← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 155: Ari Missed CES, but Eric and AI Were There. Plus, Mike Khristo Creates Marketing Plans from GitHub
00:49:23
Loading summary
Transcript174 lines
- [00:00]
A
This podcast is brought to you by Maloco. If you're an app marketer concentrating most of your budget in Google and meta, you're not alone. 88% of consumer app ad spend stays inside walled gardens. Meanwhile, there are billions of daily users spending their time elsewhere in gaming apps, dating apps, productivity tools, you name it. That's where Maloco comes in. Maloco's AI powered platform connects you to those users, reaching 2 billion daily users across 3 million apps.
- [00:29]
B
Ready to diversify.
- [00:30]
A
Visit Malocoads.com and start scaling with confidence again. Go to Malocoads.com to learn more.
- [00:51]
B
Welcome to the Market podcast. This is Ari Paparo. I am very much not in ces. I am here with Eric Ferrante, who is very much in a hotel room at ces. How is it, Eric?
- [01:02]
C
You can tell. It's good. You could tell from my background that I am in a hotel. Parts unknown in terms of specific locations. But yes, confirmed that I'm ces. Confirmed. It's been a packed week and that you have been missed. Why are you not here?
- [01:17]
B
I had a cold. You know, it wasn't like I wasn't dying from the flu. I got. I got emails from people who said. Who probably got it from secondhand sources that I was, like, on my deathbed and they're like, fried. O' Kelly apparently told somebody I was nearly dead. I got an email like, oh, are you okay? I'm like, yeah, I just have a cold and I don't want to infect everyone I know at ad tech with my cold. So I will stay at home in cold New York suffering with my cold. And it's been the most productive week of my life. I've just been, like, knocking out everything on my to do list. It's been great.
- [01:49]
C
Yeah, of course. Like the zero emails, right?
- [01:52]
B
Yeah.
- [01:52]
C
I will say that the ecosystem missed your presence.
- [01:56]
B
Oh, yeah, I bet it did. What's the vibe we'll talk about? We have a lot of news to cover, but what's the vibe generally?
- [02:04]
C
Yeah, the vibe is generally good. The vibe is generally like we're all trying to figure this thing out. All things around AI, lots of talk about just continued M and A and consolidation and big partnerships. The usual CES stuff. But it's been interesting and I kind of synthesize a lot of these announcements and conversations into a few key things we'll talk about in the news.
- [02:30]
B
So much news. Okay, so this. So I did a separate interview outside of CES that we'll have next with really interesting company called Layers. And this is an example where a company was not on my radar, but someone, I think James Borrow tweeted about it or put something on LinkedIn about how cool it was. I reached out and we have an interview with the CEO, Mike Christo. They have the best tagline. It says turn your code into users. And I just, you know, that intrigued me. It's about using. It's a tech first approach to marketing and advertising basically. That's very app specific. I think it's a really interesting conversation. Had you ever seen this company? Had you heard about it?
- [03:09]
C
I, you know, I could. It was James that's been talking it up. I think James is a, an investor advisor. I think it's part of his kind of X Snap diaspora. I saw like, like an early presentation or something but didn't have a chat with Mike. So I'm looking forward to listening to this.
- [03:26]
B
Yeah. And you know, I'm obsessed with vibe coding at this point. I have like, so I have like, I would say three or four different projects going that I'm vibe coding. And, and this is very relevant here because you can create an app but how do you make money? Right? You get a couple thousand users and you know what's next. So this is right in that alley.
- [03:45]
D
Yeah.
- [03:45]
C
It's at the intersection of two things. Vive coding and mobile app. Which as the old desktop guys, you and I need to be more exposed anyway.
- [03:54]
B
Yeah, exactly. We're the old desktop guys. As we heard recently.
- [03:57]
C
Jim Payne.
- [03:58]
B
Jim Payne, the old desktop guys. So this week I had a great interview with, with Walmart Connect. So they had a lot of announcements on Monday and I did an in depth interview with Kurum Malik. So you should listen to that. It's on YouTube. It's on this channel, if you hadn't heard it already. Walmart connects obviously a super important company in the commerce media space. Next week, I think next week we're going to have my interview with the, the Vanderhorn brother, Vanderhook brothers from Vyant who had a big announcement this week. We'll also talk about that in the news. A lot of AI stuff. So that's a great interview. And meanwhile I have to talk about my Vibe coding project, MAD db. So MAD DB had a pretty big release over this week where we added events. So there's an event calendar where the AI is picking up all the major events in the advertising space and putting them in a calendar you can add to your Google Calendar. So love to hear feedback from people. We also added a feature where if you claim your company and put in the URL of your Press News page, we will spider it, get all your news as soon as it's released, and then that gets emailed to anyone who's interested in your company. So it's free press, free SEO. So you gotta claim your company and put your news URL in there. Of course, it doesn't work exactly properly, but whatever. We're getting some kinks out.
- [05:12]
C
It's fascinating. This is becoming like a. It went from, you know, an RSS reader, you know, type thing to, like, a utility.
- [05:20]
B
I'm using it like, I used it before this. Before this recording. I went through our News of the Week and made sure they were all in there. There were a couple of articles that we thought were important that were in. In the system, mostly because having a little bit of a hard time with one or two of the sites getting the data, but also, you know, research. Like, what are the last five announcements from Walmart Connect? I was able to just quickly figure that out. So it's an experiment. It's fun. I'm enjoying it. If you have feedback, I take the feedback directly and we're trying to make it into something cool. So sign up, claim your company, get yourself read. All right, with that, let us go into it. So I'll be back with Mike Christo from CEO of Layers, and then Eric and I will be back with the refresh news of the week. All right, I'm here with Mike Christo, the CEO of Layers. Mike, thanks so much for being here.
- [06:10]
D
Thanks for having me.
- [06:11]
B
Ari, you and I are the only two people not at ces, I think. So we had to do a podcast together.
- [06:15]
D
It actually feels like kind of happy not to be there right now.
- [06:19]
B
Yeah. Yeah, I think happy is a good word for it. I was going to go, but, you know, I started sneezing and, you know, everyone says they go to CES and come back with a cold, and I didn't want to be the vector for that. I'm just sitting in my house doing work.
- [06:31]
D
I think everyone appreciates you not being patient zero.
- [06:33]
B
Yeah. Yeah, I think they do. So, yeah. So we haven't met before, Mike, but I saw your website and it said, turn your code into users. And I was like, all right, I haven't heard a good tagline in a while. Let's get them on the show and find out what that's about. What is Layers? Sorry?
- [06:55]
D
Yeah, Layers is your technical cmo. So we. I've been coding for a long time. I've been doing startups for nearly 30 years and where the most recent advances in AI, I think, have opened up a lot of opportunity for people to build a thing. Not everybody wants to build the next Salesforce. Maybe they want to build a mobile app, maybe they want to build a small SaaS tool, whatever it is. But most of the time, developers just aren't good marketers. As a developer, I can say that we're actually kind of allergic to it. I'd rather go and build another feature than actually market the thing. And so we built Layers out of our own need, trying to build mobile apps and things like that. And the idea for Layers is basically automate as much marketing for people who don't have a marketing team. And that's what the premise is. So we start with your code. That's where that tagline came from. And that's all you really have to give us. Give us your code, we'll do the rest.
- [07:48]
B
All right, I'm going to look at your C code and I'll be like, I think Meta is the right place to put some ads. I love this. This is the Vibe coding episode of marketecture, which for those of you following me, you, I'm like obsessively vibe coding at this moment. So you're going to explain to me how I'm going to turn this into a career. But let's just like, what does that mean? Target coded? I mean, what could you possibly learn from someone's code that is relevant to someone's go to market?
- [08:15]
D
It's kind of fascinating because you can learn a ton. So when you give us access to your code, whether you're building a thing like a caveman, typing your code yourself, or.
- [08:27]
B
Using Stack Overflow like our ancestors, right?
- [08:34]
D
No matter how you're building your stuff, whether you're using Lovable or Roark or anything, you can always get your code into GitHub. And so from your GitHub, which is basically the repository of where your code is, we can look at that code and we do a bunch of different things in there and we do it really quick. What we do is look for what you're building, how you're building it, who you're building it for, how you speak to them. So your onboarding screens, some of your help screens, things like that, we look at your aesthetic. So any icons, imagery, things like that, we'll pull out the colors, the aesthetic, things like that, and then we can actually put together a pretty good profile really quickly of who you're building it for and then understand how we can reach them. We then basically present to you, no pun intended, a set of layers that we recommend that you activate. And just.
- [09:19]
B
And so this is a little bit like, I've seen a bunch of AI demos and products where it's like, just give us the URL of your, like, Shopify storefront and we're going to create campaigns for you. Is this just the in app version of the same thing?
- [09:31]
D
It's a. It's a similar version of it. Where we differentiate is on kind of the. The nasty parts of advertising. So, like, what.
- [09:41]
B
What are the nasty parts of advertising that my listeners want to hear?
- [09:44]
D
The nasty parts of advertising are all these SDKs that you've been put in all the Instagram.
- [09:48]
B
That's the truth.
- [09:49]
D
Right? It's terrible. And I'll. As a developer who's done the marketing stuff, putting in the Facebook SDK is not something that's like a pleasant experience for most people. What we can do, because we have access to your code is we'll actually take that off your plate. We'll just say here, we instrumented your app, we put the Facebook SDK in, or TikTok SDK, whatever it is, and we'll instrument your app where it needs to be instrumented. So we'll take that off your plate. So as a marketer, you're not going to the developers and saying, I want you to put analytics here, here, here and here. We just do it for you.
- [10:22]
B
Oh, okay. So you're actually modifying the code too. It's not just figuring out what images to run on Meta, you're like instrumenting the whole thing.
- [10:29]
D
Yeah.
- [10:29]
B
Did you just. You should listen to my episode from Right Before Christmas with Jim Payne and his new startup. That sounded pretty cool. Did you happen to be on top of what he's doing?
- [10:40]
D
Haven't listened to that one. I'll do that right after this.
- [10:42]
B
You should. He's got an AI driven SDK. It's pretty cool stuff. So what's going on? Let's talk about vibe coding. So I'm vibe coding a lot. I have production code. Right. I have a website called maddb that I like to promote. That's like a Reddit for advertising kind of thing. So there are. There is production. Vibe code out there. Are people really, like, making money? Are there. Are there real apps that are vibe coded?
- [11:11]
D
Yeah, the. It's funny because the traditional engineer, the code artist, if you will, is still very much like, AI is slop. AI is never going to be there. Like, I've been doing this Stuff for how long? The tooling is so good and the tooling coming out every couple months leapfrogs the previous generation. And so over, you know, what we call like a winter break when, when real work slows down and people get some time to play with these tools, things like Claude code, things like some of the vibe coding tools, they've just gotten so good. Using some of the underlying fundamental models like Claude and Opus, you get really high quality code. But not only do you get the high quality code, you can have it go review the code, look for security issues, fix code that's broken. And so yes, I'm like way on the far end of the extreme of vibe coding works. You can get real quality out of it. It doesn't all look homogeneous with the purple gradient anymore.
- [12:16]
B
You can actually there's a lot of purple. It's unfortunate that my company's color is purple. So whenever we do the purple, everyone goes, oh God, lovable creat that. But it's just, it's our theme.
- [12:28]
D
They're getting better, they're getting better. And so maybe it's a small variant of the purple on your next iteration.
- [12:34]
B
Yeah, yeah. So, but you give an example. Is there an app you know of that was created by a non professional that is commercially viable?
- [12:42]
D
There's a bunch. If you go on X, which is kind of the echo chamber of AI, sure.
- [12:48]
B
My ex is all clo, all Claude all the time. It's just like one person after another talking about Claude. I don't know what happened. I don't know if it's the algo or it's reality, but it's entire entirely. Well, there's some racism and then there's Claude. Those are the only two things on X. Yeah.
- [13:06]
D
I think what happened is a really notable figure in AI, Andre Karpathy basically started gushing about Claude and that gave it the blessing that some of the code artists could jump into it and some of the purists could jump into it. There's a bunch of examples people are building in particular right now. There's this renaissance in consumer mobile apps. Utility apps, kind of single player, do one thing really well, throw a mascot on there, gamify it a little bit. There's a bunch of people building really, really fun, really engaging apps.
- [13:40]
B
Like, can you name one? What's your favorite?
- [13:43]
D
Honestly, off the top of my head, I can't name one.
- [13:46]
B
So I guess I want to push a little bit. Like because I've written production code myself, I'm not a professional developer, but My code's been used by billions of people, actually, because I'm an ad tech. So are there people who are really not coders who are able to deploy production code or is it like people like me who are like product managers, pseudo coders like yourself? Probably. Are you a developer?
- [14:13]
D
I'm a developer. I started coding back in the 90s. I'm a dinosaur. The people that I found find success there have been the non technical people. They've been former CEOs or current CEOs, they've been designers. Designers are really the fun one to watch.
- [14:31]
B
Yeah, designers are cool.
- [14:33]
D
Yeah, they can design it in figma, they can make it look exactly what they want and then they can give the Vibe coding tool access to those designs and then they don't have to try to translate what they wanted to a developer. There's a woman that I know at a previous company, Designer, phenomenal designer. You buy coded an app, it's in the app store now. She's like over the moon. Is it replacing her day job? Not yet. But is it showing her signs that she can rely on herself, bet on herself and kind of build and iterate on this? Absolutely.
- [15:05]
B
All right, let's get back to marketing, the marketing side of things. I diverted us a little bit. So. So you, you're able to, you know, investigate the code, wire it up, get the messaging out, get the, get the images. What about media planning? How do you know what's going to work? Or how does, how do you, how would you, if you were a Vibe coder, think about where you should actually spend money on customer acquisition, what the budget should be, how you should differentiate, et cetera.
- [15:32]
D
I don't have the secrets out, but meta's pretty good.
- [15:38]
B
I hear Meta is good. And you know, if you're interested in search, there's a little company called Google where they'll just, they let you buy keywords. It's crazy.
- [15:45]
D
Yeah, you can like specify there's a few things. So the way that we see it is we'll actually, we start primarily on consumer mobile apps as our primary target today. And so what we'll do is figure out what kind of app you build. Is it iOS, Android? And then obviously there's a couple important things. First of all, it doesn't matter what kind of marketing you do if you can't be found in the Axle App store, if people can't type in your app name, if you're not ranking already for those terms, it just doesn't matter. And so, or at least the organic side of distribution doesn't matter because they're not gonna find you. So Metis, obviously super important from an app install perspective, but Apple search ads are a key player. It's ranked for that. Or if you can't get ranked immediately, which you typically can't, it's by your name.
- [16:34]
B
And so by your name. Right. Cause someone hears about it, then the competitor snipes your. Your intent.
- [16:40]
D
Exactly. And in this consumer mobile renaissance, what people are finding is you can actually get a ton of distribution and downloads through organic traffic, which is a little bit counterintuitive to our world. Having been in ad tech world for far too long.
- [16:55]
B
I love organic. Organic's the way to go. Email is the best marketing channel you've got.
- [16:59]
D
TikTok Organic is blowing up. You find creators who will create content. Some people are just putting slideshows together, AI slideshows, getting millions of views, driving those downloads through comments. And that's where that App Store optimization or App Store ads really come into play. From there, it's meta TikTok ads. Those are obviously some of the key ones there. There's.
- [17:23]
B
What's your feeling on the controversial one, applovin? Probably a little too big of a minimum for your. Your type of customers. But what's your general take on applovin?
- [17:32]
D
I. I don't subscribe to the controversy. I think whatever works.
- [17:36]
B
Don't subscribe to the controversy.
- [17:38]
D
Yeah, I don't think it's controversial. It's fantastic.
- [17:41]
C
Right.
- [17:41]
D
Like, if you can get your app in front of the right user at the right time, does it actually matter what platform you use?
- [17:47]
B
Yeah. And they have a pretty high minimum though. Right. So I think. Yeah, so you probably aren't starting there. Let's go back to what you said earlier about organic. Is there an art to getting organic.
- [17:58]
D
Traffic in the App Store for App Store optimization? It's kind of the usual. This is similar to SEO. It's getting the right keywords in place, getting the right distribution through those keywords. So making sure people, when they search for those keywords are actually downloading that app. Those are all the signals that will help the App Store rank. As far as organic on social, it's a mix of art and shotgun approach. Right. It's put as much content out there as you possibly can, vary the hooks, and once you find the thing, this is, this is really the fun part about organic is as soon as you find a hook that works, you can just keep throwing at it over and over and over again.
- [18:38]
B
Like. Like what?
- [18:39]
D
What's an example we've got an. There's an example of an app that's using layers and they use. One of our layers is called Managed ugc. So when you activate that layer we actually have these agents that go out, they find and source UGC creators. They'll get those UGC traders up to speed, they'll negotiate with them and come to terms with some kind of organic distribution plan.
- [19:01]
B
Well that doesn't sound very organic. You're paying for it. It's influencer.
- [19:05]
D
They're called ugc. So they're, they're influencing in the sense that they don't have their own followings. So they'll actually go create their new TikTok account, new Instagram account, they'll warm those up and then they'll create content on behalf of this brand. We've got instances where a creator has found a hook that happens to work. Maybe it's a reaction and a demo. So it's a reactor. Face some hook and demo and it goes viral. And what they'll do is actually share with the other creators that are on that particular program saying this just worked. And so we share that content with the other creators and they'll go and replicate it, put their small spin on it. And what you'll see is this video got a million views, this one got 900,000 views, this one got 1.2 million. And it just, you can kind of keep going back to the same well over and over again until it's done. Yeah.
- [19:54]
B
So what's the minimum budget for doing something like that?
- [19:57]
D
Typically what we Recommend is a 400 base, 60 posts guaranteed and then a dollar CPM up to a thousand dollar cap per post.
- [20:07]
B
Right. So. So one of my side vibe coding projects is Recipe website. I know that's not unique, although I have a different sort of take on it. I've like, like very code centric, so I have like a code oriented recipe website. What would you recommend? If I wanted to create social excitement around that, what would the hook be like? Never saw this recipe before and then jump in.
- [20:29]
D
I'm not a hook guy, I have no idea.
- [20:31]
B
You're in the developer camp who doesn't understand marketing. You're just letting people do it.
- [20:37]
D
I don't know what the hook would be. What I would tell you though is don't settle on one. Try a bunch of things and try the things that are non obvious. The things that you would say like I don't think that's going to work.
- [20:47]
B
Yeah, so, so we talked about TikTok. You used to work At Snap. How important is Snap?
- [20:52]
D
Snap is an interesting animal. If you find the right audience on Snap, if you've got a product that really resonates with their audience, typically trending younger, you'll find. Yes, but for the most part you can find a very similar audience in other places.
- [21:08]
B
Yeah, like so. So really it's about Insta. Insta reels and TikTok and. And YouTube. YouTube shorts also.
- [21:16]
D
Yeah.
- [21:17]
B
Right. And what about any interest in Pinterest? Is that useful at all? Maybe for a fashion oriented thing down that path yet?
- [21:25]
D
I don't think. I don't think there's. It wouldn't be the starting point if I were recommending places. It wouldn't be the place I would start putting Budget. It's a place where I would.
- [21:34]
B
And Budget got it. You know, you said earlier that there was sort of a renaissance in consumer apps. I have to find the stat. But there was a stat I saw on X a couple weeks or months ago that said that like this year, 2025 was the first year the number of apps submitted had increased in like quite some time. Like the number of apps in the new apps showing up in the app store had been on decline for years and it just turned around because of this phenomenon. I think that also is a challenge though because like my stupid recipe app, there are probably twenty hundred other stupid recipe apps. Right. So is this like a fad or is this going to be a long term thing where people will be able to have sustainable business models around small side projects?
- [22:19]
D
I think the way that I see this playing out is that you end up finding a lot of more hyper targeted niche apps. And so it's not just the fitness app. It is the fitness app with some very specific goal or for some very specific niche. It's a recipe app for some specific cohort of users. Not if you want to cook a great meal, come here. But it's a more niche audience. That's really, I think, where the fun goes because now you have an audience that's using that app that gets a really bespoke experience and can get really engaged in that experience. And then they found their people. And that's really fun because now you've also built a community.
- [23:01]
B
Interesting, interesting. Have you, have you vibe coded this app or do you have actual, do you have developers working for you by.
- [23:08]
D
Coding while we're talking? Yeah.
- [23:11]
B
You got a Ralph Wiggins working in the background.
- [23:13]
D
Absolutely. I mean it's one of those periods of time where it's just so much fun just having this Little thing in the background. And then every few minutes like, hey, can I do this? Like, yeah.
- [23:27]
B
It's like having the world's best intern. I wish I had a job so I had something to do with these things. I'm trying to find makeup work to use on. On these things. So it's like, so the. So what do you see as kind of the future here? Like, if you were to prognosticate, we're having this conversation, you know, a year from now, what do you think we'd be talking about both for your company and also for this, like, this topic?
- [23:51]
D
It's hard to say. I used to have a crystal ball. I used to be able, with like some level of confidence tell you, ah, the next several years going to play out like this. Where the step change that happened when Anthropic released Opus 4. Five, in terms of the quality of code that it can output, how long it can think, how well it can reason through, building things is just a huge step forward, a huge leap forward from the previous ones. If we keep seeing that. And there's no reason to think that we're not going to see that. I don't know. And so that's like, in general, that's sort of where I see it. What I do see is a couple things. People are now empowered to go build. That in and of itself is exceptionally exciting. I am thrilled. People who have not built things before are now going and building.
- [24:41]
B
That's so cool.
- [24:43]
D
People say, oh, you're gonna, you're gonna vibe code from your car on your way to the gym, your. Your fitness tracker. Like, no, like that. That's probably not gonna happen. But the idea of building a bespoke thing for your need is definitely there. And then realizing that you're probably as special as you are, you're probably not the only one that wants that thing. There's probably other people. Is going to really empower distribution. It's going to empower people to go build a business, make a little side hustle. And I think that's the. The burden and pressure of side hustles has now been removed. If you don't need to find a technical co founder anymore. Right. You don't need to dedicate 20 to 40 hours a week. And if, you know, feel bad if you missed a day of doing this stuff, you can go build stuff. Yeah, it's wild.
- [25:37]
B
It is incredible. I think this is one of the most exciting things that's happened in my career, and I'm very excited about it. All right, well, Mike, Christo, the CEO of Layers. Thank you so much for being here. This is really interesting conversation.
- [25:50]
D
I had a great time. Thanks for having me.
- [25:54]
A
This podcast is supported by Walmart Connect.
- [25:56]
B
Hello marketers.
- [25:57]
A
At Walmart Connect, you'll find things no other retail media has.
- [26:01]
B
Not only do they have the most.
- [26:02]
A
Shoppers buying all kinds of things like cleaning supplies, video games and cookware, they get how shoppers tick and why they click. And they could help you connect with.
- [26:10]
B
Them at any point in their journey from discovery to purchase.
- [26:13]
A
Because your bottom line isn't just about views or engagement.
- [26:17]
B
It's about the most valuable truth in advertising, that your ads work. So spark innovation. Let's spark growth. Let's spark sales with Walmart Connect.
- [26:29]
C
Okay, we're back with the news of the week. This being probably one of the two biggest news weeks of the year in, in our space, even before we get into the news of the week. I don't think that Market Extra actually kicked off the news with the funding announcement. It was like the first thing that I saw this week.
- [26:46]
B
Yeah, I forgot, I forgot about that.
- [26:48]
C
You speed ran. You speed ran by dropping it first thing on Monday morning. So, so congrats on, on the seed round. We invested at Aperium. We continue to invest in the company. Tell us about it.
- [27:01]
B
Yeah, sure. So we announced a raise. We, we are doing really well. Don't like desperately need cash, but we want to expand. We have this big event, Market Extra live coming up March 11th and 12th and that's a must attend event, but it's also a big one to put on and we're looking to continue to expand. You know, we acquired six companies in the course of the previous 18 months and these are all small acquisitions but you know, if you've got a side hustle in advertising, call us. We're interested in growing inorganically and organically and it's all much better since I stopped becoming the CEO now. Now it's like a well run operation. I just in five code and add my two cents here and there. But thank you for participating, Eric. You and your LPs will be well rewarded, I hope.
- [27:49]
C
Thank you. All right, awesome. Let's get back into it. So ces, we talked about in the, in the intro, you know, generally like good vibe generally if you're I think well positioned in AI or CTV or AI and ctv, you know, you're having, I think a good start to the year. There's a bit of a looming kind of existential question around a legacy ad tech. Legacy desktop guy. Ad tech. Where Is the world moving and how can you adapt, particularly if you kind of raise at a high valuation four years ago? I think there's kind of looming questions around that. No surprise. And then on the other side, we'll talk about this with the agentic announcements. It's moving so fast that everybody's like, we need to participate, we need to be rolling out our AI, but like we have no idea where this thing is going to go. So it's like the whole thing is up for grabs. If that, you know, kind of like framework and synthesis makes sense to you.
- [28:53]
B
Yeah, I think that's accurate. I mean you'd be crazy to not be investing in AI. And how people are doing it is varying a bit. They're automating their own tools or they're looking to transact with other tools or somewhere in between. And that's really where the uncertainty is, is like which of those will be the best ROI for your customers? How much control will they want, how much will they not want to. And then you're in the environment in which the consumers are changing because of AI. And that is a big open question about how it's going to affect you.
- [29:26]
C
So many questions. Andrew Casale had a good framework that I think he's been talking about where you can split the non consumer side of AI into two things. It's either workflow automation or sell side decisioning. And if I go to, you know, the major announcements, which we'll talk into, we'll talk about in a second, it's kind of that, right?
- [29:48]
B
Yeah, I guess, yeah, I, I think of it. I put the framework in a newsletter a couple months ago which is, you know, automating a single tool is number one. Automating complementary tools like basically API integrations is number two. And transacting building a, a adversarial relationship, buyer, seller is number three. And it's interesting, number two doesn't seem to be happening that much. Everyone's doing one or three. A lot of announcements about like, hey, we have a chatbot that solves our UI problems, which is great. And then there's a lot of these experimental things like the Pubmatic announcement from before the holidays about actual transactions taking place. So that's where the action is.
- [30:30]
C
Yeah, yeah, for sure. And then even taking it up a level, you know, we're continuing to see, and maybe this is the first, you know, observation from announcements this week, the rollout of the, you know, for, for want of a better term, you know, the, the, the platform owned AI powered optimization. Right. So we started the week with Amazon, right? So, you know, they had announcement do that. They're kind of doubling down into. Into Live, which I think is very interesting. But there have two modes in the Amazon DSP now, basically like Expert and, you know, not non Expert, where, you know, it'll just be like the pmax for Amazon number two was Reddit, so. And these, with the exception of the one we're about to talk about, the. The names on these are so funny. They're just like, so intentionally generic. They're. They're colluding to make like, like, like names that, that, that are. That are just like, you know, some.
- [31:28]
B
Wait, wait, I forgot. Without looking, I can't remember if it's Reddit plus or Reddit Max, which is.
- [31:32]
C
A. Oh, no, it is Reddit Max campaign.
- [31:36]
B
Max. Okay, got it.
- [31:37]
C
Max campaign, yeah. Which I think is frankly, you know, quite interesting. If there's, you know, like, one platform that I think can benefit from, you know, just the ability to have a, you know, set it, forget it, allow us to find users. Because the fragmentation is so crazy on that platform. Like, I jump from personally, like BJJ to adopts to, like, this weird niche of, like, traveling with one B. Finding me in all these areas can be somewhat challenging unless, you know, you have the actual algo.
- [32:09]
B
Can you explain to the audience what BJJ is?
- [32:12]
C
Brazilian Jiu Jitsu. What do you think?
- [32:15]
B
I just was doing a public service for those of us listening at home.
- [32:19]
C
Stop it. Okay, so. So you had the red announcement. Then finally you had Vyant with, I will give them credit, the best name of all of these. Lattice Brain.
- [32:31]
B
Lattice Brain.
- [32:32]
C
Yeah. Yeah. So. So we've been wa. A, you know, kind of like PMAX Advantage plus for the Open Web and, And credit to the Vanderhooks. They've been early on AI across the board. They rolled out Lattice Brain, which, you know, connects all of their assets to the open Web. And it's, you know, sounds like you. You had a conversation. So I'll ask you to give us a little bit of highlights, you know, give us the goal, give us a budget, and we're going to optimize towards it, which I think is fantastic. What did you learn after talking to them?
- [33:00]
B
I. I learned that to promote this, they dropped a track. They. Did you listen to the track?
- [33:06]
C
You know what? I didn't. My schedule was so crazy, and I ran into Chris and Tim yesterday. The first thing they asked me was, what'd you think of the rap? I'm sorry.
- [33:15]
B
So there's an AI generated rap track that's pretty, pretty good. It's a little bit of a bop that we're gonna put in the podcast. So listen, when they come on, we'll play the entire song. But that's not what you're getting at.
- [33:28]
C
I was disappointed to learn that Tim or Chris or Tim and Chris did not rap.
- [33:33]
B
Yeah, that was a little disappointing. They. They feel like they have some bars in them. The. I mean they have hook in their name. So. All right, so I'm trying to make sense of this. Let's do sell side first, then buy side. So sell side, there's this progression and Eric Seoufer really laid this out well, which is you, you go from having your own inventory, sort of what used to be called walled garden, but now there's hedge gardens, whatever, it doesn't matter to the first step is are you big enough to want people to log in and run the ads themselves? And that's called, that's the gamification of publishers. Every major publisher has an am at the end of their name. Spotify ad manager, Reddit ad manager, am, AM and Pinterest ad manager. So you could go down the list and at some point it stops and you're not big enough to have your own ad manager and you maybe use Dan ads or something, something like that for self service, whatever. And now this new question is, are you big enough to also have enough AI to automate the ad manager? So if someone wants to use the ad manager but doesn't want to expend the effort because you're a small piece of their media buy, can you have a max? And that's really the Reddit play here, where Reddit is important and complicated, but too small for many advertisers to dedicate the time and effort to understand the nuances of the platform. So you have a max and you're like, okay, set it and forget it with brand safety. And you'd expect that. And anyone who's big enough to have an ad manager should also have a max as part of the suite. And then below that you get into the fragmented publisher world where you need a buy side solution. So that's my take on the sell side and where it's headed. And there's nothing sort of unexpected about that development in my mind. Mind. Does that resonate with you, Eric?
- [35:22]
C
Yeah, it makes perfect sense. It makes me wonder, you know, are we going to see more announcements from. Yeah, I would say kind of like a legacy ad tech, but, you know, scaled ad tech moving in this direction because it's really the first and only.
- [35:37]
B
Yeah, yeah. And you know, I know firsthand from my time at Beeswax, there was just so many inbound sales leads were publishers who said, I want my own walled garden. I need, I need like a bidding system and an inventory system. I don't want an ad server. I want some like algorithmically optimized thing. And the reality is that almost everyone has to build it themselves. It's not really that great an option. But yeah, I think it seems like anyone who does, let's call it a bill, a bill a year in advertising revenue is going to be in this camp. Maybe it's a little lower, maybe it's a, it's a half bill, you know, but that sounds like around the right number. And maybe there's some element of like you have to have non standard, like native ad inventory to make it worthwhile. So that's the publisher side on the buy side. I think the VIAM conversation was super interesting because what they say, and you can listen to it when it comes out is that, you know, they started with media planning, they've automated media planning and that was their big announcement a year ago. They got so much press and excitement. But now what they are trying to do is what they call outcomes, which is a complete, you know, set it and forget it for the open web outcomes, you know, results without, you know, all of the work and, and they claim they beat a human being head to head. And that's great. But I think the interesting thing is like will customers on the open web on the, or the open side of advert digital advertising choose set it and forget it or do they want the control and how is it going to break like segment by type of advertiser, by size of advertiser, by KPI. Right. And I think that they're being a little modest in saying, well we, you know, we went after E commerce, we went after like really well understood and defined KPIs. That makes sense. Right, so that's the divide on the buy side.
- [37:27]
C
Yeah, that makes sense. One of the pushback, you know, points that we've discussed in all of these as the reason why it may not actually scale to, you know, open web is because open web is generally domain of like the Fortune 2000. It's not the performance marketers who really don't care about quality per se. They just care about the outcome. The question is, you know, can you, you know, intermix and, and Viant was very careful to put this in their release you know, quality controls and brand safety. Can you kind of like bring the best of the open web from a control and safety standpoint and marry it with outcomes and not, you know, kind of optimize yourself to death and, you know, deliver $100 campaigns?
- [38:09]
B
Exactly. And there you could divide it a bunch of ways. There's CTV versus, you know, versus web, where CTV isn't as nearly as much performance despite, you know, the successes there. Most of CTV is not performance oriented and the KPIs are a little harder to measure. You have like traditional TV advertisers versus non TV advertisers. A lot of ways to divide it up.
- [38:32]
C
Agreed. Okay, so theme number one, AI optimized platforms. Theme number two is agentic. And this was where it was hot and heavy this week. I'll take through a couple of things and maybe there's one I'd love to talk about. So the first is maybe in one corner is the iab. So you have IAB laid out their roadmap. You know, if I could and I'll ask you to, you know, let me know if I'm accurate on this, you know, build on top of the existing tools. Let's not retool everything. Let's build on top of the existing tools and just have a framework accurate.
- [39:09]
B
That that is the way they're positioning it. Yeah.
- [39:11]
C
Okay. And in this corner you have AD CP and all these ad tech companies that are just like running wild and going and like putting seller agents out there and buyer agents out there and, you know, just like going really fast and having all these pilots and experiments. Magnite Pubmatic, Yahoo. I probably missed a few others. All ADD announcements around launching their agents and their pilots this week.
- [39:39]
B
Yep. Yeah, it's interesting. I have to admit, I had not had time, despite not being in ces, to fully read the IAB proposal. But I think you're characterizing it properly. There's also the history here, which I think I talked about in a previous episode, which is like the folks who are running ADCP are have a little bit of bad blood with the IB because of the way the IB rejected pre bid as a standard. So creating their own committee and funding it and having Randall Rothenberg run it is kind of interesting choices across the board very much.
- [40:14]
C
Which is why it was in this corner and in this corner. That's going to be one to watch this year. The one announcement that was pretty fascinating and it's wild because we're an investor in the company. This is Newton Research and this isn't even like the core of what they do, they build data science agents. It was an announcement that kicked off the week that Newton Research built buyer agents to execute buys for RPA and an unnamed client across NBC using Free Will, seller agent for linear television. This is so cool.
- [40:46]
B
It is cool. Yeah. I'd love to see Free Will where I used to work, get some cutting edge press. One of the big advantages Free Will has is that they're so tight with NBC because they're both owned by Comcast and it's great to see them actually do something with that. And you know, being able to schedule linear is a, is a data problem more and more and it's, it's really interesting. I love to see this, love to see more of this.
- [41:12]
C
Maybe Final question on the, on the agentic front. You put this in the show notes. So Walmart had a couple of agentic announcements. So they announced Sparky for consumers. Sounds like it's their answer to Rufus. That's pretty hilarious. Amazon Zufus.
- [41:28]
B
It is.
- [41:29]
C
And Marty for advertisers. Did you dig into this with your conversation?
- [41:34]
B
Yeah, I had a great conversation with Karim about it. I didn't want to ask him at the time, but I want to ask you. Wally was right there, man. Like Walmart Wally, like why Sparky?
- [41:46]
C
That's a great question. Yeah.
- [41:48]
B
So I don't have the data at my fingertips, but Amazon has made some noise that their Rufus is causing some huge increase in, in purchases.
- [41:57]
C
I've heard this. Yeah, yeah.
- [41:59]
B
But is that really incremental? Like I've used it once, but I was the same question I would have asked in the reviews. Like it just saved me time going through the reviews. So I'm a little skeptical about, you know, whether it's really incrementally generating a huge amount of revenue. But anyway, it's a great consumer experience being able to ask someone questions about the product. And obviously it can, you know, be used for commerce media too, where you could recommend other products potentially if someone is paying you. So that's really interesting. And then for our world, Marty, is their AI for advertisers. So they're enabling their advertisers to skip steps to automate, et cetera, similar to what Amazon's doing. And one of our favorite startups, gg, is out there with a bot that's just for Amazon. You do have to question whether single platform automation is really a business model for an independent company. Right. You know, because you know the existing. The incumbent's going to automate themselves.
- [43:00]
C
Yeah, yeah, no, I get it. Not to get too far into into Gigi. If we go back to the interview, Adam's response to, to that and, and the why. The why start with a single platform is you got to start somewhere, right?
- [43:13]
B
Yeah.
- [43:14]
C
You know, have a plan to go from there. But yeah, for sure. Platform risk is a, is, is a thing. There's a lot of partnership announcements as usual. I don't know how many we want to, we want to get into. There's a couple, but one dropped this morning that I thought was pretty interesting. So Liveramp and Publicis announced a strategic partnership. So it's Publicis AI and their data assets combined with Libramp's collaboration tech and their kind of like partner data sources to do something that, and the positioning was a little difficult because I just read an Ad Age article that gives publicists a, an advantage and it was not clear that what, what Publicis has that other holdcos don't have access to with this. I don't know if you had a different POV if you spoke to anybody about this, but it's making the rounds.
- [44:06]
B
Yeah, I have no idea. It's very hard, these data type things to really peel back the onion on what it means. And I feel like Live Ramp is a really, is in a really fraught position right now where they have obviously a very solid business, but how they fit strategically when big agency hold codes, have their own data assets and I just don't, I don't have a good explanation for how it all fits together, to be honest.
- [44:33]
C
Yeah, no, it's a, it's a question. In the Ad Age article it noted that Publicist reportedly was looking at acquiring Live Ramp. So, you know, perhaps this is a step on the path there. Perhaps this is, you know, hey, maybe we don't get married. We, you know, I don't know. I don't know what they make sense.
- [44:50]
B
I think it makes, yeah, dating relationship. I, I definitely think that Live Ramp at some point is owned by someone larger. I, I feel like they've been. They're in a very strategic spot. But the privacy backlash of someone acquiring them, especially if it was like a tech giant, a Google, Facebook, even an Oracle, etc would just be too much right now. Now because we all know what Liveramp does, but if it was put through the press and like stealing your data, selling your data, all that stuff would probably be pretty bad for anyone who acquired it. So maybe an agent Hildeco can get away with that because they already have so many data assets.
- [45:29]
C
Yeah, that's an interesting angle. Maybe one other that you know, we can touch on is. So there was the announcement of open ads by TTV last year and they did an announcement with the first wave of publisher partners that are participating on open ads. AccuWeather, Arena Group, Buzzfeed, Guardian, Hearst, Newsweek, People, Ziff Davis. So it looks like they've got this thing off the ground with some great pubs.
- [45:54]
B
Yeah, yeah. For the reminder, for people who aren't as deep as we are, Open ads is effectively TradeDesk's fork of pre bid. So it's their own version of pre bid. So for a publisher to switch, it's probably not a big deal. It's basically the same thing but slightly different. The question from the start was why would they switch? And the unanswered question is what is Trade Desk's pitch to these publishers? You'll get more money, you'll get better rates. We won't buy you unless you change. Those are kind of the open questions there, which maybe someone on the inside has the answers to. I would say it's not really a needle mover for the stock, for the companies. For anyone else, this is very much inside based products.
- [46:39]
C
Got it.
- [46:39]
A
Okay.
- [46:39]
C
Didn't I see Jeff Green respond to a question by you about this on LinkedIn a couple weeks ago?
- [46:45]
B
I don't recall that. Maybe I did ask him something, I think in LinkedIn, so maybe he did respond, I don't know. Oh, I asked the question which a lot of people have been asking, which is like if you imagine that you have the exact same demand and you have one sort of system that's meant to optimize for publishers and another system that's meant to optimize for advertisers. Won't the publisher system ultimately be better for publishers, like higher rates? Right. And I'd have to look at what he said, but I think it's a good question.
- [47:18]
C
Yeah, I saw that he replied randomly. You know how just like certain replies like become notifications. I think when Jeff Green replies to somebody might hit my notifications. I guess I'm a fan. But I think he said, you know, we should talk about this in person. Something like that.
- [47:32]
B
Well, if he wants to come on the pod, you know he is welcome. He's refused to date. So I'm not going to read his LinkedIn response.
- [47:39]
A
Nonsense.
- [47:40]
B
Unless he comes to Pod.
- [47:41]
C
Would we do, I'm putting this out there in the world. Would we do an in person POD with Jeff? I know you love your remote pod, but would you, would you break the rule for Jeff Green.
- [47:51]
B
I don't have a rule. Yeah, Jeff, you're welcome. On the pod in person. You could view it in Manhattan. We could go to Eric's house in Jersey on the shore. You can come to the stage of marketexture live with thousands of people. There is no restrictions. You have an open invite.
- [48:07]
C
Come through, Jeff. All right, well, those are the big announcements for the. For the week and a bunch of news. If you synthesize it, it's very clear where this space is going.
- [48:16]
B
Very clear. Well, get home safely and I hope everyone is listening to this on the plane, enjoys it, and had a great ces and maybe I'll see you there next year. Or at the very least, I'll see it Possible. So more construction.
- [48:30]
C
Alive comes first.
- [48:31]
B
Oh, yeah. Markstetch Live by the calendar, March 11th and 12th. Thank you.
- [48:36]
C
All right, everybody, we'll see you next week.
- [48:38]
B
See ya. Thank you for subscribing to marketecture. New interviews are added every week at marketecture tv and your favorite podcasting app.