Latest / eCommerce Podcast / How to Get AI to Recommend Your Brand Instead of Your Competitors

Transcript

How to Get AI to Recommend Your Brand Instead of Your Competitors · eCommerce Podcast

0:04
So welcome to the eCommerce Podcast. My name is Matt Edmundson, and it's great
0:08
to be with you on this very sunny day from the new recording studio. So if
0:13
you're keen-eyed amongst us, you will notice that I'm in a very different studio
0:17
to where I normally am, which is fun. I wish I could show you the setup we've
0:22
got here. It's unbelievable. But it's still a work in progress. You can't see
0:27
all the wires and the cables and everything. And it's, it's But it's we're
0:31
getting there. So anyway, welcome to the show. We talk about all things
0:36
e-commerce. If this is your first time with us, a very, very warm welcome to
0:41
you. Like you, I run my own e-commerce business, and in fact, this studio is in
0:45
our warehouse. So behind these walls, behind the office walls, we have people
0:50
busy picking and packing orders, which is wonderful. So it means we're busy,
0:55
which is always a wonderful thing. So. Uh, I do this show, I've done it for
0:59
however many years now, um, just to talk to people, either founders or experts
1:04
in ecommerce, and just pull out their stories, their expertise to help me and
1:08
you run our ecommerce businesses better. So a very warm welcome to you. Um, and
1:13
if you haven't yet subscribed, make sure you do. Make sure you do all the like,
1:18
follow, and all the— what help— however you digest the podcast, uh, then do come
1:22
and join us. And of course, if you're part of our SAM Uh, network, which is our,
1:27
our AI system. You can now just talk to AI about the podcast. Guys, I don't know
1:32
if you— we'll get around to introductions in a second, but I don't know if
1:36
you've ever actually used AI to now listen to podcasts. Have you got that far
1:41
yet?
1:46
No, I'm, I mean, uh, I, I've used them to, uh, help me with transcriptions and
1:50
kind of cut to the— cut to like important parts, but I haven't used them to
1:54
directly listen to one yet. That sounds, uh, That's very forward thinking of
1:58
you, Matt.
2:02
Yeah, something like that. Uh, I, I, we're testing it at the moment. It's quite,
2:05
it's quite fascinating.
2:09
Anyway, I'm in the same boat. And that I'll have to use, I'll have to listen to
2:12
the Sam AI podcast. I'll have to like check it out. I have not tried that yet.
2:16
It's just, well, it's just fascinating using AI to listen to a podcast and to
2:20
help you digest it. Um, one day I'll talk about how we do that. Uh, but maybe
2:24
not today. Uh, and again, another first. So not only in the first studio, but
2:28
actually, no, this is the second. I was saying this is the second time where we
2:31
have had more than one guest on the show where we're talking to 2 people
2:35
simultaneously. Uh, so guys, why don't you introduce yourself? Give us a quick
2:39
sort of elevator pitch. Warm welcome, by the way, to the, the, the eCommerce
2:43
Podcast. We were talking that we did the initial pre-call for this like months
2:47
ago, and it's taken us a while to make diaries all connect. Um, but it's great
2:51
to have you guys.
2:55
Just give us a brief intro to who you are and what you do?
2:59
Yeah, I'll go first. My name is Dave McAnally. I'm— I kind of came to this
3:04
world. I've been in the SEO space for— Alan, it's gonna sound crazy to say this,
3:10
we're coming up on 30 years here. And I've seen a lot of changes. And, you know,
3:15
what I'm excited to talk to you today about, Matt, is just some of the impacts
3:21
of AI, AI search on ecommerce. I've spent a lot of time You know, I led
3:26
agencies, big agencies, SEO practices for, you know, about 15 years. And then
3:31
I've spent about the last 6 years, you know, consulting with companies.
3:37
Yeah.
3:38
A lot of them Shopify brands. You know, that's, that's a platform that's really,
3:43
really kind of made it easy for people to get involved and bring, bring, you
3:47
know, companies to life, but also some mass ecommerce programmes on, you know,
3:52
bigger, bigger platforms. But, you know, through all that, I guess what I've
3:57
really been focused on is making sure that, you know, those things are crawlable
4:02
and accessible and those products can be discovered. And it's— I'm excited to
4:07
talk to you about some of the things that, you know, especially with AI search,
4:11
like what, what I think that's going to do in Q4 this year that are things that
4:16
a lot of people, we've just never dealt with before.
4:21
Yeah, looking forward to it. And Alan, give us a brief lowdown on you, good sir.
4:26
Yeah, similar to Dave, we've been around in the space forever. You know, I've
4:31
worked for multiple of the global ad agency holding companies. In fact, that's
4:36
how Dave and I know each other. We first met where we worked together at one of
4:41
the holding companies. You know, Dave put a number on it. I wouldn't— almost 30
4:45
years. I wouldn't say 30 years, but we, you know, we both started before, before
4:50
Google when there was AltaVista, all these other search engines that seemed to—
4:55
I remember them well. I remember showing our age right now.
5:00
Yeah, exactly. And they obviously all consolidated around Google and a little
5:05
bit maybe, you know, Yahoo and Bing. And now we're in the the AI search age.
5:10
And—
5:10
Yeah.
5:10
But in the last couple years, you know, after having worked at a variety of,
5:15
I've started a few ad tech companies and agencies, worked at some of the global
5:19
holdcos. More recently in the last 3 years, I specifically consult in the retail
5:24
media ecommerce space, working with CEOs of retail media ad tech companies. So
5:28
I'm very much in this space and I advise one of the leading SEO AI Platforms
5:33
that started a couple years ago. There's probably, you know, there's kind of a
5:37
couple, 2 or 3 of the companies that were historically in the SEO space that
5:42
have moved into the AI SEO space. And then there was, there's probably a good 9
5:46
different heavily VC-funded, you know, tech players who are in this space now as
5:51
well that have entered in the last couple years with funding anywhere from $25
5:55
million. The largest one is $100 million of funding. So they—
6:00
Oh, wow.
6:01
The VCs are not, there's no secret in this space where VCs, you know, going from
6:05
search to social to programmatic to retail media and this new AI search space,
6:09
the VCs did not hesitate to throw money at some players to, 'cause it's such a
6:14
massive growing industry.
6:18
It is, and it feels a little bit like the Wild West at the moment. Like we're
6:22
still trying to figure it out, right? But we all also feel like, We've heard
6:27
that there's gold in them there hills. And so we're trying our best to like go,
6:31
why, what do I need to do? Because we've seen these kind of things before,
6:35
haven't we? We come in, it's like the early adopters, the guys that figure it
6:40
out, tend to do really well as a result, you know. And so I can see why they're
6:44
throwing hundreds of millions into trying to figure it out.
6:48
Right.
6:49
And sort of the AI search thing. So let me ask you guys, A question that I love
6:54
to ask all our guests, and Dave, you can go first. And I'm curious to also hear
6:59
what you have to say, Alan. Obviously, you both are in this space and you both
7:04
have insight and you're both playing and you're talking to people in this space.
7:09
Thinking about our audience, the guys, you know, doing ecommerce, what's the
7:13
biggest single mistake that we're making when it comes to this space? Why are we
7:18
making it? And I guess, how do we solve it? So no small question then, Dave.
7:23
Yeah.
7:28
Well, I'll kind of put it in the dimension of thinking about AI. And one of the
7:34
things that, you know, particularly with ecommerce, I think it's really easy to
7:39
fall into the trap of just thinking about the sale. You know, like, and I look
7:44
at the world through search and, you know, it's the buy the men's shoes, buy,
7:50
you know, a woman's dress and you want to get people, you know, on the site
7:55
right away. And, you know, I think one of the things AI has done and, you know,
8:01
it's still, you know, right now, you know, it's still early days, but this stuff
8:06
grows so exponentially. So I think it's important as we're— Yeah, it does. You
8:11
know, we're talking here in early September. It's entirely possible metrics
8:17
could be, you know, 200-300-fold by come holiday season. So yeah, one of the
8:22
things that I'm— I find myself talking to a lot of companies about is what AI is
8:27
doing is kind of compressing the sales funnel, which, you know, what I mean by
8:33
that is Right now, and kind of like Alan, I work with a couple platforms that
8:38
are measuring this stuff. But I just did a study with one, one of the bigger
8:43
players. And, you know, if you ask ChatGPT a question like, you know, what are
8:49
the best hiking boots for a rainy, rainy season? And I don't know if any
8:54
listeners sell hiking boots. 97% of the time, if you ask a question like that,
9:00
ChatGPT is going to give you brands. So you're going from that real high level,
9:05
just kind of exploring, learning about a product to a shortlist. And if you're
9:10
not on that list, you know, and this applies just to ecommerce, this applies to
9:16
B2B, applies to healthcare. But that compression is going to take us from, you
9:21
know, somebody who asks a question like that isn't going to Isn't going to ask
9:26
that question of, you know, buying men's hiking boots that do blah, blah, blah.
9:32
Like all that research is being compressed into a prompt conversation. So what
9:37
you need to do as an ecommerce provider is, you know, first and foremost, you
9:42
need to understand those questions people are asking and you need to build a
9:48
strategy for it. And I think where the— to get around to your question, Matt, I
9:53
think one of the bigger mistakes people make in that regard is, you know, and
9:58
this is true whether it's Google or ChatGPT or Perplexity, your website may not
10:04
be kind of where that answer comes from.
10:10
Yeah.
10:11
And one of the things I'm seeing is real advantageous for ecommerce providers is
10:14
taking advantage of things like YouTube. YouTube is the 2nd biggest source for
10:17
Gemini.
10:21
Yeah.
10:22
Right now for where it gets answers. So all those unboxing videos, how-to
10:26
tutorials, things like that, if you're not thinking about that as a real kind of
10:29
a lever to pull in your ecommerce strategy, I think that's a big mistake.
10:33
Because those are the things when people ask those high-level questions and your
10:37
brand's going to get shortlisted.
10:41
Yeah.
10:41
That more than your website, that's going to be where they're going. And so, you
10:46
know, there is You know, whether you use a technology or you're just kind of
10:51
doing the research manually, you need to find out what those, you know, cited
10:56
sources are when those questions are answered and treat those kinds of things as
11:01
like, almost kind of like the shopping list for people. Yeah, do— or AI will be
11:06
recommending the brands earlier. That's a lot different than search, you know,
11:11
where it's like, well, it is links and you kind of go through it now. Yeah.
11:16
Well, I, as As we're talking, I'm on my phone.
11:20
Oh boy.
11:20
I'm on Perplexity. And I said, what are the best men's hiking boots for rainy
11:24
days? Uh, which I think was—
11:27
Oh, I did.
11:27
Yeah.
11:28
Yeah. And so immediately it says, well, for a typical UK walk in a rainy day,
11:33
wet grass, muddy woodlands, fells, and shallow puddles. It's very descriptive,
11:39
isn't it? I mean, I'm there. I'm, I'm there. The, the Salomon X Ultra 5 Mild GTX
11:44
is the best all-round men's choice. Light enough for regular day hikes, properly
11:49
waterproof, supportive, and grippy on mixed ground. If you prefer a tougher
11:55
leather boot, try the Lowa Renegade Evo GTX. And so it's interesting, like you
12:00
say, it's, it's, it's, it's narrowed down my question to 2 choices. And it's
12:06
given me those, I mean, it's given me a very long reason as to why we have those
12:11
2 choices if I can be bothered to read it.
12:17
Well, yeah. And you know what's interesting about that is it's, you know,
12:21
whether you're kind of a startup, because everyone, you know, product marketing
12:25
is one of those things where if you don't have a product marketing department,
12:29
somebody, whether it's the founder or whatever, somebody's doing this in some
12:33
regard. Yeah, yeah, yeah. realise it or not. But what you just described there
12:37
is, I mean, we're talking about things that are differentiators for those
12:41
products. So—
12:45
Yeah.
12:45
For like massive ecommerce brands, these are things that, you know, I find
12:50
myself talking to, you know, teams about making sure that the product marketing
12:55
really kind of understands this stuff. Like what is AI? How is it discerning
12:59
kind of what is best. And that needs to be part of— it almost kind of needs to
13:04
be part of the product strategy.
13:08
Yeah.
13:09
Now, you bring that down to like, you know, somebody who's, you know, just
13:14
selling products on Shopify, you know, you can still kind of replicate that, you
13:19
know, that engine in your business. And I think it, you know, it becomes part
13:25
and parcel to understanding what makes your product unique. And When you're
13:30
doing things like writing descriptions, even as fundamental things as that, you
13:35
know, you can use one of the kind of life hacks is, you know, if you go to
13:40
Google and look at like those people also ask things like those little—
13:45
Yeah, yeah, yeah, yeah.
13:47
Those are good indicators into, 'cause, you know, especially with Gemini, those
13:51
are good indicators into what those follow-up questions AI anticipates. So make
13:55
sure you're factoring that into how you're going to write your product
13:59
descriptions, how you're going to organise your categories. That's a really good
14:03
one, especially if you're like on Shopify, as you're categorising your products.
14:08
Those are good hints to, you know, how AI is going to think about your space.
14:12
Yeah.
14:13
But yeah, you raised a really good point there. You know, that thing took us
14:16
from a really generic question to it knew where you lived, it knew kind of what
14:20
are the attributes about those products that are going to matter to a guy in the
14:24
UK on I can't remember the colorful language he used to describe the puddles,
14:27
but, uh, you know, that kind of stuff is—
14:31
Yeah, yeah, it's fascinating.
14:33
Yeah, I want to take the discussion into a, uh—
14:36
Oh, sorry, Alan.
14:37
I said I want to take a discussion into same discussion, but just one, but just
14:41
a different, um, direction a little bit. Going back to Matt's original question,
14:45
I just want to make sure, you know, people in the audience, they're doing
14:48
something, right? Because Dave and I know 100% a year ago when we were talking
14:52
to a lot of ecommerce brands, no one was doing anything. It was 1% of the
14:56
marketplace traffic or potential.
15:00
Yeah.
15:00
And even today, like more recently, Dave and I have got on the phone with a few
15:05
people and they're really not doing anything. So, Yeah. Greek first take would
15:10
be this, like first assess what the potential is, right? So do, you know, people
15:14
like Dave and others, you could find people to say like, okay, Depending on what
15:19
my product is, what is the opportunity? Do an assessment of where you stand with
15:24
the AI search results and say, what's the pocket of opportunity that I'm missing
15:29
today? But also assuming things keep going the AI search direction in the next 6
15:33
months, what's the opportunity cost of me not doing it?
15:38
Yeah.
15:39
Because using as an example, putting our age out there again, in the early days
15:43
of search, There were brands that were, you know, Davids were beating the
15:47
Goliaths, so to speak, and they were doing it on search because in the early
15:52
days it wasn't just all about money. And I don't think it's all about money yet.
15:56
So you can find like, and we can get into the details, I'm sure Dave can rip off
16:00
a top 10 list right now of things you could be doing right now as a smaller
16:04
player going up against a bigger player. But when you use that boot example, you
16:08
know, I didn't know the brands that were listed there, but I bet one of those
16:13
brands, there was probably 2 or 3 of the brands were the big, you know, known
16:17
brands. And I bet one of them was a niche brand that probably is, I would guess,
16:21
is probably a leader in AI search in terms of really taking advantage of it and
16:25
they're showing up with the bigger players. So like, I think, you know, to use
16:29
another analogy, a few years ago when TikTok first took off, you know, some
16:34
innovative ecommerce brands, you know, went all in on TikTok and they won
16:38
because that was their strategy and they won because of it. I'm not saying it
16:42
has to be your marketing strategy or your business strategy, but I would at
16:46
least, I would make sure I would do the audit or assessment of what the
16:50
opportunity is right now for your specific ecommerce brand.
16:54
Yeah, that's a fair point. I think a lot of people have, what's the phrase I
16:59
want to get? Just sort of defaulted into AI search, right? So if you're on
17:04
Shopify, Shopify's done a few things to try and help you.
17:09
Yes.
17:09
But then of course everybody who's on Shopify have now got access to the same
17:13
things. So how's it differentiating you? Um, it's the same thing when AI writes
17:17
product recommendations that you— in Shopify that you then use on your system.
17:21
Well, it's using the same system to write the same product recommendations for a
17:25
whole bunch of the similar brands, right? So how do you stand out? Um, I think
17:30
it's a really interesting question, uh, but I, I, I, I, I do think that most
17:34
people are sort of defaulting their way into AI search. So how do we How do we
17:38
be a bit more proactive in this space? Like, what are some of the really obvious
17:42
things that we can do instantly, regardless of the size that we are, that are
17:46
gonna, you know, like a little checklist, Dave, of things that we just need to
17:50
make sure we're doing?
17:54
Well, one of the things, um, and you mentioned shop and, you know, you use
17:59
Perplexity. So Perplexity has that Shop Like a Pro programme, and that's I do.
18:04
Last I knew, Shopify does automatically opt you into that. So like people can
18:09
buy products straight in Perplexity. If you have a Shopify store, there's a
18:14
partnership with that. So yeah, that's a good example of what you're talking
18:19
about, where some of the things— but to be proactive, I really, I think there's
18:24
2 things that really move the needle for people. Number one, whether it's your
18:29
category pages or your product pages. And this is like, you know, there's no
18:34
real secret sauce here, but just a frequently asked questions thing on those
18:39
product pages that just kind of mirrors some of those people also ask.
18:44
Yeah.
18:45
That makes it really easy for those AI agents to kind of answer those follow-ups
18:50
because AI search, you know, the way I've always described is they're not
18:55
answering the one question you ask. They're anticipating the next 4 or 5 that
18:59
you might ask. So it's condensing everything, getting you answers really quick.
19:04
If you signal that in your content versus like just a generic description that
19:09
uses some of the stuff that, you know, the marketing speak language, that is
19:14
going to make it easier for them to, you know, kind of corroborate what your
19:19
product does with like what people are reviewing it or saying. So just a really
19:24
good frequently asked questions that's oriented around how people are going to
19:29
use the thing. Because you notice that Perplexity shoe example, you could really
19:34
see how— and it's true, ChatGPT or Gemini or, you know, whatever, they are
19:39
really oriented around how that user would use the thing. So a frequently asked
19:44
questions thing is just making sure you have that on there is one of them.
19:49
Yeah.
19:50
I also recommend people— this gets a little more technical. Shopify does bake a
19:54
lot of this in, but You know, you're marking up using schema markup. There's a
19:59
lot of debate right now in the community about kind of how much of value this
20:04
is. But schema, for listeners who aren't aware, is a markup language. It's been
20:08
around for a while, but it basically gives crawlers, or in this case, AI agents,
20:13
kind of identifies what things are, you know. Out of the— right out of the gate,
20:18
if you have $9.99 on a site, a crawler may not be smart enough to understand,
20:22
okay, that's the price of this product. Yeah.
20:29
Products.
20:30
Yeah.
20:30
So this language was designed to, you know, help engines kind of understand
20:35
that. And you saw it being used by like results getting kind of richer with the
20:40
picture of the product. And, you know, you might see a price under. In the world
20:45
of AI, you know, those training agents, um, you know, Google not so much, but
20:49
some of the ones like ChatGPT and Perplexity where, you know, they don't have 25
20:54
years, you know, almost 30 years of, uh, crawler advancements fused in. Um,
20:59
those things can function as like hints to help those AI agents understand what
21:04
you're, you know, the attributes about your product. Um, Google recently, there
21:09
is one specifically for frequently asked questions. Um, I recommend you use
21:14
that. Google did say they don't read it anymore, but ChatGPT still does. And
21:18
ChatGPT's Training agents are— the last stat I read is like those represent
21:23
about a third of the AI agents out there.
21:28
Yeah.
21:28
OpenAI is like all over sites right now.
21:31
Yeah.
21:32
And that's a good hint for it. You just— your goal in that technical side is to
21:37
make it kind of roll out the red carpet for that agent, so to speak.
21:41
Yeah.
21:41
This is a good way to do it. But definitely the really thinking about the use
21:45
cases in how you describe your products, how people are going to use it. Because
21:49
that's what the AI is going to grab onto.
21:53
Um, well, it's interesting you say that because the— I've clicked through on
21:58
Perplexity to its sources. So why? Why? Because Perplexity is good at telling
22:02
you why, you know, it gives you the sources of its information. Uh, the first
22:07
site is a US site, bearing in mind I'm in the UK. Um, and it's a site with, you
22:12
know, the top 10 hiking boots.
22:17
Yeah.
22:18
Type thing and some detailed descriptions on there. And then I come to a UK site
22:22
and it's really fascinating the information that's on that page because it's a
22:27
very detailed product page, right? It's got like a verdict, like we give this
22:32
shoe a score of 87 out of 100 and it's given it different scores across
22:37
different things. It's got pros, it's got cons, it's got comparisons, it's got
22:42
the guy cutting the shoe open to have a look inside. It's got like waveform
22:46
charts, like you can't see these, but like, why would you have waveform charts
22:51
for walking boots? But it's AI loves data, right? And it's, I think they, these
22:56
guys have figured out a way to present information where, like you say, like the
23:01
likes of ChatGPT or Perplexity are going to look at that and go, this is
23:06
brilliant for us because it's got everything that we need. it's got all the
23:11
comparisons, it's got all the try this, try that, don't do this, this is not who
23:15
the shoe's for.
23:21
It's on one page. Well, you raise a good point there that, um, you know, I'm
23:25
assuming those sites you looked at weren't necessarily ones that the
23:30
manufacturers of those shoes would have created. They're third parties.
23:34
Yeah, all third-party sites. I don't even know if any— clicking through them, I
23:38
don't know if any of them are what I would call a traditional e-com site versus,
23:43
you know, like the, the sort of the media sites which like do the reviews and
23:47
click through with an affiliate link type thing.
23:51
Yeah. And there's, you know, if you think about it, it kind of makes sense. When
23:57
one of the studies I did, and I did this in e-commerce back in June, if you
24:02
compare what ChatGPT is doing like this summer versus last summer, They're
24:08
naming less brands in ecommerce on average. Like, I think you got— you're using
24:14
Perplexity, but, you know, imagine it's ChatGPT. I think you got 2, you got 2
24:19
shoes. In ChatGPT, they would have kind of given you like 10.
24:25
Yeah.
24:26
Although give you like 2 or 3. But what's different is they're citing about 3
24:30
times as many sources in terms of how they get to that answer.
24:36
Yeah.
24:36
So instead of just kind of taking it from the, those manufacturers, take them at
24:40
their word, they are looking to those review sites and third parties and
24:45
corroborating what's being said. And so for marketers, you know, there's no easy
24:49
way to say it, that does complicate things a little bit because it's not, you
24:53
know, to really kind of paint that picture, it's not as simple as You got to
24:57
kind of think beyond your site. Everything I'm saying is important. You know,
25:02
you want to make sure what you're describing resonates with, you know, what
25:06
third parties are going to say. But this does mean, like, for example, you had a
25:10
top 10 site there. You need to identify who those are. And, you know, if you're
25:15
lucky enough to have a PR department, you know, you're building relationships
25:19
with those people because Alan brought up the point that, you know, it's kind of
25:23
the playing field's leveled a little bit.
25:27
Yeah.
25:28
And that means that you're not necessarily having to get your shoe reviewed in
25:32
Runner's World to be recommended by Perplexity. But you may need to build a
25:36
relationship with a blogger or, you know, some of these sites and make sure
25:40
that, you know, they're testing your shoe.
25:44
Yeah.
25:45
I'll give you an example this spring. And I'll name the name of the brand
25:50
because it's, you know, you guys could look it up. Adidas, one of the most
25:55
common ways people— one of the most common behaviours that they have a unique
26:00
return policy. And Single Bit, when people ask Google about, you know, what is
26:05
the return policy for the shoes, this is true this spring, I haven't done this
26:10
search, you know, recently. But what we were seeing was that it wasn't where
26:15
they were getting most of their answer, how they were answering that. Came from
26:20
a YouTube channel with a woman who had about 6,000 subscribers. So, you know, it
26:25
wasn't nothing, but it wasn't a, you know, huge—
26:30
Yeah, it wasn't a big channel.
26:31
Yeah. And that was driving the entire brand. That was basically, if you look at
26:35
where it was being cited in the answer, it was the biggest way for Adidas to,
26:40
you know, she was describing kind of what you got to do if you want to return
26:45
shoes online and things like that. And it was that site. Now Adidas, You know,
26:49
wasn't even mentioned in that brand. You know, like none of their own properties
26:54
were even part of that answer. Yeah, it was this, this influencer from this
26:58
girl. There's a Reddit thread and a few other things. So to Alan's point, that's
27:03
a big, that's a pretty big sea change in how you might think about how Adidas
27:08
might be thinking about, you know, their brand reputation or you know, how
27:12
products are sold. So what I would do in their case is, you know, if that girl
27:17
was saying something that wasn't accurate, which I don't believe she was, um,
27:21
but their PR needs to be looking at some of those smaller players and how well
27:26
they're represented. You know, Alan mentioned, you know, he works with a
27:30
platform that, uh, you know, there's technologies now that identify all this
27:35
stuff, you know, and then they can take your brand and say, hey, here's the
27:40
biggest influencers.
27:44
Yeah.
27:45
And to your point, Matt, um, essentially what the platform does— Oh, sorry,
27:48
what's that?
27:51
Yeah, I was just— essentially what the platform does is like instead of
27:54
keywords, it's looking at questions or queries, right? And it's, you know, for
27:58
any one brand or product, it's looking at the potential thousands of queries
28:02
that someone could be putting into the AI search engines, and then it's finding
28:06
some of the higher volume queries And in those queries, it's doing exactly what
28:10
you did, Matt. It's dissecting, well, where, what are the sources from those
28:14
queries? Where's the AI engine? And what you're finding in Dave's example is a
28:18
perfect example is what you're finding is they're not all necessarily the big
28:22
publishers or the big blog, you know what I mean? You're gonna like, so
28:25
sometimes, you know, whether, you know, you're finding these smaller, you know,
28:29
YouTube channels like that person you mentioned and your PR agent. And so what
28:33
companies are doing, quite honestly, And this is like an advantage that maybe
28:37
it'll go away at some point, but you go, you can, you know, your PR agency can
28:41
actually go and see if you can participate or like some of the brands are taking
28:45
their advertising or their media budgets and they're doing paid media budgets
28:49
towards some of these smaller niche brands that are getting a lot of juice from
28:53
the search engines. So, but it's very, very granular. Like, you know, you have
28:56
to, you know, you do that with like, 20, 30, 50, 100 sites that increases the
29:00
overall volume you're getting on your ecommerce brand. And that's where like the
29:04
devil's in the details where, you know, you can use another platform or a
29:08
consultant or a combination of both to get that incremental lift as a smaller
29:12
brand.
29:16
Yeah.
29:17
Yeah.
29:18
No, see, I remember years ago, I mean, obviously before AI, just when, but going
29:23
back to sort of how this all worked, We, um, we had a beauty company and we
29:28
looked at the source of traffic and one of the biggest sources of sales traffic,
29:33
so not just traffic to the site, but traffic that converted to sales came from a
29:38
Swedish blogger, a Swedish blogger, right? I mean, no one predicted this. And so
29:43
we reached out to that Swedish blogger and said, listen, would you like to, you
29:48
know, make this kind of like official? And they were like, no, no, it's fine.
29:52
I'm like, are you sure? Because on affiliate commission, you'd make about 6
29:57
grand a month, right? And she's like, no, no, it's totally fine. I just, I just
30:02
like you guys and what you're doing. If I can help, then it's great. We were
30:07
like, stone the crows. It was incredible, really. Just a small Swedish blogger.
30:12
And it changed in some respects the fortunes of, of the company that we had. And
30:17
so I'm a big fan of going after these um, smaller source— smaller in inverted
30:22
commas, do you know what I mean? These sort of smaller, uh, fans or people that
30:27
are blogging or writing about things. Because yeah, you just— how do you— but
30:32
how do you find, like, for AI, how do you find out where ChatGPT or Perplexity
30:37
is going? Is that a laborious manual process?
30:42
So there—
30:43
is there something you can put into ChatGPT to tell you? I'm kind of curious,
30:45
how do we do that?
30:48
Well, you saw in Perplexity there, I think you must have looked at the cited
30:53
sources and it showed you where it's getting that. Yeah, that same dynamic
30:57
exists across Gemini, ChatGPT, you know, Perplexity, whichever. There are
31:02
platforms and there's a whole cottage industry of technologies, some, you know,
31:06
like that have been around for a while and some like what Alan was describing
31:11
that have kind of come up. And they do kind of aggregate and streamline that
31:16
process. So rather than having to do it manually like you just did, you know,
31:20
you could see all those prompts and, hey, here's what's being cited, here's how
31:25
often they're— here's our questions they're answering. Um, you know, are they
31:30
saying good things or bad things about your product? Because sentiment, you
31:34
know, the AI is a little more 3-dimensional than search because now you have
31:39
It's like, it's, well, the AI has an opinion about brands.
31:44
Yeah, yeah.
31:45
Triangulating where those things are coming from. The technology will do that.
31:50
So, you know, I, and the good news is that industry is getting to the point to
31:55
where there's kind of a something for everybody. $99 a month things that will
32:00
kind of just give you the data on up to like full service. operating models for
32:05
teams type stuff that, you know, is more enterprise grade. But that's probably
32:09
something that's probably going to be part of a tech stack ecommerce brands will
32:14
need to think more about. I think, you know, before we got— just when we joined,
32:19
I think one of the things that'll be very different this quarter is— or Q4, you
32:24
know, AI overviews, which is probably the most common way people interact With
32:29
AI.
32:34
Yeah.
32:35
You know, Google's 90% of the share. Last year, Google really left ecommerce
32:39
alone with those. It was about 9% of ecommerce queries had that AI overview and
32:44
those shopping carousels and things like that. This year, it's going to be about
32:49
25%. So 1 in 4 ecommerce searches are going to have that AI in there. So, you
32:53
know, I think what does that mean for marketers? That's where You know, kind of
32:58
regardless if you're a big— I know not everyone is a big— the holiday season is
33:03
huge for them. But if it is for you, this is probably something you want to
33:08
explore, you know, make technology makers to see this stuff. Because probably as
33:12
you can imagine, Matt, if you went to every single little prompt and looked at
33:17
the sources, it would be laborious. So there are technologies that simplify
33:22
that. And there's— they're getting more and more to where there's something that
33:27
scales for every provider or every, every kind of size of business.
33:31
Yeah.
33:33
Yeah.
33:33
Have you got anything to throw in here, Alan?
33:35
Yeah, I think, um, uh, I would just add, um, it's, it's definitely working, as
33:39
Dave mentioned. Like, I think as a strategy, you need to decide, you know, is
33:44
your PR agency going to do this? Are you going to do it internally? Are you
33:48
going to use technology? And then just as I mentioned earlier, the overall
33:53
assessment of You know, how much of your budget and resources are gonna go
33:57
towards AI search? So, you know, if you haven't done that assessment, I think
34:01
it's important because it's not only is it not going away, it seems like search
34:06
is, SEO specifically is morphing into this industry. So I think it's a good time
34:10
to get ahead of it. And—
34:15
Well, how much budget do you think we should have? Because this will be one of
34:18
the questions people ask. It's like, well, I'm listening to you guys talk. If I,
34:22
you know, I'm just sort of getting my head around Meta ads and Google. Now I've
34:26
got to think about AI search. What sort of budget? I mean, I appreciate I'm
34:29
asking how long's a piece of string in many ways, but how would you determine,
34:33
maybe is a better question, how would you determine what a good budget amount to
34:36
throw in here is?
34:40
I mean, you could do it straight from an ROI assessment. I would first get an
34:44
assessment of like what, you know, find someone like Dave or someone else
34:48
internally who has the capability of doing this and find You know, basically get
34:52
assessment, okay, in the next year, what percentage of my ecommerce sales come
34:57
from, you know, come from AI search today and what do I expect it to be a year
35:01
from now? And like, what, like, you know, someone do an audit or an assessment
35:05
of like, what's my opportunity to grab more market share in the AI search
35:09
engines? And, you know, do a 3, 6, 12-month opportunity analysis, see how much,
35:13
you know, someone's willing to say that they can, you know, grab from a market
35:17
share or revenue standpoint and then, you know, put there, you know, and then
35:21
make them own it. You know, it's not much different than what you used to do
35:26
with an SEO consultant or, you know, an SEO agency.
35:30
Yeah, yeah, yeah. You know, that's what I suppose the tricky question though is,
35:34
how do we know how much search is gonna come from AI in 12 months' time, right?
35:38
I, it feels a little bit like, You know, throwing a dart at a dartboard and it's
35:43
just going to stick somewhere. And that's probably a right answer because we've
35:47
got some ideas and it's getting clearer, but it's still— I mean, David, I mean,
35:52
you're, you know, you're probably James Bond in this kind of thing with all
35:56
these secrets, tech and stuff. But it's kind of like, how did— how would we
36:00
know, right, that actually we can look at our site and we can go, well, listen,
36:05
I've had 10 sales this month from AI search. In 3 to 6 months' time, that's
36:09
going to be X amount of sales. How could I predict that with any kind of re— I
36:13
mean, I don't know if I could predict that with any kind of reasonable success,
36:18
could I?
36:22
Well, you can build a trend line. And one of the ways— so one of the cool things
36:28
Google did with GA4, Google Analytics, is you have a source medium that will
36:34
show— you used you know, you used to be able to filter all your, your, uh, your
36:40
charts and stuff in there by, um, like organic or direct. Yeah, you know, they
36:45
now have one called AI Assistant, and they launched that in May. That includes
36:51
Perplexity, uh, OpenAI. Um, it doesn't break out the AI overviews yet. Um,
36:57
that's coming though, and you can see that in Search Console, but For what we're
37:03
talking about, one of the things I like to do is so things like your, you know,
37:09
you can go into your ecommerce, you know, component of GA4, and you can filter
37:14
by that and you can see, you know, the trend of what that growth looks like.
37:20
Yeah.
37:21
That's one way you could forecast it. And that's probably the most common thing.
37:25
The other thing I like to look at too is because I think this is What really
37:30
kind of tells the story, and I've seen this, you know, you mentioned your
37:34
Perplexity example, you know, this thing went beyond just kind of recommending a
37:39
shoe for the sake of, hey, here's a brand. It really told you, you know,
37:43
corroborating from all these sources, if you're going to be doing this kind of
37:47
hiking, you want this because it does this and this and this one over here. What
37:52
that's translating to, and I've seen this across a lot of different sites, is
37:56
when people do convert directly from ecommerce, the conversion rates are a lot
38:01
higher.
38:05
Yeah, yeah, yeah, that we've noticed.
38:09
The engagement rates, people are coming to this, you know, you come from an
38:14
experience like that versus, you know, kind of cold off the street. Yeah, yeah,
38:19
you do get a more, you know, person ready to buy now. I guess since we're
38:24
talking about investment here, I would say assess that, get a sense of what the
38:29
conversions are. OpenAI has now started an advertising programme where you can
38:34
start to show your ads. And it's pretty cool how it does it because it's, you
38:39
know, kind of around the product, you know, putting a product marketing hat on.
38:44
When people are asking those kind of, what's a good hiking boot type of things?
38:49
Yeah, yeah, yeah.
38:49
You can get your product in there.
38:52
Right.
38:52
Assuming that it kind of satisfies all the elements of it, you can kind of, you
38:56
know, you can have at least have an ad show up. So I would say if you want to
39:01
test that kind of thing, get a sense of what the conversion rates organically
39:06
are right now, how much you could expect, and, you know, start small and go from
39:11
there.
39:15
Yeah.
39:17
The, um, you'll at least get a sense of kind of what the traffic looks like from
39:20
that AI filter.
39:22
Yeah, I think I said, yeah, I read yesterday, I think OpenAI had just hit $1
39:27
billion in the ad revenue for that product right there. So it's scaling quickly.
39:32
So actually, yeah, it is another topic of like, you know, kind of like in the
39:36
early days of, um, SEO, you know, it's basically finding your mixture of, or
39:41
using paid search in this case, or using AI search, paid search.
39:45
Yeah.
39:46
To maybe start to inform some of your other organic strategies too.
39:51
Yeah. That's an interesting thing.
39:53
I think it popped up like 6 months ago and it's already at a billion dollars
39:55
though.
39:58
So it's crazy, isn't it, how quickly it's growing?
40:01
Yeah.
40:02
Have you guys tested ads on OpenAI's platform?
40:08
Yeah, I've done it for a golf like golf accessory brand site.
40:13
How did you find it?
40:14
The conversions are about in line with what you see organically. I think the,
40:20
you know, the UX, it's still kind of, you know, it's early days. You're a little
40:26
limited in kind of how you control what prompts things appear.
40:32
Yeah.
40:32
But the cost, you know, it's not It's, it's more kind of my experience, at
40:37
least, it's more kind of like Bing than it is Google in terms of the cost per
40:41
interaction and stuff like that. So it's, it's, it's a pretty cheap thing to
40:46
experiment with right now.
40:50
Yeah.
40:51
You know, I just have—
40:52
yeah. I don't know, Matt, I guess for your audience, look up because it's
40:56
available in the US.
41:00
I don't know if you I think there are limitations to OpenAI's ad platform.
41:05
That's, that's what we've discovered playing around. So I think it'll get rolled
41:09
out as these things always do. And you just have to keep checking and testing.
41:13
But I think if you're in the States, definitely, you should be, you should
41:18
definitely be testing that. Because it's never going to be this cheap. I
41:22
wouldn't have thought.
41:27
Yeah.
41:28
Well, going back to the early days of Google, what a level playing field that
41:31
was in terms of, you know, what a cost per click would cost versus, you know, a
41:34
big CPM buy on TV.
41:37
Yeah, yeah, yeah. I remember when Google cost per click were cheap, uh, and now
41:42
you have to mortgage your house every time you get one. But that's, you know,
41:47
it's a different story. It's interesting though, isn't it? Because I, you know,
41:52
Google is obviously pivoting with its AI overviews and with Gemini Um, I think
41:57
organic search is down because we're now using AI. So Google seems to be losing
42:02
that space a little bit, trying to get it in the AI platform. Um, OpenAI seems
42:06
to have done a good job with it. I don't know how well Anthropic's doing with
42:11
it. I wonder if they're going more after the coders than— I don't know. It's
42:16
kind of, it's an interesting field, isn't it? Sort of, like I say, the Wild West
42:21
a little bit. Who do you see emerging as some key winners here?
42:28
Well, you know, I do say I think rumors of Google's death are greatly
42:31
exaggerated. They've had a very—
42:34
It's gonna take a long time to kill Google.
42:37
Well, you know, it's still 90% of the search market share. Google's AI share
42:42
with things like Gemini, the deal they cut with Apple to power, you know, Siri.
42:47
I think some of the moves they made with Gemini, You remember there was, uh, the
42:52
code red that OpenAI had in December because Gemini was so good. Um, I don't
42:57
think I would rule them. I, I think Google is doing actually really well in this
43:02
space. You have to remember too, um, OpenAI is having a, you know, the, the
43:07
training, the training agents that they've got are essentially having to play
43:12
catch-up to—
43:17
Yeah, yeah.
43:17
Google's been building over You know, 20-some years. So Google has got a lot
43:22
more inherent knowledge to build off of. Um, I could tell you just from a
43:26
technical standpoint, Google, you know, you know, those agents tend to be able
43:30
to handle a lot more of the complexities of the web better than OpenAI does. Um,
43:35
yeah, you know, JavaScript and things like that. And that matters because that's
43:39
going to be more access. That means that Google is going to be able to get to
43:43
data quicker, you know, and so forth. AI overviews, they were— I think it turned
43:48
out to be the right move. They did roll those out. They scaled them really
43:52
quickly, believe it or not, in healthcare.
43:56
Yeah.
43:57
Healthcare and insurance, which I wouldn't have thought. And they rolled them
44:02
out slower in ecommerce and more transactional things like travel and
44:06
restaurants, like restaurant searches, local things. So I guess to go to your
44:10
point, Matt, I don't I still look at Google as still going to be the king. They,
44:14
they, yeah, yeah. I think they made a lot of really smart moves. They were
44:19
deliberate. They didn't, they kind of took some wait-and-see moments. I think
44:23
Apple still kind of remains to be seen. I see that they're aging out there on
44:27
some of the log files I look at. I don't, I don't know kind of what they're
44:32
gonna do. But you know, yeah.
44:36
Definitely.
44:38
some proprietary, you know, activity going on. Yeah, I think your point about
44:43
Claude, um, I think their user base is definitely more B2B. Um, for e-commerce
44:48
providers, you know, what kind of role that's going to play kind of remains to
44:53
be seen. Um, but I do think, you know, the, the, you know, definitely Google and
44:58
OpenAI are kind of the two leading right now. But I do think—
45:03
Yeah.
45:04
You know, a year ago, it, you know, it seemed like there was all this talk of
45:09
Google was, you know, finally going to get overthrown. And you're right, clicks
45:13
are down about 20%. But, you know, they're answering a lot of questions, they're
45:18
sending a lot more qualified visitors. Yeah. So they're, they've improved the
45:23
traffic, they've improved the quality of the traffic that people are getting.
45:28
Yeah.
45:30
It remains to be seen how their monetization strategy is going to change on
45:32
that. Um, but yeah.
45:35
Oh, they'll find a way. We know they'll find a way. That, that we can rest
45:39
assured on. Yeah, I don't have to worry about Google making money.
45:42
I guess that's a long-winded way of me saying I still think Google's doing—
45:48
yeah, the only, uh, the only tricky part, and this is more of a question than an
45:51
answer, is, you know, when You know, a long time ago when you switched from
45:55
using Bing or Yahoo to Google, it was basically just opening up a browser and
45:59
just typing in Google as opposed to Bing.
46:02
Yeah.
46:04
As opposed to now, like, you know, I actually, just because I started in this
46:07
direction, I'm more of a ChatGPT user myself for a long time.
46:11
Yeah.
46:11
And the switching costs are a little bit higher than they were just from jumping
46:15
from one search engine to another. Having said that, Google had, as Matt said,
46:20
Google figured out how to get tentacles around me eventually and move me with
46:24
ChatGPT. If I had to buy one of the stocks, I think, well, Google, I mean,
46:28
you've seen it this year, Google stock has gone up a lot.
46:32
Yeah, yeah, you're definitely buying Google. I think if anyone's gonna collapse
46:35
first, it'll be OpenAI. That's my personal prediction. But I, you know, what do
46:38
I know? I'm not a stock market expert.
46:41
The market, Matt, the market's moving this morning as we're recording this,
46:44
right?
46:46
Yeah, yeah. It's Yeah, it's gonna be one of those. Guys, listen, I'm aware of
46:50
time. It seems to have flown by. And Dave, I know you shoot off. Is just in
46:55
closing, any final thoughts from you two before I close out the show? And also
46:59
just let people know how they reach you, how they connect with you if they've
47:04
got some questions and all that sort of stuff. What's the best way to do that?
47:10
Well, for me, find me on LinkedIn. Just Dave McAnally, my name. That's probably
47:14
the best place to connect with me.
47:22
The—
47:23
but I would say my final thought for ecommerce providers is really, you know,
47:27
think about the product marketing questions, the pain points your stuff solves,
47:32
and really drive your descriptions with that. I think if anything, OpenAI is
47:37
really making product marketing and mission critical. The mark— no offense to
47:42
the brand marketers, and but market speak and fluff and all that kind of stuff
47:47
when you're marketing to an AI agent doesn't really have the— we're not really
47:52
marketing on people's emotions when they're buying. Yeah. So that's really
47:57
critical. Don't ignore video. Unboxing videos this holiday season. I saw those
48:02
hugely influential in how AI you know, recommended products. Um, if you're not
48:07
doing that kind of stuff, this might be the year to think about getting into
48:11
that.
48:16
Yeah.
48:17
You know, on YouTube, um, because that, like I said, that's the second most
48:21
cited source on Google. Bounces back and forth between Reddit and, uh, YouTube.
48:25
So, uh, those are the two things I recommend people think about, is the product,
48:29
really getting into the pain points and the product marketing, and what, what
48:33
does your thing do that solves, you know, and orient your descriptions, and then
48:36
You know, video is absolutely part of, you know, what your sales strategy needs
48:40
to be.
48:44
Fantastic. Uh, Alan?
48:47
Yeah, same as Dave. You find me on LinkedIn. Um, glad to connect. Uh, and I
48:51
would say for the, the small to mid-sized ecommerce brands, find those niche
48:55
questions that people are asking on the AI engines where you can, you can own or
49:00
dominate those, uh, questions. Um, using as an example is David mentioned
49:04
earlier the 6,000 unique visitor YouTube site that is one of the sources for one
49:09
of the main questions for that product. So there are plenty of those out there.
49:13
There's niches that you're not finding or not exploring right now that will
49:17
enable you to beat the smaller guys, just like in the early days of SEO where
49:22
some of the smaller players could dominate. And then for the larger ecommerce
49:26
brands, just the opposite. Defend your turf, make sure you're defending your
49:31
turf. And, uh, not allowing the small to mid-sized players, you know, steal some
49:35
incremental, uh, sales or market share. Um, yeah, using the AI engines.
49:39
I like that. Defend your turf.
49:42
Yes. Um, it is what it is, you know.
49:47
Uh, it is what it is.
49:47
Playing field's gotten level and—
49:49
Yeah.
49:51
But these things are just always interesting opportunities, aren't they? And
49:55
you, you either, you either embrace it now or you Ride it out for six months and
49:59
go. Let's see where the dust settles. Depends on your personality. I'm the kind
50:02
of guy that goes, right? Let's just go and have a go and see what happens. But
50:06
Dave Allen, thank you so much for joining us, man. Really appreciate it.
50:10
Thank you much.
50:12
Thanks a lot. Take care, Matt.
50:14
Well, listen, you too, Dave. Thank you for joining us. And of course, if you've
50:18
been listening to the show, make sure you like and subscribe and do all of that
50:21
good stuff because we've got some more. More great conversations coming up, and
50:24
I don't want you to miss any fun— any of them. But that's it this week from me.
50:27
That's it from Dave. That's it from Alan. Thank you so much for joining us
50:31
wherever you are in the world. Have a great week. I'll see you next time.
50:34
Bye for now.