Subscribe
Share
Share
Embed
We share our thoughts and ideas on how to grow a business.
We're gonna call these general sites
like Reddit is a general purpose site.
Wikipedia is a general purpose site.
General sites are not
being cited very often.
It's like under 20%, 80% plus in
some, for some clients it's like
90% plus of the citations are
industry specific publications that
will never make it onto that list.
So if you are selling it, security
software or something Very niche, yeah.
Reddit's not gonna, Wikipedia
is not gonna be on there.
It's gonna be these really.
Specific like domain specific things.
If you're doing comparison shopping for
Tesla versus other cars, the information
is probably gonna skew more towards
Reddit or YouTube or consumer reports
or some of these larger sites that
have specific information about cars.
But that doesn't mean that that's what
you should do for your B2B business.
Okay.
Welcome to the Grow and
Convert Marketing Show.
If you like this, by the way, you
can keep subscribing in a few ways on
whatever podcast system you're using.
We would love likes on YouTube,
or subscriptions and comments or
ratings on podcasts would help.
Today we're gonna continue the last string
of episodes on GEO and get even more.
Into the tactical details
to help you guys out.
So we have talked about
our overall GEO framework.
Prioritize GEO.
You can check out that episode.
It's a few episodes before this.
We have also gone through an
example of implementing it using
a company that we did a demo of
our AI visibility tool tracker.
And talked about, um, the details
of how to kind of implement
this on a tactical level.
What do you do?
What content do you produce?
And continuing on that, a part of what
we talked about in, in prioritized GEO,
and that was citation outreach, which is
figuring out what sources LLMs are using
to in, in the topics that you care about.
In the responses, the LLMs
are citing these sources.
What sources are they citing?
And then saying, we said
reach out to those sources.
We do this for our clients.
It's a service we offer, we
call it citation outreach.
Um, we're partnering with another firm
to do that, kind of like link building,
but it's like the AI equivalent.
So today we wanna talk about a study we
just published that's super interesting
about patterns in those sources that LLMs.
Are citing.
And so to sort of kick it off, I am
gonna share some examples here of
LinkedIn posts and charts we've seen.
And both of these kind of say
the same thing where it's like,
where does AI get its info from?
And they've listed the top 10 or 20,
uh, like domains that AI tends to cite.
And everybody knows what the
top one is gonna be there.
It's Reddit.
Followed by Wikipedia, followed
by YouTube, Google, Yelp.
Like both of both of these are very
similar by two different sources.
And our data that we're gonna talk through
today and in the accompanying article
suggests that basing your GEO strategy
on trying to get mentioned in these
sources is a bad idea and the reason.
Is that when we look at the actual top
source citation lists of the topics
and prompts that matter to our clients
that we are looking at and tracking
via tracker, dot ai, these general,
we're gonna call these general sites
like Reddit is a general purpose site.
Wikipedia is a general purpose site.
General sites are not
being cited very often.
It's like under 20%, 80% plus in some,
for some clients it's like 90% plus.
Of the citations are industry
specific publications that will
never make it onto that list, but
for that client, it actually matters.
Yeah.
I wanted to kind of share some thinking
in terms of why, like, e even the study
doesn't make a lot of sense to me.
So, uh, I think the study is largely just
used for an infographic because it makes.
It makes it look like this is
the answer to everyone's problem.
But if you just think about the way that
the study is done, so even if you were,
so it says based on 150,000 citations
for 5,000 randomly selected keywords.
Okay?
So even if you, if you crunch all
that data, of course the largest
sites that are, that have like the
most brand recognition are gonna be
the ones that are the top sources.
Because all the industry specific sites.
Are not large enough to be shown across
all these different 5,000 queries.
So largely what I'm saying is this
makes it seem like the largest sites
influence LLMs more than other sites,
but it's just because these sites
are the most well known and get the
most traffic, and that's why you
don't see any smaller sites on here.
It's all Reddit,
Wikipedia, YouTube, Google.
Because, yeah, those, those
sites are just gonna come across
the data more often than not.
And the industry specific sites for
very specific queries are just not
even gonna be shown here because
they don't get enough traffic.
So, yeah, and set it.
Set another way, saying what
you're saying in another way.
Your goal as a brand is not to show up
on 5,000 randomly selected keywords.
Your goal as a brand is to show
up on the handful of keywords.
Keywords as kind of SEO thinking,
let's we call it in grow and convert
topics, which are like groups of
prompts that are related that your
show goal is to show up on the topic
that actually matter to your thing.
And so if you are selling.
It Security software or
something Very niche.
Yeah.
Reddit's not gonna, Wikipedia
is not gonna be on there.
It's gonna be these really specific
like domain specific things.
So, and, and we're gonna
get into the data here.
Here's what we did.
We looked@trackerdatatracker.ai, our
visibility tool for five of our clients.
It was, I think 120 different prompts.
And in Tracker, you can see, and
I'll show screenshots in a bit in
our previous videos, do what are
the common source, what are the most
highly cited sources for each topic?
And we looked at that carefully
for all of these prompts.
And basically the long and short of
it, the key data is 86% of these sites.
We're what we are labeling
industry and adjacent sites.
That's what was cited.
Only 14% were the Reddits and
Wikipedias of the world, what
we're calling general sites.
And so like just pausing here, the actual
data for our clients, four of the five
of them are B2B, one of them is B2C.
And this held for the B2C one as well.
But for our actual clients, 86%
were industry specific sites that
you will never see on a list like
this that you see in LinkedIn.
Okay.
And so, yeah, wait, go, go back.
I have some more stuff to add.
So yeah, in in essence, what we're
saying is this actually the inverse
of everything that you've been told?
Number one, so you've, you've been
told go do this general PR on Reddit
and go get on all these general sites.
But from what we're actually seeing
in the data when we're looking
at topics, the industry specific
ones are what's being pulled.
To give answers more
than Reddit and YouTube.
The other thing I don't like about these
broad studies that are being shown on
LinkedIn and like that infographic is a
lot of the tools like Profound and Peak
when when they're using examples, and
I think this is the examples on their
website or in some of the studies that
they've shared on LinkedIn or on Twitter,
they're often using brands like Tesla.
Like Samsung, they're consumer brands
and they're saying the data for Tesla
is the same as what you should do for
your B2B company, which just doesn't
even make sense because think about it
this way, if, if you're doing comparison
shopping for Tesla versus other cars.
The information is probably gonna skew
more towards Reddit or YouTube or Consumer
Reports or some of these larger sites
that have specific information about cars.
But that doesn't mean that that's
what you should do for your
B2B business because someone
searching for best design software.
Is probably not going to, like, the
information is probably not gonna
come from Wikipedia and Reddit.
It's gonna come from design sites
or other industry specific sites.
And so, yeah, I, I think largely these
studies are misleading because they're
using brands that don't represent
similar brands to your company,
and they're using data that just.
Isn't really relevant
to your business either.
Yeah.
And we're gonna get into it.
We did.
Um, I, I actually should have
shouted her out at the beginning.
This study was done by, um, Jessica.
Oh dear.
Do I even know how to
pronounce her last name?
Jessica Reitz.
At least it's how I say it.
It, I, hopefully it's right.
I know.
Hopefully she like forgives me
when she watches this, but, um.
Jessica and I did this, and we had a
really good example that she has in
there at the end that I'm gonna go over.
W comparing this kind of citation stuff
to Peloton, which is the equivalent
Benji of what you just used, uh,
Tesla and Samsung as an example.
This really like general purpose.
Everyone knows it.
Consumer facing brand to compare
to see is this like a B2B B2C
thing, but to cut to the chase.
Um, what I'm gonna show is she had this
really nice analysis where she was like,
Davis, I don't think it's actual B2B
versus B2C because we have an example
in here of an educational, like training
company, or actually it's a nonprofit.
It's one of our clients climb hire
that technically is consumer facing.
That's their audience, but it's niche.
So, um, I'm gonna mess up the
wording, but in the article, she.
Had a really nice phrasing
where she says it's not really
about consumer versus whatever.
It's about how, um, niche the brand
is and Tesla, Samsung, Peloton
that we'll look at is like, is so
general that these general site Yeah,
they're, they're all household names.
I, I think that, that,
that's the difference.
These are, these are large enough
brands where they're household names
versus, and you would do research on.
You actually look on Reddit if you
wanna decide what exercise bike
to get or what car to buy, right?
Exactly.
But if you look at YouTube
reviews or look at this, like
Amazon shows up on all of 'em.
We all know what we go to Amazon for.
It's like consumer products.
And so that's what you
cite when you do that.
So we're getting ahead of ourselves.
Let's continue this example.
So 86% of the sites are
industry and adjacent.
Here are a few examples for this trucking
company, client of ours, Toro, TMS.
For some brands that
really matter to them.
Just let me just, you don't even
have to know what these sites are.
I don't know what these sites are,
but let me read them off to you.
The top cited domains for this topic
that really matters in trucking software.
AMCs group.com.
Do you have any idea what this is?
Neither do I. Why?
Because it's a trucking
specific organization.
Trans virtual Toro.
Our own clients.
TT News is like trucking news,
dump truck dispatcher.com, and
we have some others, right?
For like.
Um, this project management client,
the top brands cited in, in this was
rike.com, product, product productive
dot e uh, io, the digital project
manager.com, Zapier, whatever.
Then for this particular other
software like you can continue
so that this is what we mean.
So this is inside tracker.
Tracker our tool like list the domain
cited and the topics that matter.
So that's like the key example there.
So in here, what do you not see
out of these three B. Two B?
The top what?
Five domains cited for three
example topics of three B. Two
B Clients you do not see Reddit.
You do not see Wikipedia.
You do not see YouTube.
Or Amazon because it doesn't make sense.
So then here's the Peloton example.
So we thought of this, Jessica and I
did as we were sort of working on this
study, is we said, is it a B2B B2C thing?
And I said, you know what?
Let's just do something.
And I, I forgot where someone had
mentioned it and a podcast where I was
listening to, so I put Peloton in there.
It has crazy brand visibility, as
you said, household name, right?
Anytime an exercise bike is mentioned,
um, it's gonna be mentioned.
And the two topics or the, the five topics
I had were all around exercise bike.
So as an example, I'm looking
at five prompts under the topic.
Umbrella of best exercise, bike for home.
Another one, virtual
spinning classes online.
Another one home exercise
bike with classes like that.
Okay, so then when we look at its top
domain cited, what do we now see in
contrast to the three B2B clients we had?
So one of 'em is a random site garage,
my reviews.com, but then right under
that youtube.com, New York times.com.
Healthline, PC Mag, target, Forbes,
Walmart, Reddit, Dick's Sporting
Goods, just intuitively as jam.
Benji was just saying, this makes sense.
That's where you go if you wanna
research consumer products.
This is what bothers me again about a lot
of the, the advice online because people
are using these household name examples
as examples for what you should do.
And so it's saying, go after
these general sites, do these.
Large scale PR campaigns, go get
in the New York Times and PC Mag
and, and all these sites, but it's
just not true for most companies.
Most companies are not these household
names, and so when we look at the
sources for a lot of our clients, it,
it's just complete opposite of this.
Yeah, and I think that actually
that, that's a good point.
It's this other thing where this, the
footnote of one of these graphs at
the beginning said, based on 150,000
citations of 5,000 randomly selected
keywords from the SEM Rush Deep Database.
When you select 5,000 random keywords from
some SEO database, statistically you're
more likely to get consumer keywords.
There are more way, way, way more
individuals, humans searching things on
anything, whether it be Google or Chat.
BT related to their personal
individual search interests.
Right, which is like consumer products
questions like why is Wikipedia on there?
Well, also what keywords
you select really do matter.
Right?
Like that.
Like we shouldn't be doing
randomly selected keywords.
Okay.
Yeah.
And what I'm saying is randomly selected
keywords is gonna bias us to just like.
The, the things you Google day to day,
like what, what movie should I get?
Like for example, I think Wikipedia
is on here because like actually five
down is like Mapbox and TripAdvisor.
Why?
Because when people are going on vacations
or traveling somewhere and then that
one after map box is open street map.
I actually don't use any of them,
so maybe it's like never even heard
of, but, but like TripAdvisor,
Mapbox, open Streete map.
Why?
A lot of the things you type into anything
is like how to get somewhere or like I'm
going on a trip and how to get, that's
what I mean is like when you randomly
select keywords, by definition, there's
more searches for consumer queries.
There's just a lot more humans googling
stuff for like their own personal life,
not for trucking management software.
It's so niche.
But if your job is a marketer for
a trucking management company,
you don't care about these
randomly selected keywords.
And, and I was gonna say.
I think Wikipedia is on here.
Second, because there's a huge
number of searches that's people
just trying to understand concepts.
It's not, it's not
business related at all.
So like you, you, it doesn't matter.
So though, what you
should do, the action is.
Figure out the topics.
And this is gonna, for those of you
that have followed everything we've
written recently, or all the videos and
podcasts we've produced on this channel,
this is gonna sound redundant, but it's
important is what you should do is start
by deciding your topics that matter
in your AI visibility and news flash.
It's gonna map very carefully to
a bottom of funnel s very well.
To a bottom of funnel SEO strategy in
both SEO and GEO, what you should do is
decide top down, these are the topics.
In SEO, you would use the
terms keywords, right?
In GEO, we would say topics that
a bunch of prompts would fall
under that matter to our brand.
We want to show up for this trucking
management software, this and
that dispatch software, whatever.
That's what matters.
Then you decide what prompts are
sort of related to that tracker.
Does that automatically for you?
Then.
You look up, okay, for these prompts
on these topics that we have decided
matter to us that we want to show up.
Four, when somebody types it into chat
T, this is what Chat t perplexity.
Google a IO.
Reviews, what have you are citing.
So to continue Peloton, the final
thing I'll say is it actually, we
have this correlation to Google
again, and I have two Google
screenshots showing on the screen now.
One is for best exercise for home.
I scrolled down, you can Google
it if you don't trust me.
I scrolled down a couple positions.
There was Reddit R cycling page.
What's hands down the best exercise
for home, why that matters.
Like that's, you would go to Reddit
to see what people are are saying
for a consumer product like that.
And so this ties back to our GEO
pyramid of our prioritized GEO
framework of even for these consumer
queries, like the big domains that
are being cited for Peloton, why?
Why are they being cited?
Because those are the domains that
show up when you Google that term.
And as we've said, these LLMs are
grounding their search in Google.
And so you can just use as kind of a hack.
If you don't use Tracker or whatever,
AI search visibility, you could just
look on Google and if you look on Google
and type like best trucking management
software or whatever like our client
would do, you're not gonna see Reddit in
the New York Times, but you will for best
exercise bike for home, you have a wire
cutter review, you have a Reddit threat.
And so these things are very
correlated and that's why tier one
and tier two of our pyramid is either
be have your own content ranking.
In Google for the terms that matter
or be mentioned in the other things,
in the other pieces that are ranking.
And as Benji said in our last
video, using that example in
Tracker, this would be an example.
If you are Peloton, you don't need
to do general purpose Reddit work.
You can literally decide what topics
matter best, exercise, bike for home.
And you can see, oh, um, this is the
exact Reddit thread that's being cited.
Or whatever for this one, it doesn't
actually show up, but this is the
exact thread that's being cited.
And then go and get
mentioned in that thread.
And that's more targeted.
So the last thing that, um, I don't
have a visual for in this episode, but
is in the article, and the article goes
into a lot more careful detail around
this, is we even looked at this from,
we categorized these prompts as bottom
of funnel, what we called mid funnel.
So bottom of the funnel was like pure.
No question.
It's a prompt where people are literally
asking like, what exercise bike should
I use or what, blah, blah, blah.
Software should I use?
Mid funnel was like something related, a
query that was related to products, but it
wasn't like, what products should I use?
It was like, what features are
important when someone is doing best?
The other, and then top of funnel was
just like general purpose questions
that were not product related.
Actually this 86% to 14%, Benji had very
little correlation to bottom, mid and top.
I think there was like slightly
more general sites for top of
funnel, like how to, or general
informational prompts and queries.
But it was, it was like barely more
so still largely like 80% plus.
Whether it be top of funnel.
Yeah.
I mean, it, it, it
actually just makes sense.
If your job as the LLM is to find the best
information to answer a question, are you
gonna go to large generic sites or are
you gonna go to very specific ones that
have detailed information about Right.
Whatever topic you're searching.
Yeah, that's exactly right.
It's like the L LMS are actually smart
enough to know, oh, it's asking about some
niche B2B thing in trucking software like.
I'm gonna go Google that.
Oh, that's what I was showing the Peloton
thing for is they're going and go.
The LLM is Googling it.
Like one, it's smart enough
to know what sites to look at.
It's not looking at New York Times
Wirecutter, Wikipedia, and YouTube
for some niche B2B software thing.
But also it's Googling it and
Google's telling it like, oh
wait, here, here's the results.
And it's trucking specific sites.
And that seems to be not just for
bottom of funnel, like even these
informational queries, what we're
calling top of funnel prompts.
It was still, it still held.
Yeah.
I mean, it just, it makes
complete sense to me at least.
Yeah.
This is a shorter episode.
That's, that's what we have.
You can read the full article, but again,
just to emphasize and, and close on a
tactical and actionable piece of advice.
It means how, how, what, how do,
how do you implement this for like
getting AI visibility for your brand
is again, you decide what topics and
prompts matter and you look at it.
If you don't use software like.
AI visibility software, whether
it be Tracker or anything
else, just do it manually.
Like literally go and go into chatt or
go into Google AI overviews, like Google
it, chatt, perplexity, whatever, Gemini.
And then write down in a spreadsheet
what domains are being cited and
just think about it and talk about
it as a marketing department or
team and, and, and look at it.
And, and you should be able to see
this and then say, okay, if we're
gonna do some kind of PR outreach.
We're gonna outreach to these specific
domains because they're being cited, not
just random stuff that like people on
LinkedIn are saying, you like this video.
Don't forget to subscribe.
You can also get the audio only
versions of these shows wherever you
get your podcasts, and you can follow
us at Grow and convert.com/newsletter
for any articles and updates
for when these videos come out.