Subscribe
Share
Share
Embed
We share our thoughts and ideas on how to grow a business.
This is like the core of the problem
statement that this episode and Rand's
blog post talks about, which is like
when you ask it once, it's giving you
six to eight different responses and he
makes a good point that I agree with.
Most regular users are
just like, okay, great.
These must be the most popular once,
and he is like, but you could just ask.
The next day or in 10 seconds later.
Or you could just be another person who
is asking that at the same time, and it's
giving you totally different responses.
And if you happen to run that prompt
once, or your AI visibility tool
runs it once, and you're mentioned,
it'll say, congratulations.
You have visibility here.
And you'll think you'll
report to leadership.
Like in chat GBT we're mentioned,
dental, CRM, look, we're good.
And then the CEO's really
happy and everyone's happy.
But if you just ran it again,
you're not mentioned, how do
these tools control for that?
Yeah, it is not the same as SEO
rank tracking, SEO, rank tracking.
There's a lot of consistency in where
you show up and that really matters.
And here it's saying if your brand
shows up, that's what matters.
Welcome to the Grow and
Convert Marketing Show.
I'm Davis, that's Benji.
We are talking today about the
extreme amounts of variance.
In AI search results, meaning when you
ask chat PT or perplexity or Google AI
overviews, or Claude or whoever, a given
question about products, and then you
ask it again in a new incognito browser
or just over and over again, every
time it's giving you a different list.
And for marketers, the challenge
that that presents is how do
you even understand or track.
Your quote, unquote, GEO or a EO
performance or your AI visibility, we have
a solution for it that we apply to our
clients that we've built in months ago
into our tool AI visibility tool tracker.
Um, but what we're gonna
do in today's episode.
Is walk through this research that Rand
Fishkin did and published on the Spark
Toro blog, and we'll look at his specific
graphs and we will talk about that so
that you guys have an understanding of it.
It's a pretty dense, really good,
really interesting set of data.
And then we'll talk about the implications
of that and we'll look at the end of how
we try to get solve it through Tracker.
Sounds good, Benji.
Sounds good to me.
Okay.
So for those of you watching on YouTube
where you can see the graphics and
all, we'll show the different graphics.
Here's Rand's blog on the Spark Tour blog.
We'll put it in the show notes.
For those of you listening, we
will explain almost all of this
verbally, so you'll, you'll be fine.
So this is this blog post.
Um, we're just gonna talk through a
few key graphics because it's, it's
for those that are not, let's say.
Into math and statistics,
it's a little bit dense.
And so I've reduced this to a few key
takeaways and graphics, which we'll
put on the screen at the same time.
Number one thing, let's
explain what did they do?
Rand and, uh, it sounds like a friend
of his Patrick O'Donnell, they gathered
600 volunteers and they looked at some,
like a few dozen or a dozen or so.
Different prompts where, where someone is
asking about a product recommendation in
either, like, examples are like dental,
CRM, uh, an e-commerce consultant in the,
and cloud computing in the B2B space to a
bunch of B2C queries that are, some of 'em
are not even that like business C like.
Best sci-fi novels, chefs, knives,
or even men's fashion Instagram
like accounts or whatever to follow.
And then a couple local ones like Seattle
Real Estate and LA car dealerships.
So what they did was for each
of these topic areas like
dental, CRM, they came up with.
12 or so different prompts of
different ways people could phrase it.
Then they asked their 600 volunteers
to run those prompts over and over
again, and they accumulated it and
they say in the thing, they get
almost a 3000 total answers sets like
from Chachi, bt, Claude, and Google.
They call it Google ai.
It sounds like it's Google AI overview.
And then they looked at all of the
different products that are recommended.
So for dental CRM, for example.
They, it, it, it, it's showing in the
graph that on average around six-ish
CRMs were recommended in each response.
Right?
And we're talking about like, and
they have 87 total runs or responses
of, of that across the three
platforms and not across, I think,
for each of the three platforms.
And, and, and so then they just
see what are the dental CRMs.
And so what's really interesting is that.
The number of total
unique brands mentioned.
Let's use dental CRM as an example.
You guys can see the graph if you're
watching on YouTube, or you can pull
it up later, um, based on the link.
But using this as an example, about
six or so responses or different
brands are recommended by the, the
CRMs between Chacha, Petit Claude,
um, and Google AI overview each time,
but over 50 different dental CRMs.
Were mentioned across the 87 times.
This was run by their volunteers
for Chachi pt. For Claude.
It, you know, the, the graph has like,
kind of some loose scales, so I'm
looking, it's about 25 to 30 different
CRMs mentioned and Google AI overviews.
It looks like 55 to 60.
And so this is the first, this is like
the core of the, the problem statement
that, that this episode and Rand's
blog post talks about, which is like.
When you ask it once you go to Google
and you say like, Hey, I'm looking
for a CRM for my dentistry practice.
Or you go to Chachi BT and ask
the same thing or whatever.
It's giving you six to eight
different responses and he makes
a good point that I agree with.
Most regular users are
just like, okay, great.
These must be the most
popular ones, and he is like.
But you could just ask the next day
or in 10 seconds later, or you could
just be another person who is asking
that at the same time, and it's giving
you totally different responses.
Well asterisk, we'll see later.
It's not actually totally different,
but if you ask over and over again for
something where it's like six responses
per message, there's actually a total
of 50 to 55 different CRMs mentioned.
Yeah.
I find this interesting from a
couple different perspectives.
So one, I actually think that this.
Blue line is probably Google AI
mode, not AI overviews, just because
I have a very hard time seeing how
there's that much variability in.
AI overviews.
I don't think they explicitly
say, but I would have to
guess that that is, yeah, ai.
No, no.
He says, he says it's Google AI overview.
Unless AI overview didn't show,
then they just asked in AI mode.
So I don't know exactly which ones are
AI mode, but by default it's AI overview.
Okay.
Yeah.
Then, then I'm very, I think what's
really interesting here is how
much variability there is in AI
overviews or in AI mode from Google.
I, I've seen a lot of the
variability on the chat GBT side.
And some of the other LLMs, but
usually the brands that are mentioned
or our rankings when at least we look
at 'em in hfs, are pretty consistent
when it comes to the AI overviews.
But maybe that's preceding some
of the stuff we're gonna talk
about later on, which is right.
Brands being mentioned overall,
which is to say what you're
saying in a different way.
I think what you're trying to say
is the consistency we see could
be that, I don't know, we're kind
of now just revealing the, the end
punchline, which is really interesting.
The most popular brands actually
are consistently mentioned.
Yeah.
So, but yeah, like the total number
of brands, like when you just sit
and think about this 'cause the chart
is pretty dense and you're like, you
know, chat GT and Google AI overview
are showing six to eight different.
You know, brands per response, but
if you just ask 87 times, actually
both of them, it's like 50 to 60
different unique brands were mentioned.
What's crazy is what we were talking about
last week, the, the JH comment in terms
of JH is another colleague in the industry
who runs growth plays, but he made this
really interesting comment on LinkedIn
about, he finds it really fascinating
how many people implicitly trust.
AI for their answers.
Yeah.
Yes.
And what's what's fascinating about
that is like if you were to run
these searches and the person's just
running a question and then getting
an answer back, and they're implicitly
trusting this as like the go-to list.
But there's actually this much
variability in the answers.
I just find that fascinating.
So 50 different companies could
be mentioned, but the results
that you're getting back from
ai, the person just trusts us.
That's the list that I should trust and,
and go do research on, but that, that
list could potentially change so many
times, just depending on how you word your
question or the day or whatever it is.
Just, it's just really interesting to me.
Yeah, he, he mentions that
explicitly and he uses one of the.
Buckets in here, cancer hospitals,
and he says like for health in
particular, you know, Google has all
of these strict requirements, right,
of like the websites and the EAT
score and all this casual people are
assuming this is like vetted somehow.
Yeah, and he, I, I think the, his wording
Rand Fishkins wording in the post, I
would say, if I wanted to, nitpick is a
little bit harsh here, but he says like,
these tools should come with a disclaimer.
I, I don't know if he's saying this
facetiously, I think he's kind of
serious that says like, there's some
randomized lottery system happening here.
No, I, he's like, I think he's
probably very straightforward.
He's, he's always been big on that just.
Holding Google or whatever search
engine accountable to the answers
because, yeah, because in particular
for health, you think like, so that
this is cancer hospitals, but this is
just an example, cancer hospitals, but
he, his point is like people ask Google
and Chachi PT in particular nowadays
for really specific health advice.
I think I did ask tragedy bt the
other day for something I forget
and, and it did say, I noticed it
said I am not a doctor, but mm-hmm.
And so I imagine they, they like put
that in there to, for like legal reasons
or lawyers told them to, to make sure
it always answers medical questions.
I think they also released
something recently where they have
it, their own health product now.
Like you can't ask it
certain health questions.
Oh, I'm guessing.
Just because there's probably gonna
be lawsuits if there is this random.
Data.
Yeah, that's coming back
are random, random answers.
So, so this is the problem to the
consumer now going to marketers.
Now the big problem is if you are
looking at your GEO performance or
you're looking at your AI visibility
and you run a prompt in one of these
like AI visibility tools, and we'll
talk about how Tracker solves for this,
we think, or at least gets is better.
But if you take one of the Outta the box
AI visibility tools that are popular.
And you run and you sell a dental
CRM, and one of your prompts
is, I have a dentistry practice.
Can you recommend a CRM to me if
you, whether you are included or not.
It's like there's six brands on
average mentioned in a chat GBT
response, and you could be one of 52.
And if you happen to run that prompt
once, or your AI visibility tool
runs it once, and you're mentioned,
it'll say, congratulations.
You have visibility here, and you'll
think you'll report to leadership.
Like in chat, GPT we're mentioned
dental, CRM, look, we're good.
And then the CEO's really
happy and everyone's happy.
But if you just ran it again,
you're not mentioned, how do
these tools control for that?
How are you controlling for that as the
marketer who's like doing GEO, whether
or not you're using a visibility tool?
That's kind of the core, the
problem statement to the marketers
that Rand sort of presents here.
So I'm, I'm gonna go into some
more detail that he present.
In this chart, there's three graphs,
chatt, Claude, and Google AI overview,
and he has two bars for each of them,
and he says, bar number one, blue.
The odds that this same list of brands
will be in two plus responses, meaning
in that dental CRM, there's like six
brands mentioned in a response when
they ran it 87 to a hundred times.
Depending on which of these,
these topics and whatever.
How many times of those 87, let's
say for dental, CRM was the list
identical between the two was the
actual brands that mentioned identical.
It was for chat, GBT 0.74%.
Less than 1%.
Less than one in a hundred.
The same brands were mentioned twice.
Claude 1.6% Google AI overview 0.8%.
So that's just another way.
It's another way of saying the chances
of any two times you run the same prompt
of you getting the same six or eight
or 10 or however many you know, brands.
The LLM gives you of it
being the same brands.
A hundred percent overlap is.
More or less 1% or lower,
2% or lower for Claude.
And then, so that's just the,
it's the same thing we just said,
said in a different, and I think
interesting way, but the, the purple
bar I think is really interesting.
I'm gonna put some words into his mouth.
So he didn't say this,
but I'm saying this.
He then looks at how many times is
not only the brands the same, but
the order that they appear the same.
In the response, it's minuscule.
It's almost never 0.1% for chat GBT 0.1%.
One in a one in a thousand
0.07% for Claude, seven in
10,000 and 0.28% for Google.
A AR reviews.
And this is this thing where
everyone's obsessed over the like.
Order of your brand.
We're number two.
We're number three.
And he, this chart basically throws
that out the window and says,
guys, the order is totally random.
Yeah.
It's, it's not the same as SEO
rank tracking, SEO, rank tracking.
There's a lot of consistency in where
you show up and that really matters.
And here it's saying, if your brand
shows up, that's what matters.
Yes.
It's not, it's not about the order.
It's about if your brand can
show up multiple times for that
same prompt or the same topic.
That, that's where you're gonna win here.
And so like, do you know, we, we try
to be pretty candid, um, and more
transparent, more candid than others
on, in all of our content, but in
this sort of show podcast as well.
And so I'll say that when we look
into Tracker, when we look at some
of ours, we do have, you know, or
you click into any given prompt.
In in Tracker, we do have the rank number
of like what brand is mentioned where.
Right, so I'm looking
at a particular prompt.
What are the best content marketing
agencies for B2C brands in perplexity?
And it shows like rank
number two, grow and convert.
It's something we should think about
and talk about is just saying what
are the brands that are mentioned?
Getting rid of the, the rank, which like
you just said, is kind of an artifact of
all of us in this space being, yeah, we're
just used to reporting like this, right?
The muscle memory and habit we're used to,
which is SEO, muscle memory, but I think
also it, it does matter to the customer,
so I, I think that even as grow and
convert now, taking ourselves out of track
or anything like that, you care where
you're mentioned in that list because.
Again, if you're mentioned number 10
and someone doesn't make it all the
way down there, it somewhat matters.
Yeah, yeah.
But what Rand is saying is that's
just noise that if you just hit enter
again, you, you could be number two.
It doesn't, it, it, it's
just, just totally random.
I, I get that.
So it would be, but it's something, it's
something to like for us to consider and
think about is even that word rank, even
that word rank is just SEO terminology.
That's not a thing.
It's like brands that are actually
mentioned, and then we'll talk
later about Tracker in more detail.
We have some.
Ideas that, that we're gonna be
thinking about as a result of this.
Um, to put into tracker of, of looking
overall, he has another chart then in
the same vein that just does it for
Google AI overviews and says, you know,
that the numbers are just minuscule.
Like we're talking about the
same list being mentioned,
sometimes 0.0047% of the time.
You get the same brands twice in this
category of luxury real estate agencies,
Seattle or whatever, um, home studio
microphones, 0.022% of the time.
It was the same list.
So that, that's the same thing.
And, and then we get to what is the
second most important or like the
big kind of takeaway that we, you
and I already kind of, you know,
let the cat out of the bag here.
Ram then does, I think, a very smart and
responsible thing and says, okay, fine.
We take the, the first plot and
the thesis of like, there's just,
it mentions all these random
brands, six brands of a response.
You ask a hundred times 50 plus Different
brands are mentioned, how do we make any
sense of whether you're visible or not?
And he says, well, let's look
at the most mentioned brands.
The most common or most mentioned
brands in each of these categories.
So going back to dental CRM, remember we
ran, they ran it 87 times and in Chachi,
bts, looking at the number one, the number
two, and the number three most mentioned
brands, and out of those 87 responses, the
most mentioned brand was in 58 out of 87.
The second 1, 54 out of 87, the
third one, 50 out of 87 for Google
AI overview for this particular one,
there's even more consistency of those.
80 out of 87% of the time, the most
common dental CRM was mentioned in 80
out of 87 of the runs, and the second
most, also 80, outta 87 and the third 66.
So he then has this big takeaway, which
is kind of what you were getting at about
like in our data from all of our clients.
When we look in Tracker, there seemed
to be a lot of consistency in Google
AI overviews is if you are one of
the top few, and I don't think.
I'm gonna say top three.
He looked at the top three, but he
doesn't say anywhere that three is some
magic number and after that it drops off.
But if you're one of the most popular
ones you act, there's actually surprising
consistency that even though the order
is totally random, and even though
the other brands mentioned there are
always be some randomness in there.
That the top few ones are just
mentioned consistently over and
over again, like greater than 50%
of the time for most of these.
Yeah.
I, I would say that consistency number,
whether it's three or 10, probably
depends on the category and how many,
how many players there are in that space.
Yeah.
And, and that can change.
Like you, you think agencies, there's
probably a lot more agencies that could
be considered top than let's say, I don't
know, some consumer product where there's.
I don't know.
One product has massive market share.
Yeah.
For, um, as an example that I think
he even mentioned that, that cancer
hospitals, there's not a lot of like top
cancer hospitals in the US to choose from.
My red box, apologies to those
watching on YouTube, I just realized
is, is shifted over slightly and it
includes the blue bar from studio mics.
That's a mistake.
But if you look at Google AI reviews,
it's like out of 71 times they ran it 67.
Oh, the three.
The three.
LLMs, Chachi, BT, Claude, and Google.
69. 69, 67.
The top cancer hospital was
in almost every single time
they ran it, 69 outta 71.
So every time except two times out
of 70, the top one was mentioned.
So this then opens the door to how a
brand should be gauging and looking at
their visibility per is when you ask.
These LLMs in multiple different ways.
Remember, they tried 12 different prompts.
When you ask multiple times and in
multiple different ways, you, you need
to be thinking of your visibility as what
percentage of those times do I show up?
And I have a couple other charts where
he literally introduces this concept.
I think he literally says in
the post, more of these AI tools
should be doing this as looking
at like a visibility percentage.
And so we wanna then transition
to Tracker and how we do it.
And, and let's get really specific
on this because I think a lot of
people listening might be like,
yeah, visibility percentage.
My AI tool does that, right?
Like peak profound and all these,
it's, you open it up and it has a
visibility percentage across the top,
but that's not what he's talking about.
That's a different percentage.
That is like across all of the
prompts that were run that that day
or that week or that time or whatever.
You know, how many of the
different prompts you could be
asking wildly different things.
How many of the different prompts are
you cited once or brand mentioned once,
and that's not what we're talking about.
He's saying if you ask the same
thing over and over again, look
at your average percentage of
those runs that you're visible.
If you're one of the top brands and
you're, and you ask a hundred times or
71 times here for canceled hospitals,
and your hospitals mentioned 69 outta
71, you're gonna have some 90 whatever
percent visibility you're doing well.
If you're mentioned twice out of 71, you
got lucky those two times, but you're
not really visible for cancer hospitals.
Let's show Tracker because I
think we hop into Build Tracker
pretty much to solve this problem.
But I think now with a
lot of this e evidence.
It just helps make the case for how we're
approaching Tracker differently than a
lot of the other AI visibility products.
Yeah, I mean we have outlined
this in previous episodes, but
we had, you know, what we call
Tracker 1.0, this is Tracker 2.0.
In between, we sort of notice this and
said, Hey, like every time we run this we
get, there's like a different response.
Now kudos to Rand for quantifying
it and collecting the data, but
we sort of anecdotally noticed it.
And our solution is this kind
of topic based organization.
So for those of you listening,
I'm now screen sharing tracker.
You can go back and, and
look at this on YouTube.
You can check out Tracker yourself.
Um, but when you set it up by default,
not as an option, but by default when
you set up your brand, the number
one thing you need to do is pick.
Topics.
That you want to track and a topic
is something we're looking at.
Grow and convert.
Content Marketing Agency, SEO Agency,
B2B, content marketing, service
B2C, content Marketing service.
So those are topics, and then
what Tracker does is it runs.
Five different prompts in that topic
that are related to that, which is five
different ways you can ask about it.
The first one in bold is always
just that exact topic name itself.
Why is that relevant?
It, it's not realistic for chat
bt, but for perplexity in Google.
A overview.
People in Google, absolutely, people
still type in things like short
keywords, B2C, content marketing service,
what I'm showing now on the screen.
But then it also asks natural language
sounding questions, which you can edit.
So for B2C, content marketing service,
zooming in on this for a second.
The second and third one are.
What are the best content marketing
agencies for B2C brands and I have
a B2C business and I'm looking
for a content marketing agency.
Any recommendations?
I think those are basically asking
the same thing they are, but in
this particular run, the first
one, what are the best content
marketing agency for B2C brands?
We are not mentioned.
Grown Convert is not
mentioned in the second one.
I have a B2C business and I'm looking
for a content marketing agency.
Any recs we are mentioned.
This is exactly what
Rand is talking about.
If you didn't do this, if you
didn't ask about B2C content
marketing in two different ways.
If we asked in the first way only, we
would say, oh, Benji, we're not mentioned.
If you asked in the second way only,
we'd be like, Benji, don't worry.
B2C.
We got that covered.
We're mentioned, we're
one of the ones mentioned.
Neither of them is a true story.
It's that there's some
frequency, some percentage of
times you're, you're mentioned.
So for Tracker, you ask the same topic.
You go at that from multiple
different angles and ways, and then
it tells you the percentage here.
Out of four different questions
about B2C content marketing.
Services.
We are mentioned one outta the four times.
So Tracker says you have 25% visibility
and you can open up the graph and
see how that's changed over time.
This is one of the most interesting
things because we noticed this early
on too, where one week we would.
Show up in shot GPT the
next week we'd fall off.
The next week we'd show up again.
The next week we'd fall off and, and I
think the Report by Rand really helps
explain why that's happening and then
also why you need to track this over
time and through different prompts.
Yeah, so in, in the dark blue
line here, which is Google,
you can see back in November.
We didn't have, we were at, you know,
25 or less percent visibility, and
recently starting December 12th and
onward we've been at 50% or over.
So you can see like an
overall trend again.
I wanna be very explicit here because we
think about this all the time, so it's
obvious to us, but, but I, but it may
not be obvious to everyone listening.
This visibility graph is not the same as
the overall brand visibility number that
most of these tools show you when you
click into your brand as like the first
graph, and it'll give you a single number.
It's Grow and Convert 34%.
That is how many different,
totally different prompts are
you visible for over time.
This is the same topic.
So this is, we were literally trying
to solve the same problem that Rand's
Post kind of presents the same topic,
B2C content marketing asked four
different ways, how many, you know.
What percentage of of those
four are you mentioned?
And so this is how you can start
to gauge, are you one of the most
mentioned brands that is gonna
be kind of mentioned regardless
of however anyone asks about it.
At any given time, you're probably
gonna be mentioned or not.
And so the other, so
I'll show you an example.
If we go to another brand that
we're, we're tracking under one
of our clients, ServiceTitan is
an absolute leader in their space.
And if you just scroll through,
look at, look at these percentages.
HVAC business management software, a
hundred percent chat, bt 80% perplexity,
a hundred percent Google AI overview,
all of them brand mentions and like we're
asking about HVAC business management
software in five different ways.
This is what we mean is when
you see this, you're like, okay.
We're one of those common ones.
We're one of the most
mentioned brands in Rand's.
Study language for HVAC
business management software.
Like we're good here.
Regardless of how you ask about
it, how anyone is in the real
world is gonna ask about it or
how many times they ask about it.
The chances of us being one of the
brands listed is very, very, very high.
And then you can look at the
graph and be like, yep, we're
sort of, they're just hovering.
I open the graph, they're
just hovering in the like.
75 to a hundred percent over
week after week after week.
This, we think is the responsible
and accurate way to report on GEO,
not you ran this one prompt once and
this week it happens to mention you.
Yeah.
Do you, do you wanna show the,
the new report that we built in
there because it, it's a good.
I would show it actually on ServiceTitan.
Okay.
So to add to this, at the top, initially
we have this classic percentage, which is,
which is more analogous to the visibility
percentage of all these other tools.
We had the same thing.
So we're showing ServiceTitan chat, GPT.
Yeah, that's, that seems to be.
There we go.
It was just a refresh issue.
Chacha vt, 74% visibility.
Perplexity.
Right now, 48% Google, 78% or whatever.
This just means whatever prompts
you happen to be tracking.
How many of them do you get a mention?
But there's some issues with this, which
is if you start tracking additional
topics or prompts that are more
ambitious, let's say something you,
some topic area that you want to have
visibility for, but you don't yet.
Right.
We, countless clients over the years
have done this where we're like,
yeah, yeah, we built our name in this
space, but now we wanna move to this.
We wanna move to the enterprise,
or we wanna move to this
other side thing, or whatever.
Or we came up with some new feature
that we hadn't talked about before.
Right there, there's always new things.
Yeah.
And so the marketing team, the concept
team should, should start tracking this
and whatever tracking tool they do.
But if you start adding those topics,
you don't have visibility yet, and all
of a sudden this percentage goes down.
And we were like, this
is artificial and weird.
And so what we've built in
is additional ways to do it.
And the one we wanna talk
about that's relevant today is
what we call topic visibility.
And what it does is it creates three
buckets and it says high visibility
for how many topics that you're
tracking are you mentioned 67 to
a hundred percent of the time.
Like in the top third, you're like
two outta three or higher percent
of the time of the prompts you're
mentioned medium is like middle.
And low is, you're mentioned at least
a couple times, and so you can track
this and see like at a topic level,
ServiceTitan, in this case, we've
only set them up with eight topics and
Tracker, they have a huge business.
We could have this be 80, right?
Sort of.
But in this particular instance, eight.
Eight outta the eight topics, four of
them, they have top tier high visibility,
meaning the majority of the time.
They are visible.
The other four, I think the simple
way to say the, the to say the high
is no matter what prompt you were
to type in around this topic area.
So if you were to ask the question of
field service management software in
multiple different ways, if you have
that high score, you're more likely to
show up as a brand mention than not.
Yeah, and medium is like,
you're pretty likely it's 34 to
66% visibility in that topic.
And then low is, you mentioned a
little bit, so you know, you can
see like for ServiceTitan it's,
it's four, the eight is high.
Four, the eight is medium.
And so they're like, they're,
they're sort of killing it.
We are very much in
agreement with Rand study.
We have observed this, we've tried to
build this into Tracker, and then we can
close with some forward looking things.
And we actually think if we go back
to Grow and Convert as an example.
That B2C content marketing service.
So that B2C content marketing service.
You know, if you asked in one way,
we were not mentioned in another way,
we were mentioned a couple details.
We're working on an overall
filter to, to change all of these
numbers to only brand mentions.
That's something to think about.
This RAND study is looking at
that generated output text.
Buy these LLMs and how many times is this
particular brand recommended or mentioned
as like, do this for dentists, CRM or
cancer hospitals or those categories?
They tested a lot of these other tools.
They're just counting your
visibility on citation as well.
We are right now, but we're
like very close to releasing em.
Just a simple toggle switch where you
count all of this for brand mentions.
Only now.
Right now you can see for those
of you looking at the screen.
We make it very explicit and clear on the
overview page, which of the prompts you
have a green filled in, brand mu, versus
which ones you only have a citation.
You'll later be able to get
these numbers and stats soon to
also only count brand mentions.
The second thing is.
We have the data already of historical,
like the historical data of running
the same prompts over and over again.
And so we're working on, in the early
stages of, of talking about and working
on how can we let in tracker someone
see some more advanced analytics
saying, okay, not just this week,
but over all of the times that I've
been tracking this topic and tracker.
What's my cumulative like over the
month or over the, you know, all time.
What's my cumulative topic?
Visibility percentage, because
you can see here for Chachi BT
for B2C content marketing agency,
that's the, the orange line.
So if I turn the other two
off, that's the orange line.
We happen to have one this week.
Last week we didn't, the
two weeks before we did.
And so like overall you can
get a more accurate picture.
And so that's kind of how
we're thinking about things.
Even what's interesting there with
the, the orange line is, is something
that we did in November, December has
now improved visibility going forward,
which I think is very interesting.
So that's kind of what you're talking
about is like if you're directionally
moving in the right place, um.
How, how could we better show
that in the, in the product that
there's more consistency on brand
mentions now versus two months ago.
Yeah, but overall, like you, everyone,
you, you guys, you need to be thinking
about AI visibility in more of this
kind of average, you know, over
multiple data sets, over multiple runs.
How, how often are we mentioning you
need to be thinking about it that way,
not just like this one prompt this day.
We happen to be mentioned, so we're good.
That's sort of, you know,
irresponsible or just very inaccurate.
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.