Latest / The Prompt / Ep. 562 | AI in B2B advertising: Craig Miller (AWS) on what’s next for marketers

Transcript

Ep. 562 | AI in B2B advertising: Craig Miller (AWS) on what’s next for marketers · The Prompt

0:00
Think about search. Traditionally, I would have gone
0:02
into Google, typed a search URL, and then seen a series of links
0:07
ranked, boosted, sponsored or otherwise, and I would engage in
0:11
that content. Today, most of us look just at
0:13
the AI synthesized search results, the AI powered result,
0:17
and that's actually reducing our visitation to the end publisher
0:22
sites who produce some of that content that ultimately rolls up
0:26
into the AI results itself. The channels by which an
0:30
advertiser or marketer can actually reach that consumer and
0:34
influence them to engage with their services, their products,
0:38
their content is starting to shift.
0:40
We've got to both rely on that traditional mix of search of
0:45
social of direct publisher or larger relationships through
0:49
programmatic at scale and now we've got to start to influence
0:53
the AI at the point of reasoning or synthesis of the of response.
0:59
The sooner you develop that practice, the earlier that
1:02
you'll be able to achieve those outcomes and benefit from the AI
1:06
that that exists even in the current state.
1:09
You'll also be ready as the new era, as we see some of that
1:13
nebulosity start to take greater shape, you'll be ready to engage
1:17
in those agentic related channels.
1:20
You'll be able to engage in new formats all through getting your
1:23
data strategy and foundations right today and then developing
1:27
your practices. So get going.
1:30
The winners of tomorrow are actually developing today.
1:34
The question starts to become as a marketer, how do I ensure that
1:38
the data that I want to influence the reasoning is
1:42
available to the AI models at the point of decision that'll
1:47
then influence the consumer itself?
1:49
And or if it's the agent making a decision on behalf of the
1:53
consumer, how do I ensure that my data is accessible to that
1:57
agent to have it make the right decision that I wanted to
2:00
influence as the marketer advertiser, this is on base your
2:06
one stop shop for solving your most pressing B2B go to market
2:10
challenges. Welcome to the On Base podcast.
2:15
I'm Gabe Roble, CEO of Demand Base, and I am here to host a
2:19
very important episode of the podcast because we're going to
2:22
focus on the future of advertising and the impact AI is
2:27
having on it. And it's important because in my
2:30
opinion, digital advertising is going to be disrupted more than
2:35
it ever has in the next few years.
2:36
And we know digital advertising is almost constantly being
2:39
disrupted. But really, LLM's Jen AI, we all
2:45
know how much it's changing the way we work, but it's really
2:49
calling some fundamental things into question in the ad world.
2:52
Notably, what does context mean, how our audience is discovered?
2:57
Does SEO and SEM even matter? And these are things that we've
3:02
all relied on to drive traffic to our websites.
3:06
And so there are a lot of questions to answer and, and
3:08
that's really going to be the focus of the podcast today.
3:10
Future of advertising. We're also going to hit on AI
3:14
use cases that are making an impact now because there's a lot
3:17
of hype. But what's working right now and
3:19
the importance of the data strategy because really data is
3:23
at the center of all this. And as my guest, I am very
3:27
honored and excited to have Craig Miller, who is the global
3:31
industry and solutions lead for advertising and marketing at
3:35
AWS. He's got decades of experience
3:37
from building the first third party ad server to leading
3:41
innovation at amazing companies, innovative companies, App,
3:44
Nexus, Xander and now AWS. He really brings, I think, of
3:49
what's going to be a valuable perspective on how data, AI and
3:52
marketing technology are converging to shape the next
3:55
chapter of digital advertising to help us answer or at least
3:59
put us in the right direction. It's really big, big questions
4:01
going on. Welcome, Craig.
4:03
It's great to have you. Thanks for having me, Gabe.
4:05
Super excited to be here and excited to talk about the topic
4:09
of the future of advertising and the how it's going to transform
4:12
with AI. So thanks for having me.
4:15
Yeah, yeah. I'm really excited about it.
4:17
I guess a good place to start. You know, we talked about your
4:20
amazing career and, and all that you've seen, you know, digital
4:24
advertising, man, as it evolved in the last 20 or 25 years and
4:28
you've had a front row seat looking back at it and
4:31
everything you've seen. What are some of the lessons
4:34
that you've learned through the past transformations and how do
4:38
you see that carrying into, you know, the AI world that we're
4:41
moving into? It's been a really exciting
4:44
evolution to watch in advertising and marketing over
4:48
the course of my career. Given that we've seen rapid pace
4:52
of innovation during that time, starting from really direct, I
4:56
would say publisher advertising into ad networks on to
4:59
programmatic advertising, social search, and now this period, I
5:03
think of this new chapter of ultimately AI.
5:07
One of the lessons that I think I've really learned in watching
5:10
that evolution is that ultimately the age-old goals of
5:15
marketers and advertisers truly remain the same.
5:18
We actually have to influence customers to engage with
5:21
products and services, right? The techniques and
5:23
sophistication by which we achieve those goals has really
5:28
grown exponentially over that period.
5:30
Our ability to reach those consumers, to manage data at
5:34
scale, to effectively decision through initially machine
5:39
learning and now through AI, to use that data to empower
5:44
intelligent outcomes in acquiring, retaining, upselling
5:48
and ultimately satisfying or supporting those customers has
5:53
been ongoing in this constant evolution.
5:56
But those goals really don't have not changed.
5:59
And so when I look at sort of the AI driven error and the
6:03
opportunities that it presents to ultimately achieve those
6:06
goals more efficiently using new formats and ultimately an
6:11
increasing number of channels like agents.
6:14
I still think that some of those lessons of sticking to the
6:18
age-old goals and powering outcomes that matter to
6:23
advertisers and marketers alike, that the solutions that offer
6:27
meaningful outcomes will be the winning solutions in supporting
6:31
those needs. So I would say stick to the
6:34
age-old goals, apply the latest in technology and innovation to
6:40
produce those outcomes and more meaningful and more efficient
6:43
ways, but don't lose sight of the goals.
6:47
Yeah, keep. Your eye on the ball, you know,
6:49
with all the bell, the the crazy bells and whistles and, and hype
6:52
out there. And I think that's really
6:54
valuable advice to focus on outcomes.
6:57
I'm pretty hopeful, although so much is still to be worked out,
7:00
that the innovation is going to enable marketers to achieve
7:05
outcomes much more directly and inefficiently than they ever
7:08
have. And, and I think that's really
7:09
the upside. And, and the excitement is just
7:12
kind of figuring out how to, how do we get there.
7:14
And I guess we, we get there by starting where we are now.
7:19
And you know, I know you're working with with so many
7:22
marketers or around using AI today to make a difference.
7:28
We also know that there's a lot of hype and we're probably in
7:31
now in the hype cycle. What do you see the best
7:34
marketers are are doing? How, what are what are the
7:37
specific use cases that are making an impact and driving
7:41
outcomes today? Advertisers and marketers are
7:44
using AI effectively today, right?
7:47
And really, I think I've looked at it as kind of an evolution on
7:51
top of the way they've done things to date.
7:54
So, you know, when we think about innovation, I don't want
7:57
to call it hype. We'll look at at AI in its next
8:01
chapter and, and really at the forefront of the pioneering in
8:04
this AI chapter. I think it's more than hype.
8:08
It's actually unsettled. The shape of the future isn't
8:12
totally determined, but what we ultimately know is it will
8:16
diversify and enrich. Well, let's look at today to
8:19
your point, right? I think advertising and
8:21
marketing, you know, customers today or leveraging AI to
8:27
increase the operational efficiencies around the
8:31
advertising and marketing that they're doing current.
8:34
And they're doing that through, you know, empowering ultimately
8:38
better outcomes through personalization techniques,
8:42
through better segmentation and targeting you, better enrichment
8:46
of data and the ability to integrate, ingest and then make
8:49
use of or activate that data within the advertising and
8:53
marketing channels that they'll reach consumers in, in the
8:56
current world. And when they start to look to
9:00
the next chapter, they'll have an increasing number of channels
9:04
by which to engage those consumers.
9:06
So the high, if there is any and I, I'm not going to call it
9:09
that, is really around this next chapter, right, of having agents
9:14
ultimately help make some of the decisions on behalf of consumers
9:19
and ultimately synthesize data to put that in front of them to
9:23
help with the decision in the discovery of those products and
9:27
services. So part of the challenge is
9:30
really how does that data become actionable at the point of
9:34
reasoning where the consumer and the agent acting on behalf of
9:39
the consumer will actually discover those products and
9:42
services. So today, great usage of AI and
9:45
personalization, really enhancing and improving the
9:50
efficiencies around the current state-of-the-art advertising.
9:54
And the the great new frontier is much greater automation on
9:58
behalf of the consumer itself in making those decisions and in in
10:02
discovering, you know, the products and services.
10:06
Hopefully that resonates. Yeah, Yeah, it does.
10:08
No. And I, I think the distinction
10:10
between, you know, hype and unsettled is a good one because
10:14
I think we all know that there's a, there, there, it's just where
10:19
are you seeing the tangible outcomes and where are we still
10:21
determining how to get them? And I think what you're saying
10:24
is in the current state, we're seeing a lot of, of great
10:27
success with efficiencies around personalization, which is a
10:33
powerful way to, to achieve that age-old goal of, of advertising
10:37
to get your audience enriching of data.
10:40
The, the, the segmentation, getting to the, you know, the,
10:43
the recommendations of what channel do you use?
10:45
So that whole flow of activating the data to the right audience,
10:49
to the right channel and personalizing, we're seeing a
10:52
lot of great AI use cases where teams are getting much more,
10:56
more efficient. Where do you see teams getting
11:00
held back? Right?
11:01
Because we, I think with our customer base, we see the same
11:05
thing. There's a a lot of our marketer
11:08
partners that are making their processes 100, you know, so much
11:12
more efficient than they were before.
11:14
AI and others are struggling more.
11:17
What do you think some of the key blockers are that teams can
11:20
overcome to get to see those benefits that that kind of best
11:24
in class we're currently seeing? Yeah.
11:26
I mean, I, I think, you know, we see a couple of different
11:29
classes of what I'm going to consider to be stumbling block,
11:32
right. 1 is really about data itself and be able to have the
11:37
right data to ultimately power the reasoning or decisioning
11:41
that AI ultimately, you know, wants to enable on behalf of the
11:45
advertiser or marketer. And so overcoming that challenge
11:49
is difficult for some organizations.
11:52
Some really have a a deep understanding of a consumer or
11:56
their targeted audience in profiles and they own that
11:59
directly and or they may work with an integrated stack that
12:03
can really help them achieve that goal of unifying the data,
12:08
enriching the data in order to make it AI ready, right?
12:13
That's one form of stumbling block.
12:14
It's not that every organization has that readiness today.
12:19
And I think, you know, there's great ways to overcome it.
12:21
We could talk about some of that later.
12:23
Other stumble lock. Stumbling blocks are really what
12:26
I'm going to call organizational readiness.
12:28
Is there the knowledge and experience or governance in
12:33
place to really enable the use of that data and AI itself?
12:38
And not every organization has really invested deeply in that
12:41
knowledge and experience and developed ultimately the the
12:45
musculature around their organization to really thrive in
12:50
an AI related. There's ways to overcome that.
12:52
We can also talk about that in in a little bit here.
12:55
But I do think that organizational readiness is an
12:58
important concern in certain industries.
13:02
We actually see also impediments to the use of some of the
13:05
personalization techniques maybe due to like SAG, Astra and and
13:09
other type regulations or agreements that ultimately may
13:15
limit the use of personalization or content, right or even data
13:20
related to that. And so there are regulated
13:23
industries and industries where agreements may in fact
13:27
decelerate the adoption of AI by those customers who could
13:31
actually benefit. And it's worth exploring how to
13:35
overcome those challenges to ultimately adopt AI control of
13:40
brand voice and style. It is another concern that we
13:44
often hear when customers are contemplating the use of AI for,
13:48
you know, generation of content, for example, for personalization
13:52
or optimization related purposes, right?
13:55
That fear of loss of control of what it means to maintain brand
14:00
style and voice. And so, you know, those are a
14:03
couple of those concerns. But when it comes down to to
14:06
driving adoption, I do think that proving the results in ways
14:12
that show that there is an optimization and a return on
14:15
investment. Again, producing those outcomes
14:18
gave that's where you know customers really need to lean in
14:22
and that starts to make the the value of overcoming those
14:26
stumbling box and reducing the impediments worth a taint.
14:30
Really good stuff. Yeah.
14:31
So I mean kind of the different, different layers of of blockers.
14:35
You've got the data layer, you know, if we don't in which we
14:39
will talk a lot more about, but if you don't have access and to
14:43
data and then the right data and kind of clean data, you know,
14:48
then the results aren't going to be there.
14:50
Then your org's readiness to actually use it and operate off
14:53
of it. Some areas where personalization
14:56
maybe isn't as much as an option, so they're not seeing as
14:58
much of A benefit. Couple follow up questions on
15:01
this. On your last two points on the
15:03
control of brand voice, I think what you're saying there is some
15:08
companies, some marketers are more willing to kind of hand
15:11
over creative, maybe even some brand type creatives to AI.
15:16
What are you seeing in terms of like the cutting edge of that?
15:19
Are, are brands like, hey, create an ad for me completely
15:23
from AI, create a home page completely from AI?
15:26
Or is it, is it, is it something less than that?
15:29
Or like I'm really interested in that because there's, there's,
15:32
there are a lot of products that you see and companies out there
15:36
that are you're starting to kind of create whole campaigns from
15:40
scratch using AI and the range of comfort zone there is really
15:44
interesting to monitor. So what?
15:45
What are you seeing there? Yeah.
15:47
I mean, I'm going to segment or really tear the market slightly,
15:51
right. And I think, you know, if we
15:53
look to the large enterprise brands who have really valued
15:57
their style and voice and market and may work with an agency
16:02
there, they may have a little less direct tolerance to turn
16:05
over the keys completely to AI. Even if their agency or in house
16:11
marketing teams are benefiting from creative generation and
16:15
creative content, you know, coming out of the AI itself,
16:19
they don't necessarily want to turn it over to the machine
16:22
directly. When we start to look at the
16:25
small, the mid sized part there there's a much greater historic
16:30
impediment or barrier to the economics of content production
16:35
that may limit their ability to ultimately engage in advertising
16:39
and marketing or personalization technique techniques.
16:42
And so working with, you know, integrated stacks, partners,
16:47
vendors or platforms that can offer turnkey solutions to help
16:52
them with creative generation really speeds their time to
16:56
market. And there are and overcomes
16:58
their inability traditionally to produce valuable content to put
17:03
in front of the consumer. And now they can actually engage
17:06
with those platforms that have AI creative generation to
17:10
originate their campaign and get going even more quickly.
17:14
And so, you know, those are kind of two tiers of the market.
17:17
I would look at the other one. Segmentation would really come
17:21
around. Is this a regulated industry?
17:23
Do I have the right to use data and personalized, you know,
17:26
healthcare, it's maybe a little more limited than say retail,
17:31
right, or other lighter weight consumer publishing segments?
17:36
Hopefully that helps Gabe. Yeah.
17:37
No, it's, it's super helpful. And I think your point around,
17:42
you know, mid sized companies or companies that haven't had the
17:45
resources to produce the content around personalization is a
17:49
really important one because we at demand base, we've worked,
17:52
we're on personalization with companies for a long, long time.
17:55
And historically, you know, the problem with the, the hope of
17:59
personalization and the, the benefit of personalization is
18:02
clear. The reality of it and all the
18:04
content that you, you need to actually do the personalization
18:07
has been what's tough historically.
18:09
And now we're at this moment with AI where we can overcome
18:12
that hurdle. And so to your point about
18:15
outcomes, probably around engagement and because you have
18:19
more content to personalize, is there's really going to be some
18:23
upside over the next couple years for, for companies that
18:25
haven't been able to personalize in in the past and enterprise
18:28
have because they have the content machine.
18:30
But to your point, other companies haven't.
18:32
And, and what an opportunity. One more thought question.
18:35
You know, you mentioned kind of wrapping all these blockers in
18:39
the point around it goes back to driving outcomes, which IA 100%
18:43
agree with. Is there any outcomes if you're
18:46
going through, you know, these different use cases that you've
18:50
mentioned around personalization and enrichment etcetera, are
18:55
there outcomes that companies should focus on that makes
18:58
clear, you know, internally for the stakeholders that they're
19:01
moving in the right direction? Yeah.
19:03
I mean again heart back to those age-old goals, right?
19:06
Pick one of those who am I? Am I looking to acquire new
19:09
customers? Am I looking to retain the
19:11
customers that I currently have, Am I looking to upsell them
19:15
right or even support them better through customized, you
19:20
know customer experience channels, contact center and
19:23
support across all of those AI is enhancing and improving the
19:28
outcomes. Which ones matter, I think
19:31
really depends on the organization that's engaging in
19:34
the marketing, the advertising, where the customer experience.
19:37
And they really should look at prioritizing the use case that
19:41
matters to them and ultimately valuing that outcome differently
19:46
for each of those stages and lean into a prioritized sort of
19:50
road map that actually enhances their ability to achieve one of
19:55
those goals. Yeah, prioritize, you know, kind
19:58
of look at your overall go to market, prioritize, you know,
20:01
look at the traditional metrics you've been tracking.
20:03
Kind of do a stack rank prioritization and kind of knock
20:06
it down with AI as you go and and make sure make sure you're
20:10
moving in the right direction. It's really good advice.
20:13
Yeah. And we mentioned earlier, make
20:14
sure you have the right data in place, right?
20:17
And that the data is accessible to achieve the enhancement to
20:20
the outcome for the prioritized valued use case.
20:23
Absolutely. I've been looking forward to
20:25
this question and it's a big one around LLMS. And so I think this
20:32
is on everybody's mind as we see the impact on search or the
20:36
potential impact on search that very cell LMS are are having
20:41
because as we are all changing, you know, I'm changing the way I
20:44
find content. Clearly that's going to change
20:48
the nature of ad channels, probably in a mild way in the
20:51
short and and midterm, and perhaps in a more pronounced way
20:55
in the long term. I know nobody has the the single
20:59
answer on this, but what do you think this means for marketers
21:03
and and how they find audiences and and how they're going to
21:06
advertise as habits change in the way you find content because
21:12
of LLMS? Yeah.
21:14
I mean, I think each and every one of us as consumers is
21:18
already enabling this change. We're seeing glimmers of it.
21:22
Think about search, right? Traditionally, I would have gone
21:25
into Google, typed a search URL, and then seen a series of links
21:30
ranked, boosted, sponsored or otherwise, and I would engage in
21:33
that content. Today, most of us look just at
21:37
the AI synthesized search results, the AI powered results,
21:41
and that's actually reducing or visitation to the end publisher
21:46
sites who produce some of that content that ultimately rolls up
21:50
into the AI results itself. So with AI starting to
21:54
synthesize those results, the channels by which an advertiser
21:58
or marketer can actually reach that consumer and influence them
22:03
to engage with their services, their products, their content is
22:07
starting to shift, right. We've got to both rely on that
22:11
traditional mix of search of social of direct publisher or,
22:17
you know, larger relationships through programmatic at scale.
22:21
And now we've got to start to influence the AI at the point of
22:25
reasoning or synthesis of response, right?
22:29
And so the question starts to become right as a marketer, how
22:33
do I ensure that the data that I want to influence the reasoning
22:38
is available to the AI models at the point of decision that'll
22:43
then influence the consumer itself.
22:45
And or if it's the agent making a decision on behalf of the
22:49
consumer, how do I ensure that my data is accessible to that
22:53
agent to have it make the right decision that I wanted to
22:57
influence it as the marketer advertiser, right.
23:01
So the in channels by which I think the marketer advertiser
23:05
will need to reach that consumer will both remain consistent with
23:10
the traditional channels and now be decorated or augmented with
23:15
this additional channel of the agent itself and AI.
23:19
And so reaching that AI, getting my data and influence into the
23:23
the reasoning model becomes the next challenge for the market
23:27
around. That's well said and an
23:29
incredibly big change in terms of how we think about context,
23:35
right. So to summarize the traditional
23:37
channels are not going away, you know publisher channel search
23:41
etcetera. They are not likely and I'm
23:44
putting words in your mouth likely not growing but will will
23:47
need to be augmented. And I think the key insight is
23:51
the context there is the point of reasoning.
23:55
So we're marketers are going to have to think about how do we
23:57
influence at the the point of reasoning, which sounds like
24:03
he's really the agents buying and selling agents, publisher
24:06
agents and the synthesis of AI in kind of a search context.
24:11
Is that right? Yeah, I, I think so, Gabe,
24:14
Right. Again, the future is TBD.
24:17
It's nebulous. We're going to see how this
24:19
plays out. But you know, look, I've got
24:22
data. I understand the consumer.
24:25
We're, we're the buyer and I want that buyer to understand my
24:29
product or services. How do I do that when it's an AI
24:34
or LOM starting to make some of the synthesis around a response
24:38
or actually make a decision about which action to take?
24:42
How do I make sure that my data, my service, is understood by
24:47
that model to help steer and influence the decision that's
24:51
ultimately made or the result that's put in front of the
24:53
consumer? It would another way to say that
24:56
be that, you know, current context or, or the model that
25:00
we've lived off is, hey, you're, you're going to place an ad,
25:02
whether it's a search ad, whether it's a display ad,
25:06
whether it's pre roll or CTV in a certain context.
25:11
And now we're going to have to augment that with placing data
25:14
at the point of reasoning. Yeah.
25:16
And or making sure that that data influences the actual
25:20
content that the consumer sees and engages with, right.
25:24
So that is my ad personalized was I segmented and targeted
25:29
using AI that all looks a little more like traditional
25:31
advertising. And then occasionally look, I
25:35
want the agent to make a decision and maybe subscribe or
25:38
buy a service product on my behalf.
25:41
How does my tastes and likes be influenced by the data that the
25:46
marketer advertiser wants to make sure I'm aware of or that
25:49
my agent is aware of? Yeah.
25:51
So using data more effectively to personalize traditional
25:55
channels and then using data to to influence the the kind of
26:00
agentic dynamic that's going on. Exactly.
26:03
Yeah. Wow.
26:05
Really exciting, really big stuff.
26:07
What we've hit on multiple times is data and the data Foundation.
26:12
You said it's critical. IA 100% agree.
26:15
It sounds like, you know, really that's going to be the
26:17
determinant on, on whether you win or lose.
26:19
Like any advice on how to develop the data strategy that
26:24
we'll win in the future? Yeah, data is absolutely the
26:27
critical ingredient. And increasingly we've got to
26:31
figure out ways as a marketer or advertiser, how to get that data
26:35
to influence that outcome that we just talked about, right?
26:38
How to influence the AI and make that AI have access to the data
26:44
and make it usable with the within the context of the AI
26:48
decision. So the data itself has to be of
26:51
sufficient quality and has to be useful to the decision itself,
26:55
meaning accessible. So it's kind of a garbage in,
26:58
garbage out problems, right? Brands have to have a strategy
27:02
to ensure that their data is useful to the point of decision,
27:05
accessible to the point of decision, and then ultimately,
27:09
they've got to unify that data in a way that powers it
27:13
effectively towards the goal that they're trying to achieve.
27:17
1 strategy for doing this and it is really to look to integrated
27:21
platforms from integrated partners and vendors that could
27:25
really help them with the aggregation and unification of
27:29
that data because we all know it's been an impediment in
27:32
advertising marketing for many, many years now, right?
27:36
The ability to access the right data and put it in into my
27:40
advertising or marketing platform in a unified and
27:43
consistent way to power. They may be traditionally
27:46
through machine learning and increasingly through AI to power
27:51
the outcome. And so brands have to have a
27:54
strategy for who are the partners that they work with,
27:59
who can really help them not have to build this from scratch,
28:03
but ultimately integrate and make use of both their first
28:06
party data and the broader data that exists out through the
28:11
larger community of data partnerships.
28:13
And then make that unified, let's call it profile or
28:18
understanding of the consumer and put that into an actionable
28:23
AI based system, whether it's agent based, we're really more
28:27
traditional based but influenced by AI.
28:30
Absolutely, yeah, 100% agree. Essentially you have to have the
28:35
data be cleaned, have to have access, has to be unified sounds
28:39
like working working with partners.
28:41
There's a lot of lot of technologies including demand
28:43
base that you know work real to to unify this data for for the
28:47
go to market. Is there any new roles or who's
28:51
doing this internally? But you know, for what did you
28:54
see at your customers or somebody within marketing that's
28:57
owning this? Is it like a CIO type or how do
29:00
you you're working? Yeah, we typically see the
29:03
marketer and the CMO's office really being the point of
29:07
origination of the use case, right?
29:09
Really want to achieve an outcome, it's getting cut
29:12
parking back to their age-old marketer outcomes.
29:14
And often times what they start to encounter is those internal
29:19
impediments to the use of their data, whether that's engineering
29:23
and technology or whether it's actually governance about the
29:26
data itself. They may start to to encounter
29:30
problems and one way that we see acceleration of this amongst our
29:34
customers is by working with the right partners who may have an
29:38
integrated stack that can really accelerate the time to market to
29:43
achieve the use case and outcome without having to overcome all
29:47
the internal build related barriers to get that use case to
29:52
market. And so working with partners who
29:56
offer integrated offerings that can help them integrate their
30:00
first party data, complement and enrich that with third party
30:04
data that may be accessible in the integrated platform and
30:08
really help put that data into the context of the advertising
30:12
or marketing decision to power the personalization and
30:16
ultimately measure and optimize the outcomes becomes I think the
30:21
right approach for most marketing organizations.
30:24
Got it. You know, so it, it starts with
30:26
the outcomes and, and the use cases that are going to drive it
30:30
identifying what the right data sets and as opposed to kind of
30:33
trying to build it internally or stitch things together, which
30:36
becomes really complex. You're, you're working with a
30:38
partner that has experience and has built modern technology to
30:42
pull together and make it accessible for that use case and
30:45
to drive that outcome. I think that's the right
30:47
strategy for most of the marketers that we end up
30:50
engaging with. Makes sense?
30:52
All right. We are, you know, demand based
30:55
in our in our audience are are really focused on B to B and you
30:59
know, there's a lot of similarities between B to CB to
31:01
B Obviously, you know, both sides of the house are are being
31:05
impacted by by AI in a huge waves.
31:09
Is there any impact though that you see as being kind of unique
31:12
to B to B versus consumer advertising?
31:15
At the end of the day, the goals are the same, right?
31:18
I want to influence a buyer to engage with my products or my
31:22
services and or satisfy and upsell.
31:26
And so you know, how we sort of treat data that the granularity
31:31
of that data may vary rather than being strictly 1 to one
31:35
consumer based and powering A1 to one personalization, perhaps
31:39
what I'm doing with B to B is grouping and cohorting some of
31:43
that data and then starting to target through that
31:46
segmentation, you know, a larger cohort of individuals.
31:51
But my goal is the same. And so the differentiation that
31:55
I really think about, Gabe and you could probably educate me
31:58
much more than I could educate you on this topic is really that
32:01
that you know much of the underpinning, the foundations of
32:06
need are very, very similar. The granularity at which we'll
32:10
do that targeting segmentation may be different.
32:13
The outcomes we're trying to produce are in fact very, very
32:16
similar as well, right. So for me the question, the
32:19
differentiation here starts to be about the granularity of the
32:22
data more so than a a difference of outcome that we're trying to
32:27
produce. Yeah, that really does resonate.
32:30
And and just to kind of add some color from from on the B to B
32:33
side. I think the big and going back
32:36
to the kind of what's working right now and how, you know, the
32:39
enrichment personalization, the the channels is something that
32:43
is really working right now that I've been an impediment before
32:46
is finding the buying group within these within an account,
32:50
right. And historically, you know,
32:53
marketing had been focused on leads, you know, who's the lead
32:57
that I'm going to give to the sales team to.
32:59
So I'm going after a person. And then what happened was,
33:03
well, don't, isn't it really an account that that buys AB to B
33:08
product? And that's what led to the
33:09
advent of ABM. And that was a big step forward
33:12
kind of a decade ago. And I think what we're starting
33:15
to see is ABM is a great set of tactics to align sales and
33:19
marketing around data to go after the same accounts and
33:23
engage them and show a lot of great outcomes.
33:26
But it's not directly correlated to pipeline, right?
33:29
And and that's what sales is most focused on is opportunities
33:33
like what I have different products within my organization,
33:36
Each one of them can be an opportunity who who really buys,
33:40
you know, these products? What is the basis of an
33:43
opportunity? It's the buying group.
33:45
And so to your point about being able to scale segmentation and
33:49
audiences, and now the technology exists to say, hey,
33:52
it's not just the account is who's the buying group for every
33:55
product? And can we align sales and
33:58
marketing around, you know, again, to your point, the
34:00
granularity of that data, which you might not see or need in, in
34:05
B to C, but in B to B becomes incredibly important.
34:08
So that's one of the, if I was going to say, hey, what are the
34:11
best in Class B to B marketers doing today that they weren't
34:14
able to identification and scaling up the buying group,
34:17
which I'm seeing making a huge impact on the outcomes with
34:20
creation of pipeline, increased bookings, those sort of things.
34:24
And it all hearts back to the data.
34:26
Gabe 100%, you know, I think, I think we were, we were talking,
34:30
you know, a couple weeks ago. It's like if there was 1 thing
34:32
that I, I can wave my magic wand on is that data strategy would
34:36
be at the top of everybody's list for, for 2026.
34:39
I think it'll be up on a lot of lists for sure.
34:42
So other practical advice, I mean, if you've really given a
34:45
lot of good information about what's working today, how to
34:49
think about data, work with external partners, if there's
34:52
somebody you know that, hey, they're just kind of
34:54
understanding, OK, now is my time to activate AI from a
34:57
marketing perspective. Is there any advice you give
35:00
them on how to start? Yeah.
35:02
I mean, you know, we already hit upon some of it.
35:04
Make sure that you have a first party data strategy and have
35:07
some data in house that you can work with.
35:10
Choose vendors and partners that have an integrated stack that
35:14
helps you take advantage not only of your data, but other
35:18
data from the broader community of providers that exist and can
35:21
be relevant to powering those outcomes.
35:24
Don't boil the ocean, right? We started with sort of the
35:27
difficulty of unification of data and the complexity of what
35:31
that can mean to enrich a profile or understanding of the
35:34
consumer. And we can, if we look at every
35:37
use case without prioritization, we often can sort of freeze and,
35:42
and lock up in our ability to pursue them.
35:44
So, you know, prioritize and choose those use cases that
35:48
ultimately will prove enough value and be achievable from
35:52
data that you have accessible through the channels that
35:56
already exist today and start to take advantage in that way,
36:00
right? Take little baby steps that
36:02
develop that organizational muscle are achievable and prove
36:06
the valuable outcomes to your stakeholder.
36:10
Kind of go from where you are today here.
36:13
Here's my current success metrics.
36:15
What can I drive prioritize that Here's my my advertising
36:19
workflow. How can I make that more
36:21
efficient and just cast start working on on the low hanging
36:24
fruit essentially? Yeah, pick and don't pick at
36:27
all, right? Make a piece of fruit at a time
36:30
and prove the value and the taste of it.
36:32
Yeah, now. AWS, we work with AWS very
36:36
closely. We have our agent base, which is
36:39
our system of agents that sits on top of our unified data
36:42
platform, which enables a lot of skill to our to our marketers.
36:47
We built that on AWS. There was a lot of amazing tools
36:50
that helped us to to launch our agents.
36:53
What are some of the things that you're launching in ad tech that
36:56
you can help a data-driven? Marketer, yeah, I won't cover
37:00
off our full road map for ad tech and for advertising or
37:03
marketing, but know that we're innovating quite a bit in this
37:07
space, right? You can really look to public
37:10
announcements that have been made around our AI related
37:13
stack, including Dead Rock. And we really think, you know
37:17
that we offer the best foundations for choosing the
37:21
right models for optimizing for cost and latency efficiency,
37:25
right? For ultimately securely
37:27
customizing with your data the AI itself for being able to run
37:33
AI responsibly in an integrate enterprise grade manner here on
37:38
the AWS cloud and now increasingly to be able to
37:41
deploy and operate agents at scale.
37:44
And that's all part of our larger, you know, AI and machine
37:48
learning stack here at AWSI won't go into the details of
37:52
that. I don't want to bore, you know,
37:53
our listeners today gay, but do you know that we're going to
37:56
continue to pioneer and make quite a bit of innovations in
38:00
this area when it comes to advertising and marketing
38:04
specifically and and to the, you know the likes of yourself at
38:07
demand base. We're really powering, I think
38:11
some of the foundational, undifferentiated heavy lift
38:15
around the data that helps both, you know, integrate, ingest,
38:20
enrich, identify and collaborate around data in ways that start
38:26
to allow you to unify foundationally.
38:29
And then power the AI stack here at AWS to be able to drive AI
38:36
based insights, AI based actions and power those outcomes through
38:42
next best actioning, next best offering and personalization.
38:46
And then when it comes to the the sort of final mile of the
38:49
stack, we work with partners and customers like yourself who
38:53
really build those applications on behalf of the marketer and
38:57
advertiser and really bring them to market at scale.
39:00
And so I really think and we think at AWS about democratizing
39:06
the use of AI and ML through our enterprise grade stack.
39:11
But we do that to remove the undifferentiated heavy lift.
39:15
And then we let the AI shine when it marries with data on
39:19
companies like yourself and from our marketer and advertiser
39:23
customers who want to take advantage of that to power their
39:26
outcomes. That's where they start to look
39:28
to partners. The other role that that AWS
39:31
really plays is around, you know, I'm going to call it that
39:35
organizational readiness, right? The series of partners that we
39:38
have with both their integrated stacks.
39:40
But then also, you know, our transformation and consultancy
39:44
practice partners who can really help come in and work with
39:49
marketers and advertisers to get their AI strategy right, to help
39:53
choose the right partners and then start to integrate those in
39:57
ways that can enable the transformation if they can't
40:00
find a single off the shelf provided for them.
40:03
So we play a a variate of roles and won't cover off all of our
40:07
road map, but do you know that we play in every one of those
40:10
layers from data ultimately to the AI itself and then a little
40:15
bit in that final mile of the application through which
40:18
customers can I hope that resonates.
40:22
I, I mean, I can attest to the to that value, I say 100% that
40:27
that AW has been an amazing partner.
40:29
I'm really proud of the pace of innovation that demand base has
40:33
and I think it's only increasing.
40:35
And we've worked with you across those different levels that kind
40:38
of the heavy lift data side and then adding our, you know, have
40:41
you lift the AI side and then adding all the data that we
40:44
have. We have all B to B to assets.
40:45
We bring in the first party data in the ways we talked about for
40:49
unified data platform and then having the tools to build the AI
40:52
on top of that. And we've certainly taken
40:55
advantage of your transformation services, the days up in Seattle
40:59
that have really been game changing for us.
41:01
So you guys, you guys are a huge partner and add a ton of value
41:05
in the ecosystem and it's appreciated.
41:07
Now partnership, Gabe, thank you.
41:10
Yeah, yeah, likewise. So we did, you know, the
41:13
practical advice for kind of the practitioner if you were going
41:16
to look from like a business perspective, if we hit back on
41:21
like, you know, maybe it's not hype, maybe it's uncertainty or
41:25
to be determined in a future. What would your advice be to CM
41:31
OS and CR OS that are navigating this transformation?
41:35
Yeah. I mean, you know, we kind of hit
41:36
upon some of the difficulty that CM OS can encounter.
41:40
And you know, what I would say is don't hesitate to get on the
41:43
bike and try to figure out how to ride it, right?
41:45
Because you ultimately need to build organizational knowledge
41:50
and that organizational operational musculature to
41:54
overcome those challenges. And it's not too late to start,
41:57
right? It's never too late to start,
41:59
but get started today because the sooner you develop that
42:02
practice, the earlier that you'll be able to achieve those
42:05
outcomes and benefit from the AI that that exists even in the
42:09
current state. You'll also be ready as the new
42:14
era, as we see some of that nebulosity start to take greater
42:17
shape, you'll be ready to engage in those agentic related
42:21
channels. You'll be able to engage in new
42:24
formats all through getting your data strategy and foundations
42:28
right today and then developing your practices on top of that.
42:33
So get going. The winners of tomorrow are
42:36
actually developing today. Yeah, there's God, there's an
42:39
urgency, and I know I did. You know, I think I see in the
42:42
market that it can be daunting, but things are being figured out
42:46
every day and just getting the muscle, like you say, to be a
42:49
part of the transformation, even if it changed dramatically,
42:52
picks up momentum for your organization.
42:54
So I think that's great advice. Last question.
42:57
This is a tough one. Thinking five years out, which
43:01
doesn't sound that far in the future, but given the face of
43:05
change, it might as well be 50. We're looking at 20-30.
43:09
What does advertising look like and what what will have changed
43:13
and what will stay the same? Yeah, the need for data is going
43:16
to stay the same that it going to be ubiquitous, right.
43:20
The the age-old goals stay the same.
43:23
The the channels and the mix of how we'll achieve them and how
43:28
we'll employ and use data to produce those outcomes is
43:33
ultimately going to transform itself, right?
43:36
We'll still see the traditional mix of channels like we have
43:39
with every one of the eras of advertising that we talked about
43:42
from an evolutionary spectrum. And now we'll start to augment
43:46
those with brand new approaches to engaging those consumers and
43:50
influencing their outcome. Across every one of those eras,
43:54
we've seen the growth of data, the growth of the utilization of
43:58
that data to power outcomes, increase in its potency.
44:03
And I think this next era is going to produce very similar
44:06
results where we'll both decorate what existed and
44:11
produce better outcomes with the addition of the new channels,
44:14
right. So increased efficiencies
44:16
throughout still relying on data and then the advent of you know,
44:21
additional agent based type advertising and marketing
44:25
scenarios at scale going from the what we may have considered
44:30
to be nebulous and and or hype based today to a reality within
44:35
three to five years. Yes.
44:37
So I mean, the theme that's really emerged here and I feel,
44:41
you know, being ACEO of my organization, which I even feel
44:44
more urgent about after this conversation, is the importance
44:48
of data. The importance of data, both in
44:51
the outcomes that you can drive today through making current
44:55
channels more efficient, and then the importance of data and
45:00
driving the new channels and the outcomes of the future.
45:03
So I I think that really emerges as as the top, top theme for me.
45:07
I really like that the hitting on outcomes.
45:09
I think a lot of times we focus on the means to the outcomes,
45:13
but now with these tools, we can meet the even more laser focused
45:17
on outcomes. We can do that through, for
45:20
instance, driving more engagement because we can still
45:23
scale personalization so much more than we used to be able to.
45:26
But it's again, it's a like focusing on the outcomes is
45:29
really the thing that connects the past to the future.
45:32
So really energizing discussion. I feel like we could keep
45:36
talking about this. We should probably do it again
45:37
in, in six months or a year to see, see how things have changed
45:41
and, and what we're excited about then.
45:42
I, I have a feeling we'd still talk about data then.
45:45
But thank you for joining us today.
45:47
For listeners that want to get in touch with you, what's the
45:49
best way? Reach out to your AWS accountant
45:52
and really drive a conversation about your use cases for
45:56
advertising or marketing, including with AI.
46:00
And they will engage myself and my tea to ultimately come on
46:03
your behalf and guide, advise, help shape some of your strategy
46:08
or enable your solutioning, including through partners like
46:10
Demand Base. So reach out through your AM
46:13
team. Awesome.
46:14
And we have a, we have a great AM team.
46:15
I'll just give that pluck at AWS.
46:17
So it's great to talk with you. I know listeners are going to
46:20
love this. I was energized by the
46:22
discussion. I think we're at such an
46:24
exciting period for for advertising and marketing and
46:28
it's really a joy to be a partner with you and let's see
46:30
where things go. Gabe, thank you so much for
46:32
having me today. It indeed was a thrilling
46:35
conversation. I'll look forward to the next
46:37
chapter of this. Thank you.
46:40
Thanks for joining us for On Base.
46:42
If you enjoyed this episode, please be sure to share, rate,
46:45
review and subscribe on Spotify, Apple, YouTube, or wherever you
46:49
get your podcasts. This show is brought to you by
46:51
Demand Base, the only account based experience platform built
46:55
for sales and marketing alignment.