← Next in Media
Loading summary
Transcript83 lines
- [00:00]
A
I always like to start with think of this math problem. You're a media company and you have movies or shows that you're promoting, or you're a retail organization, you have products. You have to create X amount of assets in X amount of formats. And that can be video, that can be still imagery, that can be 3D times the amount of languages. And so there has to be some level of automation in that process and efficiency that's going to push us into that next stage. So that means today we're mostly reactive on the creative side. So we push it out. We really don't know how it's going to perform. One of the advantages we see from the Adobe side is everything in our models has been legally obtained. This can really improve the level of advertising, the relevance to the consumer, all the things we've talked about for years. I feel like with the creative piece that we talked about today and the data foundation that we already have, we're really moving to the direction we talked about. I don't see a world where a CMO just goes, oh, I'm just going to hand over my whole budget to net of. Let them figure it all out for me.
- [01:09]
B
Right. Tell the CFO like, I don't know. This week on Next Immediate, I had Sam Garfield, head of digital strategy at Adobe, on to talk about the future of the creative tech business. Sam and I discussed whether brands are really ready to let machines make thousands of ad variations for them, whether we're all going to start getting personalized messages at every moment. Sam and I also broke down what all these big shifts means for creative executives, media agencies and the ad tech landscape at large. Lots to get into. So let's get started. Hi, everybody. Welcome to Next to Media. I'm Mike Shields. My guest this week is Sam Garfield. He's the head of digital strategy for CMT Data and AI platforms at Adobe. Hey Sam, thanks for being here.
- [01:43]
A
Hey, Mike. Happy to do it.
- [01:46]
B
Psyched to talk to you because I think it's an interesting moment for Adobe because I think for those of us who cover who have been in the space for a long time, Adobe went hard into advertising. It seemed like about a decade ago with tubemogul and getting into ad tech buying. And then it seemed like maybe the company was going to pull out of the business to some degree. Obviously things, a lot of things have changed with AI. So kind of give me, give me like the state of affairs right now for Adobe in the digital ad world.
- [02:11]
A
Sure. You know, it's funny if you say Adobe from A consumer perspective, everybody immediately goes to PDF and photo. But you're right, Adobe has a long history with enterprises, specifically Fortune 500 and even some of the middle market where Adobe really does three different things for these enterprises. The first would be like the creative layer. So that's Creative Cloud. We have a foundational AI model called Firefly. The second area would be how do you orchestrate that content supply chain? Once you have the creative, we have tools like workfront and asset management, AM assets. And then the last piece is the data and experience layer. And so we have customer data, platforms, analytics. When you put all of those tools together, you essentially have an operating system for a marketer. And so that's really where Adobe has been leaning in is how do you put that all together? Especially now I'm starting to see the data and creative. I'm sure we're going to get into this coming together. And in that world, Adobe plays a huge role as powering that operating system. I can't think of another company that can go from the ideation of an asset, the creation of that asset, the variations that you need, but then the application of data, the activation of that asset into the channels and then the measurement. So you kind of see almost an end to end ability to operate. Or as you say in the past, maybe Adobe's made specific roles in specific areas. We might see it much more across.
- [03:54]
B
Yeah, well I was going to ask you because the list of products you mentioned, that feels like it's a lot of constituencies are always going to say like, and it's probably hard to answer this in a nutshell, but who is your customer? Is it the creative guy at big agency? Is it the. Do you just kind of work directly with the brand and kind of give them a full stack? Does it really vary?
- [04:13]
C
Like who do you, who do you work with?
- [04:15]
A
Yeah, it's a great question. It's all the above. We're definitely on the list for CMOs, so major brands, chief marketing office, marketing officer, organizations and then agencies as well. This week we announced a big, you know, expansion of our WPP relationship. Agencies are now looking at how they've been take advantage of this operating system as well. And then publishers, we work with all the publishers on various aspects of what I described. So when we used to use Adobe
- [04:51]
B
products at Adweek, when we were putting Adweek, the print publication out back then.
- [04:55]
A
Yeah, but when you have that many products, you're really across the ecosystem trying, I think about it as that operating layer. So it's not world guarding and it's not just a dsp, it's really across the entire operating system across a marketer, advertising.
- [05:12]
B
So this has been a subject for number of years like when it, you know, there have been companies that do creative optimization in the, in ad tech and they were, it was, it was kind of tied to the transaction world, buying and selling less like the, I think the brands and creative guys that seems to be shifting but like where are we as an industry overall when it comes to doing what we want to be doing? As there was a creative optimization in tech versus maybe a couple years ago.
- [05:40]
A
It's definitely the early days. I always like to start with think of this math problem. So you're a media company and you have movies or shows that you're promoting or you're a retail organization. You have products, right. You have to create X amount of assets in X times X amount of formats and that can be video, that can be still imagery, that could be 3D times the amount of languages. Most of these companies are global times the amount of channels. So you have to get on meta, you have to get on TikTok. They all have different size, aspect ratio
- [06:11]
B
always changing, always growing.
- [06:13]
A
Mike, think about that math formula on how many assets you have to produce just to maintain kind of the status quo. And so there has to be some level of automation in that process and efficiency that's going to push us into that next stage and we're there. That's where we're at now. We're is we're laying that foundation on how can we create asset variations and localization using generative AI or AI, right. Technology to be able to make that, to be able to create that foundational layer. I like to use the analogy to data because we all started, at least I did on the data side 10 years ago. Audiences remember how manual it was to pull together audiences and we were kind of testing them and we didn't really know what we were doing.
- [06:59]
B
And then yeah, trying to mix and match things, it didn't always work together.
- [07:02]
A
And we built the foundation and now it's like, okay, I can, you know, it's really running. You can have audiences in real time, you have multiple audiences. I think we're at step one on that creative side on laying that foundational layer that we're going to look back in three to five years and go, okay, we're there now and we. What being there means.
- [07:22]
B
Yes, because what it sounds like, and I don't mean to like reduce what you're describing, but it sounds like you're Starting with, let's just make production a lot easier getting this stuff made and out there. But where, I guess the. I think you're leaning towards a world where there's a lot more intelligence and automation of creative. You know, we dream about the, the world where there's five zillion ads and they're all being optimized at real time. Like where, where do you think things might be going?
- [07:48]
A
Look, I, I think, you know, there's, there's definitely an element of intelligence that you brought up that we should talk about. So that means today we're mostly reactive on the creative side. So we push it out. We really don't know how it's going to perform. Then we find out and maybe we make an adjustment, maybe we don't. There's going to be a lot of pre scored content intelligence.
- [08:10]
B
We're learning on the client's dime a lot of times. Right.
- [08:12]
A
So you can, you can move that process up now through some of the automation and tools to do kind of pre scoring. And then on the variation side, look, are we ever going to get to one, you know, 10,000 variations per campaign? I don't think so. I think it's going to be variations at scale and you're going to have that operating system that's going to have some level of modular assembly, predictive scoring and automation that's going to enable whatever the right number of versions for that asset is. And I don't know that yet.
- [08:45]
B
Yeah, there's a lot, obviously tons of excitement around this. A lot of. We're all kind of theorizing. When you talk to marketers, it's probably going to range by how conservative a brand they are, how new they are, are they ready to go, you know, quote unquote, all the way and like let the machines run wild with their assets or is there still a level of like, we got to really have a lot of checks, amounts here, Be careful.
- [09:09]
A
Yeah, I think so. A couple of things on that one. Look, everybody has to be very sensitive to how this gets rolled out. Especially there's compliance and privacy issues and things like that. One of the advantages we see from the utility side is everything in our models has been legally obtained. And so if you're going to put out a finished asset, you want to have confidence that those images and everything
- [09:35]
B
that you've tried, your fear or copyright violations.
- [09:37]
A
Yeah. So there's a certain element of that and then there's a certain element of all this technology is great, but if you can't integrate it into your workflows, then you kind of, all the efficiency kind of goes out the window. Can now you have another three systems you're working in. And so yeah, it's building. The people in process are always behind the technology and how do you integrate it into the workflow? So we spend a lot of time on working with companies on what's the. Org model, how do, how do you shift and what do you have to shift to be able to take advantage. And we really look at it. It's not like you're obviously the humans are the creatives and this is just automating some of that work that I described up front in the math problem that people are having a hard time keeping up with. And it's not necessarily value added to be able to make 10,000 different formats for 10,000.
- [10:29]
B
No, that's manual labor. That's not really like your creative genius coming through. Just variations. So, okay, what, what about how do. How does what you do jive or not with, you know, like the, we talk about this all the time. Like the big platforms have offered some, some elements. Elements what you're describing. And they would sort of say, like, don't worry, we'll do all the creative for you. We'll take, give us your, your budget and your data and we'll do all this. Do you. Are you trying to like out Meta? Meta. Are you trying to. Does it compliment what they're doing or, or in TikTok or whoever? How does that work?
- [11:07]
A
Yeah, I think like, like I said up front, I think our relationships mostly are with brands and agencies and still hearing they want to control their brand imagery, they want to control some level of their data, they want to control some level of where they're showing up and how they're showing up. So I think there might be an element, platform by platform, where somebody says, within this platform I'll be able to optimize better using your AI technology. And I think that's. But we're really looking, like I said in the beginning, we're the operating system across a brand marketer, an agency and a publisher. And so you want to do a lot of that work on the brand agency side to make sure that you're comfortable with the assets you're creating, how you're showing up and then where you want to have those assets being placed versus I don't see a world where a CMO just goes, oh, I'm just going to hand over my whole budget to Meta and let them figure it all out for me.
- [12:07]
B
Right. Tell the CFO Like, I don't know, it's. It's going to work, I think.
- [12:11]
A
Yeah, Yeah, I don't see that either.
- [12:13]
B
Okay, let's take a quick break. I'm here with Josh Melick. He's the SVP of Product and Data at App Science, a division of sao. Hey, Josh, thanks for being here.
- [12:21]
C
Yeah, nice. Thanks for having me, Mike.
- [12:23]
B
So, Josh, you guys have made household graph a big differentiator in what you do. Why is that?
- [12:28]
C
You know, from the very beginning, data and data and audiences has been what we focused on. We combine mobile data with CTV data and we think that gives a really good picture of audience. I wouldn't say that we focus on every element or every type of segment, but there's plenty. There's quite a few segments that we really focus on that are important to our clients and advertisers. Multicultural audiences, younger generational, cultural shifts, or a lot of these areas that are really important to the advertisers we serve. And we've focused on that from the very beginning.
- [13:02]
B
So how central is the household graph to the media that you run?
- [13:05]
C
We use our data on like something like 95% of our campaigns. So it's very central, it's very important to what we do now. We do bring in outside data as well. And actually, I think often the most interesting audiences or the most impactful campaigns use something, use our data along with some data by the advertisers, perhaps some of their first party data or some other data as well. And we think that's really where it's powerful. And yeah, all of our campaigns are using something like that.
- [13:33]
B
All right, great stuff. Thanks, Josh, back to the show.
- [13:36]
C
Thanks, Mike.
- [13:37]
B
You guys have some recent research on creative intelligence. Can we talk about that and what you found, what we're trying to uncover?
- [13:45]
A
Yeah, so we did some. We sponsored some research with Winterberry Group on creative intelligence. And that means really the ability to understand why consumers engage. And so the collection and analysis of effectiveness data, creative effectiveness data, and then being able to build that into optimization and continuously improve campaigns for effectiveness and engagement. This is the new frontier. Right? We spent the last five, eight, ten years on data and how we can scale data. And now we're realizing, hey, creative isn't a fixed cost. It can actually be a performance driver. And really you need the same way we have the audience layer, which we spend so much time on, and the media layer, which is where and when your media shows up. Now the investments, they're showing 23% increase in investment in this creative intelligence area. And it's really about how do you capture that and make it more systematic to understand. Is it the format of your content? Is it the tone, is it the imagery? Like, what is driving effectiveness and conversions in that, in that creative world?
- [14:58]
B
And are you mostly talking open web, display, social, or could this work on television? Like, where, where are you seeing this?
- [15:05]
A
This could work all the above. Like, we're working, we're working across the walled gardens who are feeding in some of that data. We're working across CTV and retail media networks. We're really looking at that holistically of how do you bring that data in and that it's really across all those channels to be able to understand. And it could be by channel. Like, it could be there's creative effectiveness that works on, you know, on display, that doesn't work in video.
- [15:34]
B
Have there been, have there been surprises or general things you've pulled out yet? Or is it really going to depend on the campaign?
- [15:40]
A
It depends on the campaign. It's still the early days where I would say we're, we're all kind of testing and learning things right now. But we do see this as a, an area that as we learn and continue to build those blocks, it's really going to become more of an automated kind of function. But right now it's, it's very early days. I'll come back. All right. Back in another time when you're ready.
- [16:06]
B
You're on the show again. People, you've talked a lot about the brands or how willing they are to. To kind of let the. Let. Let general AI go do its thing. What people always wonder, do the creative agencies like this? Do they. Are they totally threatened? Like, what is your experience? Because, you know, if you talk to one person there those days, their days are numbered. And others are like, no, this is going to help them do the job so much better. Like, what do you see?
- [16:29]
A
Definitely the latter. All the agency we're working with, all the major agency holding companies, we're all leaning in, I think. So Adobe as the creative and data operating system is definitely like the. One of the foundational elements of what the agency is building into their proprietary platforms.
- [16:51]
B
That was going to be my next question, because they all want to build their own thing right these days, but
- [16:55]
A
they're building it with a differentiated approach on top of operating systems that they might not want to build. They might not want to go out and build a. Create. Start from scratch and build a creative operating system.
- [17:07]
B
They're not going to be Adobe or so.
- [17:10]
A
Exactly. So it makes Sense to kind of partner with somebody like Adobe and then build your differentiated solution. Integration of data, integration of the agents, whatever it is on top of that operating system. And they're leaning it, I think, you know, like I said in the beginning, everybody is I think leaning into how do we automate some of the rote things that really don't need somebody sitting there with a, a lot of creative talent but you know, doing auto version. So those are the kinds of things it's like, okay, if I have an operating system that can do that, then I can really do the work that I want to do, which is how do I come up with the next best creative idea? You know, where do I look for those, the optimization and how do, how do I choose the right audience and how do I pre test some of this? Like all the cool stuff we've been talking about. That only happens when a lot of this kind of tasks get automated and opens up time because right now people are swamped with that original math problem. They just can't like look their up to be.
- [18:16]
B
Oh, we say oh that's how you learn. But actually like it's just kind of miserable. And there's, you could learn a lot more about actually being strategic and getting into the business. Okay, what about. Because there's the, this future you're talking about with this like far more in creative intelligent levels of automation world. At the same time the media buying world is going through its changes. Are you going to like, do you, do you envision those things coming together? Like will creative, will the creative OS work with the rest of ad tech? Or are you going to. Or they will they be like one thing? How does that look?
- [18:50]
A
I see coming together, I mean you're always going to have the economies of scale of media agencies and media buying, which makes sense. And then I think, you know, I'm definitely seeing it at Adobe, the creative and data sides coming together. So what happens is, you know, you get that great creative intelligence we've been talking about, but what audience does that go to? Like how. And, and as soon as you ask that question, you're bringing the data integration layer into the creative intelligence layer. And it's like, okay, if I, if I only have the creative side and I don't have the data or the audience intelligence, I'm missing a huge piece.
- [19:28]
B
Doing one in a vacuum is not, doesn't make any sense.
- [19:30]
A
Yeah, it doesn't make any sense. So it has to come together. I think like I said, the data side is pretty far advanced because we spent so much time and effort laying that foundation over the last eight years, 10 years. But the creative side is going to quickly catch up because the technology is moving very, you know, much faster than I think we did on the data side. And then I think where things are headed, think about this world, Mike is. And always on marketing. Instead of always on marketing campaign versus one and done campaign. So we think of campaigns, they very literally like I do X, Y and Z and then I launch a campaign
- [20:09]
B
and there's a flight and then it
- [20:10]
A
ends and you know, it ends and then I go measure it and then I go back and try to do another one. If, if these systems are automated in the way that I think they will be and integrated, you'll be able to have more of an always on campaign. And I think about it, you ever been to a network operations center and a major company there, there's a bunch of screens and a bunch of things going on and somebody's monitoring it 24 7. You can have, you can think of a marketing world where that's going on, where there's personalized campaigns happening all across the board, you know, being automated. And of course you're going to have a human involved in that.
- [20:46]
B
You can, you can make decisions and see spot trends and move stuff around. Like they've kind of dreamed about that dashboard of the future.
- [20:52]
A
This is the future we can dream about, right?
- [20:55]
B
What, okay, what about does this mesh with, how does this marry up with this whole Agentic thing we're talking about? First of all, I guess is Agentic a thing yet? We're talking a lot about it. Like are you seeing it, using it and where does that change what you're talking if it does?
- [21:13]
A
Yeah, yeah. I think Agentic's early days. There is no full stop solution for Agentic in the advertising world. It's within platforms. Like you say Meta has an Agentic solution within their platform. Many platforms coming up, but there is no end to end Agentic powered solution. And so I think right now you're seeing it in specific places. Adobe strategy around Agentic is to build it into our platforms where it makes sense. And so if you want to get started with Agentic, there has to be an easy way in the platform you're already working in.
- [21:51]
C
Right.
- [21:52]
A
You don't have to build a lot
- [21:53]
B
like you're talking about a lot of changes at once to try and change the way that these companies operate fundamentally and then throw in Agentic. That seems a lot to take on.
- [22:01]
A
Seems to be a lot to take on. And so well, where we're seeing adoption is within the tools that people are already using. Hey, I want to pull an audience. Well, usually it would take me a week to go in and go, hey, give me a query on anybody who hasn't bought from my retail shop in the last 30 days. Overlay that with high income spending. Now that you can do that using an agent pretty quickly. So it's, how do you, within the existing workflows across an enterprise platform where you have this, you know, end to end, what are the things that I, that I can accelerate?
- [22:34]
B
We talked about the creative side, the brands. What about you? Like there's a lot of talk about how the media agency of the future is going to be unrecognizable or maybe not that many people. And it's, it's all, you know, it's going to be an algorithm, algorithmic driven world. Where do you see that? What is the potential impact and how does that maybe connect with what you're doing on the creative side?
- [22:56]
A
Yeah, I think, I mean we covered the agencies before. I do think that they're building differentiated solutions and these platforms that they're building are going to have a lot of the elements of what we discussed. And so I think, I don't think that's a big overhaul from a people perspective. I think the people are going to be doing higher value tasks.
- [23:20]
B
Different stuff.
- [23:21]
A
Yeah, like you said, strategic. Start thinking about, okay, what, what strategic campaign did I run? What's this audience that I put up against now? Like, I think some of the thinking like moves up front and then on how do we do this at scale versus say there's a lot of, you know, everybody's running now. I'm just trying to get the campaign out.
- [23:45]
B
Yeah, there's been a lot of research coming out about how much brands care about how they're showing up in the various LLMs, how they get discovered, how brand discovery is changing and whether or not they can, like someone can actually work the system or control that. Like can you talk about what you guys are doing there, what you're seeing?
- [24:04]
A
Huge issue right now. We also do our own proprietary research. Traffic is down, generally 40% going to brand sites and publishers. Yeah, obviously from a publisher perspective, huge problem. Your ad revenue is taking a hit. And then brands, like you said, they don't, they're worried about using that direct relationship with the customers or how are they even showing up. So I think, you know, there's a, there's a couple things that, that they can do and we're really Leaning into Generative Engine Optimization. So we're working on LLM Optimizer, which does two things. One, it monitors how you're showing up and that's basically just what's going on. What's going on?
- [24:48]
B
Yeah.
- [24:48]
A
And then second, it helps with restructure, structure your content in a way that is going to be picked up by an LLM. And that's really doing two things. One is it's making sure that you're more visible in the LLMs, but it's also, it's feeding some of the other sources that the LLM might pick up things from. So there can be, so you're not
- [25:12]
B
blind, you can see some of that and get, and, and, and steer.
- [25:15]
A
You can see the improvements. Yeah. And so it's like structuring your content as, you know, as a product. There's like product information that you have to get out there that gets picked up. There's, you know, it's kind of FAQs, there's, there's ways to structure the content where it's like authoritative information about your products and kind of, you know, the company that gets picked up in a way that now you see that with.
- [25:46]
B
And as those, as ChatGPT and others start rolling in, they say, they're saying advertising is not going to be, it's not going to affect the results. Do you think most brands are going to have to shift a bunch of their search ad strategy that direction to maintain what you're talking about or is it unknown right now?
- [26:04]
A
I think that's a, I think that's a huge unknown. I think the first issue is, is, is ensuring that you're showing up as a brand that you want to. And then also the referral traffic that's coming from those generative engines has to like, if you're getting X amount of search traffic, you want to keep some level of that up in the new world.
- [26:27]
B
Right.
- [26:28]
A
Advertising is interesting. I don't you tell me how that's going to play out. Our. Is it going to be, I guess it depends on how they do it. How integrated is it going to be? You know, we're used to a certain level of here's the information I want. Is it going to be in the middle of that? Is it going to be in the top of that? Is it search?
- [26:44]
B
You know, it's. Search is such a utility historically. Like it's different, it's a definitely different mentality and use case. I think it's a great unknown like that. You, you might assume, okay, they're going to tailor this in a smart way and make it. But these, the way you interact with these things is different. And are you going to be as receptive, less receptive about it? I don't know.
- [27:05]
A
Right. It's going to be interesting. I mean, we're living in interesting times.
- [27:10]
B
Yes.
- [27:12]
A
I mean, you can see how it's all like, we sit here today and we talk about it and you see it all on the horizon and you can see how this can really improve the level of advertising, the, the relevance to the consumer. All the things we've talked about for years. I feel like with the creative piece that we talked about today and the data foundation that we already have, we're really moving into the direction we've talked about for a long time.
- [27:40]
B
Yeah. It is a fascinating. It does feel like we're headed towards a fulfillment of some things we've been waiting for. Let's see how, let's see how it goes. Sam, awesome, awesome conversation. Fascinating stuff right now. Great moment. So thanks so much for your time here.
- [27:52]
A
Yeah, happy to do it and look forward to seeing you.
- [27:54]
B
Let's do it again. Yeah. Thanks again to my guest this week, Adobe, Sam, Garfield and my partners at Tabio. If you like this week's episode, please take a moment to rate and leave a review. We have lots more to bring you, so please hit that subscribe button and we'll see you next time for more what's next in media. Thanks for listening.