SER analysis

for example? Uh or simrush versus chat

GPD for SEO research. Uh what is SEO

dwell time or explain dwell time for SEO

optimization or best blogs to follow?

>> And so you get this information and then

what do you do with this information?

>> So you you would then take this back um

let's pick one here. Let's let's pick

this long one here. Um, difference

between Google Search Console and Google

Analytics Sessions. So, this is a common

question when people go to Google Search

Console, they see they've had 5,000

clicks, they go to Analytics, and it's

showing 2 and a half thousand clicks

from Google, and they're like, "How how

could this be?" Most of the time, that's

because of cookies. But what we'll do is

we'll take this prompt um and we'll put

it back into perplexity.

>> I want to I want to just I want to make

a shout out to David. Everyone should go

and follow David on X. And if you have

an SEO question, Dave at David onx and

David's going to answer it. David is

extremely responsive on X

>> and very generous with his time.

>> Thank you. Yeah, I'm here to help uh

help crush myths and solve problems. And

what's interesting here is it's it's

taken the prompt, it's gone to Google,

and look, it's quoted my blog post.

as the primary result.

>> So you put it so you put it in and you

got Okay. So would you uh would you look

at maybe the queries that

>> Yeah.

>> that Perplexity thinks would be used and

then maybe you realize that you don't

actually cover this topic on your page

and then you're like oh this is

something that I should add into the

page that I made.

>> Correct. So earlier I was having a look

um because I do have a page for YouTube.

Um, so what are the what are the best

YouTube channels for SEO? And I just

wanted to see if this would be an

example that would come up. Um, so here

are some SEO channels. AFS Patel and I

noticed I didn't have your channel.

>> I know this is garbage. This is a

garbage list. What's going on?

>> But if you look here, I'm actually cited

here.

>> You are

>> four times. Yeah.

>> Right. All right. And I just updated my

blog post 10 minutes ago because I

noticed

>> it was looking

>> see the way it added 2025

>> and if I come back here.

>> So you added me into your blog post.

>> I changed my blog post to this earlier

and saved it and then I got Google to

>> I am. I'm number five. Thank you for

that.

>> There you are. So hopefully when

somebody does the search later today,

you'll be in that list. And I literally

got

>> I literally crawled this 10 minutes ago.

So

>> Oh my gosh.

>> I think it's actually I think it's

actually even in my history.

>> Yeah, there it is. There.

>> I'm extremely flattered. Thank you.

>> Uh request indexing and updated and I

reranked in Google straight away for

that search phrase.

>> Awesome. Um, another great way, and

actually a lot of geo tools do this, if

you if you're wondering about, well, I'm

used to ranking for best plumber in

Boston or uh best SEO agency in uh

Seattle or whatever it is, and you're

like, what are people typing in today?

Uh, a trick I picked up is, and let's

let's do this in chat GBT cuz I think

that it it's quite good at this. Uh, so

let's just zoom in.

Um,

search Reddit for

my ICP.

So, this is if you're using Chat GPT and

it knows your personality and it knows

your business and it understands the

type of uh work you're doing. um for

people looking for cyber security

agencies

and what questions they're asking that

also you can make typos in your

contention

>> that they're asking as prompts.

>> Have you ever have you used have you

used dev tools to view chat GPT query

fanouts? Um I I have there is a tool

from um SEO Pub who's a very active

member of

>> Oh, I had him on I had him on the show.

Yeah.

>> Oh, okay. Great. Mike Freedom.

>> Yeah. Mhm.

>> Okay. Excellent. Um so yeah, he's got a

a tool. I keep forgetting to save it,

but I will send it to you so you could

add.

>> But even even just using Chrome DevTools

to do it. Chris Long had a really good

uh he made a good tutorial on it and

then I made that tutorial as well on uh

Instagram and it did really well. But

it's cool.

So, here's a great way to uncover what

um what people might be searching for

that you didn't think about and using

Reddit to help. Um

>> so, wait, can you go back to your

prompt?

>> Yes.

What um what people ask when shopping

for a cyber security agency. So it went

to RMSP. It went to R asknetsec,

it went to our cyber security and cyber

insurance security incidents, backup and

restore processes. Uh are they do they

have sock 2 um compliance for example?

>> Yeah.

>> Um and so how do you currently assess

the maturity of your cyber security

vendors? What criteria do you ask? What

do you think of a v vendor incident?

What's worst case? So, these are things

that I think a lot of um people don't

think to ask when they're building

marketing content. And this is a great

way to actually not spend 10 days on

Reddit reading every single question and

answer. You can use the LLM and get it

to summarize. Look, this is the heart of

what people are asking. And then,

>> can you scroll all the way up to your to

the prompt that you put in? I just want

to read it out loud. Search Reddit for

my for my ISP for people looking for

cyber security agencies. and what

questions they're asking as prompts.

That's excellent.

>> And and there's a there's so much

learning here and it really takes out

the bias um that I think a lot of us as

marketers put into content.

>> Oh yeah.

>> U you know there's there's a lot of um

you know we need to say this or we need

to look like this. At the end of the day

you need to be present for what your ICP

is searching for. And if that's not the

heart of SEO then I don't know what it

is.

Yeah.

Oh, if you were to give an overview for

how to rank in LLMs,

just like a generalized overview,

how would you how would you summarize

it? I would start at my different LLM

tools, whether it's Grock or um or um

Perplexity or Claude, and I would just

start thinking what what are people

searching for? You don't have to be um

you don't have to have a subscription. I

don't have a subscription to uh to

Claude. It's it's free. And I would just

start trying to think about what I'm I'm

ranking for in Google. So um top SEO

agencies in New York. I think a better

question is to what extent should people

be focused on ranking in LLMs versus

trying to do the best SEO that they can

knowing that the more SERs that they

appear in the more they'll be covered in

LLM.

>> Absolutely. Uh that that's that's why I

I would start here, right? Um in the

sense that I'm ranking first here simply

because I'm ranking in in this list of

results,

>> right? Right.

>> And I'm ranking in here twice cuz I'm in

here and I'm in clutch.

>> And you gave the example, you changed

your page, you submitted it to Google

Search Console. So then the LLMs are

going to pick it up.

>> Exactly. And you can you can literally

take a a search where you don't appear,

go and create content for it, publish

it, and come back and test. And you need

to test it a few times so that you can

figure out how the query drifts and

changes over time. and then you can

account for that or modify or add to

your content so that you can be ever

present.

>> So you're looking at the searches that

Perplexity is doing so that you know the

language the right language to use in

your content.

>> Exactly. And here's uh Claude and Claude

doesn't use Google. It uses um Brave

Search, which is a German um search

tool, but it's 100% uh Pagrank clone,

and it works almost exactly the same way

as Google,

>> right?

>> And I think I said on on our um first

cast uh nothing has beaten Pagrank, and

>> yeah,

>> this stands over that, right?

>> Yeah. Yeah. Yeah. that for for anyone

who hasn't listened that one that so

actually I think that's my highest

performing guest episode now was our

episode oh

>> wow that's crazy

>> yeah so so uh for people who haven't

listened that's episode 8446

of the show the SEO playbook that

actually works page rank topical

authority and real results and that one

is like a 2hour

banger episode that people just love and

love and love and love Yeah. And this

this episode too, you shared some cool

stuff and I and especially just seeing

the language, showing how to see the

language that here. So, do you do you

know the uh the dev tools way to do it

in ChachiPT?

>> I don't. Do you want to show me?

>> Yeah, I'll show you. So, can you go to

Chachi PT? So, this is how to find query

fanouts in Chachi PT. So search

something that chat chatpt will actually

search the web for like fall or maybe

maybe winter 2025 fashion trends.

>> I found if you ask it for anything now

or 2025 it it does a web search. So

let's you said what fall

>> fall winter whatever 2025 fashion prints

cuz yeah date specific it's a lot more

likely to to do a search.

>> Mhm. And we're going to this is to see

the actual language that ChachiPT is

searching with.

>> Okay.

>> So, we have this. And what's so funny is

if if you ask ChachiPT what it searched,

it won't actually tell you the exact It

doesn't know if you say, "What did you

search to find this?" It doesn't know.

It It'll tell you what it thinks it

searched, but it won't actually tell you

what it searched. So, what you have to

do is you have to rightclick inspect.

>> Okay.

Um, no, no, not view page source in in

>> Yeah,

>> in speak.

>> Yeah.

>> Okay.

>> Yeah. And then, uh, can you increase the

size of this?

>> Yeah, I think so.

>> Yeah. There we go.

And go to go to uh network.

Good. And so now go to the URL and copy

the string. The string that starts after

C forward slash.

Okay. And paste that into the filter or

the search the wait is it the filter or

yeah the filter. Paste that into the

filter and then refresh the page.

And now Yeah. And that so open this up

on the bottom.

or Yeah. Yeah. Yeah. Perfect. So now

click click on the uh on the orange

bracket.

The or the orange bracket. Yeah. Yeah.

So click good. And now go to response

and then do a search within this for a

queer as in qer

and then it'll show what it actually

queries.

>> Yeah. Yeah. And there this is what it

this is what it searched.

Ah, amazing.

>> Yeah. Yeah. So, I got this from Chris

Long shared this on LinkedIn and I'm

like, that's cool. And then I So, I I

learned it and I made videos on it.

>> That's so interesting cuz like one of

the things I found so funny is when I'm

like in So, there's a lot of dead web

happening on on Reddit. There's a lot of

subs that are allowing bots and uh

companies have five bots in one

conversation, you know? So the one bot

will come in and go like, hey, we have a

case study of 400 uh data points, right?

And we analyze this and this and this.

Now, there's no reference to the data.

There's there's just the the summary of

the case study. And I think that's got

to be a red flag, right? If you if you

if you can't point to where the data is

and then you'll see like other bots

coming in and saying, "Oh yeah, this is

exactly what I thought." Or and you see

the bots talking. It's like the dead web

in real life. It's like bots talking to

bots and people reply to it. And I think

that's awful because the bots are never

going to respond to you and the same

bots are having the same conversation in

other Reddits that allow them to have

that conversation. Um, but if you

actually start to challenge the people

who spread this information, um, if you

go to Perplexity and say, "How do you

work?" Or like you said, if you ask

CatchPT, how does it work? It either

makes it up or Perplexity will actually

go and ref go and do a Google search and

come back and it'll it'll share what a

blog post tells it how it works. And you

know, I think we I said as well to you

before that these aren't research tools.

And I think that's a real danger, right?

They're just going to Google and they're

like, if you ask Perplexity how it

indexes content, it's just going to

Google and say, "How does Perplexity

index content?" and it's giving you a

bloggers version and that might or might

not be wrong. Chances are it's actually

wrong. And I know that because I can

insert my own blog post in there and

change it and I know that I don't know

how it actually works. So I know that

it's not an actually valid research

tool.

>> Yeah, it's it's it's really scary.

Actually, I had a I had Alejandro

Meerhans. He's a link builder on this

show. That was a really popular episode

and I talked with him. We just went

totally off the rails and I was talking

with him about the fairmy paradox.

>> I don't know if you're you're familiar

with the fairmy paradox.

>> Yeah. So the fing fairy paradox is like

the universe has been around for such a

long time and it's so big that it's just

inevitable that there would be and and

the rate of spread for any like

sufficiently intelligent uh civiliza

civilization

would eventually be enough that by this

time like we would just be bombarded

with signals from alien ci

civilizations. Statistically speaking,

like if you do the math, statistically

speaking, we should be bombarded with

signals from alien civilization. And

we're not. And the question is why? And

and so I had this thought. I'm like,

well, what if it's because every

upcoming civilization invents generative

AI and then they pollute their knowledge

base with AI slop that is hallucinated.

[laughter]

And now no one can get the correct

answer to important things anymore. And

prog progress just stagnates. But like

it seems it, you know, it seems like

funny and and far-fetched and silly, but

the reality is every year more

more uh pages or more research papers

are being are are citing hallucinated AI

content.

No one's no one's you know people will

try to avoid it but it is still

happening and it's still happening in

research papers and it's happening all

over the open web and like that AI slop

as a service like look how we can game

Google with the with many thousands of

unchecked AI generated uh pages

it one that gets caught eventually but

two when it's not caught then people are

learning off of this hallucinated

AI content. And uh

>> and that's what's happening on on

LinkedIn, right? There's a lot of

wellto-d do marketing people who want to

help other marketers and say, "Look, I

figured out how it works. It's citing

Reddit or it's citing this, but they're

actually operating on in misinformed

content." It goes back to, and we see

this all over the place where people

say, you know, I'm doing my research.

They're watching videos on YouTube. I

doubt that's how Harvard or NASA work,

right? But I don't think they bring in

engineers and they sit and watch YouTube

and call it research. So I think that

just by sitting down and having a chat

with chatbt, I don't think that's

research and I think people need to be

careful of that.

>> Um you raised an interesting point about

the firmy paradox something I I I

thought as a when I first read uh Dr.

Steven Hawings book as a teenager. Um,

if you look at one of the interesting

paradoxes is that if if you were an

Alpha Centura, which is the closest star

group to us that may support a a sort of

like life supporting planet, if they had

a telescope big enough to see Earth,

they would still see dinosaurs because

that's how many billions of years light

is taken across. And so even the closest

stars to us, they were only close

billions of years ago. And so if they're

moving at close to the speed of light,

they're actually even trillion. So I

think it it's actually getting harder

and harder for life to even catch up

with us unless they can get around the

speed of light. And I think that's the

problem. Either we can get past the

speed of light or we can't. And then

that's why we've actually never had any

contact.

>> You think so? You think that's the it's

the speed of light limitation is the

solution to the fairmy fairmy paradox?

Well, NASA's already put the um Voyager

um was it was it Voyager that left the

solar system 10 years ago? Um it it's

moving at like a million miles an hour,

right? Cuz there's no gravitational

force. There's no you know it's in a

vacuum. So it it's literally it it used

um Pluto's gravitational push to spin

out into the universe. It it'll take

tens of thousands of years to get in to

touch anything else. And that still

hasn't left our solar system, right? Oh,

sorry. Yeah. So, um, so we can already

once we get out of the gra the the

gravitational pull of the earth, we can

move at a million miles an hour. It's

just not fast enough.

>> I don't remember the to be honest, I

don't remember the exact uh the the

exact reason for the fmy paradox. If it

was like if if it was like you measure

the rate of progress of any civilization

and by this amount of time they should

be they should I don't even know if the

fmy paradox supposes that civilizations

should have faster than light speed. I I

I can't say that I remember that. Um and

maybe it doesn't. And so like yeah sure

that would be because the universe is

expanding. So

>> yeah that that would be my take. And and

I think we're also seeing um with with

the with the um James Webb telescope,

we're seeing

>> Oh, yeah.

>> Um galaxies further further up.

>> Way further up.

>> Yeah. This is two SEOs discussing

>> I know

>> discussing [laughter]

discussing the universe science fiction,

>> I think. But one of my takeaways that I

I would I would love people to do is

like, you know, if someone tells you

something and you can't go and prove it,

but also if you can't disprove it,

right? Like I I see a lot of SEO like I

I got into a debate with someone about

not using metad description and and this

person got really upset that you have to

have a metad description. I'm like I'm

pretty sure you don't need a metad

description. I have pages that rank for

what is Google that don't have metad

descriptions. I don't my my my my team

know that I don't allow meta

descriptions on my own blog as much as

possible. Wow.

>> Um because Google writes them in, right?

And I'll I'll demonstrate it.

>> I recommend them as a best practice, but

I also acknowledge that they're not

necessary at all.

>> So here's here's ranking, right? Apple

here. I am right above Google help above

bug Google and this page doesn't have a

meta description because it's grabbing

the text from the page

and it it ranks just fine and it's it's

it's one of my top pages and the content

is

>> it's so funny

>> relatively terrible. Um,

but I I see a lot of people saying like,

I um I always put an image or I use five

um I've got five bullet points or I've

got at least six headings. And I'm like,

have you tried it without? Right? What

happens if you try it without doing it?

And I think people are almost getting

superstitious and they're like, oh, I'm

afraid to do it. And it's like, why? Why

are you afraid to do it? You shouldn't

be afraid to try it without an image or

with an image, right? And I think if

you're if you're if you're saying,

"Well, it works all the time. That's why

it works." You're not really being very

scientific about it.

>> Some people are are in the stress of,

"Oh my god, I need to get results now."

That's when the most problems are caused

is when people feel that stress of I

need to get results yesterday.

>> Absolutely.

>> Yeah.

>> But if you if you if you want if you

need to get results, right, it it tends

to be a series of results, right? So if

you've got a SER tracker, you're

probably tracking more than 50 queries.

So if you've got a 100 search phrases

and each search phrase requires 10

pieces of content, you're now at what

a,000 10,000 pieces of content that you

have to write and publish, you now need

to reduce the number of steps. So, if

you've got um and someone wrote this on

Reddit, an everything SEO bagel, right,

where you've got everything in the page,

you got bullet points, images, schema,

all tags, everything, right? The more

things you do, the longer that's going

to take. So, if you really want to chase

long-term results, I believe you have to

bring it down to the minimum viable

product.

Can you uh we were talking about this

before the call. Can you share your

opinion on if schema is necessary or

not? And if so, how necessary is it to

show up in LLMs? Cuz a lot of people

will say, you you've probably heard it.

Yeah. If you want to get better LLM

visibility, use schema. It's so

important.

>> It it's so not important. Um, you know,

if if you take like all of the content

that I've been ranking today, like what

is Google? There's no schema in here.

Uh, the SK the schema for a a blog

article contains nothing. It just

contains the publisher, right? It just

contains and you also see there's no

publisher on this page. There's no eat

in this page, right? Um, it there's no

schema for this post that tells you

anything that is in the post. You need

the post. the schema doesn't replace the

post and the schema doesn't have

anything in it that tells you more about

the post than the post does. Uh where

schema is really really useful is if

you're trying to do something like um

flight departures from New York today or

let's just do PBI. Um if Google wants to

grab things like this review data or if

it wants to um grab this information

from a page, you absolutely need schema.

you you're going to struggle to get in

here without schema.

But that's a very very specific case

where trying to put, you know, writing

out New York versus NY, the time, you

know, using 24-hour clocks or 12-hour

clocks, using codes, using airline um

airlines full name or their short code.

That's what schema is really, really

good for, and that's what you need

schema for here. But for the most part,

99% of content, the schema doesn't add

value to the to the content. Um, and I

think you you you were covering um uh

you said on before we started the call

that Google's deprecating more schema

types.

>> Yeah. They're nutrition facts and events

like uh events near me.

>> Both of those are being deprecated.

>> If you can't get into Google, then

chances are you're not getting into

LLMs, right? So for an LLM to prefer

schema, it's got to have all of the

results. It then has to be able to stack

the results with schema versus uh no

schema. And then it has to try and pick

which what what it of what it has left,

what's the right content. And Google

keeps saying, you know, if we, you know,

if schema helps us delimit content, if

um

if uh sort of like having um a certain

amount of other things helps us find

content, those are signals, but we're

page speed, for example, um they're

never going to show a page because it's

faster over a page that's better, right?

Because scammers can put in schema.

scammers can make pages get a 100 uh%

score on CWVs for example. So this idea

that schema sort of is magical and shows

some sort of technical proficiency that

is some sort of trust signal it doesn't

it doesn't work right because scammers

can do it too. Uh people that want to

present misinformation can do it too.

People who want to hack Google results

can do it too. It's not a particularly

it's not a third party control that adds

value and validates something. And so I

I just think it's nonsense. The thing

that annoys me a lot about schema is

that marketing agency SEO agencies will

charge a lot of money for unnecessary

schema implementation because it it's

just it's a pain to implement. It's a

pain to implement and it's a pain to to

maintain. And so agencies will Yeah. And

you but for a blog post is like pretty

that's like pretty basic but but

>> but still like agencies will charge uh a

lot of money to people who don't know

better saying like you need this to get

results and yeah and and then people are

like then people will go with an agency

and then they'll have a bad taste in

their mouth about SEO. Yeah, I paid this

agency all this money. I didn't get any

results. And it's just there's so many

shady SEO agencies out there that will

do things like that. And uh there's a

problem that

>> it's not just SEO agencies. I think it's

it's it's web designers as well. Um

there are a lot of web developers uh

coming on to Reddit saying, "Oh, you

know, someone will say I'm not ranking."

And it'll be because of like crawl

discovered, which we covered the last

time. Um and I'm like, "I'll I'll put a

link to this page and make it rank."

which means all of the people saying

it's content quality, that's gone. Uh or

people saying it's too thin, that's

gone. If you can get a link to that page

and Google ranks it, then it couldn't

have been a content quality issue. There

are people that uh reply to these posts

and say you don't have enough schema.

I'm like, I don't have any schema. There

really are there are guys and they're

they're not shy about hiding the fact

that they're

>> at schema agencies. So schema agencies

have started to pop up, which is

>> terrifying. And I think they think well

schema is a structured way of like

creating entities and relationships

between entities and I'm like maybe but

that doesn't mean it's not

[clears throat] and I think someone um

used schema to take ownership of a of an

entity in Google right so just by

publishing who they were and then adding

uh one of their assets in the schema

Google started publishing that in um one

of their knowledge graphs and I think

people are overreading in between the

knowledge graph and Google Google's

supposed understanding understanding of

the internet. You know, I think Google

doesn't understand the internet. That

would be impossible.

>> What should people be focusing on

instead of schema?

>> I think people should be uh right back

to topical authority. They should be

looking at what are they ranking about?

What are the keywords they're getting

clicks on? Um how do they understand

their search console information?

Search. If you're looking to to figure

out your first blog post, Simrush,

Google Ad Planner, they're the great

places to start for guessing, but they

only have access to maybe 5% of the

keyword universe, right? Like how many

keywords are being used every day. Um,

but if you actually want real data,

you've got to take a guess. You've got

to publish. I call it, you know,

establishing a beach head. So, you

establish a beach head, you get a page

onto page one, you will now start to see

more keyword information than you'll

ever see in a keyword research uh tool.

You will see lots of different

variations, longtail, you'll see

keywords that you're ranking on page 8

for, which means they're actually very

high volume um queries. And if you start

to see queries,

>> can you talk about that a little bit

more? So, you said you'll see keywords

that you're ranking on page 84, which

means they're very high volume.

>> Yes. If you if you're on page 8, let's

say you're position 76 or 72 and you see

130 impressions, that means that a very

small percent of people, like less than

1% of people will actually go all the

way to page eight to look for a result.

If you're if you're on page eight and

you're seeing impressions, that means

the page one impressions are going to be

10 to 100 times higher.

>> And we call it in in in

>> So what do you do? So what do you do

when you even when you notice that?

What's the next step for an SEO?

>> You start looking at the Will I bring up

Search Console and actually show you?

>> Yeah, sure. That'd be great.

>> All right, let's let's do it.

>> Let's do it.

>> Um, that's this is exactly how I got to

the top for what is Google LLC, right?

Um,

>> does that has volume behind it?

>> It does. It Now, if you go to Seamrush,

it'll tell you there's no volume, right?

>> Well, yeah. I mean it's hard to trust uh

a lot of SEO tools for

>> well they they get their data from they

get their data from ads right and ads

only stores the data if um if people are

buying it and so if you look at it

they'll tell you like what is Google LLC

has a search volume of 300 per month uh

actually it's a lot higher um and if you

look down here there's on on just the

word Google LLC there's actually all of

these variations in all these languages.

>> Mhm.

>> And so what you can do is you can see

most of them are in the top 10 and then

suddenly Google's LLC is 20. You can see

Google LLC company profile I'm ranking

16. So I now have topical authority to

expand to this and I could go and write

a blog post and I'd probably go to the

top three straight away. Um, and

>> when do you when do you make a uh when

do you make a new post for something

that an existing page is ranking for

versus add that to the existing page?

>> Great question. So, this comes down to

whether or not the page is is an exact

match or a broad match. So, if this was

on a page just called Google, I would

probably struggle to get this on a new

page and the new page would probably

either cannibalize if it was successful

or if it was unsuccessful, it would just

render out. Um, and so if I was using a

high um high volume root keyword like

the word Google, I would put this in

page. If I had a very specific page, in

this case it's an FAQ that directly um

matches what is Google LLC, I would have

to do this as a new page. Um, if I was

ranking 9 10 11, I would probably do it

as a H2 on the same page.

>> Yeah.

>> And that's kind of my rough rule of

thumb.

>> Yeah, David, you're awesome. We could we

could this this episode was supposed to

be 30 minutes. We could go on for

>> for much longer.

>> Well, that's fantastic.

>> Yeah,

>> I think just you're such a great host

and and you bring out the best in

everyone you speak to and um maybe we

can do another one uh when our new uh

when my new work from home office opens

up. We'll we'll do a

>> there. We have one. So, we have uh one

scheduled for December with Mr. Barry

Schwarz. Yeah, that's going to be a

really excited for that. and you have a

list of just you're you're welcome back

anytime and I'm I'm looking forward to

the other things that that we're going

to cover. This was a lot of fun.

>> Why don't we get some people to leave

some comments in your in in the uh on

your on your video and tell us what they

want to hear about and we'll try and

cover those.

>> Yeah, that'd be good. That's a good

idea. People feel free to to leave us

some comments. If you're listening on

Spotify or Apple Podcast, please go to

the YouTube and just uh drop drop a

comment or two. All right, David. Oh,

and uh so where should everybody find

you for for people who aren't already

following you?

>> You can find me on X. You can find me on

LinkedIn. Um and you can find me on

Reddit as Primary Position SEO.

>> Cool. Cool. All right. Thank you, David.

And for everybody watching and

listening, this is episode 859

of the Edward Show. Yeah.

>> 859 days in a row. Thank you.

>> Thank you. Thank you. 859 days in a row

doing this podcast.

I will talk to you again tomorrow.