Copy the formatted transcript to paste into ChatGPT or Claude for analysis
All right. So, um, our next speaker, um,
is Monik Bond. Uh, Monica and I actually
hung out a few months ago in Toronto and
we're really deep on SEO and, you know,
the the depth of his knowledge really
impressed me to the degree which I knew
he had to become a speaker and actually
a sponsor as well here. So, go check out
the booth in between breaks. So, let's
give it up for Monik. Monik Bond,
everybody.
How we doing?
>> Good morning, Toronto. How we feeling?
>> Yeah. All right, let's get some energy
in here. Where's my clicker? Do I get
one? I have a laser pointer. Who do I
not like?
Just kidding. Um, the clicker would be
great though.
>> That's okay. We like Ian. All right,
guys. Today
I thought what would be most helpful for
me to share with you. And this year has
been a lot of things. A lot has changed
on Google. Looking back a year ago, SGE
was an experiment. Today, AI overviews
are the new norm. It's hard to overstate
how much the search landscape has
changed. And in the last 12 months, I've
made 11 white papers. We've done a lot
of SEO research. And I thought that the
way that I could be most of service to
you all today was to share with you
those white papers. So this is our
agenda for today. We're going to go
through some selected local SEO research
papers. We're going to do a little SEO
mythbusters which is always fun. We'll
look at semantic SEO and how we can use
principles from that in pain media.
We'll do some LLM introspection. We'll
and then we'll look at some backlink and
content research papers. We'll talk
about aic SEO and then we'll take some
questions. Okay, does that sound good?
>> Yes.
>> All right, let's roll. We're cooking.
So, I'm Anik Bond. I'm the founder of
Search Atlas and Linkraph. We work with
a lot of the brands up here. You might
be asking like, why would Search Atlas
still have an agency? And the thing is,
the agency is really what allows us to
know, separate theory from fact, what
actually works from things that are just
concepts. So, that's the beauty of
having the agency. And our goal is never
to have more than 100 clients. We have
about 90 something clients. And in order
for us to take on a new one, someone's
got to go because the goal is not to
scale the agency infinitely. It's really
a quest for truth and knowledge. And I'm
happy to share everything that I can
with you guys, especially on my open
source GitHub repo. So here is what you
can find on my GitHub. You can find
patent brain, a GPT with every Google
patent. You can see some of the research
we've done around visualizing topicality
for websites and more and more. So
patent brain was quite cool. We took all
the Google patents and my wife was
telling me, "Monic, you got to stop
doing this research. Like, you're not
hanging out with us." And so, it's like,
"How can I do this research faster?" So,
we took all the research papers and
patents, dropped them into an LLM, and
here we go. 9,525
patents. I think it's a little bit more
than that because new patents have come
out. We also have one for Bing, and we
also have one from uh OpenAI. And we've
learned a lot. We've been able to use
this to ask some interesting questions
like the ones up here. But this is
really, I think, a great launchpad for
your curiosity. If you ever wondered
about what do the patents say about what
Ian was talking about, CT manipulation,
methods of filtering, click signals,
measuring topical authority, which they
actually call topical relevance, all of
that can be found in these patents. And
so, let your curiosity be your guide.
I'll show you some of the things that we
were interested in learning about. Let's
start with some local SEO questions. So,
we've always known that Google Maps is
really volatile, right? We see that we
can observe that in our heat maps. But
people don't really do intraday
local SEO rank tracking. And so the
experiment was to take it was about 50
different businesses, some open 24/7,
some not, and track their rankings on an
hour-to-our basis. And we did that for a
total of 13 days. We started with these
heat maps. We're able to see how the
heat maps transform over time. And then
we're monitoring and seeing how the flux
looks for those businesses. And we tried
to find a pair of competing businesses
that were in the same market competing
with each other but had different open
and close times. We also looked across a
multitude of markets to understand does
New York City for example behave
differently than Toronto than Dallas. So
here's what we learned. Two coffee
shops, the one on the left, the one on
the right, um different businesses, same
area. Look at that volatility.
That was the first surprise is how
volatile these rankings are. At the best
here, we see position three and at the
worst we see position 15. You guys know
that local SEO heat map tools, I don't
think a single one of them besides the
one we have in search atlas actually
lets you pick what time of day you're
pulling your rankings. So, if the server
decides to run your ranking report at
3:00 a.m., your rankings are trash. They
don't mean anything. And so looking at
this, we realized how sensitive these
local SEO rankings actually are. And we
can see a really strong temporal signal
here. What we find is that for this
business, their best rankings are in the
mornings from 8:00 a.m. to 9:00 a.m.
That's when they're open and their
competitors are closed. And you can see
the same trend with their competitor.
And what you what you seem to find here
is that there is a open for business
advantage. If you're open while your
competitors are closed, you're going to
be getting better rankings because being
open is a huge component of the ranking
signal. So much so, and it applies to
all industries. So, here's another
example of this. On the left, Walgreens.
On the right, Preston Road. Walgreens is
open 24/7. Here, we can see the dynamics
here. And look at how they win the best
rankings at midnight, right? Because
that's when the otheries are closed.
They get their worst rankings in the
morning throughout the day until 11 p.m.
because that's when their competitors
are open. So, this was really
interesting to see. Again, a lot of
volatility. And you can also see here
there is a weekday versus weekend
seasonality as well to these rankings.
Very interesting. But what about law
firms?
Who picks a law firm just because
they're open, right? These cases go on.
This isn't a cup of coffee. It's not
some it's not some pharmaceuticals. This
is a different type of local business.
But what was interesting is that we also
see a similar trend with the rankings
for law firms and also for plumbers and
also for roofers and for car
dealerships.
You guys get the picture. Pizza as well.
Okay. So, a lot of useful insights here.
I'm gonna keep it moving. There's a lot
I want to share. So we're going to keep
rolling through these slides. Uh also
what was interesting is that we saw that
the branded searches significantly less
volatile which is what we expected than
generic uh exact match searches. We also
see and this is important. All right.
How many of you guys are agency owners?
Pop your hands up. Yep, that's me too.
How many of you guys are in-house
marketers?
All right, decent clip. So when we're
tracking our rankings, especially in
local, the word the word order really
matters, right? So if you say coffee
shop versus coffee shop near me versus
coffee shop Brooklyn, you're going to
see completely different rankings. That
word order can dramatically affect the
ranking behavior, which is interesting.
I don't think it's actually a sign that
the search engine is doing what it's
supposed to, but that is what we've
observed.
Next, let's look at sector specific
ranking dynamics. So we wanted to better
understand from the original research we
did on local how much of the local SEO
ranking equation could we actually
decode. Could we actually build a linear
regression equation of all the
components and try to actually get to
predictive uh predictive ranking so we
could basically take a GMBB analyze all
of its signals and then pl plug them
into this equation and then project
their rankings. That was the goal and
these were the variables that we used.
So here we have um GBP profile fields on
the left. So these are the ones that are
within our control. On the right we have
variables that are not in our control
directly. So the distance to the to the
user obviously we know that this
matters. Uh but we also have other
variables as well about the profile
including its its total number of
reviews, its rating, the uh semantic
distance between the target keywords and
the reviews. We even looked at something
called web score. So when we use patent
brain, we learned about a variable
that's very important for local SEO,
which is web score. What is web score?
Well, something about the target keyword
that we're looking at in local and how
that's expressed on the root domain, the
actual website. So here we're actually
studying semantic distance of that
keyword to the website. We also looked
at authority measurements. Some people
talk about backlinks being important for
local. We wanted to see if that actually
shapes out. So here's kind of what we
learned at a high level. This is the GBP
ranking factor matrix. And what you can
see right away is that all these blobs,
they all kind of look similar in the
sense that that variable on top has a
lot of explanatory power. That's
proximity. But you also see that across
industries, it is not the same. There
are some industries that have very
different ranking mechanics than others.
This is an interesting insight because I
think a lot of people tend to think that
in local the ranking equation is the
same for all industries. What we what we
were able to see is that we were able to
disprove that. It's not the same
globally across all sectors. What we
find is proximity is number one across
all industries. The relevance of the
reviews for the keyword you're ranking
for is actually much more important than
your review score and your count of
reviews. So, if you're ranking for a
keyword, try to find a way to get that
into your review copy from your
customers. Very important. And then if
you look a couple down, you see right
there, keyword URL score, URL content
score. So, that is the semantic distance
of the website's content to the keyword
you're ranking for. And there it is. Web
score is one of the top five ranking
factors. So, then we looked at some of
those odd local businesses that don't
really fit the mold of a local, right?
like a law firm. Now, what's strange is
who in here would pick a law firm just
because they're across the street from
you? I hope no one raises their hands,
right? Google in their infinite wisdom
decided still that this was important
for law. And so, I don't know maybe why
that could be maybe to help some smaller
firms get advantage or get an
opportunity in local. I think that could
be why versus big law firms that would
otherwise rule the roost. But here we
see something that kind of violates what
we would have expected. Then we look at
car repair. How many of you guys have
been rearended or had a bad experience
at a car repair shop? Raise your hand.
I'm raising my hand twice because my
Tesla Model X was out of commission for
6 months. And when I saw how much the
insurance company had to pay for that,
it was like $25,000 for a little bump in
the back of my car. Right? A lot of
people have bad experiences with car
repair. And so is it any surprise here
we're seeing reviews are actually more
important than distance. Here it's not
about how close the car repair shop is
to the user. It's actually on their
merits and what people say about that
car repair shop. So very interesting.
Now you guys shouldn't be trying to
memorize any of this if you're
interested. This was published by search
engine journal. They picked up our
research on this. You can find the full
list of that grid and the ranking
factors by industry online. Just go look
looking for this research. Um yeah, so
really interesting research what we were
able to to determine from this. What we
also found is that the authority factors
were actually not anywhere on that list,
which is interesting. So it doesn't
matter if the back links are high
authority. It actually matters more the
anchor text. It's a little secret. And
that clue was given also in the patents.
So it doesn't really matter how
authoritative the links are. It's more
important that uh you're getting good
anchors.
Also a lot of people think that EMDs in
local it's a thing. I don't agree. I
don't see that in the data. So putting
the keyword in the name of the GMBB I'm
really glad that's not a factor because
otherwise every coffee shop would be
called coffee shop near me. And is that
any fun?
Right. Awesome. So there's more research
that we're going to do. We'll be
publishing more on this. Obviously, the
findings on this have helped us produce
more case studies, more case studies,
more case studies. 300 more case
studies.
I hope no one um has any seizures here.
If you do, you might want to close your
eyes, but these are over 300 case
studies. We have over 3,000 case
studies. Design team just couldn't
figure out how to get that GIF in with
all 3,000 case studies. Next stop,
10,000 is our goal. All right, going in
circles. Correlation study. This is a
bit of mythbusters. How many of you guys
have tried the geo circle strategy?
Raise your hand. Ian, put your hand up.
I know you have. It's good if you tried
it, right? Because we all have to learn.
Like, does this work? This strategy I
always thought was kind of weird and I
wanted to know if it would work. So,
this on this episode of Mythbusters,
Steve, by the way, this is a great photo
of Steve, isn't it? He cleans up good.
Um, so here is a moving company in
Southwest Florida. This is where they
started. And here you can see the
ranking progression as we built these
geo circles. And I feel so sorry for the
person on my team that had to build
these by hand. It took them like weeks.
They were doing this for like two or
three weeks.
TLDDR no impact. Thank goodness
because building geo circles is not a
lot of fun. But we really found that it
had no impact on ranking behavior at
all. whether it was a business that
already had good signals uh or new
businesses really there was no impact at
all from running these geo circles that
we could discern. Um and for anyone
who's been building these geocles on
Fiverr find a new job like do something
else with your life. Uh obviously people
will buy some of this stuff but it's
just not effective SEO. Now I want to I
want to talk about something else we've
studied. So obviously we all we all kind
of care about what's happening with AI
overviews. This is a recent 50 million
SER panel from the last 30 days from our
data sets and what we're seeing is a a
very slow but steady increase in the
percentage of AI overviews we're seeing
on Google. Our projection for this time
next year is 27 to 30% incidence rate of
these AI overviews. I think Google's
going slow, right? because they have to
figure out how do they protect their
core business which is actually not
organic it's ads right that's how
they're in business so they don't want
to cannibalize that when we look at
where we see these AI overviews we see
them informational queries more often
than anything else but we also see them
encroaching into commercial so one of
the things that we're going to be
monitoring very carefully is the
expansion of AI overviews
beyondformational
into these other realms of commercial,
transactional, navigational, because
obviously
those of us that are building
theformational content, we can't get
those clicks, which means we can't
retarget those customers. All it is is
potentially a brand play, right, to some
limited extent. But essentially, we've
been trying to monetize search as an
industry collectively through the
commercial and transactional queries,
but we're now going to become more and
more deprived of those clicks as well.
So, the other thing we're looking at is
how often are we seeing ads inside the
AI overviews? And we do see, as you can
see, a small uptick in that percentage,
but it's not that great. And there are
some reasons for this that we're going
to get into. But first, I want to ask
you guys a question. What percentage of
queries on Google are unmonetized?
Is it a raise your hand if you think
it's a 50% unmonetized queries on
Google?
Thank you. One man, one brave soul. Is
it B? 30%. Raise your hand
seems to be more of a popular one. Okay.
And how many of you guys think it's C?
Higher than 80%.
Okay. C is correct. Higher than 80%.
Which creates the of course perverse
incentive for Google to try to get us to
buy those queries.
Okay, this is really important because
what they're doing is they're doing this
blind bid system where they're saying
just let us do it for you, right? Give
us the keys, give us your credit card,
we'll decide what keywords you bid on,
right? And and so they want us to do
broad match targeting. They don't want
us to pick our keywords anymore because
they're really trying to figure out how
do they monetize that 80% of the search
landscape that they are not monetizing
currently.
Which brings us to the three laws of
Google economics. Number one, zero-click
search is never going to peak. Every
year, it'll always get bigger and
bigger. The second law is that Google
will always send more% of their traffic
to ads and monetize more uh of their
search. The open web hit their peak
about two years ago before SGE came out.
And then the third is that Google will
continue to consolidate its demand into
its own properties. So that's maps,
that's YouTube, Gmail, discover, Chrome,
Android, etc. and also potentially
through that Reddit deal where they're
basically getting all the Reddit data to
train Gemini and the LLM plus paying
them 60 million a year. So I think we
can expect that this is how the
landscape will continue to evolve. And I
call this digital Darwinism because in
this landscape we have to fight to
maintain or otherwise expand our search
visibility online in order to just stay
where we are because we're losing more
and more of those clicks. Which means I
would say as an industry we need to
think more holistically about search,
right? This is an SEO conference.
Most of us probably don't run paid
media, but we're going to have to. And
so that brings us to this, which is from
diversion intense, unified conversions,
semantic mirroring, and paid search
optimization. So the idea here is can we
construct semantic clusters at all parts
of the funnel from top of funnel to
bottom of funnel and build clusters the
way that we've been doing in search for
SEO. So let's take for example Starbucks
premium coffee beans right we take that
product or service and we create
clusters around it and our goal is to
have really neatly defined semantic
clusters in the middle and that will
allow us to build really optimized ad
copy on the right and if we do this
right we should be able to really create
high performance Google ad campaigns so
I fig you guys' be more interested in
the details so let's look at some of
those if the product we're we're selling
is this Starbucks house blend of premium
coffee beans. Some of our clusters are
going to be really focused around let's
say these three semantic identities.
One, coffee beans, which is more of the
head. So in that we see keywords that
all contain coffee beans, those tokens.
But then there's also this coffee
premium context where now we see coffee
and premium in every single keyword,
right? And interestingly, we see Arabica
coffee as well as entering as a type of
premium coffee that is obviously not
lexically the same, but semantically
there's a proximity there, which is why
it enters this cluster. Coffee roasting,
coffee blends, you guys get the picture.
Each of these clusters has a certain
search volume, average CPC, and semantic
distance to the core concept. And from
that, we're able to then create these
clusters. So, I wanted to show you guys
how we thought about it. So you can take
and process your query network to build
these clusters. Each cluster should have
a very distinct semantic identity and a
clear intent. And what's cool about this
is that it allows us to optimize what
keywords are we buying, which is the
opposite of Google's blind bid system
and optimize how much we pay based on
the return on ad spend. All right,
you'll see this not as theory but in
reality, the actual metrics that we got
from this in just a moment.
The other powerful thing about creating
these very distinct semantic identities
is that when we create ad copy, we're
able to mirror the intent and the
semantic identity in our ad copy itself.
So that creates a higher click-through
rate. And this is what we call semantic
mirroring. Um,
and one thing that's important here
obviously is when you're building ad
campaigns. Curiously, how many of you
guys also manage the paid media
campaigns at your firm? Nice guys.
Awesome. Lot of people. Good. So, if as
you guys know, you want to max out the
SER real estate and so this is a lot of
work if we're building a lot of
campaigns, especially as many as we're
creating here. So, we have 14 ad groups
per campaign on average, eight and a
half keywords per ad group, 12 ad copies
per ad group. median keywords, search
volume 300, and the median cluster has
about 37 uh 35,000 clicks per cluster,
which is a a good kind of goldilock
zone. The other thing that we're also
looking at is the traffic temperature.
So on the left we have cold, on the
right we have hot. When we're cold,
there's different levels of cold. So
there's people who are solution aware,
some people are problem aware, and some
are just symptom aware. But the goal is
can we also create campaigns that target
people at every stage of this funnel.
Right? And so for each component of this
there is a strategy within ads that we
can reflect. And one of the things that
we notice is when we target especially
in conquesting there are lots of
opportunities because a lot of
competitors are not bidding on that
semantic space around their brand. You
can get very lowcost clicks if you
target that semantic space that your
competitors should be bidding on for
their own brand but are not. Couple
dollars per click. They're already
solution aware. They're shopping around.
They're about to buy your competitor and
then just at the end they see you. Huge
opportunity. Some of our best performing
campaigns. So what do we learn? Median
conversion rates increased 50%.
We didn't even touch the landing page.
All right, this is not CRO. This is
about buying people who are more high
intent. People who are searching longer
tail, right, that are inside these
semantic clusters have way more intent
than the ones that are just searching
for coffee beans, right? So, we see a
massive increase in our conversion
rates, which is great because every year
we have a very high inflation of CPCs
and CPMs. The other thing that we see is
that the CTRs are about 30% higher. And
one of the things you're able to
completely eliminate are campaigns that
have no CTR. So really poor performing
campaigns. You can see that salmon chart
is essentially the control group which
are campaigns in the account and the
teal are the ones that are built in the
strategy. So pretty cool. Next I I'm
going to try to wrap up soon. Uh I want
to ask this question which is what is
the end state of the open web? And as we
all know, Google still is the king. But
look at chat GBT. They have a 3% share
of search. And what's fascinating is the
growth rate here is 65% year-over-year
while Google's at a humble four. So it
doesn't take a rocket scientist to draw
that line out. And what I believe is
we're going to see open web is going to
become information artifacts. Our sites
are going to become information
artifacts which are fodder for LLMs. We
will start to see the open web decline
as a destination. we're going to be
seeing less and less clicks. And the
idea of building our websites is really
to create these components enter the
composable web where our websites as
information artifacts are going to be
providing these components for the LLMs
to drive the AI mode answer canvas not
just on Google but everywhere.
And there's going to be new laws of
search. So every unmonetized query will
be monetized. The open web is becoming
an extraction layer, not a destination.
And we're going to start to see that
Asians will outnumber humans in query
volume. I think we can say this is
already beginning to happen. So it's
important for us to move out of an SEO
and GEO mind space to more of a holistic
search mindset. And as we enter 2026,
we're almost at the end of the year. I
believe next year we're going to see aic
clicks, conversions, and narrative
inclusion become more of the emerging
KPIs. We can also look forward and
project agentic GDP. And the numbers
honestly shocked me to see that uh the
agentic GDP is going to significantly
outpace um what we're at today at a
humble 5%. It's going to grow very
quickly. And this is going to create a
new opportunity for us as search
marketers. Something that I'm calling
agentic visibility optimization where we
actually are optimizing for inclusion in
these agents. And I wanted to give you
guys a framework that you can use to
think about this opportunity. There's a
level one data layer where we provide
our facts and build these artifacts.
There's a narrative layer, there's a
cognitive layer, there's a reputation
layer, and then there's a measurement
layer where we do recall and visibility
analysis.
And the core pillars as I see them are
number one for narrative inclusion, we
have a process called narrative
induction, which is how we put our
information into the LM through guided
inquiry and directional prompting.
There's also knowledge seeding where we
can bring our brand's factual data into
the the models latent space and drive
rag into the base model. There's also a
process of reinforcement learning that
we can encourage through our
conversations that we have with with
chat. And then there's also an
opportunity for us to build a reputation
architecture.
So I think it's going to be a really
important concept. And I I actually for
one don't want to see the open web
decline. AI mode is a black hole. It
sucks the world's information in and it
uses it to create Google's answer canvas
depriving the open web of its traffic
and clicks. But there is an opportunity
for us to build a white hole which is a
CMS synthesizer. Takes our brand our
source context our story and it
constructs that information artifact. So
instead of serving one page to a
thousand people a thousand different
times, we can do what AI mode is doing.
And I think this is a necessary next
step for all CMS's to step into. If
anyone is interested in building this,
I'd be happy to cut the first check
because I think it's such an important
technology. And um I'm probably way over
time, but I'm just going to try to get
through this super fast. I also wanted
to share with you guys our research on
how do we get included in the LLMs. And
so this is a 5.5 million um
query data panel. And here you can see
that it's not the sites that have the
highest traffic or domain power, domain
rating or domain authority that are
visible in LLMs, but we do see that
there is a slight preference for domain
authority. I believe the reason for that
is common crawl uh common crawl because
they're using sorry I don't know what
happened there. Um they're using domain
authority to figure out what sites to
crawl. You'll also see a lot of times
sites in common crawl that have no
traffic. The other thing we were looking
at is pair wise agreement of citations
across LMS. Here you see Gemini and
OpenAI actually have a high degree of
correlation in their answers. Why? I
think OpenAI train their data on
Google's data. Right? I think that's the
most plausible explanation. We can also
see when it comes to bias and favoritism
in LLM citations, Gemini strongly
prefers its own properties, but so does
OpenAI. it actually since it was trained
on the same data. We see a bias as well
in OpenAI for Google's assets versus
perplexity.
Some other interesting insights,
but the content age one is probably the
most insightful where we see that
content that is highly fresh has the
highest incidence rate and visibility.
And that is especially true in OpenAI
more so than Gemini or Perplexity.
Perplexity seems to be less bothered by
content freshness as compared to OpenAI
and Gemini is somewhere in the middle.
case studies to show that the framework
is effective. Um, how many of you guys
have been to Snackus?
It's a delicious place. You guys got to
jump on the nacho navigation deck. It's
so much fun. So, this was was a creative
concept that we had. Can we create
something that doesn't exist inside
Google's knowledge graph? And so what we
learned from this is that we created all
these pages online about snackus to see
if we could narrative induct Google into
believing in this. What was so
interesting and we also did it with chat
GPT. You guys can go to Google right
now, type it in. You can go to chat,
type it in and see what comes up. But
what's interesting is that we learn that
by seeding these queries, we're teaching
something to Gemini and all LLMs. But
there is a second layer factual uh
review where they're reviewing the
information.
Um here's a study on comparing different
authority metrics. We see that the
trafficbased metrics are way more highly
correlated on page one than the old
school domain rating domain authority
legacy metrics which are not as
correlated. So that's an important
insight for anyone doing link building.
We also study topical relevance to
understand how um impactful topical
relevance is as a ranking signal. Uh
trying to understand how Google draws
its topical boundaries. What we found is
that the more specific the topicality,
the more likely it is to be where Google
is drawing these topical boundaries. Uh
so for example, not SEO but be more
specific like backlink research software
or link building software or content AI
software. That's more likely to be a
topical boundary. And we can see when
you compare this versus other factors,
the topical score is almost twice as
powerful as the as the search atlas
domain power trafficbased metric, which
is twice as powerful as the href's
domain rating classical page rankbased
metric. This is a really important
insight because it leads us to really
understand that on Google to win we have
to master topical relevance. There is a
equation behind that and I'm happy to
share more about it. We also have done
some research into site distance, site
radius. Uh this code is on my GitHub. We
also were able to prove that this is an
effective ranking strategy. Uh pruning
content from websites using this
technology. We can also use it to
compare site A to site B to understand
which one has in more in-depth
topicality.
Uh and I want to conclude with this
analysis which I think is my magnumobus
where we were taking 43,000 sites and
testing a micro optimization framework
to measure the impact on search
visibility. So here you can see these
these sites exactly what we did on page
what the actions were to optimize them
technical issues metadata heading
structure semantic enhancements and what
we were able to see is that these sites
saw significant improvement in their
primary SEO KPIs here is what we saw in
the data panel in their search
impressions. So after 30 days after
deploying these micro optimizations
massive improvement of 107%
median traffic improvement of 81% and
you can see that there are outcomes that
are radically higher. Uh so there's some
that actually see better results. I
think part of that has to do with the
opportunity how much optimization was
done on the site. Um and and some
especially smaller sites can see bigger
gains. We also looked at what strategies
were the most powerful in this to move
the needle the most. And we studied that
by each SEO KPI. And what we were able
really able to find is that if you do
all of these, you see the best results,
but some drive outcomes more than
others. And this behavior generalizes
across all sizes of sites, medium,
small, and big.
But does it work? Okay. 152 case studies
here u for an orthodontist that we were
working orthodontal agency we're working
with that has over 150 orthodontists in
the country what you can see is when you
implement the framework you can see how
quickly we see a response inside search
console so a lot more case studies here
to show that uh this effect is very uh
it's very easy to see and measure um and
and observe
uh one of the ways we're able to do this
is by going universal and being able to
publish into the CMS itself. So now we
have support for the top 20 CMS
platforms including all those you see
here where now we can publish blogs,
landing pages, update metadata and
manage those sites remotely seamlessly.
And this is fully LLM and crawler
friendly. I also built an MCP that you
guys can enjoy and use for free. Uh the
goal being to bring SEO data into your
workflows and systems. It's a community
MCP. We're never going to charge for it.
Um, and if you want it, you can plug it
into your coding editor of choice,
whatever that might be at that URL
there. Uh, it'll give you SEO metrics,
keyword data, backlink data. Um, if you
do have a Search Atlas subscription
though, it gets kind of cool. There's
over 138 tools that you can access. Uh,
your usage quotas are essentially
unlimited. Um, and there's a lot of
things that you're going to be able to
do to basically replicate some of our
experiments with your website in an
agentic fashion. This guys is what I
what uh uh Mr. Deutsch calls the
beginning of infinity uh where websites
are no longer going to be your
destinations. Search is really not going
to be about classical traditional SEO
and we're seeing a change in the click
economy. The ones the brands that I
think are going to be the most
successful in this brave new world are
not the ones that are fighting for the
traffic and the clicks, but they're the
ones looking to engineer their
visibility into every answer engine,
every surface, every agent. one word,
one pixel, one bite at a time. If you
guys are interested in this type of
work, if this was on your bingo card, uh
we are looking for product managers and
SEOs who like building products and
doing research like this. If that's your
jam, come find me or meet my team
outside at our booth. And thank you guys
so very much for your time. [applause]
All right.
>> Um, first question. Um, is there a
chance we can get this deck after the
fact or
>> Steve? Is that on the table or
>> just ask me?
>> Okay, awesome. Um, second question. the
local data you were illustrating
earlier. Have you guys done any
experimentation to see if those ranking
factors actually differentiate from
geograph like geographic to other
geographics like from province state
like country level sort of thing?
>> Yes, that was a big question we had to
to look at as well. And so initially we
broke up the uh GBPs by geography and
looked and see if those features were
normalized across all geos and they seem
to be although the data panel we're
going to do more research into that and
I don't I didn't put the charts in the
presentation but we will be sharing more
on that for sure. Yeah. So it is kind of
a universal I think globally but each
industry has a different set of factors
to it.
Good question. [applause]
My question is, have you and not to put
[snorts] me in on the spot, but have you
experimented with CTR manipulation for
local and does it work?
It does work. Yeah, it does if you do it
right. I can't say too much more than
that, but it is an effective strategy
for sure. Historical data is one of the
things we see in the patents that they
use to really create the topical
relevance equation. So the way that that
equation works is essentially it's an
engram analysis of your topical share of
your topical search volume. So if you
look at your keyword data for your
website and you take all your keywords
and you kind of map them into their
topical boundaries and look at your
overall share of search within your
topical boundary that defines on Google
who is the topical authority number one
2 3 4 5. If you're trying to figure out
what keywords live within that topical
boundary, all you really have to do is
take those keywords, search for them,
and look at which keywords share the
same domains on page one. That
effectively is going to help you
understand those topical borders. There
is a product in our platform called
topical dominance and search atlas that
you can use to run that report if you're
interested in looking at what that data
says for your website and your topical
neighborhood. All right, I think that's
our last question and
[applause]
AI Overviews rewriting ranking rules. Topical dominance importance. Local SEO volatility. LLM inclusion frameworks
Nikhil Kamath
