August 3, 2025 • 30 mins
In this episode of The Digital Marketing Podcast, hosts Daniel Rowles and Ciaran Rogers explore one of the most pressing questions facing marketers today: what does SEO look like in a world dominated by AI search and generative answers? With Google rolling out AI Overviews and AI Mode, and tools like ChatGPT and Gemini changing how people search, the rules of SEO are shifting fast. Some publishers report traffic drops of up to 60%, while others in e-commerce are holding steady. So is SEO really dead, or just evolving into something new? Daniel and Ciaran dig deep into what’s happening, separating hype from reality and providing clear, practical advice for marketers trying to adapt. They explain why understanding Generative Engine Optimization (GEO) and AI Overviewsn (AIO) is critical, and how you can stay visible when clicks are disappearing. What you’ll learn in this episode: Why AI Overviews are changing user behaviour, and pushing organic results further down the page How share of voice and brand sentiment are becoming as important as traditional ranking factors Why advocacy and positive online mentions may now trump old-school keyword tactics The growing importance of Wikipedia, Reddit, and forums in shaping AI-driven answers Why updating and referencing your best content is more powerful than ever How no-click searches are rewriting what “success” looks like in SEO Practical ways to benchmark your visibility Why prompt-like queries and longer conversational searches are the new normal How to build trust through Google’s E‑E‑A‑T principles (Expertise, Experience, Authoritativeness, Trustworthiness) The Future of SEO Episode Mindmap The hosts also explore pitfalls to avoid, from snake-oil SEO sales pitches to rushing into poorly thought-out Wikipedia pages that could backfire. They emphasise a smarter, evidence-based approach: focusing on authenticity, advocacy, and quality content that’s updated, referenced, and truly useful. Key Takeaways: SEO isn’t dead, but competing for the top three AI-sourced links is tougher than ever. Building advocacy and brand mentions is critical to visibility in AI search. Quality and authenticity (with evidence and references) beat quantity every time. Think conversational: searches are getting longer, more natural, and more complex. Following Google’s E‑E‑A‑T framework is the best hedge against AI-driven disruption. If you’re a marketer, publisher, or business owner trying to make sense of the shifting SEO landscape, this episode provides both clarity and actionable steps to future-proof your strategy. 📥 Access the show notes, tools, and links at: 📩 Join the free newsletter for access to monthly live sessions:
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Ciaran (00:00):
Hello and welcome back to the Digital Marketing Podcast,
(00:03):
brought to you by Target Internet.
My name is Kieran Rogers,
Daniel Rowles (00:07):
and I'm Daniel Rolls.
Ciaran (00:08):
and today we are discussing the future of SEO.
Daniel Rowles (00:21):
Okay, so we have had multiple conversations recently in
podcasts and in various articles aboutwhat's happening with search engine
optimization and the big conversationhas obviously been about GEO, which
is generative engine optimization.
And then there's been lots of debates.
Is it really a thing?
Is SEO just expanding?
(00:42):
We've got AI overviews.
And then we talked a little bitabout answer engine optimization,
which was, and anyway this all seemsto be settling down a little bit
Ciaran (00:51):
Is it
though?
Daniel Rowles (00:52):
Well, no.
If I'm honest,
Ciaran (00:54):
I would say the SEO is dead.
Mongers have gone intooverdrive with it all.
Daniel Rowles (00:59):
that's normally me.
Ciaran (01:00):
I know, right?
But no, it's everywhere now.
you, maybe you were like the Heraldthat saw it coming first, but there's
a lot of panic and there's a, andthere's a lot of lawsuits flying around.
People are getting that antsy about the
Daniel Rowles (01:13):
So let's talk about that.
'cause you mentioned this tome as well, so that people
are suing Google on the basis.
They're not as getting as manyclicks as they used to get.
Ciaran (01:21):
yeah.
Yeah.
Daniel Rowles (01:22):
That seems slightly wacky.
Ciaran (01:24):
of European businesses.
I'm not gonna go into it 'cause Idon't wanna get embroided in the legal
Daniel Rowles (01:29):
Yeah.
Dragged into a legal case.
Ciaran (01:30):
not really.
But if you search on it,I'm sure you'll find it.
And I think it's really interestingthe basic premise of this is across
the board, a lot of businesses,particularly content marketing
focused businesses and publishers.
In particular, and I think that particularconsortium is made up of, a lot of
Daniel Rowles (01:47):
Right.
Ciaran (01:48):
have seen their organic year on year traffic absolutely decimated.
And there's all sorts of numbers flyingaround between like 40 and 60% drops
in some cases depending on kind of thekeyword landscape that you're looking at.
And what I found particularly interesting.
Is that there's also a lot of reportsthat actually businesses just aren't
seeing those levels of drops withine-commerce, So this is much more top
(02:12):
of funnel kind of content research.
And that makes perfect sense becausewe now have AI answers, you know,
that's been launched across the world.
But I think my view on this isthat actually what's actually
happening is because of that,people's behavior is shifting.
Radically, like before AI answers,you had to work a bit to get answers.
(02:33):
You typed in your search query andyou got your 10 blue links, and then
you did, you clicked around and youvisited quite a few places and you
actually, it was a journey, wasn't it?
You sort of learn a fewthings along the way.
Sometimes when you wantedto learn something.
Actually, it helps to have to workfor it rather than it just being
handed to you on a plate like
Daniel Rowles (02:50):
Well, let's talk about that then for a second,
because we've got AI overviews
And it started off peoplesaying, these are useless.
And then they got better because the largelanguage models got better, and Google
particularly got much better at this.
And then we started to seethat in some circumstances was.
Answering people's questions.
The thing is, like you've said to mebefore, they are, they're sourcing
(03:11):
maybe three different links in there,three websites where this information
has kind of come from to some extent.
Ciaran (03:16):
Yeah.
Daniel Rowles (03:16):
And what's interesting to give you a quick example, is that
if you search, what is data scraping,we always rank number one for that.
Now, when you search,what is data scraping?
There's an AI overview.
We're still number one below that,but we're actually listed in the
websites where the informationhas come from in the first place,
Ciaran (03:32):
Yeah.
Daniel Rowles (03:33):
so, so what that tells me is quite interesting is that actually
by answering questions and exactlyto your point, is there's a bit of a
killer strategy still at the moment oflike actually working out what are the
questions that people are searchingfor, answering those in a really
effective way that's still gettingpicked up by the search engines, but
it's getting picked up to some extentby the large language models as well.
Ciaran (03:53):
here's the problem.
You're gonna see a traffic drop, Soacross the board, in the accounts that
I look at, I've seen lots of this.
You see an in quite a big increasein impressions year on year.
But you're seeing a big drop in clicks andmore recently a drop within the rankings.
And that flows for me because I knowthat when you get a drop in click-through
(04:13):
rate, it affects your rankings,like your average ranking overall.
So that kinda makes sense.
Why aren't people clicking?
Because actually for the majorityof them, they've had their itch
scratched by the AI answers.
Daniel Rowles (04:23):
Right.
Ciaran (04:24):
You know that is enough, so not, so let's not forget if you're
on a mobile device, that lovely AIanswers, just push those 10 blue
links even lower down the page.
we all know how lazy, lazy fat thumbs weall have when we're on the mobile device.
Like we just, it's effort to flick up
Daniel Rowles (04:42):
right.
Ciaran (04:42):
you know, if I've had the information itch
scratched, why would I need to?
Scroll down.
This is the thing.
It's pulling together andactually it's quite smart.
Before we had to get quite creative.
Sometimes you'd need to rearrangethe words or add in additional you
know, you know, words to filter
Daniel Rowles (04:59):
Huh,
Ciaran (05:00):
Now
Daniel Rowles (05:00):
a great stat on this
Ciaran (05:01):
Yeah.
Go on.
Go on.
Daniel Rowles (05:02):
so that the AI overviews are only including the exact phrase
you search for in 5% of cases.
So the point being is that theyunderstand, well, this phrase
really means this phrase as well,and that's the same as this.
And therefore if I give you the answerto this, it's is what you're asking for.
You've just asked for it ina slightly different way.
Ciaran (05:21):
Yeah.
Daniel Rowles (05:22):
So that kinda shows that exact, this phrase, I'm gonna
optimize for this phrase, and I'm gonnarank for this phrase is kind of gone.
Ciaran (05:27):
I'm gonna throw in another thought for you all as well, and I
haven't got, unfortunately, I haven'tgot anything that backs this up, but
it makes logical sense to me if you aregetting, as a user, if you're getting
used to having conversations with Google.
Much like if you use chat GPT or Gemini,you have conversations with those chat,
chat entities as well and Google'sresearch actually does back this up.
(05:49):
The number of words people aresearching on in any given search
has increased quite dramatically
Daniel Rowles (05:54):
Oh no, The numbers are there to back up.
Exactly.
There's all that stuffthat came out from Google.
Put it into the show notes.
Ciaran (05:58):
so here's the thing.
If I'm a publisher and for the last 15, 20years I've been dining out on optimizing
for certain key phrases and questions,when people stop asking those, 'cause
actually they're phrasing 'em slightlydifferently, I'm not getting picked up.
So, you know, is that, is, is the thing Ithink's kind of unfair to, to, to Google.
(06:18):
And I mean, they're big en, they'rebig enough to fight their own battles.
They don't really need me to doit, but I'm gonna do it anyway.
You know, it's not Google's fault that.
had to innovate, they've had tobring in this new technology.
Let's face it, if they don't, they'regonna get wiped out by other competitors
within the marketplace who will.
Right.
And I think actually, from what I've seen,they've done a pretty good job of this.
Like it's good, it's usable.
(06:38):
People are using it, thedata's there to support that.
if I'm on Google, is it really my faultthat the whole marketplace is now using
the web and search in a different way?
Is it really my fault that, you know,certain businesses have over optimized for
certain phrases that have been dining out?
No, not really.
That's just life.
Every marketplace has that.
(06:59):
People come in, new innovations happen.
People lose their share.
Other operators like still stealtheir thunder that's just how
business has always worked.
So I kind of alarming that, like peopleare going down the lawsuit route.
Maybe there's an angle onthis that I haven't kind of
clocked, and if so, us a shout.
Daniel Rowles (07:18):
Well, I do know a little bit more about it.
So it's the fact that the publishers,what's been happening is that
publishers rely on the fact thatGoogle was swallowing up some of
their stuff, sourcing it in thesearch results, but then people were
clicking through to the website.
So they kind of felt, well, we'veinvested in creating this content,
but actually at least we aredriving some traffic value from it.
But what's actually happening now is that.
They're getting all that stuff inGoogle itself, and they're not just
(07:40):
seeing the value from it and they thinkit's a break in that relationship,
Ciaran (07:43):
Okay.
Daniel Rowles (07:43):
but, which I get, but it, that's not their fault specifically.
It's 'cause the operatingenvironment has changed.
Ciaran (07:47):
Yeah.
Daniel Rowles (07:48):
let's take a step back for a second.
So we've got AI overviewsshowing up in the search results.
Search engine optimization seemsto be expanding to include.
This concept of generativeengine optimization.
I wanna show up in chat, CPT.
But the thing is we don't know ahuge amount that's about how these
large language models are working
Still.
(08:08):
But we do know a couple of things.
LLMs are basically built onneural networks essentially.
And a neural network is a, somethingthat's been around artificial
intelligence for a long time.
And it's the connections between things.
Okay.
As in if a particular phrase is mentionedin lots of different places and there's
positive sentiment about it, and there'sall these kind of interconnections.
This is probably relevant.
(08:29):
Okay, so what that leads to isshare a voice and brand sentiment
being seen as really important.
But actually that's not as advanced aswhat Google will be doing a long time.
'cause Google have that next layerof user experience to say, well we
think this is the most relevant thing.
But if a user clicks through toit, do they have a good experience?
(08:51):
Do they stay on that webpage?
They to scroll to the bottom.
So it's potential that thelarge language modules are
kind of putting stuff up front.
In their search results that mightbe mentioned a lot might seem to be
good, but isn't necessarily a gooduser experience and so on as well.
Now lots of people would argue, actuallyGoogle's been showing stuff at the
top of the search results isn't agood user experience for a long time.
And you know, Rand Fishkin has beensaying that, you know, for a long time.
(09:13):
But what it does mean is when I use atool like the HubSpot AI search grader,
which has actually got a lot betterand it's trying to kind of estimate
how visible you are in Chatt PT andGoogle Gemini and places like that.
So two things they say is,where have you been mentioned?
What's the brand sentiment?
The brand sentiment analysis ispretty clever 'cause it'll say
these are the things that peoplethought were good about you.
(09:35):
These are the things that peoplethought were potentially negative
about you, but it's sourcing stuffthat maybe isn't very reliable.
So what they're looking at a lot isstuff that's mentioned in forums.
So Reddit.
Cure those kind of places?
Well, we know thatReddit is the Wild West.
There's some good user stuff in there.
There's also some prettywacky stuff in there as well.
(09:57):
But it also, if there's a lack of,like if you've got online reviews.
But you've got, you know, mentions you'vearound other places they can be picked up.
So places like App Sumo, whichI'll give an example of places
like Wikipedia and so on.
Ciaran (10:10):
Yeah.
Daniel Rowles (10:10):
So let me give you the app Sumo example.
I bring that one up.
We did an app Sumo promotion, andif you're not familiar with it,
you go into App Sumo, you say,right, we can do a special offer.
We want a load of users just so wecan test out our product platform.
And they take.
More often than not 70% of any moneyyou make and you're supposed to give a
lifetime subscription for like a 10thof what people would normally pay.
(10:33):
Right.
It's a pretty, it's a prettybonkers model when you look at it
from a commercial point of view.
'cause you have to sella lot to make it work.
But what it does give you isa user base and a test base.
Right.
So we did an app Sumo promotion.
I can't say that I'm entirelypleased with the results, but let's
skirt around that for a moment.
Ciaran (10:50):
you being very excited before it went live, and then you got frowny.
That's all I'm
Daniel Rowles (10:55):
Yeah.
Ciaran (10:56):
When
Daniel Rowles (10:57):
Well, I'll tell you what was interesting was that my
commercial director had said tome, this is a really dumb idea.
We've got a reasonablywell established product.
We are not the rightlife stage of business.
And I think that's completely right.
It's like if you're a brand newstartup and you needed a test user
base, this makes perfect sense ifyou're as established as we were.
And because she'd said tome, this is a bad idea.
And then I was like, no,it's gonna be brilliant.
We're gonna get thousands of users.
And then we did it and it was a, itwas like we, we've basically given
(11:20):
away about 50,000 pounds worth of ourproduct for about a thousand dollars.
Ciaran (11:24):
do you
T-shirt she gave you saying?
I told you so?
Daniel Rowles (11:27):
No I've got so many of them now that you know, but
anyway, you should probably, peoplelisten to your commercial directors.
Ciaran (11:33):
Yes.
Daniel Rowles (11:34):
so anyway, you get a load of reviews from these people,
and these people weren't inherentlyour target user base, right?
So we had some kind of strange, wehad loads of really nice reviews.
I mean, it was still like 95% positivereviews, but because the large
language model is looking for thepositive and the negative to analyze
the brand sentiment, when you go in.
And I went into the AI search grader,the HubSpot tool, put our brand in
(11:56):
and it says, most people love you.
You are brilliant, it's great.
Here's positive stuff.
But some people think an outcome, theseapp Sumo reviews from how many five
years ago that it was kind of thing.
Because we aren't on many platforms likethat by nature of what we kind of are.
That will now haunt us forevermore.
So the only way you can really dealwith it is say I need to be mentioned
in more places by more people.
Ciaran (12:17):
Yeah.
Daniel Rowles (12:18):
advocacy, and I've been telling this to a
few training courses recently.
We were talking about advocacybefore, like, you know, it's not about
influencers, it's about advocates.
It's those everyday people thatsay nice things about your product.
You want a thousand truefans, all that kinda stuff.
It's never been more true because yes,Google's looking at Reddit more now and
places like that, but actually theselarge language models are as well.
So if there were two quick wins in this.
Ciaran (12:39):
yeah.
Daniel Rowles (12:40):
I would say the first one is advocacy.
Get nice people saying stuffabout your more places.
And the other one is answer more questionsin depth and think about that the way
that people are changing their searching,
Ciaran (12:50):
and like I've seen this on Wikipedia and be
Daniel Rowles (12:53):
right.
Ciaran (12:54):
there's a big drive at the moment.
People have clocked that alot of the AI systems really
rely on Wikipedia for facts.
And it is quite good for that, right?
Because they have quite a robustsystem of editors, like human editors
has the double check stuff anddouble check edits when they're made.
I think for like a big organizationlike Take Brew Dog for example,
(13:15):
have a look at Brew Dog's Wikipediapage, there's a ton of stuff.
On their page is quite balanced.
There's some positives.
Plenty of negatives as well, you know,overall they've got quite an established
like brand page on Wikipedia, right?
So something, you know, negative butrelatively small happens, it's quite easy
for that just to get ignored or missed.
'cause it's not relevant to thebig entity that they are, like
(13:38):
the editors will overlook it.
However, if you are a new company.
Without much history and somebodygoes, oh, let's get a Wikipedia page.
'cause then we'll exist as an entityand the LLMs will trust us more.
Yeah.
True or true, have to go throughthis process of submitting an entry.
You have to kind of.
(13:58):
Make sure it's balanced.
Like it can't be overly on promotion.
You have to submit it to an editor.
editor's gonna look through your immediatekind of recent history, just do a few
checks to check what you're saying.
And if they find anything negativethat you haven't included there,
they're gonna slap that in there.
Yeah.
Or worse ban you for like making a,an unbalanced, unbiased submission.
Daniel Rowles (14:21):
Right.
Ciaran (14:21):
And then once it's out there, remember this page is public property,
so anybody can make a. An edit.
And I think the problemis when you don't have.
I suppose it's a bit like, let'ssay Kieran decides that he's
important enough to deserve a pagein like the Encyclopedia Britannica.
Like it is it, that's a nonsense.
Like I'm not, I'm just not at that level.
(14:42):
I doubt I ever will be.
Right?
But let's say I decided I was gonnapush it not be very much verifiable fact
that's interesting enough that you couldput into that Britannica page, right?
And so what is there isgonna be a bit skewed.
if there's any one thing that'sno, like verifiably notable
and it's a bit negative, that'sdefinitely going in there.
(15:03):
Right?
And now you've got something that'sfactually true, but kind of biased
because it's only focusing onthe things that are big enough or
epic enough or verifiable enough.
And that's the risk with this.
So
Daniel Rowles (15:15):
Be careful what you wish for.
Ciaran (15:16):
Yeah.
Daniel Rowles (15:16):
Yeah.
Ciaran (15:17):
careful what There's there's a concept really here that actually really.
Wise people actually spenda lot of time thinking about
the outcomes of their actions.
Fools just rush in and don't thinkabout the outcomes that their
actions and bad stuff happens.
Right,
Daniel Rowles (15:34):
Well talk, talking of which
Ciaran (15:35):
Yeah,
Daniel Rowles (15:36):
I I decided that I was worthy of a Wikipedia page, so.
Yeah.
Yeah.
So because, well, becausepublished authors quite often
you are a published author.
I've got quite a few, like I'm sitting,I'm, I've the humble break behind me
that no one can see is I've got thebig bookshelf with all my books on.
It
Ciaran (15:52):
are behind him and they're extensive.
I'm
Daniel Rowles (15:54):
is
Ciaran (15:55):
on half of one of those books.
Daniel Rowles (15:57):
half of one.
That's all right.
Ciaran (15:58):
Well, I wrote half of it.
Daniel Rowles (16:00):
But
Ciaran (16:00):
count?
I
Daniel Rowles (16:01):
you are on two of them.
'cause there's one in another languageI think as well, but there's like
14 languages, books, blah, blah.
Anyway, so I thought, well thatmust be a good opportunity.
Ciaran (16:07):
in
Daniel Rowles (16:07):
Anyway.
I've been,
Ciaran (16:08):
Isn't it so
Daniel Rowles (16:09):
I think it is.
Ciaran (16:10):
I think it's in Russian too.
Yeah.
Daniel Rowles (16:12):
So there's a few language.
There you go.
You are worthy of Wikipedia page.
Look, you in multiple languages.
But the point being is there was that, andthen there was like I published Imperial
College academic papers and a few otherthings, and I thought, come on this would
be really good from an SEO point of view.
Ciaran (16:23):
yeah.
Daniel Rowles (16:23):
I went on to Fiverr and I found this guy that like
specializes in in doing Wikipedia pages.
And he got.
Well, he was actually, he was really good.
I'm not gonna diss the guy, but he I spentlike an hour and a half writing an email.
Here's all the links to all my staff.
Here's all my books, here's theconference of the judging that I do, the
things that I'm members of, and blah,blah, blah, blah, blah, blah, blah.
(16:44):
And after an hour, if I submittedit and within 45 seconds, he'd
responded to me going, you reallydon't deserve Wikipedia, Paige.
I said, that's brutal.
Alright, well thanks for that.
Ciaran (16:55):
been there though, right?
He
Daniel Rowles (16:56):
Well, that was the point.
Ciaran (16:57):
is.
Daniel Rowles (16:57):
That was the point.
Ciaran (16:58):
the thing.
This is the thing.
And I think that's the danger with this.
And let's say you get your company thereand it only mentions one thing about you.
And that thing's a positive thing.
Everyone's happy, right?
Until something not so goodhappens and it gets reported on
in the press, and then you're justwaiting for that, that to drop,
Daniel Rowles (17:13):
Right.
Ciaran (17:13):
It could happen at any point.
They might.
Here's the thing, when you firstsubmit, everything's double checked.
Once you've got a page.
what I've seen, I've found quite a fewexamples of companies that had all sorts
of, you know, advertising standards,agency rulings against some stuff never
mentioned on the Wikipedia page, butit's 'cause the Wikipedia, your page
was created maybe six, seven yearsbefore that happened, So, but at any
(17:36):
point that could catch up, and when itdoes, very difficult to get rid of it,
and that's permanently on your recordand that is totally gonna get skewed.
So you could be in a situation where,yeah, the large language models are
using it and they're now only citingone thing that's a bit negative.
They're trying to give a balance viewand actually it doesn't, so be careful
Daniel Rowles (17:53):
Okay, so we've got Share of voice is important.
Brand sentiment is important.
Be careful your brandis showing up something.
So we've got increasingly got thisno click environment where you're
not as getting any clicks, andthat's because people are searching,
getting the answers they want.
You might get more directvisits, you might get more brand
searches, so keep an eye on that.
Ciaran (18:12):
important.
See, people are still thinking about you,
Daniel Rowles (18:15):
Right.
Ciaran (18:16):
it in a way that you can measure.
They're not clicking andit's the clicks that we can
Daniel Rowles (18:22):
Well, you mentioned to me off platform research.
I thought that was an interestingway of thinking about it.
Ciaran (18:25):
do it all the time.
Like before I Google stuff now I'll goto my large language learning model and
I'll do my research there and I'll refinemy search down for two or three things.
Sometimes just one thing.
I was telling Daniel, I've justbought my first Windows pc.
I'm so excited about it.
Secondhand, it was old.
I needed to use some Windows software fora project that I'm doing, so I instantly,
(18:49):
I decided, right, I'm gonna try out theagents feature in chat, GPT and I set
So I've just bought my first Windows pc.
Very excited about it.
It's secondhand.
I needed it for a personal projectworking on, I needed some Windows software
that was only available on Windows.
So I set my agent to find me the cheapest,best spec laptop on eBay for under 150.
(19:14):
It came up with absolute sto.
It did all this work for me.
Now, previously I would've beenclicking through, you know,
Daniel Rowles (19:21):
From Google search after search.
Yeah.
Ciaran (19:23):
gonna get a trail of it.
eBay's gonna get a massive trail ofwhat I've been looking at and thinking
it's gonna start showing me ads basedon what, I didn't have any of that.
Because actually what it did is itfound me a really lovely, like X one
carbon think pad from 2019 for 135 quid.
I'm like, get in there.
That's a two, 300 pound quid model.
(19:43):
It's got decent memory, decenthard drive, SSD, loving it.
So I just went straight onand bought it straight away.
So suddenly all of that clickand evidence data that previously
would've been used to stalk thehell out of me and mark it to me.
Just irrelevant.
Didn't even happen.
Right.
And actually my decision process from,okay, I need this to finding the right
(20:05):
solution, was literally about 10 minutestry and reach somebody in that window.
Daniel Rowles (20:11):
So it's a complete change in search behavior.
Ciaran (20:12):
just shows Actually, I think people do, and that's, so,
lots of people are doing this andlots of different levels, not just
with agents, but just with, youknow, the chat services in general.
And so they're refining itbefore they even start searching.
And that is a massive change.
You know, people used to do halfa dozen searches to refine, they
just use Google for doing this.
Now they're using othertools as well as Google.
(20:33):
So you've got a double effect on that.
And I think, you know, we'rekidding ourselves If you think
people are still behaving the sameway they were this time last year.
So when you're doing those yearon year comparisons, like, you
know, you have to factor this in.
Daniel Rowles (20:46):
Let's make a little view actually on as well, why the stats
are being so confusing about this.
And I'll give you an example.
I thought what I'll do is a littlestudy to try and work out how do
people actually behave when they getan AI overview versus when they don't.
And was it like, and my stats werecompletely misleading when I worked out.
So I did, every time I did a trainingcourse for a while, I got people
(21:08):
during the training course to buddyup with someone else and just go.
One of you search for something andthe other one of you observe them
and then you see their behavior,which in my head was a lovely test.
Okay?
It's completely flawed.
I'll explain why the second I say to you.
Go and search for something.
People in that kind of classroomenvironment go to a search engine
(21:29):
and they put in a search phrase.
Which in their head is like a two orthree word phrase, but that's not what
people are actually doing anymore.
What they're doing is they'reinteractively asking questions and
discussing and debating, but I'vejust said, go and search them.
I put them in a mindset, which iswhat they were doing two years ago.
So what I found in that scenario isthat 82 people I tested this out on,
(21:50):
and it was only about 5% of them thatgot an AI overview answer and then
didn't click through to our website.
They just got what they wanted, right.
Ciaran (21:58):
Yeah.
Yeah.
Daniel Rowles (21:59):
We chose if everything was just flat and Google is just
having AI overviews now, you'd alreadybe losing 5% of the clicks, right?
But that's without all this change inbehavior of me going into chat two BT
and researching something, my searchterms getting longer because the fact
that I'm used to now inquiring in a moreconversational tone of voice as well.
So you'd layer on that on topand you start to get to the 30,
(22:20):
40% we're talking about as well.
Ciaran (22:22):
When you just had 10 blue links, it made no sense to give
it a three sentence question.
'cause you, it justwouldn't give you better
Daniel Rowles (22:31):
Yeah.
Yeah.
Ciaran (22:32):
But now it does and we've learned how to prompt and actually
Google's learned how to accept.
Prompt level inputs.
And actually increasingly peopleare just talking to Google,
like they're using voice.
And it's much more conversationalbecause the, it can deal with that now.
Like it couldn't, five years ago itwas you could do it, but you wouldn't,
you weren't gonna get the result, youweren't gonna get anything back out of it.
(22:55):
Now, you definitely
Daniel Rowles (22:56):
So two, two points in this, first of all is if you haven't
seen it, we haven't mentioned itbriefly before they've rolled out AI
mode in search, which is basicallyyou click the AI mode button and it
drops into Google Gemini basically.
And essentially you're havinga contextual conversation.
IE you ask it a question and thenyou can ask it follow up questions.
So that's a completely different wayof searching, and it's clear to me what
they're kind of doing is saying, well,people are using chatt pt. But if they've
(23:20):
got that built into Google search,hopefully more people will come to us.
So that seems like acompetitive move to do
Ciaran (23:25):
Yeah.
Daniel Rowles (23:26):
that.
The second one from this is that.
If you think about previous search page10 blue links, you click through to them.
What would generally happen is you'd aska question, you click on the first one.
It's like, oh no, this isn't what I want.
You click on the second or the third andyou go through four or five of them and
eventually you get the answer you wanted.
'cause they trying to work it.
Well, actually, if you are takingthat away from me and you are doing
that work for me, which is thetheory, it's not necessarily that
(23:47):
they are, but that's the theory.
They're working outwhat is the best answer.
Ciaran (23:50):
yeah.
Daniel Rowles (23:51):
I don't need to do that dance, but what that also means.
If that was five clicks toget to the right answer, four
of those clicks were trash.
As in, I got your website, Ididn't like what I saw and I left.
And if anything, it was a negativebrand touchpoint to some extent because
I got your website and you didn'treally gimme the answer I wanted.
So actually, we are worried aboutthis loss in clicks, but what is it?
(24:11):
A loss in quality clicks
Ciaran (24:13):
The
Daniel Rowles (24:14):
and
Ciaran (24:14):
suggests not, but I think certain keyword landscapes, you know,
Daniel Rowles (24:20):
Yeah.
Ciaran (24:21):
the world's moved on.
It's changed and wholesections are being decimated.
But
Daniel Rowles (24:26):
Right.
Ciaran (24:26):
is out there too, right?
'cause
Daniel Rowles (24:27):
Yeah.
Ciaran (24:28):
literally crying in the street over it.
Daniel Rowles (24:30):
Yeah.
Yeah, that's it.
So, so look, if I were to try andtake away some of this, like what
on earth can I practically do now?
I would just go, like I said amoment ago, there's a couple of
things you want to talk about.
Share a voice and brand sentiment.
Where can you get mentionedonline by your advocates?
How can you drown out thenegative with positive stuff?
That means be really nice to yourcustomers and encourage them to engage.
(24:50):
About you as much as possible as well.
Build that advocacy, encourage thoseconversations as much as possible.
Benchmark yourself now to workout where you're being mentioned.
So you've got tools like ai,search Grader from HubSpot.
You've got way K, which standsfor what AI knows about you.io.
So way, k.io.
(25:10):
They're tools that are basicallysaying kind of like social media
monitoring it slightly different.
Like where have you beenmentioned, where are the brand
mentions of you online as well?
So Aldi, where you are now, andthen think about is that good stuff?
If it's bad, how might you get rid of it?
But also how are you gonnabuild up positive mentions?
So brand advocacy hugely important.
I really work on that.
And then the other piece is,okay, well we know that people are
(25:32):
searching in more complex questions.
We need to answer questions and weneed to think about those deeper
questions and really do more researchand speak to our target audience.
So I think that's a reallygood starting point.
One thing you mentioned to me, whichI thought was interesting, which
is the recency of LLM training.
Now
Ciaran (25:46):
Yeah.
Daniel Rowles (25:47):
you were making about this was that an LM is updated every month,
three months, whatever it might be.
So, and what they're gonna do,they're not gonna be rebuilding
the whole LLM each time.
They're adding on what's new.
Ciaran (26:00):
Yeah.
Daniel Rowles (26:00):
So actually the point you were making that it's gonna favor
Ciaran (26:04):
There's a
Daniel Rowles (26:05):
logically,
Ciaran (26:06):
there's a potential weak point you can exploit there.
I think if you are publishing highquality, well referenced, and here's
the thing, you can't just churn stuffout guys, which says and claim stuff.
You have to back it up
Daniel Rowles (26:20):
right?
Ciaran (26:21):
supporting links that show you know what you're talking about.
and this is a bit more trustworthythan everybody else just.
Standing on a soapbox, spouting stuff.
So I think it's, it is one of thethings I've been playing around with.
I think like tools likeGemini are fantastic.
Helping you to find stuff thatsupports what you're saying,
that's gonna be like taken by.
(26:42):
But in a way, our commercial writing hasbecome much more aligned with what they
would do in academia where you cite.
You know, studies and
Daniel Rowles (26:50):
You have to reference everything.
Yeah.
Ciaran (26:51):
You do.
Otherwise it's just gonna bemore dross for the grist and it's
just gonna kind of get ignored.
And publishing regularly on recent stuffand recent events, that's potentially
somewhere where you could get attention.
'cause not everybody's doingthat 'cause it's hard work.
And that's it's a strategy thatI'm exploring at the moment.
I can report back on, on howwell that works, but my hunch is
(27:12):
that'll work a little bit better.
I think also there's a greatopportunity to, you know, we've
always talked about, you know,take your diamonds and polish them.
You know, actually, if you've not recentlyupdated your best stuff, maybe it's time
to, maybe you need to revisit that and.
Daniel Rowles (27:27):
I really agree with this.
Yeah, because I thinkthis is a big opportunity.
Ciaran (27:30):
and reference it properly.
Whereas previously wewouldn't have done that.
And it just a few hyperlinks inthere can make all the difference.
But make sure you are quoting, youknow, well respected co content.
So suddenly I think your links to otherpeople's content are much, much more
important than they potentially have been
Daniel Rowles (27:48):
S so a fo.
Ciaran (27:49):
reason.
But remember, you are now competing for.
For the top three links, not the top 10.
Right?
So there's a lot less people that aregonna win with this and I think that's
the key take home for me on all of this.
don't think SEO is deadfor of course it isn't.
There'll always be what like thingsthat you can do to help and exploit you
know, the systems that are out therethat indexing everything for sure.
(28:12):
But it's become a lot harder toinfluence it and be wary at the moment.
There are a lot of absolute snake.
Oil salesmen out there trying tosell you like solutions to things
that are not really well tested.
And I think, you know, brandsneed to be wary of this.
Everybody's coming out of thewoodwork with amazingly sounding
(28:32):
like products that claim they dothis, but where's the evidence?
That's the key quote.
Take a step back.
Where's the evidence?
Look, I know they're sellingsomething that you want because
you've got a real problem here.
But please just check.
Do some like sense checking beforeyou leap in and start spending
tens of thousands of pounds on it.
Daniel Rowles (28:49):
Brilliant advice.
Well, let's finish on a positive.
I've got a really positive noteto finish on here, which is a lot
of the stuff we've just spokenabout is actually demonstrating.
Your expertise, your experience,your authorit, and that builds trust.
Well, that's eat, that'sGoogle's eat, right?
And Google.
Ever since a lot of this AI generatedcontent came out and said, you must
(29:10):
now demonstrate, eat with your content.
Add referencing, add case studies, addexamples, make your content different.
You know, give answers toquestions, your experience.
So actually you can do really well.
SEO by thinking about those principles.
And we'll put links through in theshow notes, target internet.com/podcast
to loads of other stuff we'vedone about eat previously.
(29:30):
But actually that's gonna help you froma generative engine optimization if you
wanna separate the two things as well.
Now what we have done recently iswe've done all of our members and
our newsletter subscribers getthese monthly update sessions.
So if you're not a newsletter subscriber,target internet.com/newsletter.
We did one on AI agents.
(29:51):
And we did one on the latestGoogle algorithm updates as well.
So if you're not signedup to newsletter, sign up.
'cause you'll get those every month.
It's one hour live session online,but members have got the whole library
of the recordings of those as well.
What we've also got from members comingup is a half day masterclass on using
AI agents and a half day masterclasson latest SEO techniques as well.
(30:12):
So, either get signed up asa member to target internet.
The annual membership is on areally good offer at the moment.
Or just get signed up to thenewsletter if you just want those
monthly updates 'cause we're talkingabout this stuff a lot and it gives
you a chance to ask some questionsand those kind of things as well.
Thanks for listening to theDigital Marketing Podcast and
we'll see you again very soon.
I.
For more episodes, resources toleave a review or to get in contact,
(30:34):
go to target internet.com/podcast.
Hello and welcome back to the Digital Marketing Podcast,
(00:03):
brought to you by Target Internet.
My name is Kieran Rogers,
Daniel Rowles (00:07):
and I'm Daniel Rolls.
Ciaran (00:08):
and today we are discussing the future of SEO.
Daniel Rowles (00:21):
Okay, so we have had multiple conversations recently in
podcasts and in various articles aboutwhat's happening with search engine
optimization and the big conversationhas obviously been about GEO, which
is generative engine optimization.
And then there's been lots of debates.
Is it really a thing?
Is SEO just expanding?
(00:42):
We've got AI overviews.
And then we talked a little bitabout answer engine optimization,
which was, and anyway this all seemsto be settling down a little bit
Ciaran (00:51):
Is it
though?
Daniel Rowles (00:52):
Well, no.
If I'm honest,
Ciaran (00:54):
I would say the SEO is dead.
Mongers have gone intooverdrive with it all.
Daniel Rowles (00:59):
that's normally me.
Ciaran (01:00):
I know, right?
But no, it's everywhere now.
you, maybe you were like the Heraldthat saw it coming first, but there's
a lot of panic and there's a, andthere's a lot of lawsuits flying around.
People are getting that antsy about the
Daniel Rowles (01:13):
So let's talk about that.
'cause you mentioned this tome as well, so that people
are suing Google on the basis.
They're not as getting as manyclicks as they used to get.
Ciaran (01:21):
yeah.
Yeah.
Daniel Rowles (01:22):
That seems slightly wacky.
Ciaran (01:24):
of European businesses.
I'm not gonna go into it 'cause Idon't wanna get embroided in the legal
Daniel Rowles (01:29):
Yeah.
Dragged into a legal case.
Ciaran (01:30):
not really.
But if you search on it,I'm sure you'll find it.
And I think it's really interestingthe basic premise of this is across
the board, a lot of businesses,particularly content marketing
focused businesses and publishers.
In particular, and I think that particularconsortium is made up of, a lot of
Daniel Rowles (01:47):
Right.
Ciaran (01:48):
have seen their organic year on year traffic absolutely decimated.
And there's all sorts of numbers flyingaround between like 40 and 60% drops
in some cases depending on kind of thekeyword landscape that you're looking at.
And what I found particularly interesting.
Is that there's also a lot of reportsthat actually businesses just aren't
seeing those levels of drops withine-commerce, So this is much more top
(02:12):
of funnel kind of content research.
And that makes perfect sense becausewe now have AI answers, you know,
that's been launched across the world.
But I think my view on this isthat actually what's actually
happening is because of that,people's behavior is shifting.
Radically, like before AI answers,you had to work a bit to get answers.
(02:33):
You typed in your search query andyou got your 10 blue links, and then
you did, you clicked around and youvisited quite a few places and you
actually, it was a journey, wasn't it?
You sort of learn a fewthings along the way.
Sometimes when you wantedto learn something.
Actually, it helps to have to workfor it rather than it just being
handed to you on a plate like
Daniel Rowles (02:50):
Well, let's talk about that then for a second,
because we've got AI overviews
And it started off peoplesaying, these are useless.
And then they got better because the largelanguage models got better, and Google
particularly got much better at this.
And then we started to seethat in some circumstances was.
Answering people's questions.
The thing is, like you've said to mebefore, they are, they're sourcing
(03:11):
maybe three different links in there,three websites where this information
has kind of come from to some extent.
Ciaran (03:16):
Yeah.
Daniel Rowles (03:16):
And what's interesting to give you a quick example, is that
if you search, what is data scraping,we always rank number one for that.
Now, when you search,what is data scraping?
There's an AI overview.
We're still number one below that,but we're actually listed in the
websites where the informationhas come from in the first place,
Ciaran (03:32):
Yeah.
Daniel Rowles (03:33):
so, so what that tells me is quite interesting is that actually
by answering questions and exactlyto your point, is there's a bit of a
killer strategy still at the moment oflike actually working out what are the
questions that people are searchingfor, answering those in a really
effective way that's still gettingpicked up by the search engines, but
it's getting picked up to some extentby the large language models as well.
Ciaran (03:53):
here's the problem.
You're gonna see a traffic drop, Soacross the board, in the accounts that
I look at, I've seen lots of this.
You see an in quite a big increasein impressions year on year.
But you're seeing a big drop in clicks andmore recently a drop within the rankings.
And that flows for me because I knowthat when you get a drop in click-through
(04:13):
rate, it affects your rankings,like your average ranking overall.
So that kinda makes sense.
Why aren't people clicking?
Because actually for the majorityof them, they've had their itch
scratched by the AI answers.
Daniel Rowles (04:23):
Right.
Ciaran (04:24):
You know that is enough, so not, so let's not forget if you're
on a mobile device, that lovely AIanswers, just push those 10 blue
links even lower down the page.
we all know how lazy, lazy fat thumbs weall have when we're on the mobile device.
Like we just, it's effort to flick up
Daniel Rowles (04:42):
right.
Ciaran (04:42):
you know, if I've had the information itch
scratched, why would I need to?
Scroll down.
This is the thing.
It's pulling together andactually it's quite smart.
Before we had to get quite creative.
Sometimes you'd need to rearrangethe words or add in additional you
know, you know, words to filter
Daniel Rowles (04:59):
Huh,
Ciaran (05:00):
Now
Daniel Rowles (05:00):
a great stat on this
Ciaran (05:01):
Yeah.
Go on.
Go on.
Daniel Rowles (05:02):
so that the AI overviews are only including the exact phrase
you search for in 5% of cases.
So the point being is that theyunderstand, well, this phrase
really means this phrase as well,and that's the same as this.
And therefore if I give you the answerto this, it's is what you're asking for.
You've just asked for it ina slightly different way.
Ciaran (05:21):
Yeah.
Daniel Rowles (05:22):
So that kinda shows that exact, this phrase, I'm gonna
optimize for this phrase, and I'm gonnarank for this phrase is kind of gone.
Ciaran (05:27):
I'm gonna throw in another thought for you all as well, and I
haven't got, unfortunately, I haven'tgot anything that backs this up, but
it makes logical sense to me if you aregetting, as a user, if you're getting
used to having conversations with Google.
Much like if you use chat GPT or Gemini,you have conversations with those chat,
chat entities as well and Google'sresearch actually does back this up.
(05:49):
The number of words people aresearching on in any given search
has increased quite dramatically
Daniel Rowles (05:54):
Oh no, The numbers are there to back up.
Exactly.
There's all that stuffthat came out from Google.
Put it into the show notes.
Ciaran (05:58):
so here's the thing.
If I'm a publisher and for the last 15, 20years I've been dining out on optimizing
for certain key phrases and questions,when people stop asking those, 'cause
actually they're phrasing 'em slightlydifferently, I'm not getting picked up.
So, you know, is that, is, is the thing Ithink's kind of unfair to, to, to Google.
(06:18):
And I mean, they're big en, they'rebig enough to fight their own battles.
They don't really need me to doit, but I'm gonna do it anyway.
You know, it's not Google's fault that.
had to innovate, they've had tobring in this new technology.
Let's face it, if they don't, they'regonna get wiped out by other competitors
within the marketplace who will.
Right.
And I think actually, from what I've seen,they've done a pretty good job of this.
Like it's good, it's usable.
(06:38):
People are using it, thedata's there to support that.
if I'm on Google, is it really my faultthat the whole marketplace is now using
the web and search in a different way?
Is it really my fault that, you know,certain businesses have over optimized for
certain phrases that have been dining out?
No, not really.
That's just life.
Every marketplace has that.
(06:59):
People come in, new innovations happen.
People lose their share.
Other operators like still stealtheir thunder that's just how
business has always worked.
So I kind of alarming that, like peopleare going down the lawsuit route.
Maybe there's an angle onthis that I haven't kind of
clocked, and if so, us a shout.
Daniel Rowles (07:18):
Well, I do know a little bit more about it.
So it's the fact that the publishers,what's been happening is that
publishers rely on the fact thatGoogle was swallowing up some of
their stuff, sourcing it in thesearch results, but then people were
clicking through to the website.
So they kind of felt, well, we'veinvested in creating this content,
but actually at least we aredriving some traffic value from it.
But what's actually happening now is that.
They're getting all that stuff inGoogle itself, and they're not just
(07:40):
seeing the value from it and they thinkit's a break in that relationship,
Ciaran (07:43):
Okay.
Daniel Rowles (07:43):
but, which I get, but it, that's not their fault specifically.
It's 'cause the operatingenvironment has changed.
Ciaran (07:47):
Yeah.
Daniel Rowles (07:48):
let's take a step back for a second.
So we've got AI overviewsshowing up in the search results.
Search engine optimization seemsto be expanding to include.
This concept of generativeengine optimization.
I wanna show up in chat, CPT.
But the thing is we don't know ahuge amount that's about how these
large language models are working
Still.
(08:08):
But we do know a couple of things.
LLMs are basically built onneural networks essentially.
And a neural network is a, somethingthat's been around artificial
intelligence for a long time.
And it's the connections between things.
Okay.
As in if a particular phrase is mentionedin lots of different places and there's
positive sentiment about it, and there'sall these kind of interconnections.
This is probably relevant.
(08:29):
Okay, so what that leads to isshare a voice and brand sentiment
being seen as really important.
But actually that's not as advanced aswhat Google will be doing a long time.
'cause Google have that next layerof user experience to say, well we
think this is the most relevant thing.
But if a user clicks through toit, do they have a good experience?
(08:51):
Do they stay on that webpage?
They to scroll to the bottom.
So it's potential that thelarge language modules are
kind of putting stuff up front.
In their search results that mightbe mentioned a lot might seem to be
good, but isn't necessarily a gooduser experience and so on as well.
Now lots of people would argue, actuallyGoogle's been showing stuff at the
top of the search results isn't agood user experience for a long time.
And you know, Rand Fishkin has beensaying that, you know, for a long time.
(09:13):
But what it does mean is when I use atool like the HubSpot AI search grader,
which has actually got a lot betterand it's trying to kind of estimate
how visible you are in Chatt PT andGoogle Gemini and places like that.
So two things they say is,where have you been mentioned?
What's the brand sentiment?
The brand sentiment analysis ispretty clever 'cause it'll say
these are the things that peoplethought were good about you.
(09:35):
These are the things that peoplethought were potentially negative
about you, but it's sourcing stuffthat maybe isn't very reliable.
So what they're looking at a lot isstuff that's mentioned in forums.
So Reddit.
Cure those kind of places?
Well, we know thatReddit is the Wild West.
There's some good user stuff in there.
There's also some prettywacky stuff in there as well.
(09:57):
But it also, if there's a lack of,like if you've got online reviews.
But you've got, you know, mentions you'vearound other places they can be picked up.
So places like App Sumo, whichI'll give an example of places
like Wikipedia and so on.
Ciaran (10:10):
Yeah.
Daniel Rowles (10:10):
So let me give you the app Sumo example.
I bring that one up.
We did an app Sumo promotion, andif you're not familiar with it,
you go into App Sumo, you say,right, we can do a special offer.
We want a load of users just so wecan test out our product platform.
And they take.
More often than not 70% of any moneyyou make and you're supposed to give a
lifetime subscription for like a 10thof what people would normally pay.
(10:33):
Right.
It's a pretty, it's a prettybonkers model when you look at it
from a commercial point of view.
'cause you have to sella lot to make it work.
But what it does give you isa user base and a test base.
Right.
So we did an app Sumo promotion.
I can't say that I'm entirelypleased with the results, but let's
skirt around that for a moment.
Ciaran (10:50):
you being very excited before it went live, and then you got frowny.
That's all I'm
Daniel Rowles (10:55):
Yeah.
Ciaran (10:56):
When
Daniel Rowles (10:57):
Well, I'll tell you what was interesting was that my
commercial director had said tome, this is a really dumb idea.
We've got a reasonablywell established product.
We are not the rightlife stage of business.
And I think that's completely right.
It's like if you're a brand newstartup and you needed a test user
base, this makes perfect sense ifyou're as established as we were.
And because she'd said tome, this is a bad idea.
And then I was like, no,it's gonna be brilliant.
We're gonna get thousands of users.
And then we did it and it was a, itwas like we, we've basically given
(11:20):
away about 50,000 pounds worth of ourproduct for about a thousand dollars.
Ciaran (11:24):
do you
T-shirt she gave you saying?
I told you so?
Daniel Rowles (11:27):
No I've got so many of them now that you know, but
anyway, you should probably, peoplelisten to your commercial directors.
Ciaran (11:33):
Yes.
Daniel Rowles (11:34):
so anyway, you get a load of reviews from these people,
and these people weren't inherentlyour target user base, right?
So we had some kind of strange, wehad loads of really nice reviews.
I mean, it was still like 95% positivereviews, but because the large
language model is looking for thepositive and the negative to analyze
the brand sentiment, when you go in.
And I went into the AI search grader,the HubSpot tool, put our brand in
(11:56):
and it says, most people love you.
You are brilliant, it's great.
Here's positive stuff.
But some people think an outcome, theseapp Sumo reviews from how many five
years ago that it was kind of thing.
Because we aren't on many platforms likethat by nature of what we kind of are.
That will now haunt us forevermore.
So the only way you can really dealwith it is say I need to be mentioned
in more places by more people.
Ciaran (12:17):
Yeah.
Daniel Rowles (12:18):
advocacy, and I've been telling this to a
few training courses recently.
We were talking about advocacybefore, like, you know, it's not about
influencers, it's about advocates.
It's those everyday people thatsay nice things about your product.
You want a thousand truefans, all that kinda stuff.
It's never been more true because yes,Google's looking at Reddit more now and
places like that, but actually theselarge language models are as well.
So if there were two quick wins in this.
Ciaran (12:39):
yeah.
Daniel Rowles (12:40):
I would say the first one is advocacy.
Get nice people saying stuffabout your more places.
And the other one is answer more questionsin depth and think about that the way
that people are changing their searching,
Ciaran (12:50):
and like I've seen this on Wikipedia and be
Daniel Rowles (12:53):
right.
Ciaran (12:54):
there's a big drive at the moment.
People have clocked that alot of the AI systems really
rely on Wikipedia for facts.
And it is quite good for that, right?
Because they have quite a robustsystem of editors, like human editors
has the double check stuff anddouble check edits when they're made.
I think for like a big organizationlike Take Brew Dog for example,
(13:15):
have a look at Brew Dog's Wikipediapage, there's a ton of stuff.
On their page is quite balanced.
There's some positives.
Plenty of negatives as well, you know,overall they've got quite an established
like brand page on Wikipedia, right?
So something, you know, negative butrelatively small happens, it's quite easy
for that just to get ignored or missed.
'cause it's not relevant to thebig entity that they are, like
(13:38):
the editors will overlook it.
However, if you are a new company.
Without much history and somebodygoes, oh, let's get a Wikipedia page.
'cause then we'll exist as an entityand the LLMs will trust us more.
Yeah.
True or true, have to go throughthis process of submitting an entry.
You have to kind of.
(13:58):
Make sure it's balanced.
Like it can't be overly on promotion.
You have to submit it to an editor.
editor's gonna look through your immediatekind of recent history, just do a few
checks to check what you're saying.
And if they find anything negativethat you haven't included there,
they're gonna slap that in there.
Yeah.
Or worse ban you for like making a,an unbalanced, unbiased submission.
Daniel Rowles (14:21):
Right.
Ciaran (14:21):
And then once it's out there, remember this page is public property,
so anybody can make a. An edit.
And I think the problemis when you don't have.
I suppose it's a bit like, let'ssay Kieran decides that he's
important enough to deserve a pagein like the Encyclopedia Britannica.
Like it is it, that's a nonsense.
Like I'm not, I'm just not at that level.
(14:42):
I doubt I ever will be.
Right?
But let's say I decided I was gonnapush it not be very much verifiable fact
that's interesting enough that you couldput into that Britannica page, right?
And so what is there isgonna be a bit skewed.
if there's any one thing that'sno, like verifiably notable
and it's a bit negative, that'sdefinitely going in there.
(15:03):
Right?
And now you've got something that'sfactually true, but kind of biased
because it's only focusing onthe things that are big enough or
epic enough or verifiable enough.
And that's the risk with this.
So
Daniel Rowles (15:15):
Be careful what you wish for.
Ciaran (15:16):
Yeah.
Daniel Rowles (15:16):
Yeah.
Ciaran (15:17):
careful what There's there's a concept really here that actually really.
Wise people actually spenda lot of time thinking about
the outcomes of their actions.
Fools just rush in and don't thinkabout the outcomes that their
actions and bad stuff happens.
Right,
Daniel Rowles (15:34):
Well talk, talking of which
Ciaran (15:35):
Yeah,
Daniel Rowles (15:36):
I I decided that I was worthy of a Wikipedia page, so.
Yeah.
Yeah.
So because, well, becausepublished authors quite often
you are a published author.
I've got quite a few, like I'm sitting,I'm, I've the humble break behind me
that no one can see is I've got thebig bookshelf with all my books on.
It
Ciaran (15:52):
are behind him and they're extensive.
I'm
Daniel Rowles (15:54):
is
Ciaran (15:55):
on half of one of those books.
Daniel Rowles (15:57):
half of one.
That's all right.
Ciaran (15:58):
Well, I wrote half of it.
Daniel Rowles (16:00):
But
Ciaran (16:00):
count?
I
Daniel Rowles (16:01):
you are on two of them.
'cause there's one in another languageI think as well, but there's like
14 languages, books, blah, blah.
Anyway, so I thought, well thatmust be a good opportunity.
Ciaran (16:07):
in
Daniel Rowles (16:07):
Anyway.
I've been,
Ciaran (16:08):
Isn't it so
Daniel Rowles (16:09):
I think it is.
Ciaran (16:10):
I think it's in Russian too.
Yeah.
Daniel Rowles (16:12):
So there's a few language.
There you go.
You are worthy of Wikipedia page.
Look, you in multiple languages.
But the point being is there was that, andthen there was like I published Imperial
College academic papers and a few otherthings, and I thought, come on this would
be really good from an SEO point of view.
Ciaran (16:23):
yeah.
Daniel Rowles (16:23):
I went on to Fiverr and I found this guy that like
specializes in in doing Wikipedia pages.
And he got.
Well, he was actually, he was really good.
I'm not gonna diss the guy, but he I spentlike an hour and a half writing an email.
Here's all the links to all my staff.
Here's all my books, here's theconference of the judging that I do, the
things that I'm members of, and blah,blah, blah, blah, blah, blah, blah.
(16:44):
And after an hour, if I submittedit and within 45 seconds, he'd
responded to me going, you reallydon't deserve Wikipedia, Paige.
I said, that's brutal.
Alright, well thanks for that.
Ciaran (16:55):
been there though, right?
He
Daniel Rowles (16:56):
Well, that was the point.
Ciaran (16:57):
is.
Daniel Rowles (16:57):
That was the point.
Ciaran (16:58):
the thing.
This is the thing.
And I think that's the danger with this.
And let's say you get your company thereand it only mentions one thing about you.
And that thing's a positive thing.
Everyone's happy, right?
Until something not so goodhappens and it gets reported on
in the press, and then you're justwaiting for that, that to drop,
Daniel Rowles (17:13):
Right.
Ciaran (17:13):
It could happen at any point.
They might.
Here's the thing, when you firstsubmit, everything's double checked.
Once you've got a page.
what I've seen, I've found quite a fewexamples of companies that had all sorts
of, you know, advertising standards,agency rulings against some stuff never
mentioned on the Wikipedia page, butit's 'cause the Wikipedia, your page
was created maybe six, seven yearsbefore that happened, So, but at any
(17:36):
point that could catch up, and when itdoes, very difficult to get rid of it,
and that's permanently on your recordand that is totally gonna get skewed.
So you could be in a situation where,yeah, the large language models are
using it and they're now only citingone thing that's a bit negative.
They're trying to give a balance viewand actually it doesn't, so be careful
Daniel Rowles (17:53):
Okay, so we've got Share of voice is important.
Brand sentiment is important.
Be careful your brandis showing up something.
So we've got increasingly got thisno click environment where you're
not as getting any clicks, andthat's because people are searching,
getting the answers they want.
You might get more directvisits, you might get more brand
searches, so keep an eye on that.
Ciaran (18:12):
important.
See, people are still thinking about you,
Daniel Rowles (18:15):
Right.
Ciaran (18:16):
it in a way that you can measure.
They're not clicking andit's the clicks that we can
Daniel Rowles (18:22):
Well, you mentioned to me off platform research.
I thought that was an interestingway of thinking about it.
Ciaran (18:25):
do it all the time.
Like before I Google stuff now I'll goto my large language learning model and
I'll do my research there and I'll refinemy search down for two or three things.
Sometimes just one thing.
I was telling Daniel, I've justbought my first Windows pc.
I'm so excited about it.
Secondhand, it was old.
I needed to use some Windows software fora project that I'm doing, so I instantly,
(18:49):
I decided, right, I'm gonna try out theagents feature in chat, GPT and I set
So I've just bought my first Windows pc.
Very excited about it.
It's secondhand.
I needed it for a personal projectworking on, I needed some Windows software
that was only available on Windows.
So I set my agent to find me the cheapest,best spec laptop on eBay for under 150.
(19:14):
It came up with absolute sto.
It did all this work for me.
Now, previously I would've beenclicking through, you know,
Daniel Rowles (19:21):
From Google search after search.
Yeah.
Ciaran (19:23):
gonna get a trail of it.
eBay's gonna get a massive trail ofwhat I've been looking at and thinking
it's gonna start showing me ads basedon what, I didn't have any of that.
Because actually what it did is itfound me a really lovely, like X one
carbon think pad from 2019 for 135 quid.
I'm like, get in there.
That's a two, 300 pound quid model.
(19:43):
It's got decent memory, decenthard drive, SSD, loving it.
So I just went straight onand bought it straight away.
So suddenly all of that clickand evidence data that previously
would've been used to stalk thehell out of me and mark it to me.
Just irrelevant.
Didn't even happen.
Right.
And actually my decision process from,okay, I need this to finding the right
(20:05):
solution, was literally about 10 minutestry and reach somebody in that window.
Daniel Rowles (20:11):
So it's a complete change in search behavior.
Ciaran (20:12):
just shows Actually, I think people do, and that's, so,
lots of people are doing this andlots of different levels, not just
with agents, but just with, youknow, the chat services in general.
And so they're refining itbefore they even start searching.
And that is a massive change.
You know, people used to do halfa dozen searches to refine, they
just use Google for doing this.
Now they're using othertools as well as Google.
(20:33):
So you've got a double effect on that.
And I think, you know, we'rekidding ourselves If you think
people are still behaving the sameway they were this time last year.
So when you're doing those yearon year comparisons, like, you
know, you have to factor this in.
Daniel Rowles (20:46):
Let's make a little view actually on as well, why the stats
are being so confusing about this.
And I'll give you an example.
I thought what I'll do is a littlestudy to try and work out how do
people actually behave when they getan AI overview versus when they don't.
And was it like, and my stats werecompletely misleading when I worked out.
So I did, every time I did a trainingcourse for a while, I got people
(21:08):
during the training course to buddyup with someone else and just go.
One of you search for something andthe other one of you observe them
and then you see their behavior,which in my head was a lovely test.
Okay?
It's completely flawed.
I'll explain why the second I say to you.
Go and search for something.
People in that kind of classroomenvironment go to a search engine
(21:29):
and they put in a search phrase.
Which in their head is like a two orthree word phrase, but that's not what
people are actually doing anymore.
What they're doing is they'reinteractively asking questions and
discussing and debating, but I'vejust said, go and search them.
I put them in a mindset, which iswhat they were doing two years ago.
So what I found in that scenario isthat 82 people I tested this out on,
(21:50):
and it was only about 5% of them thatgot an AI overview answer and then
didn't click through to our website.
They just got what they wanted, right.
Ciaran (21:58):
Yeah.
Yeah.
Daniel Rowles (21:59):
We chose if everything was just flat and Google is just
having AI overviews now, you'd alreadybe losing 5% of the clicks, right?
But that's without all this change inbehavior of me going into chat two BT
and researching something, my searchterms getting longer because the fact
that I'm used to now inquiring in a moreconversational tone of voice as well.
So you'd layer on that on topand you start to get to the 30,
(22:20):
40% we're talking about as well.
Ciaran (22:22):
When you just had 10 blue links, it made no sense to give
it a three sentence question.
'cause you, it justwouldn't give you better
Daniel Rowles (22:31):
Yeah.
Yeah.
Ciaran (22:32):
But now it does and we've learned how to prompt and actually
Google's learned how to accept.
Prompt level inputs.
And actually increasingly peopleare just talking to Google,
like they're using voice.
And it's much more conversationalbecause the, it can deal with that now.
Like it couldn't, five years ago itwas you could do it, but you wouldn't,
you weren't gonna get the result, youweren't gonna get anything back out of it.
(22:55):
Now, you definitely
Daniel Rowles (22:56):
So two, two points in this, first of all is if you haven't
seen it, we haven't mentioned itbriefly before they've rolled out AI
mode in search, which is basicallyyou click the AI mode button and it
drops into Google Gemini basically.
And essentially you're havinga contextual conversation.
IE you ask it a question and thenyou can ask it follow up questions.
So that's a completely different wayof searching, and it's clear to me what
they're kind of doing is saying, well,people are using chatt pt. But if they've
(23:20):
got that built into Google search,hopefully more people will come to us.
So that seems like acompetitive move to do
Ciaran (23:25):
Yeah.
Daniel Rowles (23:26):
that.
The second one from this is that.
If you think about previous search page10 blue links, you click through to them.
What would generally happen is you'd aska question, you click on the first one.
It's like, oh no, this isn't what I want.
You click on the second or the third andyou go through four or five of them and
eventually you get the answer you wanted.
'cause they trying to work it.
Well, actually, if you are takingthat away from me and you are doing
that work for me, which is thetheory, it's not necessarily that
(23:47):
they are, but that's the theory.
They're working outwhat is the best answer.
Ciaran (23:50):
yeah.
Daniel Rowles (23:51):
I don't need to do that dance, but what that also means.
If that was five clicks toget to the right answer, four
of those clicks were trash.
As in, I got your website, Ididn't like what I saw and I left.
And if anything, it was a negativebrand touchpoint to some extent because
I got your website and you didn'treally gimme the answer I wanted.
So actually, we are worried aboutthis loss in clicks, but what is it?
(24:11):
A loss in quality clicks
Ciaran (24:13):
The
Daniel Rowles (24:14):
and
Ciaran (24:14):
suggests not, but I think certain keyword landscapes, you know,
Daniel Rowles (24:20):
Yeah.
Ciaran (24:21):
the world's moved on.
It's changed and wholesections are being decimated.
But
Daniel Rowles (24:26):
Right.
Ciaran (24:26):
is out there too, right?
'cause
Daniel Rowles (24:27):
Yeah.
Ciaran (24:28):
literally crying in the street over it.
Daniel Rowles (24:30):
Yeah.
Yeah, that's it.
So, so look, if I were to try andtake away some of this, like what
on earth can I practically do now?
I would just go, like I said amoment ago, there's a couple of
things you want to talk about.
Share a voice and brand sentiment.
Where can you get mentionedonline by your advocates?
How can you drown out thenegative with positive stuff?
That means be really nice to yourcustomers and encourage them to engage.
(24:50):
About you as much as possible as well.
Build that advocacy, encourage thoseconversations as much as possible.
Benchmark yourself now to workout where you're being mentioned.
So you've got tools like ai,search Grader from HubSpot.
You've got way K, which standsfor what AI knows about you.io.
So way, k.io.
(25:10):
They're tools that are basicallysaying kind of like social media
monitoring it slightly different.
Like where have you beenmentioned, where are the brand
mentions of you online as well?
So Aldi, where you are now, andthen think about is that good stuff?
If it's bad, how might you get rid of it?
But also how are you gonnabuild up positive mentions?
So brand advocacy hugely important.
I really work on that.
And then the other piece is,okay, well we know that people are
(25:32):
searching in more complex questions.
We need to answer questions and weneed to think about those deeper
questions and really do more researchand speak to our target audience.
So I think that's a reallygood starting point.
One thing you mentioned to me, whichI thought was interesting, which
is the recency of LLM training.
Now
Ciaran (25:46):
Yeah.
Daniel Rowles (25:47):
you were making about this was that an LM is updated every month,
three months, whatever it might be.
So, and what they're gonna do,they're not gonna be rebuilding
the whole LLM each time.
They're adding on what's new.
Ciaran (26:00):
Yeah.
Daniel Rowles (26:00):
So actually the point you were making that it's gonna favor
Ciaran (26:04):
There's a
Daniel Rowles (26:05):
logically,
Ciaran (26:06):
there's a potential weak point you can exploit there.
I think if you are publishing highquality, well referenced, and here's
the thing, you can't just churn stuffout guys, which says and claim stuff.
You have to back it up
Daniel Rowles (26:20):
right?
Ciaran (26:21):
supporting links that show you know what you're talking about.
and this is a bit more trustworthythan everybody else just.
Standing on a soapbox, spouting stuff.
So I think it's, it is one of thethings I've been playing around with.
I think like tools likeGemini are fantastic.
Helping you to find stuff thatsupports what you're saying,
that's gonna be like taken by.
(26:42):
But in a way, our commercial writing hasbecome much more aligned with what they
would do in academia where you cite.
You know, studies and
Daniel Rowles (26:50):
You have to reference everything.
Yeah.
Ciaran (26:51):
You do.
Otherwise it's just gonna bemore dross for the grist and it's
just gonna kind of get ignored.
And publishing regularly on recent stuffand recent events, that's potentially
somewhere where you could get attention.
'cause not everybody's doingthat 'cause it's hard work.
And that's it's a strategy thatI'm exploring at the moment.
I can report back on, on howwell that works, but my hunch is
(27:12):
that'll work a little bit better.
I think also there's a greatopportunity to, you know, we've
always talked about, you know,take your diamonds and polish them.
You know, actually, if you've not recentlyupdated your best stuff, maybe it's time
to, maybe you need to revisit that and.
Daniel Rowles (27:27):
I really agree with this.
Yeah, because I thinkthis is a big opportunity.
Ciaran (27:30):
and reference it properly.
Whereas previously wewouldn't have done that.
And it just a few hyperlinks inthere can make all the difference.
But make sure you are quoting, youknow, well respected co content.
So suddenly I think your links to otherpeople's content are much, much more
important than they potentially have been
Daniel Rowles (27:48):
S so a fo.
Ciaran (27:49):
reason.
But remember, you are now competing for.
For the top three links, not the top 10.
Right?
So there's a lot less people that aregonna win with this and I think that's
the key take home for me on all of this.
don't think SEO is deadfor of course it isn't.
There'll always be what like thingsthat you can do to help and exploit you
know, the systems that are out therethat indexing everything for sure.
(28:12):
But it's become a lot harder toinfluence it and be wary at the moment.
There are a lot of absolute snake.
Oil salesmen out there trying tosell you like solutions to things
that are not really well tested.
And I think, you know, brandsneed to be wary of this.
Everybody's coming out of thewoodwork with amazingly sounding
(28:32):
like products that claim they dothis, but where's the evidence?
That's the key quote.
Take a step back.
Where's the evidence?
Look, I know they're sellingsomething that you want because
you've got a real problem here.
But please just check.
Do some like sense checking beforeyou leap in and start spending
tens of thousands of pounds on it.
Daniel Rowles (28:49):
Brilliant advice.
Well, let's finish on a positive.
I've got a really positive noteto finish on here, which is a lot
of the stuff we've just spokenabout is actually demonstrating.
Your expertise, your experience,your authorit, and that builds trust.
Well, that's eat, that'sGoogle's eat, right?
And Google.
Ever since a lot of this AI generatedcontent came out and said, you must
(29:10):
now demonstrate, eat with your content.
Add referencing, add case studies, addexamples, make your content different.
You know, give answers toquestions, your experience.
So actually you can do really well.
SEO by thinking about those principles.
And we'll put links through in theshow notes, target internet.com/podcast
to loads of other stuff we'vedone about eat previously.
(29:30):
But actually that's gonna help you froma generative engine optimization if you
wanna separate the two things as well.
Now what we have done recently iswe've done all of our members and
our newsletter subscribers getthese monthly update sessions.
So if you're not a newsletter subscriber,target internet.com/newsletter.
We did one on AI agents.
(29:51):
And we did one on the latestGoogle algorithm updates as well.
So if you're not signedup to newsletter, sign up.
'cause you'll get those every month.
It's one hour live session online,but members have got the whole library
of the recordings of those as well.
What we've also got from members comingup is a half day masterclass on using
AI agents and a half day masterclasson latest SEO techniques as well.
(30:12):
So, either get signed up asa member to target internet.
The annual membership is on areally good offer at the moment.
Or just get signed up to thenewsletter if you just want those
monthly updates 'cause we're talkingabout this stuff a lot and it gives
you a chance to ask some questionsand those kind of things as well.
Thanks for listening to theDigital Marketing Podcast and
we'll see you again very soon.
I.
For more episodes, resources toleave a review or to get in contact,
(30:34):
go to target internet.com/podcast.
Popular Podcasts
Fudd Around And Find Out
UConn basketball star Azzi Fudd brings her championship swag to iHeart Women’s Sports with Fudd Around and Find Out, a weekly podcast that takes fans along for the ride as Azzi spends her final year of college trying to reclaim the National Championship and prepare to be a first round WNBA draft pick. Ever wonder what it’s like to be a world-class athlete in the public spotlight while still managing schoolwork, friendships and family time? It’s time to Fudd Around and Find Out!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy, Jess Hilarious, And Charlamagne Tha God!