Brad Kugler, CEO of DirectMail2.0, Who’s Mailing What!, and DM20.ai, was invited by USPS to participate in this webinar which presents the research of Winterberry Group’s latest white paper on AI, performance marketing, data, and omnichannel. Since Brad built the industry’s first AI platform for direct mail performance, he was a key contributor in this conversation. What are your favorite takeaways?
Transcript:
1
00:00:03.890 –> 00:00:05.140
Ray Van Iterson: Good afternoon.
2
00:00:05.290 –> 00:00:14.680
Ray Van Iterson: My name is Ray Van Iterson. I lead marketing strategy at the Postal Service, and really want to thank you for taking the time to join us here today.
3
00:00:14.700 –> 00:00:32.570
Ray Van Iterson: At the Postal Service, we have a mission to bind the nation together through communication and commerce, and part of that mission is to work on education and making sure that information flows to everybody. For communication in recent years.
4
00:00:32.570 –> 00:00:36.019
Ray Van Iterson: Omnichannel has become key in terms of
5
00:00:36.020 –> 00:01:00.700
Ray Van Iterson: digital and analog communication working together, and in more recent years, AI has become a key element of that communication. So what we wanted to do today was to really have a conversation about some of the leading developments in how AI is driving performance in terms of communication, in all types of communication
6
00:01:00.700 –> 00:01:24.050
Ray Van Iterson: that you do. So, as part of the Postal Service mission to do that, we have helped make possible a study by the Winterberry Group on delivering performance, in the era of AI. And so, today, what we wanted to do was to first have a three-part conversation. First, Jonathan Margulies, managing partner at the Winterbury Group.
7
00:01:24.050 –> 00:01:25.720
Ray Van Iterson: We’ll speak about
8
00:01:26.040 –> 00:01:50.979
Ray Van Iterson: what he has… what the research that he has just been doing. Then secondly, we’re going to add in a couple of practitioners who deal with performance marketing and AI on a daily basis, and we’ll have a roundtable discussion among the four of us. And then lastly, we’ll open it up to Q&A to all of you, and so we’ll deal with your questions as well. So, if you have questions
9
00:01:50.980 –> 00:02:06.200
Ray Van Iterson: at any point during the webinar, please don’t hesitate. Put them into the Q&A. If it’s something that we can incorporate in the earlier parts, we’ll do that. Otherwise, you’ll be first in line for the questions that we’ll have at the end.
10
00:02:06.200 –> 00:02:20.830
Ray Van Iterson: If we don’t get to every question, we’ll be in touch with you afterwards. Again, this webinar is being recorded so that you’ll have a chance to share it with friends and colleagues and listen to it again if you so choose. So.
11
00:02:20.920 –> 00:02:25.509
Ray Van Iterson: Without further ado, Jonathan, take us away and tell us about the research.
12
00:02:25.510 –> 00:02:35.030
Jonathan Margulies – Winterberry: Oh, thank you very much, Ray. Appreciate being here, and I appreciate everyone joining us as well today. Give me a moment here, and I’m going to share
13
00:02:36.140 –> 00:02:41.010
Jonathan Margulies – Winterberry: slides, It’s gonna take us through some of the content.
14
00:02:42.580 –> 00:02:56.020
Jonathan Margulies – Winterberry: Which everyone should be seeing now. Great. So as Ray mentioned, my name’s Jonathan Margulis. I’m a managing partner at Winterberry Group. By way of quick introduction, we’re a strategy consulting firm based in New York.
15
00:02:56.020 –> 00:03:13.850
Jonathan Margulies – Winterberry: And we work exclusively across the different… the disciplines of marketing, advertising, technology, data, and analytics. It’s a bit of a mouthful, I grant you, but we stay very, very close to the kind of disruption that’s happening in marketing circles, how that’s impacting brands, publishers.
16
00:03:13.850 –> 00:03:24.149
Jonathan Margulies – Winterberry: And how that’s impacting folks that are really across the entire marketing ecosystem, as you think about folks that live in the supply chain, that provide data to fuel marketing campaigns and the like.
17
00:03:24.260 –> 00:03:41.789
Jonathan Margulies – Winterberry: Many folks know us through our advisory work, but I would probably say far more know us through our published research. We’ve been keeping tabs on where disruption is happening in and around marketing ecosystem for the better part of two decades.
18
00:03:41.790 –> 00:03:59.399
Jonathan Margulies – Winterberry: And have tracked a lot of the more transformative developments, in this world over that time, as we think through the impact of technology, the proliferation of data, the emergence of new media channels, and a host of
19
00:03:59.460 –> 00:04:15.739
Jonathan Margulies – Winterberry: different innovations and considerations, some that have really proven meaningful and some somewhat less so. Part of our role is always to try to make heads or tails of exactly what’s happening and how that’s going to impact all sorts of folks in the future.
20
00:04:15.780 –> 00:04:20.610
Jonathan Margulies – Winterberry: The last few years, we’ve been working with, with Ray and his team at the USPS.
21
00:04:20.930 –> 00:04:28.230
Jonathan Margulies – Winterberry: to publish insights, really tracking the evolution of the direct mail channel, and, and have,
22
00:04:28.440 –> 00:04:47.659
Jonathan Margulies – Winterberry: advanced a series of research under the Delivering Performance banner. Gotta get a sense of how the role of mail is changing, and how brands are looking to tap into that channel, along with many others, to drive a wide range of change in their own business. One of the,
23
00:04:47.890 –> 00:04:56.250
Jonathan Margulies – Winterberry: One of the observations that we’ve been able to reach after many, many years of doing this is that when
24
00:04:56.250 –> 00:05:08.229
Jonathan Margulies – Winterberry: a certain development or innovation sort of reaches critical mass in marketing circles. You can kind of track just how much traction it’s getting by the amount of infographics that, that
25
00:05:08.230 –> 00:05:20.780
Jonathan Margulies – Winterberry: various folks, analysts like myself, for example, are putting to work to try to explain what’s happening. We saw that happening years ago, when multi-channel marketing became a theme that marketers were really
26
00:05:20.780 –> 00:05:39.219
Jonathan Margulies – Winterberry: hot to pursue, the idea of putting many channels to work in the service of their marketing objectives. We saw it a few years later, around the emergence of big data, the notion that information was proliferating from many different sources, online as well as offline, and it was incumbent on brands to
27
00:05:39.250 –> 00:05:48.250
Jonathan Margulies – Winterberry: put it to work. We’re seeing it today in and around the discipline of AI, and with… with…
28
00:05:48.360 –> 00:06:03.260
Jonathan Margulies – Winterberry: The understanding, again, being that technology is advancing, at such a pace, and coming to support just so diverse an array of use cases that, the stakes are growing higher than ever before, and it’s incumbent on marketers to get a handle around.
29
00:06:03.260 –> 00:06:14.389
Jonathan Margulies – Winterberry: exactly what the technology can do and exactly what the technology can’t do. One of the things we’ve realized, both as thinking through the impact of AI within individual channels, like direct mail.
30
00:06:14.400 –> 00:06:31.870
Jonathan Margulies – Winterberry: And as a general force, a platform to drive transformation and marketing in all different ways, is that often in these cases, even though the technology presents a lot of promise, it presents more questions than definitive answers, certainly when it’s in its formative stages.
31
00:06:32.530 –> 00:06:37.969
Jonathan Margulies – Winterberry: And as we’ve been thinking about exactly how to tackle this in the context of our delivering performance
32
00:06:38.330 –> 00:06:57.469
Jonathan Margulies – Winterberry: series. Those questions over the last 12 months or so have really… have really taken on a very defined form to them. How’s the technology different than that which came before? What is it actually doing in practice today that allows marketers to achieve results that… that…
33
00:06:57.470 –> 00:07:00.019
Jonathan Margulies – Winterberry: You know, were previously not… not possible.
34
00:07:00.080 –> 00:07:15.260
Jonathan Margulies – Winterberry: Is AI something that should represent an opportunity and a target for investment, or is it something that is actually going to be more disruptive to brands as they think about their work and their prospects to advance what they’re aiming to do?
35
00:07:15.260 –> 00:07:31.619
Jonathan Margulies – Winterberry: And maybe most importantly, if AI is something that does, offer transformative to the positive opportunity, how do we lean into it? How do we optimize for it? How do we plan? How do we strategize? How do we make it part of our business as usual?
36
00:07:31.750 –> 00:07:49.249
Jonathan Margulies – Winterberry: So with those as sort of North Stars, as questions that we’re still wrestling with, and that we hear folks, from the brand community, from the marketing supplier community, asking us about, we sought out to understand a little bit more around how AI is working in practice, both
37
00:07:49.300 –> 00:07:53.490
Jonathan Margulies – Winterberry: Specifically, with respect to direct mail and other
38
00:07:53.550 –> 00:08:02.829
Jonathan Margulies – Winterberry: If you will, performance-oriented channels, but more broadly, as we think about the technology and the tools that brands are going to need to leverage in the future.
39
00:08:03.240 –> 00:08:21.560
Jonathan Margulies – Winterberry: We published our results in a, in a white paper a couple of weeks ago. That white paper is available for free download, and I’ll present, a link later on that, by which you can get access to the entire data set. But I’m going to spend a few minutes now, reviewing a few of our, of our
40
00:08:22.010 –> 00:08:31.819
Jonathan Margulies – Winterberry: high-level, conclusions, which I think will frame nicely, some of the conversation we’ll have with practitioners, later in this discussion.
41
00:08:32.100 –> 00:08:38.729
Jonathan Margulies – Winterberry: There is a lot that can be said, needless to say, about AI and its impact, so we won’t have the time
42
00:08:38.730 –> 00:08:54.389
Jonathan Margulies – Winterberry: Just in this… this introduction to review even all the findings of our… of our research process, but… but needless to say, there’s much that’s going to keep, gonna keep practitioners busy for the next year, both untangling the technology and the potential of what it can do.
43
00:08:54.590 –> 00:09:14.119
Jonathan Margulies – Winterberry: The first conclusion of our research, you know, as surfaced to us by our panel of over 250 brand marketers, is that at the end of the day, no matter what any specific tool, tool set, data set, can actually accomplish, one of the North Stars of
44
00:09:14.180 –> 00:09:21.170
Jonathan Margulies – Winterberry: of AI implementation today is really being put to work to further what we see as a macro marketing trend.
45
00:09:21.170 –> 00:09:41.439
Jonathan Margulies – Winterberry: that’s been gathering steam for a number of years, and that is the pivot of brands to really embrace performance-oriented marketing objectives and use cases as the centerpiece of their effort. That is to say, to move their dollars, to move their strategic emphasis, their tactical efforts.
46
00:09:41.440 –> 00:09:48.719
Jonathan Margulies – Winterberry: Towards, any kind of initiative that can support specific, defined.
47
00:09:48.870 –> 00:09:52.789
Jonathan Margulies – Winterberry: Bottom line business objectives. Customer acquisition.
48
00:09:53.070 –> 00:10:04.789
Jonathan Margulies – Winterberry: incremental sales, retail traffic, you name it. As opposed to more general, sort of amorphous, brand-building, awareness-oriented use cases.
49
00:10:05.010 –> 00:10:25.000
Jonathan Margulies – Winterberry: This has been gathering steam for a number of years. This is one of the key reasons we know that brands continue to invest in direct mail, in digital, and certain other marketing channels, and it’s one that is increasingly driving how they engage with technology. And the reason is simple. Those bottom line objectives.
50
00:10:25.000 –> 00:10:27.210
Jonathan Margulies – Winterberry: The demand for accountability
51
00:10:27.240 –> 00:10:37.000
Jonathan Margulies – Winterberry: The understanding that technology exists to both substantiate performance and help optimize it is very, very central to the marketer’s job today.
52
00:10:37.240 –> 00:10:55.099
Jonathan Margulies – Winterberry: Perhaps not surprisingly, a vast majority of brands tell us that they’re putting greater emphasis, and in concert with that, more spending every year on performance-oriented efforts and the channels that, that go along with them.
53
00:10:55.100 –> 00:11:05.239
Jonathan Margulies – Winterberry: It doesn’t mean that awareness and branding is no longer important. In fact, for a lot of reasons, I think it can make the argument that those kind of use cases are
54
00:11:05.260 –> 00:11:23.719
Jonathan Margulies – Winterberry: developing their own performance addressability as well. But, it does… it does change the… the direction, the strategic priority, of the marketer, and… and it certainly does crystallize what they’re looking for in terms of tools and solutions to support their work.
55
00:11:24.070 –> 00:11:31.730
Jonathan Margulies – Winterberry: Part and parcel to that, the understanding in marketing circles is growing increasingly that performance
56
00:11:31.890 –> 00:11:50.969
Jonathan Margulies – Winterberry: isn’t an objective that any one marketing channel can accomplish in a silo. And as a matter of fact, just as the distinctions between awareness and performance are beginning to dissolve, as technology proves itself addressable to all sorts of different media channels.
57
00:11:50.990 –> 00:12:06.090
Jonathan Margulies – Winterberry: What marketers tell us is that increasingly they expect, most of the marketing channels in their mix to play a role in supporting their performance aims and to be able to substantiate the kind of incrementality they deliver.
58
00:12:06.160 –> 00:12:18.700
Jonathan Margulies – Winterberry: More and more, though, as our research over the last few years has sort of borne out, there’s essentially a core four channels of sorts that are really playing a disproportionate role in these performance efforts.
59
00:12:18.780 –> 00:12:34.620
Jonathan Margulies – Winterberry: Online display advertising, paid social media, direct mail, and paid search are sort of the bedrocks of this, of this performance, initiative and what marketers rely on to help support their business needs.
60
00:12:35.140 –> 00:12:46.850
Jonathan Margulies – Winterberry: And they tell us that, that that’s specific in an individual channel context, and they tell us that really weaning into those performance outcomes is important as they think about how they’re going to be leveraging
61
00:12:46.850 –> 00:13:05.410
Jonathan Margulies – Winterberry: all of these channels, on an ongoing basis into the future. Ultimately, the objective isn’t necessarily to spend more or spend less in any one channel, so much as it is to really optimize what they’re able to accomplish by way of that customer acquisition or those incremental sales.
62
00:13:05.410 –> 00:13:08.669
Jonathan Margulies – Winterberry: And use all the tools at their disposal to do so.
63
00:13:08.940 –> 00:13:23.509
Jonathan Margulies – Winterberry: AI, however you define it, as a generative tool, as an agentic tool, as a basket of technology that’s simply disruptive, is beginning to play a real central role in helping support, not just
64
00:13:23.510 –> 00:13:28.709
Jonathan Margulies – Winterberry: Incremental performance, but to support the performanceification, if you will.
65
00:13:28.710 –> 00:13:44.599
Jonathan Margulies – Winterberry: Of all those channels at disposal. There are a lot of reasons why. What it’s also doing, perhaps counterintuitively, is it’s actively changing the role and the impact of certain channels, certain media, that have been sort of core to how performance has been
66
00:13:44.970 –> 00:13:57.180
Jonathan Margulies – Winterberry: practiced for several years leading up to now. Right now, for example, we’re seeing, and perhaps many of you have experienced this, a tremendous amount of disruption in digital circles.
67
00:13:57.680 –> 00:14:09.530
Jonathan Margulies – Winterberry: Where AI, or at least in this case, the advent of AI-driven overviews that are now, in many cases, presented as a default, search
68
00:14:09.720 –> 00:14:16.560
Jonathan Margulies – Winterberry: output, is actually disrupting the ability of companies to generate digital traffic.
69
00:14:16.610 –> 00:14:25.590
Jonathan Margulies – Winterberry: Leveraging the same paid tools that they have been, that they’ve been familiar with and they’ve been comfortable using for a number of years right now.
70
00:14:25.590 –> 00:14:40.140
Jonathan Margulies – Winterberry: That has effectively, in many cases, marginalized their ability to rely on these tools to generate the kind of scale, the kind of clicks, the kind of purchase behavior that those particular media channels have been associated with.
71
00:14:40.210 –> 00:14:50.549
Jonathan Margulies – Winterberry: Over the last few years. What it’s also done in concert, unintended consequence, but perhaps a positive consequence nonetheless, is it’s really crystallized
72
00:14:50.550 –> 00:15:01.309
Jonathan Margulies – Winterberry: The priority of brands to begin building out long-term infrastructure, long-term media mix, that’s focused on really using every available channel
73
00:15:01.680 –> 00:15:06.219
Jonathan Margulies – Winterberry: Whether it’s digital or traditional, whether we think of it as paid or owned.
74
00:15:06.650 –> 00:15:25.160
Jonathan Margulies – Winterberry: in sort of an orchestrated mix to… with AI as an underlying infrastructure to help optimize, their… their ability to support those performance use cases that are core to their… to their business needs and core to their objectives. And when we… when we…
75
00:15:25.310 –> 00:15:32.620
Jonathan Margulies – Winterberry: asked our panel to really talk more about what AI was going to do to them… do for them, they said, well, it…
76
00:15:32.620 –> 00:15:47.059
Jonathan Margulies – Winterberry: In short, it can do a great many things, but what we need the technology to do, if it’s going to help us be more efficient, help us going to be more effective, is gonna… it’s going to help streamline, standardize, and optimize our performance,
77
00:15:47.180 –> 00:15:59.640
Jonathan Margulies – Winterberry: In ways that are specific to our business needs, and move past, sort of, the first-generation challenges that we’re seeing onboarding, and sorting out the role of different technology.
78
00:15:59.810 –> 00:16:11.090
Jonathan Margulies – Winterberry: Not surprisingly, as they sort of wrestle with both those first-generation challenges and think more around what the long-term future holds for the technology and its roles.
79
00:16:11.150 –> 00:16:21.880
Jonathan Margulies – Winterberry: Brands are beginning to make substantial, substantial progress up the maturity curve, both in terms of understanding what AI technology can do.
80
00:16:21.880 –> 00:16:31.920
Jonathan Margulies – Winterberry: And then leveraging these tools to generate real and meaningful impact, once again, across all those use cases that, that
81
00:16:31.950 –> 00:16:34.759
Jonathan Margulies – Winterberry: Are important to them in their marketing efforts.
82
00:16:34.910 –> 00:16:44.380
Jonathan Margulies – Winterberry: I think… I think some of the… the most important sort of black and white data that came out of our research effort, is really represented in
83
00:16:44.450 –> 00:16:55.220
Jonathan Margulies – Winterberry: by way of that maturity evolution. First and foremost, of course, being that brands, perhaps for the first time since we’ve been asking questions along these lines over the last 3 years or so.
84
00:16:55.220 –> 00:17:08.260
Jonathan Margulies – Winterberry: Are reporting that they’re generating meaningful business, performance, both in terms of effectiveness, that is to say, how much they’re selling, how many new customers are coming into their funnel.
85
00:17:08.270 –> 00:17:26.359
Jonathan Margulies – Winterberry: how much traffic they’re able to generate, but also in terms of efficiency, which is to say how much waste they’re able to remove, how fast they’re able to stand up campaigns and cycles, et cetera, et cetera. When we’ve asked questions like this in the past, what we often heard from
86
00:17:26.359 –> 00:17:42.160
Jonathan Margulies – Winterberry: brands was that, you know, they were optimistic about the future of AI, but that they hadn’t yet been able to leverage the technology to generate meaningful returns, particularly at scale, and particularly across both effectiveness and efficiency levers.
87
00:17:42.390 –> 00:17:54.840
Jonathan Margulies – Winterberry: That’s changing. It’s changing across media channels. It’s changing in a particularly pronounced way within individual channels, like direct mail that are well-established.
88
00:17:54.850 –> 00:18:05.529
Jonathan Margulies – Winterberry: where there’s perhaps a pretty clear north store for how the media is used and how people can put it to best effect. But where there’s also, I think, a tremendous amount of opportunity to drive
89
00:18:05.530 –> 00:18:18.890
Jonathan Margulies – Winterberry: innovation and better harmonize, the effort, along with how other media are, are deployed. And once again, with sort of an eye on supporting effectiveness and efficiency.
90
00:18:18.900 –> 00:18:35.489
Jonathan Margulies – Winterberry: Our panel told us that that’s happening today, that that’s happening in widespread respect, and there’s a tremendous optimism in a sense, that they’re going to continue to leverage those kind of initiatives in the foreseeable future.
91
00:18:35.830 –> 00:18:41.869
Jonathan Margulies – Winterberry: One of the… one of the, you know, along the same lines, one of the… the most sort of, like, startling,
92
00:18:41.980 –> 00:18:46.099
Jonathan Margulies – Winterberry: Findings from the research for us was… was just how much
93
00:18:46.340 –> 00:18:59.780
Jonathan Margulies – Winterberry: pure incremental value, that is to say, additive value that marketers can attribute directly to their deployment of new agentic and generative tools, is actually being realized today.
94
00:19:00.120 –> 00:19:12.710
Jonathan Margulies – Winterberry: we do a lot of surveying on a lot of topics in marketing circles here. It’s very, very rare to see 93, 94% of a panel agree with anything, but in this case, 93%
95
00:19:12.710 –> 00:19:26.610
Jonathan Margulies – Winterberry: plus percent folks are telling us they’re seeing incremental value, through the AI deployment to direct mail. It means something real is very, very happening. And it’s happening across a lot of different disciplines. It’s happening by way of
96
00:19:26.610 –> 00:19:41.779
Jonathan Margulies – Winterberry: better targeting. It’s happening by way of creative optimization. It’s happening because brands are helping streamline cumbersome workflows and clean up organizational processes, harmonize efforts with other channels.
97
00:19:41.780 –> 00:19:48.250
Jonathan Margulies – Winterberry: a host of different ways. There is, as yet, no, if you will, killer app, no single application.
98
00:19:48.250 –> 00:20:03.299
Jonathan Margulies – Winterberry: that AI is proving uniquely adept at solving for in a way that no other technology can address. Instead, brands are saying that this is technology because it offers so much that can do a great many things.
99
00:20:03.350 –> 00:20:13.470
Jonathan Margulies – Winterberry: If there was, though, a discipline that did sort of represent the lion’s share of the focus and attention in these early days of adoption, it’s probably around the discipline of data.
100
00:20:13.560 –> 00:20:31.890
Jonathan Margulies – Winterberry: Leveraging the technology to better wrangle the voluminous array of audience data that is available, becoming available on a regular basis, both from more traditional terrestrial sources and is generated from behavioral indicators online and things of that nature.
101
00:20:31.950 –> 00:20:50.819
Jonathan Margulies – Winterberry: It’s, it’s being leveraged to support measurement and attribution use cases. It’s being leveraged to help improve data quality, segmentation, and do a lot of other things that are not particularly new as we think about how direct mail programs are built, constituted, and deployed.
102
00:20:51.270 –> 00:20:57.979
Jonathan Margulies – Winterberry: Limited as well as large scale. But it’s helping move these processes along substantially faster.
103
00:20:58.080 –> 00:21:15.659
Jonathan Margulies – Winterberry: With greater granularity, and with greater, accuracy, quality levels, that, once again, you kind of see naturally help support both effectiveness and efficiency outcomes that really do move the needle in meaningful ways for, for brands of all shapes and sizes.
104
00:21:16.160 –> 00:21:27.219
Jonathan Margulies – Winterberry: That movement up the maturity curve, that, like, substantial advancement from an era of experimentation to one where we’re actually leveraging this technology.
105
00:21:27.260 –> 00:21:34.999
Jonathan Margulies – Winterberry: As sort of a business-as-usual, in-flight, resource, is really, really important.
106
00:21:35.000 –> 00:21:50.120
Jonathan Margulies – Winterberry: And it’s important as well because it kind of speaks to what the long-term pivot in performance circles is, the long-term role of the technology happens to be. And that is the ability of brands to leverage
107
00:21:50.150 –> 00:22:03.100
Jonathan Margulies – Winterberry: these tools as a mechanism to help essentially bring media together, to realize the opportunity sort of inherent in omni-channel marketing. That is to say.
108
00:22:03.160 –> 00:22:11.170
Jonathan Margulies – Winterberry: The orchestration of many different media channels, the deployment of different sources of audience data, so that
109
00:22:12.100 –> 00:22:14.139
Jonathan Margulies – Winterberry: Programs can be optimized.
110
00:22:14.550 –> 00:22:23.649
Jonathan Margulies – Winterberry: Audiences can be engaged with the… with truly the right message at the right time, to the right… to the right audience member.
111
00:22:23.950 –> 00:22:37.930
Jonathan Margulies – Winterberry: And that, business results can be optimized, not just for the brand itself, but so in service of delivering a better experience to the customer. And so when we asked our panel, like, are you bullish on that? Do you see…
112
00:22:37.930 –> 00:22:51.370
Jonathan Margulies – Winterberry: That, that effectiveness, that efficiency, that harmonization, that optimization as something that is going to continue to pace your priorities and drive the way you allocate budgets and resources internally.
113
00:22:51.450 –> 00:23:08.609
Jonathan Margulies – Winterberry: The answer to that was… was pretty resounding. There’s a wide level of expectation that those improvements are coming, that more use cases are growing addressable to AI in the future, and that for many brands, that continued sort of march up the maturity curve.
114
00:23:08.610 –> 00:23:16.449
Jonathan Margulies – Winterberry: you know, from a world where we were just a few years ago, where brands were really thinking about AI as a tool to take out some
115
00:23:16.520 –> 00:23:35.350
Jonathan Margulies – Winterberry: Costly workflow issues, or avoid certain workarounds in their process, eliminate duplicative manual effort and the like, to a world where we were focused on sort of a set of narrow creative optimization to where we are today, which is really all about putting data to work and empowering how it’s used.
116
00:23:36.180 –> 00:23:55.210
Jonathan Margulies – Winterberry: So many brands now have, in their line of sight, a future where AI is used as a tool to really drive omnichannel orchestration, right? To better put to work many channels, both paid and owned, in the service of delivering better experiences, more impactful offers, and ultimately.
117
00:23:55.210 –> 00:23:58.909
Jonathan Margulies – Winterberry: Helping optimize how marketing dollars are spent and how marketing
118
00:23:58.980 –> 00:24:04.460
Jonathan Margulies – Winterberry: Experiences are realized and valued by, by customer audiences.
119
00:24:04.790 –> 00:24:15.050
Jonathan Margulies – Winterberry: There is… there is a great deal more content in our paper, which, as I mentioned before, you’re welcome to come and download.
120
00:24:15.050 –> 00:24:31.650
Jonathan Margulies – Winterberry: when it’s convenient. There’s much as well to discuss, with our panel over the next few minutes from these themes and many others. And so, I am going to pause at a moment and turn the microphone back to my friend Ray, and I think we’re going to have a good conversation around so many of them.
121
00:24:33.380 –> 00:24:57.980
Ray Van Iterson: Thank you very much, Jonathan. That was incredibly interesting, and I, you know, can’t wait to have the opportunity to really dive into it more. Now what I’d love to do is to add a couple more folks, to this conversation. I feel honored to introduce Brad Coogler, who’s the CEO of Direct Mail 2.0, and Brett Fox, the VP of Marketing for Post
122
00:24:57.980 –> 00:25:18.199
Ray Van Iterson: pilot, who are, as you said, the people who are dealing with this on an everyday basis. And so, just to get things going, Brad, could you start by introducing yourself and telling us about your relationship with AI? And then, Brett, if you could please, follow up afterwards.
123
00:25:18.520 –> 00:25:25.220
Brad Kugler – DM2.0: Sure thing. Thanks, Ray. Happy to be here. I don’t know if it was intended, but I’m… okay, you were sharing your screen still.
124
00:25:25.980 –> 00:25:37.690
Brad Kugler – DM2.0: I’m Brad Coogler, I’m the co-founder and CEO of Direct Mail 2.0. We have 3 platforms that enhance direct mail. Direct Mail 2.0, which is our legacy platform.
125
00:25:37.950 –> 00:25:47.070
Brad Kugler – DM2.0: who’smailingWhat.com, which is a competitive intelligence platform where we pick up thousands and thousands of pieces of direct mail, and
126
00:25:47.070 –> 00:26:04.860
Brad Kugler – DM2.0: share that information with advertisers and direct mail agencies and manufacturers. And then our latest is direct mail, DM2.ai, where we actually take the data from both platforms, and we’ve built a propensity model to predict how
127
00:26:05.270 –> 00:26:12.920
Brad Kugler – DM2.0: how your mail will do based on past results, and we just launched that this past spring out of beta, and
128
00:26:13.060 –> 00:26:23.879
Brad Kugler – DM2.0: It’s very exciting what we can do. We’re learning a lot. We are not, Sam Altman here, but we’re doing what we can. So, happy to talk about that, but,
129
00:26:24.070 –> 00:26:25.360
Brad Kugler – DM2.0: That’s what we do.
130
00:26:26.550 –> 00:26:27.480
Ray Van Iterson: Right. Brett?
131
00:26:27.480 –> 00:26:31.720
Brett Fox – PostPilot: Yeah, nice to meet you all. I’m Brett Fox, VP of Marketing over at Postpilot.
132
00:26:31.760 –> 00:26:46.959
Brett Fox – PostPilot: A little background on me. My job is more than just marketing. It’s really on the go-to-market and operations side, so I spend most of my day working with AI tools and figuring out how to implement them in our own business, as well as selling a product that layers on top of AI.
133
00:26:46.960 –> 00:27:04.099
Brett Fox – PostPilot: So it’s very much just the ins and outs of my day. My background is demand gen and performance marketing, so I’ve spent a lot of my days in meta ads and LinkedIn ads and all those different channels, kind of dealing with the same challenges other folks have.
134
00:27:04.100 –> 00:27:21.449
Brett Fox – PostPilot: And one of the reasons I ended up here at PostPilot is the value is just very clear and obvious to me as we continue to run into a commoditized digital world with AI. Physical is becoming more and more important, more and more valuable as AI continues to commoditize
135
00:27:21.450 –> 00:27:34.020
Brett Fox – PostPilot: digital, so I think that there’s a really great opportunity for physical to bridge that gap, and ultimately create unique experiences with, with direct mail specifically. On the post-pallot side.
136
00:27:34.100 –> 00:27:52.539
Brett Fox – PostPilot: like I mentioned, we do have some AI products. Acquisition AI launched earlier this year. That is intelligent targeting based on your brand’s customers, and who’s bought from you, and the most recent purchasers and site visitors, in order to target your best new and future… and previous customers.
137
00:27:52.630 –> 00:28:05.800
Brett Fox – PostPilot: And in addition to that, benchmark recommendations with AI, sort of similar to what Brad mentioned, but looking at benchmarks from folks in your space and creating recommendations for your campaigns, to ultimately optimize those over time.
138
00:28:06.870 –> 00:28:10.460
Ray Van Iterson: Thank you both. So let’s get things going.
139
00:28:10.460 –> 00:28:30.380
Ray Van Iterson: Jonathan, one of the things that I was, impressed by or influenced by when I saw your research was 71% of respondents have increased their use of performance marketing in recent years. With the AI boom, do you think that that’s going to accelerate or slow that, and where would that shift come from?
140
00:28:31.950 –> 00:28:51.520
Jonathan Margulies – Winterberry: I do, and I guess this is a bit of a chicken or the egg question, Reagan, like, do brands increase their focus on performance because AI allows them to do it, or do other folks inside the enterprise, let’s say the CFO, folks in the finance suite, folks in the
141
00:28:51.560 –> 00:29:01.320
Jonathan Margulies – Winterberry: CEO Suite, essentially drive, you know, continued emphasis on performance, because the understanding is that these kind of tools allow for
142
00:29:01.320 –> 00:29:15.290
Jonathan Margulies – Winterberry: measurement, optimization, and enhancements, they allow for accountability, right, in marketing investment that wasn’t necessarily possible beforehand. I don’t know that you need to have necessarily a…
143
00:29:15.290 –> 00:29:30.050
Jonathan Margulies – Winterberry: The black and white answer to that question to reach the same conclusion, which is that, hey, at the end of the day, if we have the tools that allow for better, faster, and over time, more accurate optimization of how we’re putting resources to work.
144
00:29:30.530 –> 00:29:35.230
Jonathan Margulies – Winterberry: we’re going to lean into that. We’re going to continue to do that work to help
145
00:29:35.230 –> 00:29:53.190
Jonathan Margulies – Winterberry: substantiate and concentrate on what works, and we’re going to allow the marketer to be successful. I think the other interesting thing that this kind of technology does, and I want to be clear, too, AI offers a great many things, not perfect technology, right? So much of…
146
00:29:53.210 –> 00:30:08.720
Jonathan Margulies – Winterberry: What makes it unique from predecessor tools is that learning capability, the ability of the tools to essentially learn from their own experience in a way that is traditionally, you know, been part of human intelligence.
147
00:30:08.820 –> 00:30:25.259
Jonathan Margulies – Winterberry: But I think in this case, what the tools do is they allow all sorts of media and touchpoints that traditionally haven’t been thought of as performance addressable, or capable of driving that kind of performance, to
148
00:30:25.310 –> 00:30:39.750
Jonathan Margulies – Winterberry: to be accountable in all sorts of new ways. So, long answer, but in short, yes, the ability to optimize, for sure, is going to have more enterprises looking to shift
149
00:30:39.750 –> 00:30:50.040
Jonathan Margulies – Winterberry: Dollars towards those channels and those tools, in the… in the name of driving better performance, but it’s also going to raise the bar for all sorts of other media channels.
150
00:30:50.060 –> 00:31:05.909
Jonathan Margulies – Winterberry: to become accountable in their own ways. I wonder if we have this conversation two years down the road, whether we can see, the clear-cut distinction between that which is performance and that which is maybe not so performance that we can still have today. It’s a good, it’s a good question.
151
00:31:06.500 –> 00:31:14.570
Ray Van Iterson: Perfect. Brad, in, you know, for your own… within your own company and your clients, are you seeing more of a shift to performance marketing with AI?
152
00:31:15.840 –> 00:31:20.070
Brad Kugler – DM2.0: Absolutely. I think that those type of specifics are…
153
00:31:20.240 –> 00:31:31.900
Brad Kugler – DM2.0: it’s easier to get more granularized now with AI. There’s so many tools out there that can run instant modeling on who would be the best segments, based on intent, based on…
154
00:31:31.940 –> 00:31:49.349
Brad Kugler – DM2.0: browsing history based on past responses to ads. This is something that was really only available to, say, maybe the Fortune 100, Fortune 500 companies, because the data needed and the ability to sort of assemble it for use
155
00:31:49.410 –> 00:32:06.170
Brad Kugler – DM2.0: was really restricted to a much larger set of advertisers who had the resources, whereas I think AI makes it more available to almost everyone. Even small businesses now, through some of the smaller AI tools, have an ability to segment and
156
00:32:06.430 –> 00:32:16.710
Brad Kugler – DM2.0: market directly to those who would be most interested. So, I think it’s becoming more… more widespread just because of the ease of use and the spread of the technology.
157
00:32:17.270 –> 00:32:19.290
Ray Van Iterson: Makes sense. Brett, what do you think?
158
00:32:19.290 –> 00:32:38.160
Brett Fox – PostPilot: Yeah, no, I would second that. In our case in particular, it’s allowed us to help build that prospecting engine. So, direct mail, historically, people would look at these campaigns as one-off campaigns. You’d send a huge blast out to several hundred thousand prospects, and you’d hope for a return. You’d wait 6 to 8 weeks to get that data back.
159
00:32:38.160 –> 00:32:40.540
Brett Fox – PostPilot: And there’d be a lot of guessing involved.
160
00:32:41.140 –> 00:32:58.169
Brett Fox – PostPilot: AI has helped us build out Acquisition AI, as I mentioned before, which has ultimately allowed us to basically create kind of what you’d expect meta to be, but for direct mail, so now you give us a daily budget. We go and find your best customers, and we optimize day after day, or week after week, based on how you set your budget.
161
00:32:58.170 –> 00:33:10.320
Brett Fox – PostPilot: In order to optimize towards those customers that are buying. So using holdouts and incrementality reporting, we can see, here’s who’s buying from you, let’s go find more people that look like that and improve that audience over time.
162
00:33:10.380 –> 00:33:17.089
Brett Fox – PostPilot: So yeah, people are doing it, and I think, Brad, something you mentioned is, yeah, this used to be available to the Fortune 100.
163
00:33:17.180 –> 00:33:31.560
Brett Fox – PostPilot: that was kind of PostPath’s mission to begin with, was helping make the direct mail process more accessible to smaller companies. So we… our founders came from the D2C space. They ultimately wanted to make that process a lot more simple.
164
00:33:31.560 –> 00:33:45.609
Brett Fox – PostPilot: So that started with having a print facility and creating an account manager system to be able to make that process faster, but now it’s turning into bringing in AI tools to help also push that to kind of the smaller companies so they can run more powerful campaigns.
165
00:33:45.640 –> 00:33:50.700
Brett Fox – PostPilot: So yeah, I think that’s exactly what we’re trying to deliver, and…
166
00:33:50.970 –> 00:33:53.760
Brett Fox – PostPilot: I think that’s what we’ve achieved so far.
167
00:33:54.500 –> 00:34:19.030
Ray Van Iterson: Perfect, thanks. The other key stat that really amazed me was, Jonathan, that the 93% of companies that you’re seeing that are, already receiving incremental value from it, you know, I think a couple years ago, when we saw the stats, it was like, you know, well, we’re testing it, but the bleeding edge, you know, it’s not making money yet. What do you think the main shift has been that allows people to now be receiving
168
00:34:19.060 –> 00:34:20.620
Ray Van Iterson: value today.
169
00:34:20.800 –> 00:34:31.689
Jonathan Margulies – Winterberry: I mean, there… there’s… there are at least three, sort of, core… core factors there. The first, of course, is just experience. Trial and error. Much as we saw in, sort of, like, past
170
00:34:31.690 –> 00:34:40.280
Jonathan Margulies – Winterberry: paradigm shifts in this world, right? Like, big data. When your CEO, your CMO, your CFO is barking at you to have a strategy.
171
00:34:40.280 –> 00:34:52.240
Jonathan Margulies – Winterberry: to put big data to work, or in this case, put AI to work, yet the incentive is pretty strong to get out there and test, and learn from those tests around where there’s value to be realized, and where there’s
172
00:34:52.239 –> 00:34:56.520
Jonathan Margulies – Winterberry: Perhaps more… more smoke than fire out there, and marketers are doing that.
173
00:34:56.590 –> 00:35:12.999
Jonathan Margulies – Winterberry: The second, I think Brett and Brad could speak to this far better than me, which is to say there’s been a huge groundswell of investment and focus from the technology community, the supplier community, to stand up products and solutions that support… that are AI
174
00:35:13.820 –> 00:35:20.469
Jonathan Margulies – Winterberry: centered, that are built on AI technology, but are focused on outcomes, focused on very specific applications.
175
00:35:20.570 –> 00:35:28.589
Jonathan Margulies – Winterberry: Acquisition programs, better measurement and attribution, better coordination of, of,
176
00:35:28.850 –> 00:35:40.350
Jonathan Margulies – Winterberry: Targeted mail, let’s say, in conjunction with digital targeting and the like, that are real easy for marketers to test, to stand up quickly and understand what works and what doesn’t.
177
00:35:40.350 –> 00:35:50.149
Jonathan Margulies – Winterberry: And then finally, these tools themselves, and by tools, I mean AI platforms, the LLMs, the chat GPTs, the quads, the perplexities of the world, and the like.
178
00:35:50.150 –> 00:35:59.690
Jonathan Margulies – Winterberry: Increasingly are serving as platforms for a lot of custom development inside brands. You know, businesses that are investing in the development of custom agents.
179
00:35:59.690 –> 00:36:12.250
Jonathan Margulies – Winterberry: that lean on their own organizational resources, information, and otherwise to support highly defined use cases. We’re still sort of early in the game, but virtually every enterprise that I know
180
00:36:12.250 –> 00:36:28.010
Jonathan Margulies – Winterberry: has been investing real money to go stand up their own systems that help support individual applications, and that’s driving higher-level sophistication, but also real-world results that I think you see, you know, represented in the survey feedback.
181
00:36:28.710 –> 00:36:35.480
Ray Van Iterson: All right. Brad, what are you guys seeing in there? Are you… are you… is there that shift in performance and effectiveness?
182
00:36:36.050 –> 00:36:54.440
Brad Kugler – DM2.0: You know, it takes more… it takes more time to get that data back, and I’ll be honest, we’re working a reseller model in terms of… through our printers. I’m not privy to their exact ROAS, because our customers will resell our product, mark it up, and
183
00:36:54.850 –> 00:37:04.580
Brad Kugler – DM2.0: we don’t have a lot of direct communication with the clients, the advertisers, so I’m probably not a good one to answer that. I could only answer it through…
184
00:37:04.790 –> 00:37:20.270
Brad Kugler – DM2.0: repeated mail or marketing, but again, that’s not solid enough data, but let’s put it this way. The things that we are doing cannot be hurting, okay? At the… at the worst case, maybe they’re not dramatically effective, but they sure…
185
00:37:20.270 –> 00:37:31.500
Brad Kugler – DM2.0: do not take away from the effect of mailing. If anything, it could only enhance it. Without doing any of it, worst case, you stay the same, is just my logic-based response to that.
186
00:37:31.780 –> 00:37:36.660
Ray Van Iterson: Absolutely. And Brett… Brett, what about your internal and client results?
187
00:37:36.660 –> 00:37:42.070
Brett Fox – PostPilot: Yeah, could you… could you give me that question again? I just wanted to make sure that I interpreted what John said correctly.
188
00:37:42.490 –> 00:37:55.260
Ray Van Iterson: Yeah, no, so a couple years ago, when people were looking at what their returns were from AI, a lot of the early returns were, yeah, we’re not making money on this. 93% of the people in the survey now say they are, so what’s changed?
189
00:37:55.750 –> 00:38:02.449
Brett Fox – PostPilot: Yeah, no, I think the tools are getting better, and I… again, I’m working in these daily, and like.
190
00:38:02.540 –> 00:38:17.049
Brett Fox – PostPilot: you’re constantly getting new models that come out, you’re constantly finding people that are learning how to use these tools better, and companies are hiring folks to go work with these tools specifically to figure out how to improve their effectiveness in the company. That’s, again, a lot of what I’m doing on a day-to-day basis.
191
00:38:17.110 –> 00:38:26.999
Brett Fox – PostPilot: So I think it’s becoming familiar with this new, like, in a way, new employee that you’re working with, and figuring out how to leverage it throughout the business most effectively.
192
00:38:27.000 –> 00:38:38.860
Brett Fox – PostPilot: And the improvements of the models over time. So, yeah, I think that totally resonates from an improvement and effectiveness perspective, and like I said, we’re kind of dogfooding when it comes to that.
193
00:38:39.180 –> 00:38:41.040
Ray Van Iterson: Makes sense. Yeah, so…
194
00:38:41.040 –> 00:39:03.119
Ray Van Iterson: Let’s drill down just a little bit further. Let’s, you know, one of the key benefits that Jonathan was talking about was sort of the ability to use… wrangle data and target more. Brad, can you talk a little bit more about, you know, exactly what you’re doing on a day-to-day basis, what people are now able to do, in terms of the targeting, and how that’s changing?
195
00:39:04.420 –> 00:39:05.800
Brett Fox – PostPilot: You said Brett, right.
196
00:39:05.820 –> 00:39:06.570
Ray Van Iterson: Yeah, you…
197
00:39:06.570 –> 00:39:07.109
Brett Fox – PostPilot: Sorry, the chief.
198
00:39:07.110 –> 00:39:08.479
Brad Kugler – DM2.0: It sounds very similar.
199
00:39:08.480 –> 00:39:09.870
Brett Fox – PostPilot: Yeah.
200
00:39:09.870 –> 00:39:11.400
Brad Kugler – DM2.0: I wonder, too.
201
00:39:11.400 –> 00:39:15.670
Brett Fox – PostPilot: That’s what I was alluding to before with Acquisition AI. So…
202
00:39:15.810 –> 00:39:25.199
Brett Fox – PostPilot: Again, I… before, people would be purchasing datasets from an outside company, and that company would go and try to curate some sort of audience to be able to send to.
203
00:39:25.330 –> 00:39:38.879
Brett Fox – PostPilot: Now, you have the ability to go use your own company’s purchase data, so whether you’re a Shopify brand or what have you, that data is available to you to go create audiences from that’s going to be best
204
00:39:39.030 –> 00:39:43.229
Brett Fox – PostPilot: Built around your own set of… performance, so…
205
00:39:43.260 –> 00:39:58.039
Brett Fox – PostPilot: you’re now more in control than you’ve ever been. You don’t have to rely on some outside party doing some analysis. You can go and get that feedback immediately. You can go build those audiences within a few minutes, or maybe an hour, with some back and forth.
206
00:39:58.040 –> 00:40:08.819
Brett Fox – PostPilot: So again, with Acquisition AI, in our case, it’s always constantly trying to improve day after day. After you get a purchase, you’re gonna get that signal back into the system through Shopify or your platform.
207
00:40:08.820 –> 00:40:12.699
Brett Fox – PostPilot: We’ll receive that purchase, we’ll see the AOV,
208
00:40:12.700 –> 00:40:22.529
Brett Fox – PostPilot: We’ll see the purchase value, and you’ll be able to optimize towards that over time, just again, like you would with Meta or a different paid… paid digital channel. And that’s powered by AI. So…
209
00:40:23.120 –> 00:40:39.359
Brett Fox – PostPilot: I think that’s kind of just the beginning of it. It’s gonna become more powerful and more predictive as you get more data, so I think that’s maybe the takeaway I would have, is be patient with it, too. Like, you’re not gonna necessarily launch and have unbelievable results right away.
210
00:40:39.360 –> 00:40:46.849
Brett Fox – PostPilot: But launch, get some feedback, get some feedback, get some feedback, and that… those results are gonna pile up over time, with the help of
211
00:40:46.850 –> 00:40:48.619
Brett Fox – PostPilot: the predictive analytics and AI.
212
00:40:49.190 –> 00:41:03.970
Ray Van Iterson: And, Jonathan, with that improvement in the ability to target, are you seeing any changes in terms of how companies are using different channels, or, you know, approaching omnichannel in a different way?
213
00:41:04.710 –> 00:41:17.399
Jonathan Margulies – Winterberry: Well, they’re thinking about omnichannel as something that’s more than theoretical. I think that’s the biggest change in the environment right now, and if you ask about how it’s actually being operationalized on a regular basis.
214
00:41:18.250 –> 00:41:24.920
Jonathan Margulies – Winterberry: orchestration is a word we use a lot, right? We use a lot to describe more… something that is still…
215
00:41:24.920 –> 00:41:40.820
Jonathan Margulies – Winterberry: more vision than reality. The end goal essentially being, hey, can I maintain close to real-time visibility into how our channels are performing, how we’re load balancing our dollars and our efforts across all these audiences, across all these channels?
216
00:41:40.840 –> 00:41:49.999
Jonathan Margulies – Winterberry: And then optimize as close to real time as we possibly be, so that we, again, we hit the right customers when they’re
217
00:41:50.000 –> 00:42:09.639
Jonathan Margulies – Winterberry: then they’re in a position to make a purchase decision, and we really optimize how we’re putting our dollars to work. That’s beginning to happen. It’s beginning to happen, I think, first and foremost, rather than from an execution standpoint, from a measurement and attribution standpoint, right? The notion being that it sort of starts with understanding
218
00:42:09.640 –> 00:42:18.380
Jonathan Margulies – Winterberry: The incremental impact of each dollar of spend, and if we understand that, then we have the platform of intelligence to go and actually drive that next
219
00:42:18.380 –> 00:42:24.629
Jonathan Margulies – Winterberry: spend decision. And so we… we see that, we see that in the emergence of… of… of…
220
00:42:24.860 –> 00:42:26.020
Jonathan Margulies – Winterberry: different…
221
00:42:26.050 –> 00:42:35.299
Jonathan Margulies – Winterberry: attribution tools, tools that are really focused on unified experience. We see it a lot from the perspective of creative, increasingly.
222
00:42:35.330 –> 00:42:40.969
Jonathan Margulies – Winterberry: And… and the… the goal for many brands to advance beyond, sort of.
223
00:42:41.000 –> 00:42:52.630
Jonathan Margulies – Winterberry: basic, sort of, first-generation approaches to personalization or variation, the ability to test simple, simple creative concepts, A to B,
224
00:42:52.630 –> 00:43:05.610
Jonathan Margulies – Winterberry: towards a much more holistic approach built around creative intelligence. This is to say, if AI, like all these other things, allow us to develop deeper insight into how
225
00:43:05.890 –> 00:43:25.220
Jonathan Margulies – Winterberry: different media touchpoints, offers, creative performs, then we can optimize that. And so, I think those two dimensions right now, the measurement and creative intelligence, are where we’re seeing the most movement in terms of actually
226
00:43:25.670 –> 00:43:33.310
Jonathan Margulies – Winterberry: acting in an omni-channel way. There are all sorts of other orchestration opportunities, but I think that those are still a little bit down the road for most brands.
227
00:43:33.310 –> 00:43:56.769
Ray Van Iterson: Yeah, I think, you know, when I’ve been talking to people, what I’ve been hearing is that, you know, in many cases, if you don’t know who your audience is, you have to, you know, do spray and pray, you know, just broad brand analysis, whatever may be the least expensive. And now, when people are able to identify those, you know, who are the true folks who are likely to respond
228
00:43:56.770 –> 00:44:06.149
Ray Van Iterson: Companies are able to really focus on that impact and targeting omni-channel messages right to that person, because they know that that’s where the value.
229
00:44:06.150 –> 00:44:06.680
Jonathan Margulies – Winterberry: that’s…
230
00:44:06.680 –> 00:44:07.200
Ray Van Iterson: is.
231
00:44:07.200 –> 00:44:07.760
Jonathan Margulies – Winterberry: That’s exactly it.
232
00:44:08.760 –> 00:44:22.729
Ray Van Iterson: Now, Brad, I know one of the things that you guys are really doing a lot with AI on is on the creative side. Can you talk a little bit about, you know, how AI is helping you and your clients on really optimizing messaging and look and feel?
233
00:44:22.950 –> 00:44:37.530
Brad Kugler – DM2.0: Yeah, so I know we’ve spent most of the time on targeting, and I want to focus a little bit on the creative side. Having access to over half a million pieces of creative across some 80,000 campaigns…
234
00:44:37.640 –> 00:44:49.419
Brad Kugler – DM2.0: We’ve actually tracked mail to 2 billion pieces of mail to almost, I think, every household in the United States. So what we’ve been able to do is digitize the creative.
235
00:44:49.590 –> 00:44:52.219
Brad Kugler – DM2.0: And then correlate that to.
236
00:44:53.050 –> 00:45:05.299
Brad Kugler – DM2.0: engagement. So, did certain aspects of the creative, or the keywords, or the text in that creative, tend to lift the engagement, or was the engagement depressed at a benchline? So.
237
00:45:05.570 –> 00:45:11.030
Brad Kugler – DM2.0: We think that’s invaluable to helping people drive and create better creative. So.
238
00:45:11.030 –> 00:45:28.560
Brad Kugler – DM2.0: Combining the targeting aspect, which is also key, to the creative aspects of what actually caused people to engage. Again, I don’t necessarily know the ROI, but I know if they engaged. Did they go to a website? Did they click? Did they scan a QR code? That’s telling data.
239
00:45:28.560 –> 00:45:31.530
Brad Kugler – DM2.0: And we have that across 160 verticals.
240
00:45:31.540 –> 00:45:44.350
Brad Kugler – DM2.0: So, essentially, what we’ve done is we’ve made buckets based on vertical, geography, type of mailing, and we’re able to regurgitate that in recommendation form.
241
00:45:44.350 –> 00:45:51.699
Brad Kugler – DM2.0: to the person prior to launching his campaign. So before he drops $5,000, $10,000, $15,000,
242
00:45:51.700 –> 00:45:58.089
Brad Kugler – DM2.0: Well, run it through this, let’s see what you’re planning to send, and who you’re gonna send it to, and when. Timing is key, too.
243
00:45:58.150 –> 00:46:11.790
Brad Kugler – DM2.0: And is there anything that can be done to increase that engagement rate by 1%? So, you open that engagement rate by even 1 or 2%, that affects the ROAS or the ROI.
244
00:46:11.790 –> 00:46:24.770
Brad Kugler – DM2.0: to an exponential degree down the line. So I’m looking for micro-improvements here. And again, this is something we’ve just been doing for the last 60 days in a commercial basis, beyond a
245
00:46:25.140 –> 00:46:27.369
Brad Kugler – DM2.0: a beta form, so…
246
00:46:27.480 –> 00:46:39.910
Brad Kugler – DM2.0: you know, I like to repeat what I’ve heard many times. The AI that we have right now is the worst AI it will ever be. So, as these data sets get bigger, and they’re learning and training
247
00:46:39.910 –> 00:46:49.479
Brad Kugler – DM2.0: sets get larger and iterate more quickly, this only proves to get better. Is it going to be an instant change between someone
248
00:46:49.480 –> 00:46:57.429
Brad Kugler – DM2.0: not making money on a campaign, or, you know, 10X-ing their ROI. I don’t know, that may be a bit of a…
249
00:46:57.450 –> 00:47:06.930
Brad Kugler – DM2.0: A high expectation, but… we will get incrementally better with every iteration, and additional data used for training. So…
250
00:47:07.160 –> 00:47:20.079
Brad Kugler – DM2.0: I look at this as an investment more long-term than an immediate ROI, you know. And again, what are we charging to model a campaign? It could be as low as $20. So, would you invest $20?
251
00:47:20.460 –> 00:47:28.210
Brad Kugler – DM2.0: before you’re gonna spend 5 or 10 grand, just to get an idea if you’re on the right track? I would, and that’s why we think the value prop is pretty good here.
252
00:47:29.160 –> 00:47:48.110
Ray Van Iterson: So, before I ask my one last question, this is going to be our last question of the roundtable. We’re about to throw it open to the Q&A. We’ve had a few questions already thrown in, but if you can send some more in, we’ll be on your questions in just a minute. So, my last question for the panel, for me, is that
253
00:47:48.140 –> 00:47:59.980
Ray Van Iterson: Jonathan’s last slide was showing that maturity, was saying that the future is going to be the ability to orchestrate everything, both omni-channel together. So I’m going to say that, you know.
254
00:48:00.050 –> 00:48:14.209
Ray Van Iterson: Brett, starting with you, if you were trying to prepare for the future, what should people be doing today so that they’re ready to be able to really take advantage of the next generation of AI, as Brad was just talking about?
255
00:48:14.620 –> 00:48:30.500
Brett Fox – PostPilot: Yeah, well, first, I know you had question marks on the… on that slide. I don’t think it’s that far away, like, I think… I think it’s gonna be here before we know it, which, Frank, again, as a performance marketer myself, like, it’s exciting. But what would I be… what I would be doing, and maybe what I am doing today,
256
00:48:30.600 –> 00:48:37.509
Brett Fox – PostPilot: I think it’s really important to be building out the foundational layer across your AI. And what I mean by that is.
257
00:48:37.760 –> 00:48:53.590
Brett Fox – PostPilot: Your design system, your documentation about your products, your… all the things that need to be standardized across your marketing and across your business need to be stored somewhere, and they need to be maintained, and they need to be reflected on.
258
00:48:53.590 –> 00:49:11.739
Brett Fox – PostPilot: I would spend time building that out. That way, as these programs progress, and as these tools progress, you’re gonna have a source of truth to be able to have them derive from. Otherwise, these tools like to drift, they call it, so they’ll start to pull from random sources, and they’ll make stuff up, and they’ll think they’re doing a good job.
259
00:49:11.780 –> 00:49:29.469
Brett Fox – PostPilot: But if you can be really critical and say, no, refer to this documentation, because it’s been blessed, then you’re going to end up having a really great foundation to be able to build from. So that’s what I would spend the time doing, and frankly, exactly what I’m doing on a day-to-day basis in a lot of cases. So, hopefully that’s a helpful tip.
260
00:49:30.160 –> 00:49:32.960
Ray Van Iterson: Okay. Brad, what do you think people should do to prep?
261
00:49:32.960 –> 00:49:36.329
Brad Kugler – DM2.0: I think it’s a great question, and
262
00:49:37.000 –> 00:49:45.489
Brad Kugler – DM2.0: I think Brett is on the right track, but I’ll go one step further. It’s the data layer that is the basis and foundation for all AI, and
263
00:49:45.750 –> 00:50:00.850
Brad Kugler – DM2.0: I see, if I’m looking 3 to 5 to 10 years down the line, you’re gonna have individual companies that have their data that must be standardized, normalized, and in a situation that can be open to AI. So then you have
264
00:50:01.190 –> 00:50:07.920
Brad Kugler – DM2.0: an application layer that may be a specific type of company that comes in and reads your data and
265
00:50:08.060 –> 00:50:13.540
Brad Kugler – DM2.0: Basically puts it into use for a specific application.
266
00:50:13.620 –> 00:50:30.400
Brad Kugler – DM2.0: in that application layer, or maybe in between the application and the data layer, you’ll have the LLMs, the big guys that we all know their names. Anthropic, OpenAI, Gemini. Those guys will be the compute
267
00:50:30.480 –> 00:50:41.049
Brad Kugler – DM2.0: between the application output and your own data layer. So, all of compute will be organized by a company’s own proprietary data.
268
00:50:41.100 –> 00:50:51.769
Brad Kugler – DM2.0: The compute layer that will probably fall to the top 3 or 4 names that are pretty much in the news now, and then the infinite amount of application layers for specific
269
00:50:53.720 –> 00:50:57.980
Brad Kugler – DM2.0: action items that you want to do. So if you’re going to be a guy who’s going to target
270
00:50:58.060 –> 00:51:15.199
Brad Kugler – DM2.0: creative, or you’re a guy that’s going to target specific types of segmentation, you’ll contact those companies, they will then, in turn, use these large LLMs, and they will connect seamlessly to your data. So, I would ensure that your data is either
271
00:51:15.310 –> 00:51:26.680
Brad Kugler – DM2.0: MCP, Alliant, which is what these large learning models use, and there are skills that will be connected through Claude or Codex.
272
00:51:27.020 –> 00:51:43.779
Brad Kugler – DM2.0: I believe there’s going to be 3 layers. The data layer, the LLM layer, the compute layer, and then the application layer. And that’s probably where things are going. So if you can organize your data layer and make it as normalized and accessible as possible, you’re halfway there.
273
00:51:45.130 –> 00:51:49.810
Ray Van Iterson: And Jonathan, what did the research say? What do you think that they should do to prep?
274
00:51:49.810 –> 00:52:09.390
Jonathan Margulies – Winterberry: I mean, I agree with everything these guys have said right now. The one sort of overarching thing I think I’d add to it is bear in mind, like, no matter where you see yourself as sitting on the maturity curve, and no matter how ambitious you may want to be towards chasing down that long-term
275
00:52:09.390 –> 00:52:21.639
Jonathan Margulies – Winterberry: omnichannel, orchestration, you know, this is the holy grail. We’re truly able to optimize across every dimension, every audience member, every media channel, every penny of our media spend.
276
00:52:21.700 –> 00:52:24.920
Jonathan Margulies – Winterberry: To optimize results, right?
277
00:52:25.230 –> 00:52:37.380
Jonathan Margulies – Winterberry: The technology can help do that faster, do it with greater accuracy, do it in a way that sort of learns from its experiences and even its mistakes to improve.
278
00:52:37.380 –> 00:52:46.390
Jonathan Margulies – Winterberry: But at the end of the day, those functions, you still need to be clear-eyed around the use cases, the media channels. What are you looking to accomplish?
279
00:52:46.390 –> 00:53:00.229
Jonathan Margulies – Winterberry: There’s not going to be one single tool, one single solution that you plug in, and the next day, you’re humming, right? It’s still up to the brand at the end of the day to take a good black and white look.
280
00:53:00.260 –> 00:53:17.559
Jonathan Margulies – Winterberry: at your resources, your channels, potential gaps in your current approach, and say, we need to think about, sort of, a phased approach, and what’s important to us to accomplish over time. And that’s where we need to gear our effort. Just, you know.
281
00:53:17.560 –> 00:53:30.240
Jonathan Margulies – Winterberry: we’ve made so much progress as a… as a marketing community, as a use… as a… as a tech… as a community of technology users over the last few years, but that’s… that hasn’t come in one fell swoop. It’s come…
282
00:53:30.240 –> 00:53:36.530
Jonathan Margulies – Winterberry: Based upon, you know, a series of phased initiatives and a whole lot of trial and error.
283
00:53:36.530 –> 00:54:00.629
Jonathan Margulies – Winterberry: And so I think the folks that will succeed are the folks that expect some of that trial and error, that are prepared to go about, sort of, the phased implementation, and are clear-eyed that the technology is going to serve to advance the use cases. The technology is not just going to be, you know, a single solution in a nicely packaged box. If you’re thinking about it that way, then it’s much easier to understand the role
284
00:54:00.630 –> 00:54:11.939
Jonathan Margulies – Winterberry: for the data, for the creative, for the optimization, the application tools, and the like, and to put it to work for, I think, pretty substantial impact sooner rather than later.
285
00:54:12.580 –> 00:54:19.050
Ray Van Iterson: Perfect. So, speed around, we’ll get to a couple of audience questions. First one is.
286
00:54:19.620 –> 00:54:27.940
Ray Van Iterson: Each of the companies are now offering, you know, multiple different variations in terms of the tools that they’re offering.
287
00:54:27.940 –> 00:54:44.289
Ray Van Iterson: which one… how should we approach trying to figure out which one should we be going with? Should we be going with, you know, the fanciest one that just got released, or which may be using up lots of tokens, or should we be using somewhere a little lower at the end? Brett Fox, what do you think about that one?
288
00:54:44.290 –> 00:55:00.400
Brett Fox – PostPilot: Yeah, so it depends on what you’re trying to achieve with it. And it, like, Claude is really good about kind of recommending, or at least telling you, here’s what this specific one’s for. So there’s Sonnet, there’s Opus, and the most recently, I’m forgetting the name, it was the part of the Mythic release, but,
289
00:55:00.400 –> 00:55:04.730
Brett Fox – PostPilot: My recommendation is you can tweak it based on if it’s a low…
290
00:55:04.730 –> 00:55:20.149
Brett Fox – PostPilot: energy task, if it’s something that’s pretty simple, repeatable, use a lower model, use some… use it to save some tokens. Now, you are going to have a different output if you were to go apply it to a newer or more powerful model, so you have to know that there’s going to be some changes.
291
00:55:20.240 –> 00:55:35.120
Brett Fox – PostPilot: That being said, I’m not changing the model I do for every little thing, and, like, maybe we’re not necessarily focusing on saving tokens, I think we’re more so trying to figure out how to make all this work in the best way possible.
292
00:55:35.220 –> 00:55:39.600
Brett Fox – PostPilot: So… I would lean towards trying to be more…
293
00:55:39.700 –> 00:55:57.440
Brett Fox – PostPilot: consistent and predictable, but if you are cost-conscious, then yeah, use a lower model and test that, and if you’re not getting the outputs that you want, then yeah, try to scale it up, and it’ll cost a little bit more, and you’ll be a little bit less efficient from a tokens perspective, but you’ll have a better output. So.
294
00:55:57.440 –> 00:56:11.010
Brett Fox – PostPilot: I think you have to play with it and see what makes most sense, but if you’re doing a small writing task, like Sonnet, for example, is probably fine, and a lower version of it, but if you’re doing something that’s ingesting a ton of data, or a larger exercise like that, then
295
00:56:11.010 –> 00:56:18.099
Brett Fox – PostPilot: Opus 4.8, the most recent one on Opus, is probably a better fit. So yeah, you’ll have to play with it and kind of find the right use cases.
296
00:56:18.610 –> 00:56:32.799
Ray Van Iterson: Okay, so I’m just going to limit one question per person, just so we can get through a few more. Brad Cugler, how should businesses responsibly draw the line between black box services and human in the loop as they move up the orchestration curve?
297
00:56:33.210 –> 00:56:35.369
Brad Kugler – DM2.0: Wow, difficult one.
298
00:56:36.680 –> 00:56:55.480
Brad Kugler – DM2.0: You know, and in the end, it comes down to cost and ROI. You know, what are you paying for the black box versus what are you paying for a human, and what is it returning? So, I think this is a trial and error thing. This is… we’re still relatively at the beginning of this adoption of this AI methodology, and…
299
00:56:55.840 –> 00:57:14.599
Brad Kugler – DM2.0: I would say, this is trial and error, A-B test. If you want to do automated targeting for new clients with a black box, and there’s lots of them out there, it’s almost overload every time I look on my screen, or you want to hire a couple of SDRs and have them bang the phones and the email and do it the old-fashioned way.
300
00:57:14.830 –> 00:57:16.980
Brad Kugler – DM2.0: I think you gotta try both.
301
00:57:17.310 –> 00:57:32.399
Brad Kugler – DM2.0: I don’t know that there’s a clear decision in that. And not every AI black box is equal to the other AI black box. There’s a lot of guys selling smoke and mirrors out there that throw the word AI on their black box that they’ve been using for 6 or 7 years, and they’ve added a little AI reporting.
302
00:57:32.400 –> 00:57:42.460
Brad Kugler – DM2.0: And I’ve seen a lot of that stuff going around, too. So, buyer beware, a lot of snake oil people out there, and they’re trying to make a quick buck and ride this wave. That’s what I found.
303
00:57:43.290 –> 00:57:58.400
Ray Van Iterson: And last question to Jonathan. Jonathan, within your presentation, you showed that the number of clicks that people are getting out of their paid searches dropping through the floor because of AI. What should marketers do in response to that?
304
00:57:58.930 –> 00:58:09.719
Jonathan Margulies – Winterberry: I mean, at the end of the day, you need to keep in mind you have one budget, you have one set of objectives, there may be some complexity to them, there may be multiple layers of them, but it…
305
00:58:09.800 –> 00:58:27.489
Jonathan Margulies – Winterberry: it’s incumbent upon you to recognize that the market conditions are changing, and it’s your obligation to optimize. AI can and ideally should, by virtue of its learning abilities, be a force for you to help reach decisions along those lines.
306
00:58:27.490 –> 00:58:30.740
Jonathan Margulies – Winterberry: Faster, in ways that help you optimize.
307
00:58:30.740 –> 00:58:32.020
Jonathan Margulies – Winterberry: But…
308
00:58:32.020 –> 00:58:41.159
Jonathan Margulies – Winterberry: But don’t take for granted the fact that market conditions are set, and they’re always going to be that way. If we were having a conversation
309
00:58:41.420 –> 00:58:57.079
Jonathan Margulies – Winterberry: 3 or 4 years ago about… about threats to… to paid search as a source of digital traffic, people would have probably said we were insane, just given how the trajectory and the dollars were flowing into that media for, you know, the better part of
310
00:58:57.260 –> 00:59:13.900
Jonathan Margulies – Winterberry: of 30 years already. And yet, here we are today, and what we hear from brands over and over and over again is, I need to rationalize. I need to take back more control over how I’m putting dollars to work to do all sorts of things from a marketing perspective.
311
00:59:13.900 –> 00:59:25.309
Jonathan Margulies – Winterberry: That’s… that’s the responsibility. AI, as we’re sort of defining it very, very generically here, can be a real force for helping you optimize and manage
312
00:59:25.310 –> 00:59:31.709
Jonathan Margulies – Winterberry: across a wide array of media. I think you should be doing that, but recognize that that landscape
313
00:59:31.710 –> 00:59:50.700
Jonathan Margulies – Winterberry: and the usefulness of tools along those lines is always going to be changing, and so managing them is now an active part of the marketer’s set of responsibilities, but it’s an important one in most enterprises, and thus, I think it’s something that most… most
314
00:59:50.700 –> 00:59:58.050
Jonathan Margulies – Winterberry: Brand managers, most marketing executives should embrace as part of their responsibility going forward.
315
00:59:59.210 –> 01:00:10.450
Ray Van Iterson: Thank you all, and, you know, I think you all… you all deserve a round of applause. I apologize that you’re not able to hear it from all of, you out there, but really appreciate it, Jonathan, Brad, and Brett.
316
01:00:10.600 –> 01:00:28.149
Ray Van Iterson: Last thing I just wanted to do is to tell you, the audience, if you enjoyed this, we’re going to have another webinar in a couple of weeks. This next one is a little bit different. This is about, how to prepare for the upcoming holiday season. We’re going to be talking about,
317
01:00:28.180 –> 01:00:51.749
Ray Van Iterson: impact and how you can use Omnichannel for Black Friday and Cyber Monday. We’re going to be talking about updates in the mailing promotions, including a discount for new mailers that we’re launching next year. Come and join us on Friday, June the 26th at 1 o’clock p.m. You can either scan the QR code, or I’m putting
318
01:00:51.930 –> 01:01:06.270
Ray Van Iterson: the link in the chat, you can click on there, and we will look forward to seeing you then. So, thanks again to our August panel, really appreciate it. I’ve learned a lot, and I hope that everybody else has too. Thanks again.
319
01:01:34.850 –> 01:01:36.039
Brad Kugler – DM2.0: Were you off the air?
320
01:01:38.360 –> 01:01:40.169
Ray Van Iterson: No, we’re still on…
321
01:01:42.550 –> 01:01:43.250
Brad Kugler – DM2.0: *****
322
01:01:44.010 –> 01:01:44.800
Ray Van Iterson: Nope.
323
01:01:47.800 –> 01:01:48.670
Steck Stoecklin – USPS: Excellent.
324
01:02:58.430 –> 01:03:00.880
Brad Kugler – DM2.0: Hey, Jonathan, are you able to share that deck?
325
01:03:01.520 –> 01:03:02.550
Jonathan Margulies – Winterberry: I’ll send it to you.
326
01:03:02.550 –> 01:03:05.040
Brad Kugler – DM2.0: Alright, thanks. It’s great. I mean…
327
01:03:06.020 –> 01:03:08.869
Brad Kugler – DM2.0: We’ll obviously give credit if we use anything.
328
01:03:09.190 –> 01:03:09.760
Jonathan Margulies – Winterberry: For sure.
329
01:03:10.280 –> 01:03:16.240
Brad Kugler – DM2.0: So, do we need to hang on here for some after wrap, or are we… Oh, okay.
330
01:03:16.240 –> 01:03:18.140
Ray Van Iterson: Give me one more sec.
331
01:03:18.530 –> 01:03:19.140
Brad Kugler – DM2.0: Okay.
332
01:03:50.730 –> 01:03:54.139
Jonathan Margulies – Winterberry: There’s a lot of downloading happening.
333
01:03:55.020 –> 01:04:06.810
Ray Van Iterson: I’ve never quite mastered the how do you end the webinar part, so I was going through one by one and uninviting people, which may not be the best of customer experience, but it did work, so…
334
01:04:06.810 –> 01:04:07.180
Brad Kugler – DM2.0: I mean…
335
01:04:07.180 –> 01:04:21.710
Ray Van Iterson: very much. I was… I was ruder today than I think I’ve ever been in terms of sending messages. Hey, you know, gonna be a little shorter kind of thing. But we… we had so much to try to get through, so I really appreciate you guys, taking the time. I thought there was awesome information.
336
01:04:22.250 –> 01:04:22.900
Brad Kugler – DM2.0: Good.
337
01:04:22.900 –> 01:04:23.800
Jonathan Margulies – Winterberry: Thank you, guys.
338
01:04:23.800 –> 01:04:24.410
Brett Fox – PostPilot: Thank you.
339
01:04:24.410 –> 01:04:36.389
Brad Kugler – DM2.0: It was great. So, yeah, I mean, even if you just end the webinar, it will continue to download and record. You don’t have to wait till everybody leaves the room, as far as I know.
340
01:04:36.390 –> 01:04:48.130
Ray Van Iterson: Yeah, I just wanted to give the other… yeah, no, I could have just ended it, but then… but, wanted to thank you guys in person, but I will now have to figure out how I can cut this from the end of the, of the recording, so…
341
01:04:48.130 –> 01:04:51.519
Brad Kugler – DM2.0: No problem. Are you gonna post this somewhere? Yeah.
342
01:04:51.520 –> 01:05:00.790
Ray Van Iterson: We will… we will post this. It’ll be on something called Postal Pro, that you can do it. I’ll send it, and there will be an email going out to everybody with the recording.
343
01:05:01.140 –> 01:05:06.889
Brad Kugler – DM2.0: Great, thank you very much. Alright guys, thank you for a good chat. I’ll see you guys.
344
01:05:06.890 –> 01:05:09.039
Jonathan Margulies – Winterberry: Take care, guys. Bye-bye.
