← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 105: Live from CES, plus Professor Matz from Columbia B-School talk about psychology in advertising
00:49:53
Loading summary
Transcript127 lines
- [00:00]
Eric
What do planking, Gangnam Style hashtags and viewability have in common? They were all popular in 2012, but while the first three fizzled out, viewability has lingered. Adelaide thinks it's time to retire that relic. That's why they created au, a modern metric you can use to plan, buy, optimize and even trade media. Edweek calls AU the attention economy's most widely recognized metric and it swept the measurement category at last year's Ads Exchanger Awards. Don't get stuck with a fanny pack of metrics. Ask your agency about switching to AU. You can also visit Adelaide metrics.com that's adelaidemetrics.com.
- [00:42]
Professor Sandra Matz
This podcast is brought to you by Azerion. Azerion is one of the world's leading full stack digital advertising partners for 300 plus agencies and 2500 plus brands. Their proprietary platform called Azerion Infinity, seamlessly integrates the capabilities of owned and operated dsp, ssp, DMP and ad server technology to drive incremental performance, value and efficiency across the full marketing funnel. At Azerion, they offer custom audience activation, leveraging cookie identity data and contextual signals to gauge users across digital ctv, olv, doo h audio in game channels everywhere else. Choose between DIRECT or PMP activation, leveraging world class trading teams to deliver top of plan results or alternatively access audiences through a data connected PMP as a market leader in the sell side curation category. That's Aerium, the sponsor of this podcast.
- [01:42]
Eric
Welcome to the Mark Dexter Podcast coming right at you from Las Vegas ces. Eric, you having fun here?
- [01:49]
Unknown
I'm having zero fun. I've been in meetings hourly since I've arrived here. The meetings have been great but I have not had any fun.
- [01:59]
Eric
I'm not busy and I'm also not having fun. So I'm a double whammy. Market is busy so you may have seen our various posts on LinkedIn, etc. We are running a program called Suite Suites where we're going into people's suites and recording what they're up to. We have MIQ, OpenEx, Yahoo, a bunch of others. I'm not going to name all of them, which is going to make our sales guy unhappy, but there's a really great content. It's really fun. They're short. I'd recommend watching them, especially if you know the people involved. It's always fun to watch people goofing around on camera and a little serious stuff too. So this week we have an interesting guest so we're a little change of pace so we Have Professor Sandra Matz from Columbia Business School, and she has a new book out called Mind Masters and what she does. She's a professor of management, but she really focuses on statistical analysis of consumer behavior. And she talks about using psychology in advertising and marketing. I found this conversation super fascinating. Did you learn stuff from this, Eric?
- [02:57]
Unknown
Yeah, this was one of my favorite interviews we've ever done. Like, the whole time I was just, like, writing things down. My mind was going crazy. She talks about, you know, basically, like, the five personality traits. Openness, conscientiousness, extroversion, agreeableness, and neuroticism, and how that plays into advertising and marketing. And, you know, how if you play this out, you know, you could use these in a way to help influence purchases, help influence sales. It's so cool to, you know, really apply research to this stuff that we. That we do.
- [03:29]
Eric
And there's the AI overlay, because AI can help you produce content for those people, but also can analyze their behavior and put them into categories. It's really interesting.
- [03:37]
Unknown
The AI discussion on how, you know, you can use chat GPT to basically build these models was the best part.
- [03:43]
Eric
Yeah. Well, we'd love to get your feedback if you like this kind of content. This is a little bit out of our normal wheelhouse. It's a little more marketing, a little broader. So if you like that stuff, let us know. And then we had news of the week from cs. Obviously, a million announcements we'll go through, including Comcast, new Universal ads. So let's get right to it.
- [03:59]
Unknown
Let's go.
- [04:02]
Professor Sandra Matz
All right, so, Professor Matz, thank you so much for being here.
- [04:05]
Unknown
Well, thanks so much for having me.
- [04:07]
Professor Sandra Matz
So I saw you speak at Columbia Business School a couple of months ago. I'm an alumni of Columbia, and they are holding various ad tech and marketing events. And it was really interesting to hear you speak. So I reached out and I said, well, hey, why don't you come on the podcast? And you said, well, hold on. I have a book coming out, so let's time it that way. So you got a lot going on. So, Professor Matz, so you, first of all, what is your area of study that's to do with psychology? So tell us about that.
- [04:33]
Unknown
Yes, well, you got to do that. You got to get the marketing right. That's why we're doing this now when the book comes out. So totally on topic. Yeah. So my background is actually in psychology and computer science. So it's trying to study human behavior not by giving people questionnaires or having them press a Couple of buttons in a lab, but by observing their data. So anything from what you post on, on social media to your credit card swipes, to the sensors that are embedded in your smartphone that capture everything from where you live, who you meet, who you talk to your network, and how we can translate those digital traces into something that's more meaningful on a psychological level.
- [05:10]
Professor Sandra Matz
Oh, wow. And presumably, let's do the book pitch right now. So what's the book called? Where is it available and what's it about?
- [05:17]
Unknown
The book is called Mindmasters, the Data Driven Science of Predicting and Changing Human Behavior. It's Harvard Business Review and I think it's available like on all of the major retailers, Amazon, Barnes and Nobles. So the way that I think about it is it's like the human story of data. Right. I think oftentimes we think of data as just predicting markets, as kind of predicting outcomes. But I wanted to see who's behind the data. What can we learn about the people who generate the data by observing all of these traces? So kind of diving into predicting your personality, predicting your values, and then also thinking about applications. So once I can get these insights, to what extent can I use them to potentially change your behavior, change the way that you think, feel and behave?
- [06:02]
Professor Sandra Matz
All right, so that's a really big topic, psychology. Someone's happy, sad. I'm not really sure how to get my arms around that, but I know that there's something called. Is it the Big five? Is that what it's called?
- [06:12]
Unknown
Yeah, it's like the, I would say the most popular personality model that we currently have in science. I think a lot of people know mbti, which gives you a cute little label. And they are kind of nice. They're not great from a psychometrics point of view. So the Big five essentially gives everybody a score on these five dimensions. So we have openness, conscientiousness, extroversion, agreeableness and neuroticism. And you can imagine this is like this five dimensional space and you're like a little dot somewhere in this space and it just says something about how you're likely or most likely to behave in this world.
- [06:49]
Professor Sandra Matz
So it is the Big five. Has it been peer reviewed? Is it empirically proven that it's real? It's not just a Myers Briggs test or something, Right?
- [06:58]
Unknown
Exactly. So the reason for why a lot of scientists are using it is that it's been around for almost 30 years now. And it works across cultures. So the same structure actually shows up across cultures, across different languages, across different social Demographic groups, and it also predicts behavior across a whole bunch of domains. So that's why it's so popular. If you want to see what correlates with vocational interests, what correlates with music preferences, where it correlates with mental health and physical health, the big five is just this relatively pragmatic framework that predicts behavior across all of these areas.
- [07:33]
Professor Sandra Matz
Yeah. So, okay, now that you just begged the question or asked me to ask the question, which is what's the application to advertising and marketing?
- [07:40]
Unknown
So the way that I think about it is actually it's an equivalent to what we used to do for centuries. Right. If you think about any offline conversation that you have, it's to some extent customized to who's on the other side. You don't talk to a three year old in the same way that you talk to your boss or that you talk to a good friend of yours. What we do very intuitively as humans is we try and gauge who's on the other side. What do they care about? How can I frame this in a way that's most persuasive? Every toddler gets really good at figuring out how to convince mom to give them something versus dad. And I think as we move to the online world, that got lost a little bit because it was essentially now a number in the system. We talk to everybody in the same way. TV advertising. And now we're gradually shifting back to saying, okay, I can use the data that you generate to make predictions about who you are and not just on a behavioral level. So it's not just that I know what you bought and maybe kind of the places that you've been to, but on a much more holistic psychological level. And then I can use these insights to say, okay, maybe I can offer you different products, because as an extrovert, you might be interested in different activities than an introvert. But I can also change the language. If I want to market a beauty product to you, for example, I would probably sell it very differently to someone who I know thrives in these exciting social environments versus someone who kind of prefers to be home by themselves.
- [09:02]
Professor Sandra Matz
Tell me about the study you did with Facebook data, because I saw it when I saw you at the presentation. I thought it was just fascinating what you were able to do there.
- [09:12]
Unknown
Yeah, there's a couple of studies, but I think one of the first ones was actually in the context of beauty products. This was a study, and it's almost 10 years ago now that we did this, but essentially what we tried to do is we were approached by a beauty retailer, and their ask was, well, we have all of these products. How do we pitch them to different types of personalities? So why is it that people buy beauty products for all kinds of reasons. So what we said is like, okay, can we divide an audience into women that are more extroverted? So those are the people, like the women who kind of thrive, again, on social interactions. They love to be the center of attention, and then more introverted ones who prefer the peace and quiet of their own comfort, if you want. So you can imagine, we just teamed up back in the day with a creative team of the beauty retailer, and we came up with different ads for these different groups. So the extroverted ads would always be like, women on the dance floor. And I think the copy said something like, dance. Like, no one's watching, but they totally are. So playing with their need to be the center of attention. And then you have the introverted ones playing. It's like, usually one woman in front of the mirror or like in a quiet setting, I think the copy would say something like, beauty doesn't have to shout. So you can see it's pretty simple, right? It's just taking the motivations and playing into these motivations. And I think that's the interesting part is how do you get to extroverted and introverted women? Right, Exactly. So it's not something off the shelf that your Facebook targeting platform allows you to do. If anything, I think they've changed it. They've made it a lot more restrictive since some of the major scandals. So you used to be able to target sexual orientation, political ideology. That's off the table. And you were never able to target personality. Even though I should say they have a patent from 2015 Facebook to predict personality from their data. So the fact that they were very surprised when Cambridge Analytica happened. But what you can do is you can. Essentially, what we know from science is that certain interests are associated with certain personality traits. So if you're extroverted, you're more likely to, like, going to parties, like something like beer pong, cheerleading, all of the fun stuff. And then if you're introverted, you're more likely to like computers, reading manga, role playing. So there's like, hold on.
- [11:22]
Professor Sandra Matz
That's fun stuff too. Like, hold on. I object to the characterization. Right, okay.
- [11:28]
Unknown
Anyway, keep going so you can sort yourself. Right? And that's the interesting thing is that that's something you can target, like on these platforms like Facebook, Google, you name it. So now, because we knew empirically that some of these interests are related to extraversion and some of these interests are related to introversion. We could essentially define these audiences based on the interest that people had. And it's an extremely crude way of targeting. Ideally we would get all of the data of the people, we would run it through our models and then have a personality profile for each person. What we did in this case was just say, okay, if someone is connected to the interest of partying, that person is in the extroverted bucket. If someone is interested in reading and books, that person is in the introverted bucket. So extremely crude way of targeting. And then we just had our extroverted and introverted audiences. Extroverted, Introverted ads. And the idea was if there's a match, if an extrovert gets an extroverted message, introvert gets introverted, they would be more likely to purchase. And that's exactly what we found. So what we found is like a 50% increase in this case the conversion rate. So how often did women buy something compared to how often the ad was shown? And that's a big lift if you think about the fact it's super crude targeting. And the only thing that we changed was the ad that people saw on Facebook. So once they go to the beauty retailer's website looks exactly the same. So it's a pretty substantial effect just by playing into people's psychological motivations.
- [12:59]
Professor Sandra Matz
So not everyone has the benefit of, well, not every platform has the benefit of Facebook's data knowing someone's interested in partying or something. Do you think this is applicable to more anonymous environments like the open web or maybe what Netflix knows about what people watch? Do you think those same techniques can work?
- [13:18]
Unknown
Yeah, so there's, I think a couple of things. One is that oftentimes people have first party data, right? So you might have existing customers that, you know, you kind of, you collect some first party data on. It used to be the case that it was really difficult to translate that into psychology. Right. So like you, if you didn't have a proprietary model that translates your data into something like the big five, that was it. Now with Generative AI, there's essentially an open source tool that you can just pop and it's remarkably accurate. You can just pop any data and say, hey, here's the purchases of someone, here's the websites that they visited, here's everything that I know about them. Who do you think that person is? When it comes to the big five personality traits, it's not perfect. So the same way that our original models were Never perfect, it makes mistakes, but it's pretty damn accurate. And I think that is totally democratizing access. Right, because like you can use first party data in a way that you couldn't use before. And you can also say, well, here's the options that I have for targeting people on the platform that I'm using. Which ones do you think that I should pick to reach an audience? That is extroverted. And because ChatGPT has read entire Internet, they have like a pretty good sense of what that could be because they know what the big five are associated with. So there's a way that we can do this now at much larger scale without having to have these proprietary models.
- [14:41]
Professor Sandra Matz
What's the tool called? Do you remember?
- [14:43]
Unknown
Oh, the tool. I'm just talking about ChatGPT.
- [14:45]
Professor Sandra Matz
Oh, you're just talking about ChatGPT.
- [14:46]
Eric
Oh yeah.
- [14:46]
Professor Sandra Matz
I thought you was a plugin or something.
- [14:48]
Unknown
Any, any off the shelf. So that's, I think, the remarkable thing. But any off the shelf. LLM knows everything about psychology, knows how to translate your behavior into something that is. And it also means we can. We talked about the big five, right? So we are only always limited by, well, we're going to use the big five because that's what we have data for. It doesn't really matter. Today I can just say, well, right now I feel like I should be predicting your values or tomorrow I should be predicting your moral foundations. If I want to advertise a more political space. Those models can predict whatever you ask them because they know everything about psychology.
- [15:24]
Professor Sandra Matz
So there might be a parameter that's not one of the big five that's still useful, like how religious someone is, I guess. Could be one example.
- [15:31]
Unknown
Yeah. Or if you think about the health space, there's. There's something that we call prevention versus promotion. To what extent are you trying to get better at what you're doing or to what extent are you trying to prevent you from getting worse? So there's very much depends on the context which ones might be the most interesting.
- [15:50]
Professor Sandra Matz
Is this better or worse for privacy?
- [15:54]
Unknown
Well, what is the comparison? What's the counterfactual?
- [15:57]
Professor Sandra Matz
Okay, so yeah, so let's say the comparison is cookies that tell you everything the person did across a lot of different websites and activities. And now we're in a world where maybe you can identify the person specifically, but you can make a judgment about what they're like.
- [16:14]
Unknown
I think it's a great question because I think for me the main difference is perception. Was it always extremely intrusive and invasive? If I had access to all of your data. Absolutely. Because a computer could make sense of millions of data points. A computer doesn't necessarily need five personality domains, dimensions to understand who you are. But what I think is different is that people suddenly realize, hang on, you can take all of my spending records and you can translate it into an understanding of whether I'm impulsive or neurotic or open minded or like, you name it. I think just people's perception, it feels a lot more intrusive because that's how we think as humans, right? We think of humans not as like, oh man, it's like GPS records and it's purchases. It's like a person, it's dependent on the person.
- [17:01]
Right? Because for one person this might be intrusive. For another person it's like, whoa, will the ads and offers I get be that much more relevant and tailored to me based on all of this data? That it's actually more of a service. So I think it's fascinating that it depends on where you sit on the sideline.
- [17:19]
And I think that's true. So I think when I think of as intrusive, I'm mostly talking about there's someone who has a better understanding of who I am and what I want. And then the second step, you're absolutely right, is what is that a good thing or a bad thing? Right. Because like, and you said it like there's, there's just no way that we can consume all of the information that's online. So there needs to be some way of filtering like what we see and how we get to the content that we want to see, the products that we, that we like. And right now we're doing this purely based on past behavior. Right? So if I look for a camera on Google, that camera equipment for that camera is going to follow me around for the next four weeks and I don't want it anymore. Or I've already bought in like personality. What I like about it is it sits kind of in a sweet spot. It's not exactly like your past behavior to future behavior. But we're saying, well, hang on, maybe camera is an indication that you are like more artistic or open minded and then you like to explore. So now once I have that, I can still recommend things that are somewhat in that realm but not exactly tied to that camera. So I think personality is a way in which we can still let you explore a bit more, but in a realm that you still kind of show some affinity and it's not just completely random. So I totally Agree. I think there's so many use cases where this is actually extremely helpful.
- [18:36]
Professor Sandra Matz
What practically should our listeners who are involved in day to day advertising be taking away from this? Is there something they can go back to their desk and do? I know that's not your realm as an academic, but. But actually it'd be interesting to hear what your top of your mind thoughts are.
- [18:54]
Unknown
I mean, so the one thing that I always pitch to practitioners is that if you can make it a two way street. So like I think oftentimes what we do is we just advertise online and we find people and it depends a little bit on whether you already have a connection. Right. But like the moment that you have customers that are willing to engage with you, there's a risk of like doing all of these passive predictions and just broadcasting to people because first of all it feels like a little bit more creepy. So you're just collecting, harvesting data and you don't really have any transparency and you're going to make mistakes. As I said, those models, they're pretty accurate on average, but they constantly make mistakes and predictions. And that's frustrating because if you get it wrong and you think that person is someone who they're not, that just means that you're not maximizing your efficiency because you're sending them stuff that they're not going to buy and they're having a terrible experience because they're now seeing stuff that they don't want. So they're much more likely to leave you if you can. You can actually have consumers engage with them and say, look, here's like we're trying to help you find the stuff that you're interested in. And you see this I think happening a lot more with applications with like you go through like an experience and they really kind of show you that they want to improve the service and here's how they're doing it. So you have like interactions with consumers which reduces the risk of being seen as kind of creepy and just you losing your reputation and trust. Plus you're going to be much better because you have the user input in addition to the predictions that you can make. So I think that for me is the most promising way forward in which we can again move it from it's just intrusive in a negative way to actually there's a lot of understanding and it's helpful.
- [20:34]
Professor Sandra Matz
It seems to me like the first step would be to understand what the overlay is between psychological traits and your product or offering. I mean, it could be that your product is Only for extroverts, so that split isn't useful. But if you could take, I don't know, product reviews or interviews with customers or whatever sort of text heavy content they've given you, and then ask ChatGPT to figure out what the personality attributes are, that seems like a really helpful first step.
- [21:03]
Unknown
There's different ways, right? So you can either say, well, here's what we know. ChatGPT, tell me, what do you think is the personality affinity? It's probably going to do a decent job. You could also just say, well, who are the people who have bought this product before? And if we can have their personality, we can essentially just kind of create this personality profile of all of the products that you have. What I think is perhaps even more interesting when it comes to personality is that you're not limited to like one affinity group. Right. Oftentimes you can just think of ways in which you could make a product exciting for different groups of people. And because it's like it's personality, right, we can all deal with it. Once you give graphic designers a description of what it means to be extroverted or agreeable or neurotic, they will have a sense of like, okay, how do I make this camera interesting for someone who's. How do I make this product interesting for someone who's open minded? So for me, the fact that we were traditionally just trying to figure out who's the 1 target audience for this product, I think almost breaks down a little bit because we can now go, maybe that's the one that we start with. With limited resources, that's probably what we want to go for. But the moment that we have more resources and we want to try and reach a wider audience, we can actually say, okay, here's a group that we haven't traditionally attracted, but how do we make it attractive?
- [22:20]
Professor Sandra Matz
Right. I'm wondering what a designer would say if you asked them to make an ad for neurotic people. Is neurotic. Neurotic has sort of a negative connotation in commonplace usage. Are you, is that the case in the way you're using it?
- [22:35]
Unknown
So neuroticism is actually one of the interesting traits because there it's not entirely clear how you want to do it. Right. So people who are neurotic are usually the ones who fluctuate more in their emotions. And as you said, it's not always a bad thing. So the completely emotionally stable people are essentially psychopaths. It's true.
- [22:54]
Right.
- [22:54]
So they has no fluctuations, no empathy. So being neurotic on the extreme is just we know that it's related to a lot of health outcomes, so it's oftentimes bad for the individual. But what you can do as a marketer is say, well, how do I help you kind of reduce that level of anxiety and reduce that level of stress. Right. So it's essentially a way of helping you get out of that state that you might have a tendency to be in and to feel better and a lot more calm. So I think there's a lot of value in that space that marketers can add.
- [23:24]
Professor Sandra Matz
What do you think people are going to get out of it? What do you hope people get out of it?
- [23:28]
Unknown
Hopefully many things. I hope people get to page 15.
- [23:35]
Professor Sandra Matz
Is it an acrobatic.
- [23:37]
Unknown
Oh, my God. It's super interesting book. And then. No, but I think what is a couple of lessons. I think one is just like, look, here's what your data says about you. And it's not just like bits and pieces. It's not just like what you post on social media, but it's really this window into your psychology. And the second piece that was important for me in the book, and it's actually coming back to what Eric was saying, is that it's not just all bad, but if we, if we look at the media coverage it's all about, well, it's being used to manipulate elections. There's so much potential. Right. Think, like, if you could take the marketing context, how do you make it more relevant? Like it's saying, like it's only spam if it's not relevant. But then also instead of just getting people to spend more, can you use it the same marketing techniques to get them to save more? Can you use it to help people feel better when it comes to mental health? So I think that the big part of the book that I try to get across is an understanding of who you are. Obviously has a potential for abuse, but it also has enormous potential to be used in our best interest.
- [24:39]
Professor Sandra Matz
All right, well, thank you so much, Professor Sandra Mats for joining us from Architecture. The book is called Mind Masters. I'm going to buy it next week when I'm back from ces. We're recording this the week before. It's great to have a different perspective on this show.
- [24:55]
Unknown
Thank you.
- [24:56]
I'm pre ordering it right now, actually. This is so fascinating. I'm doing this. This is so fascinating. I already have questions for it.
- [25:03]
Professor Sandra Matz
We'll be back in a minute with the news of the week. Thank you.
- [25:09]
Eric
What do Planking Gangnam Style, Hashtags and viewability have in common. They were all popular in 2012. But while the first three fizzled out, viewability has lingered. Adelaide thinks it's time to retire that relic. That's why they created au, a modern metric you can use to plan, buy, optimize, and even trade media. EdWeek calls AU the attention economy's most widely recognized metric, and it swept the measurement category at last year's Ads Exchanger Awards. Don't get stuck with a fanny pack of metrics. Ask your agency about switching to AU. You can also visit adelaidemetrics.com@adelaidemetrix.com.
- [25:50]
Unknown
Great. Where are you?
- [25:52]
Eric
I'm in the Planet Hollywood, which is, you know, not the best hotel. I've been having some major wi fi problems. I'm giving it one star here on the podcast. Where are you?
- [26:00]
Unknown
I'm sorry, I'm at, I'm at the Cosmopolitan where much of the, the action is taking place here at, here at ces. I think we've had two different experiences from our, our initial prep call. Yeah, I've been, I've been doing 10 meetings a day. And you have not.
- [26:16]
Eric
I've not. I've been really not slacking. You could call taking it easy. You called slacking. My team, the architecture team, is working their asses. Us atg, Jeremy, we got Nicole Pruis, who's our spokesmodel for the week. And we're doing, we're in people's suites, we're taking videos. It's awesome. But I don't have anything to do. I'm just in the way. I'm not a producer, I'm not the talent. So I'm just leaving them to do what they do. And then I wander around the Aria lobby trying to have random meetings. It's pretty low key. I don't have any regrets though. No regrets.
- [26:47]
Unknown
I'm going to send you 11 calendar invites for today and we'll fill that schedule. So, okay, you've been meandering around the Aria lobby. For those who are unfamiliar with ces, basically the Aria and Cosmopolitan are where all of, all of the meetings are held. All of the ad tech and, you know, related people hang out and connect with each other. And there's in the area, there's the space where companies do meetings in their suites with your run ins with people. What's the Vodcheck? What are people talking about?
- [27:20]
Eric
I think a lot of ctv, a lot of retail media, it's kind of always the, that's always what everyone wants to talk about. I think that we'll get to the news but Comcast announcement of Universal ads is the one I'm most interested in. I think that was a really interesting announcement and there's just a lot going on in ctv. Last week's announcement two weeks ago the about Google allowing fingerprinting. We covered that or Alan Chappelle covered that. This week's Monopoly Report newsletter. So I would read that. Just a lot of like jockeying. I think CTV has got to a maturity state where people are now jockeying for share instead of inventing things from scratch.
- [27:57]
Unknown
Yeah, yeah, that makes sense. So a couple things. Number one, people are still talking about curation a lot. I think the Audigen deal woke a few people up in terms of the opportunity and there's still a lot of talk about and this is maybe a little bit skewed towards legacy platforms and some of these next gen programmatic startups. How can we get into this game in a differentiated way? I think curation is going to continue to be a big conversation in 2025. That's sort of thing one. Thing two is surprising to me where I think the CTV to retail media ratio in terms of my conversations of what people are focusing on, what people are building towards, it seems to me like 70, 30 towards CTV which is a big shift. Like everybody was trying to get into retail media for a long time and now everybody's trying to figure out their CTV strategy. So that was surprising to me and then just there's a general sense of optimism which feels really good.
- [29:02]
Eric
Obviously retail media is still super exciting but I think people have a hard time getting into it. Retailers have very long lead times so it could take years for them to build a program and on the buy side it's just incremental. Each, each single retailer they work for, work with is just another project kind of. So it's less dynamic, it's. But the money's there.
- [29:25]
Unknown
Yeah, yeah, I agree. I continue to come away thinking that this, this concept that you know we've talked about on the POD for as long as we've been doing the pod, some sort of aggregation play, some sort of way to you know bring like non endemic to demand to retailers. You know some way to aggregate retailers into a curated or network like vehicle seems to be like the, a bigger opportunity this year than it was last year and then it was in 23.
- [29:53]
Eric
Two different things. There's a non endemic where we see a little stuff going on and then there's the network of especially in the grocery category. So I think, I think KROGER'S is kind of doing something. Kroger's is doing something as a network. But yeah, you just have this fragmentation problem where not everyone is Uber, not everyone is, you know, doordash. How do you help those, those retailers?
- [30:15]
Unknown
But then it's the number three, which. And is it Andrew Lipsman that talked about on the pod a couple weeks ago? The Wawas of the world. More and more of those are coming online. And you know, that's, that's, I think the other aggregation plate, Ari is now showing me live from Las Vegas, a Wawa sandwich. Rapper Wawa.
- [30:35]
Eric
I have the Wawa backgrounds here, I guess. Yeah. Out of a strange confluence of events, I happen to have Wawa wrapping paper, the stuff they put the sandwiches in in my hotel room in Vegas, even though there's no Wawa within 3,000 miles of air. And you know, I don't know, things get weird when you do a pod every week. People listen and they start doing weird stuff like bringing you Wawa swag in Vegas.
- [30:59]
Unknown
That's awesome. You have fans. I think it's great. The fans love you.
- [31:03]
Eric
The fans love me, I guess.
- [31:05]
Unknown
All right, news of the week, let's get it kicked off. I agree that maybe, I mean, maybe it was like the first announcement, but I think probably the most interesting one and one that's certainly on everybody's lips is this Comcast universal program. Why don't you take us through it?
- [31:20]
Eric
Yeah. So Comcast announced universal ads and effectively it's SMB self service TV ads, which is a theme we've talked about on this pod quite a bit because we're both involved with Vibe and TV Scientific respectively. And you know, Comcast has enormous assets here. They have free will, they have beeswax. So they have the tech to make it work. They have the inventory from NBC. They have the existing salesforce with effective. And there's always this concept that as streaming disrupted linear viewing, the sales channel would also be disrupted and changed. And Effective, the traditional local advertising group still has a pretty big ticket price. Like that is for like car dealerships and chain restaurants. They're spending tens to hundreds of thousands of dollars on local ads. There's a different market for direct response, very small businesses, hundred dollar budgets. And also people like app install companies or shopify merchants. And those folks have traditionally not been part of the TV ecosystem. And Comcast is putting its foot out there and it's going to do that. They also part of the announcement was that they also are going to do it with some other supply sources. I think was it Box that they were doing it with Warner Brothers, Discovery, amc. Yeah, exactly.
- [32:41]
Unknown
And Roku. And Roku. We should talk about that in a second.
- [32:44]
Eric
Yeah, yeah, that's kind of interesting because Roku announced their own self serve direct response system just two months ago. Yeah, yeah, we'll come back.
- [32:53]
Unknown
Yeah, I think it's great. So a huge validation of this thing that we've been talking about and the SMB thing you've been talking about in particular and validation for Vibe, which you're on the board of. I think that's awesome. Kind of thought A thought B is in the piece. I think I read it on CNBC or whatever we'll link to in the newsletter. There was a lot of talk about bringing Facebook advertisers, Google advertisers to ctv, which again, I agree, is the prize here. Question for you. Given your knowledge of Comcast. So they have free will, they have Beeswax. They talk about some sort of gen AI capability to help build the ads, which I think is great. Remove one more level of friction. Do they have the AI algorithm chops to really drive performance or is this a little bit of a leg in? Hey, Mr. Mrs. Dry Cleaner. 50 bucks, 100 bucks, we'll build an ad, let you run on TV, make it hyperlocal.
- [33:53]
Eric
Yeah, I don't know. The AI thing is a mystery to me. I don't know if they're. They're licensing it for somebody or they're using off the shelf. It's very unclear from the announcement. I will point out the Freewheel has, I don't know, I want to say a thousand engineers in Beijing. Freewheel is not an understaffed company. It's not being starved for resources by its big corporate partner like some ad tech companies. So they have the capability if they want to use it. You know, I think I got a lot of congratulations because everyone just assumed this was running on Beeswax, which is my baby. So I think that's you.
- [34:25]
Unknown
I'm 34 years later.
- [34:27]
Eric
Yeah, I know. I'm pretty sure it is running on Beeswax, so I'm feeling pretty good about that.
- [34:31]
Unknown
Congrats, buddy.
- [34:32]
Eric
I do want to raise a question which is it's a little unclear if this thing is actually live or not. Is this like a plan for a plan or a plan that's almost there or is it half there? I'm getting kind of mixed news on that, so I'd love to know when this thing is fully operational. Death Star, instead of just an announcement at CES So I'd keep my eye on that. One other thing I want to mention is they brought on sort of a product lead who we know, James Burrow from who we had on this pod. Maybe a year ago I experienced sort of west coast product guy who worked at Snap, who understands retail, who understands algorithms. Definitely the kind of hire you'd want to run something like this.
- [35:14]
Unknown
So they not only brought on James, brought on James and his team that he's, he's, you know, kind of assembled to do a few things together, starting with the Snap business. They built the, the Snap ad business, which, you know, ultimately was a wiz and is a billion dollar company. They then built Market, James, Dan and Sean and sell that to Verishop. We were an investor in Market, went quiet for a couple months and now they're the team behind the ads platform at Comcast, which is awesome and to your point, there's no better team for this thing. So it's going to be fun. We should keep an eye on it. And I have a meeting with someone later on at Comcast. Maybe I'll give you a little intel about the level of external availability on this one. So let's talk about Roku. So one is specific to this. I think this is so interesting and indicative of what Jay Askanasi was talking about last year on the pod, which is a shift from Roku into really aggressively pursuing more open strategies and really trying to be ecosystem embracing because they have lots of inventory. So I thought it was interesting that they're just pursuing lots of different channels for monetization, this being one of them.
- [36:25]
Eric
Yeah, they have sort of a half open approach. Like they're not one guard, they're not fully open, sort of metered, who gets access and what you're doing. I think it's pretty smart. It's sort of a little bit like what Netflix is doing. I think Netflix assumes most of its sales are going to be direct or using its tech, but you could also buy it from the trade desk. I like that approach, especially for challengers. I'd love to see Amazon move to that approach. Let people bid from other DSPs, although they don't seem to be heading that direction. I think it's really smart for CTV type activities where they're very different strata of customers from SMBs, the mid levels to the traditionals. Might want to use the trade desk or someone else to buy your inventory.
- [37:02]
Unknown
Exactly. Like, you know, especially when you're a Roku, you have a lot of inventory. You know, having various channels is a is a must. And they had their own announcement. The Roku clean room, I think it's called Roku Data Cloud.
- [37:15]
Eric
Yeah, that's interesting. I mean Roku, like a lot of device manufacturers has a lot of information about consumers. So presumably they want to open that up and let people get performance based on their files with the payment information, credit card info. But noctia bad info out. So publisher focus Clean room sure makes sense.
- [37:34]
Unknown
Yeah, I think it makes a lot of sense. It was super cool. Final on the CTV front, there's a big, big merger, right. Disney is combining their Hulu and Live TV offering with Fubo which will create.
- [37:50]
Eric
Like I guess the second largest after YouTube TV. Yeah, yeah, this is, this is Hulu's virtual MVPD, the VMPD. So it's the live TV on Hulu, because you have to pay retrans fees is not the most profitable part of their business. I think they have to pay the channels. They put on that quite a bit. And we know that YouTube TV has been raising its rates pretty aggressively, which implies that it's not maybe not the best business or the most sustainable business where they have to pay all these fees. So maybe Disney wants to A like move some of that to a different operating company and B, like fubu TV is also kind of in a too small to compete world. So combining them make me makes a strong second player.
- [38:35]
Unknown
Ever watch Fubo?
- [38:36]
Eric
I have never watched Fubo.
- [38:38]
Unknown
I've landed on it a few times. You know, just oftentimes when you're like looking for a movie from like the 80s to watch.
- [38:44]
Eric
Oh yeah, yeah, you get the carousel of.
- [38:46]
Unknown
All right, it's available here. Like Fubo comes in handy. They, you know, kind of come in clap sometimes.
- [38:51]
Eric
Yeah, all right, I'll buy it. I mean I have, I have, I have. What do you use? I have Hulu. I pay for Hulu, but I don't pay for their live. I use YouTube TV live.
- [39:00]
Unknown
We have everything that Hulu offers. So we have Hulu Live.
- [39:04]
Eric
There was an interesting article that I don't think we talked about on the POD a couple weeks ago from ad exchanger, where YouTube TV is really tied into the rest of the Google Ad tech ecosystem. And as a result, sometimes there are ads available through DV360 that you wouldn't expect, like live football. So you might have like a live football game on a premium channel like NBC. And the resold ads that YouTube has the rights to can end up available at DV360 for kind of lower CPMs than you might expect. So it's like a weird backdoor to getting that inventory. And that just kind of, I think also emphasizes the fact that YouTube is not fantastic partner for many of its broadcast companies because its sales channels are very opaque and aren't optimized to the same CPMs that they get direct.
- [39:55]
Unknown
You put in the notes here and I'm sure we'll link it out. There was an article in the Information, which I'm not a subscriber to, that was a deep dive on just how Google is I think behind, to put it politely, on streaming. Any insights on that one?
- [40:10]
Eric
Yeah, so Kathryn Perloff is a reporter who was at Adweek and she's really making a name for herself from really no knowledge at all about advertising to really deep coverage. So she moved the information and it's one of her first pieces I'm quoted in it. It's sort of a jumping off point on the fingerprinting announcement from a couple weeks ago. So Google loosens its fingerprinting restrictions and that was widely seen as a way to play catch up in streaming because in streaming there's less identity. There's no cookies, there's no sandbox. So they did that and then that kind of opened the question of like where are they in ttv? And the answer is they're not in great shape on the buy side. DV360, obviously it has exclusive access to YouTube but is not used very widely for other video for other ctv. They're certainly well behind on the trade desk and identity has a lot to do with that. And then on the sell side, you know, their video ad serving product is also ran to Freewheel and also to Springserve and a bunch of other guys. At this point it's mostly used for sort of web video. Not as much for long form. Their biggest customer was Disney and Disney has built their own and migrated off it, I think maybe not migrated entirely. So. So it's really not been a winning area for them. And Katherine explores that and some of the reasons why.
- [41:27]
Unknown
Do you think Google will ever get acquisitive again? Just given how much things have evolved and changed with CTV and retail media and kind of like agentic AI which is coming next? Or do you think that just given the to be determined future of the Google business with you know, the lawsuits, everything is like on hold?
- [41:48]
Eric
I think that there was a rumor last year about HubSpot and I think you would look for stuff like that that's not in their wheelhouse right now because their wheelhouse is under attack for legislation and for the judiciary. But things that are sort of acquisitive, that would help their businesses are clearly not part of a monopoly concern. I'd like to see them sort of emulate Microsoft's acquisition strategy where they buy LinkedIn and they buy GitHub and then they buy Minecraft. And these are all sort of weird, interesting businesses that are profitable, that work with the whole, but aren't doubling down on what they're already near monopoly on.
- [42:25]
Unknown
Here's an interesting one. Experian launched a third party data marketplace in 2025.
- [42:31]
Eric
You explain this one to me. I'm not capable.
- [42:35]
Unknown
Okay, well, Oracle Ads is gone, which means no, whatever the kind of legacy name for bluekai and some of those other assets, there's still a need for unique third party data. Seems like Experian just took a look at the market and said, hey, let's bring the third party data marketplace back. And they're partnering with interesting companies attain. They're partnering with Sarkana, they're partnering with a couple of others that are these kind of like next generation data companies that have interesting and in many cases like commerce based data to bring to bear.
- [43:12]
Eric
I mean, third party data marketplaces. My understanding is that they're not great businesses because you're a middleman, you're taking, you know, a slice off of somebody else as a standalone business.
- [43:24]
Unknown
That's correct. But under experience.
- [43:26]
Eric
Yeah. Under Experian they already have the identity, they already have a lot of other things. So I don't know.
- [43:32]
Unknown
I don't know. I like it. I like bringing that stuff back. Good to see it. And then one other thing I think is interesting. So a lot of people are saying that the year of 2025 is going to be the year of M and A. And people expect all sorts of craziness and the market's coming back. All that remains to be seen. I was quoted actually in a Marketing group piece and I confirmed that just with our portfolio companies, particularly some of the top ones and notable ones, they're all getting a lot of inbounds. I think one of the things that we're going to see, I know one of the things we're going to see, the unfreezing of the Series A and Series B markets. There were very few ad tech series B, series A's in 2024 and 2023. We've covered them on the pod here. Last week, prior to ces, Rembrandt portfolio company of ours raised a, raised a sizable series A $23 million. Speaking of third party data marketplaces. So the lead was Superset, which is Tom Chavez, former Crux and Rembrandt is founded by Omar Tawakal, former Bukai CEO and a couple interesting strategics including the trade desk participating. Rembrandt is an amazing company. They help with next generation ad formats like insane product placement and they're moving into ctv, you know, so, so A, you know, I think it's direction very interesting. But B, you know, my sense is like the series A, Series B is coming back, which I think is going to make a very interesting marketplace dynamic. And this is probably the first that we'll, we'll see of these types of deals.
- [45:09]
Eric
Yeah, that's really exciting. I'm happy to hear it. I know it feels like there's a lot of companies who got seeds a year, year and a half ago and are having a lot of traction. That's what should happen, right? That's what you expect, I think. Also on Rembrandt, specifically, this goes to my comment earlier about jockeying for share and innovation in ctv. The ability to enhance your creatives is now an area where people are spending more time Rembrandt. But also I think this week Curve, which is a competitor, Rembrandt, they did a deal with Magnite to bring sort of similar AI driven overlays into ctv. So it's interesting those two announcements happen the same week.
- [45:48]
Unknown
Absolutely. Shout out to Curve. Where Rembrandt is focused on is a little bit less like overlays and some of those interactive elements. What Rembrandt is doing. It's very fascinating. I think this is something that folks should pay attention to. They take video that is in post production. So formerly it was like creator videos, YouTubers, TikTokers, so on and so forth. They would then use AI models to find appropriate placements to put a product placement or an in scene placement. So they kind of prove that out. And the next phase and basically why they raised the series A is you could do post production in anything. You could do post production in movies, in sitcoms, in effectively anything. So their vision is CTV and this is where it aligns with the Curve needs new innovation around ad formats. So it's really focused on that. So I think they're going to push the envelope on what we might see in terms of kind of next gen ad format. Super cool company. Final note, this was last week, the Honey thing. I think it's fascinating just from the standpoint. So for those who are not familiar, Honey got called out by a YouTuber for effectively gaming last click attribution and doing a whole bunch of other stuff. Ari, your article on it was a must read. The thing I think is fascinating about this, and this isn't a CES thing per se, is how all of a sudden attribution is being, you know, kind of brought to the forefront and, you know, the entire world is learning about how kind of effed up, you know, some of the stuff that, you know, is endemic to ad tech is the.
- [47:25]
Eric
Affiliate world is like, you know, teenagers learning about sex. They're like, they think they're figuring it all out for the first time and no one's ever experienced this before. And the adults are like, oh man, this is bad. It's like, this is not the first time that there's some video, some YouTuber is like, you won't believe that. Well, Philia, it's a ripping you off. So I wrote this article about Honey that wasn't that controversial, I don't think. I just basically said, this is a scam and you should consider turning it off. And I got a lot of feedback. Some of my newsletters just go off like a dead fish. And other ones, like I get 10, 20 emails about them. And this was one of those where everyone had a point. And the main point people made to me was that there are a lot of these sort of coupon sites and coupon oriented affiliates and Honey's just the only one that is right in your face in the browser. So they get a lot of attention. But it's part of the affiliate world. And e commerce vendors are kind of used to it. They think it improves conversion maybe. And it's not new for the merchants. It's like part of the ecosystem and they kind of know about it, but maybe they do. Maybe they're doing smart things. I have a really hard time thinking this is incremental revenue for them, that it's probably not incremental, it's negative incremental. But I still haven't heard from anyone who's like, run a test or done anything that would prove that something like Honey is work.
- [48:49]
Unknown
I would assume. I mean, PayPal spent over a billion dollars on this thing. So I would, I would hope that, you know, this, this thing, this thing works.
- [48:56]
Eric
Was it 4 billion or something? I think it was 4. Yeah, you're right.
- [48:59]
Unknown
You're right, it was 4. It was 4 billion. I would love if someone, you know, took you so you ended your newsletter, you know, saying, prove out the incrementality to yourself to shut off honey, shut off everything. You know, wait a week. Wait a. Wait a month and report back. I would love. If somebody actually does that, reports back. We can. We can certainly talk about it. Yeah. All right, Well, I got a. I got a meeting, so I gotta. I gotta hop. All right.
- [49:21]
Eric
All right. Well, this was a great episode, and, you know, it's good to see you at ces, even as virtually from our hotel rooms.
- [49:29]
Unknown
Yeah, yeah, yeah. I hope to see you at some point later on the week.
- [49:33]
Eric
All right, thanks, everybody.
- [49:35]
Unknown
Take care, everybody. Bye. Bye. Thank you for subscribing to marketecture. New interviews are added every week at marketecture TV and your favorite podcasting app.
- [49:53]
Professor Sandra Matz
Thank you for listening to the Market podcast.
- [49:55]
Eric
New episodes come out every Friday and.
- [49:57]
Professor Sandra Matz
An insightful vendor interview is published each Monday.
- [50:00]
Eric
You can subscribe to our library of.
- [50:02]
Professor Sandra Matz
Hundreds of executive interviews at marketecture tv.
- [50:05]
Eric
You can also sign up for free for our weekly newsletter with my original.
- [50:08]
Professor Sandra Matz
Strategic insights on the week's news at News Market tv.
- [50:13]
Eric
And if you're feeling social, we operate.
- [50:15]
Professor Sandra Matz
A vibrant Slack community that you can apply to join@adtechgod.com.