top of page

The Attention Ledger: Matt Krepsik and Vas Bakopoulos on Benchmarking Creative in a Crowded Market

16 minutes ago
74 min read





What happens when AI makes it possible to create more advertising, faster and cheaper than ever—but attention gets harder to earn?


MediaRadar CEO Matt Krepsik and MMA Global’s Vas Bakopoulos return to Signal & Noise for a deep follow-up to our live conversation at Advertising Week. This time, the discussion goes beyond AI adoption and into the economics, measurement, creative strategy, and competitive dynamics reshaping advertising.


At the center of the episode is new research examining how brands spend and what they say across 10 industries, six channels, 20 quarters, and roughly two-thirds of U.S. advertising spend. The findings reveal a market where brands are producing more creative, packing more messages into ads, and increasingly trying to balance brand building with measurable performance.


Matt and Vas unpack why message complexity has risen, why some categories produce dramatically more creative than others, and why “more” does not automatically mean “better.” They explore the tension between attention and clarity: an ad can stop someone in their tracks and still fail to leave behind a meaningful message.


The conversation also takes on one of advertising’s oldest problems: measurement. Are marketers actually creating new demand, or simply taking credit for customers who were already going to buy? Matt explains the attribution trap through a simple club-promoter analogy, while Vas discusses MMA research showing why brand impact can continue paying dividends long after a campaign ends.

Then the discussion turns to AI.


AI can generate enormous volumes of creative, analyze competitive signals, and optimize delivery at a speed humans cannot match. But Matt draws an important distinction between intelligence and wisdom. AI can process information; marketers still need context, taste, judgment, and courage. As Vas argues, AI raises the floor—but human choice has to raise the ceiling.


The group also digs into:

  • Why retail produces roughly 9x the creative volume of pharma

  • Why automotive advertising has shifted from features to lifestyle and experience

  • The return of brand messaging after years of performance-marketing dominance

  • Why incrementality matters more than clean attribution

  • How creative quality can compensate for lower share of voice

  • The “first-second strategy” required to win attention in fast-moving feeds• Whether brands will eventually advertise to AI agents instead of people

  • Why authenticated audiences may become more valuable as the open web declines

  • The danger of AI-generated sameness, creative fatigue, and optimization toward the mean


The episode closes with a challenge for marketers: use intelligence to sharpen judgment, not replace it.

Or, as Matt puts it: “Intelligence should sharpen your judgment, but not replace your courage.”

Signal & Noise is hosted by Brett House and Rio Longacre.



Transcript:


Brett (00:01.3)

Hey everybody, welcome back to Signal and Noise. This is Brett House, joined by my co-host Rio Longacre. And today we're having a friend of the pod back for his third visit, Matt Krepsick, CEO of Media Radar. Round number three. Yeah, he's t he's t he's tied the record for most. And Vaz Bacopoulos, who leads industry research research at MMA Global. and for those on the


Rio Longacre (00:07.232)

Everyone?


Rio Longacre (00:13.356)

Round number three. That that's is that a record or has he tied the record? I think he's tied the record.


Krepsik (00:14.878)

Round number three.


Rio Longacre (00:27.19)

And Voz has done two, but that but very quickly he's done two. So we're honored Voz.


Brett (00:30.162)

Yep. Yep. And for those that haven't seen the advertising week live in front of a studio audience recording that we just did with these two wonderful gentlemen, it was phenomenal, a ton of fun, and well worth l you know, looking into. So for those that don't know Media Radar, I'll give a quick intro. It's a media and creative intelligence company, enabling advertisers and publishers to monitor advertising spend, analyze media trends, been benchmark industry category movement. and the MMA, which recently rebranded.


Vas (00:31.173)

Catching up.


Brett (01:00.428)

from from the marketing or from the mobile media association to the marketing and media alliance. It's a global and non-profit community of chief marketing officers and senior marketer leader marketing leaders advancing marketers' ability to create value. Greg Stewart, another friend of the pod, he's been on.


Rio Longacre (01:18.572)

He was on he that was a great episode with Greg. He's so funny.


Brett (01:21.362)

He's he is he's a funny guy, right? You know, and Matt, for those that don't know Matt, he was a a Zamboni driver. I think I think you get you begin your like executive career at at the hockey rink, right?


Rio Longacre (01:30.84)

This is true, right?


Krepsik (01:32.71)

It's true. So I think like any executive, you gotta start driving a Zamboni, you know, and figuring out how to like cut the ice. I mean it's it's a high high risk high risk, high exposure job. There's a lot of expectations from the figure skaters and the hockey players out there to have a really nice smooth ice surface.


Rio Longacre (01:42.158)

Cut your teeth, drive it to Zamboni.


Brett (01:52.67)

Yeah. And your dad ran a a a rink, right? Or a


Krepsik (01:57.293)

no, he ran a window company as well. Yeah, then I ended up running an ice rank, yeah. That was yeah.


Brett (01:59.264)

okay.


You ended up running a nice rank, right? And


Rio Longacre (02:03.941)

so it's your rink and you just when I guess when the Zamuni guy didn't show up, you just jump on there and


Krepsik (02:09.481)

So I started as a Zamboni driver and then eventually needed somebody to kind of run the rink. And that's also how I met my wife. She was a figure skating coach. And she was like the and when I took over running the ice rink, she became like the skating director. And we've since never worked together again.


Rio Longacre (02:12.173)

Okay.


Rio Longacre (02:28.198)

You you learn your lesson.


Brett (02:29.127)

You're a skate you're a skater yourself though, who's the better ice skater? The ice hockey players are like, we're always the better ice skaters.


Krepsik (02:31.749)

I I I'm not gonna comment on that one. I'd like to stay I'd like to keep


Vas (02:34.705)

Wise.


Rio Longacre (02:36.175)

Ha.


Krepsik (02:39.635)

Yeah. but yeah, she's still


Rio Longacre (02:40.791)

They can check harder, but I don't know if they can do this the spins or the pirouettes or whatever.


Krepsik (02:44.189)

Yeah, the spins and the jumps, it's a lot a lot of hard work, a lot of technical skills there. So so I had to go I had to go I had to go get another career and so had to start back at square one and got into my advertising in me.


Brett (02:44.434)

Yeah.


Brett (02:49.447)

Yeah, yeah.


Brett (02:57.703)

Yeah, and he was you spent you know fi fifteen, sixteen years at Nielsen. That's where you and I met originally. you were the CTO and then CEO of Quotient before joining Meteorator as this as the CEO. And and Vaz, you've also got a background that's spanned Digitas Publicis, right? Cantar. and you've been at at MMA for quite a while at this point, right?


Vas (03:13.712)

Yeah.


Vas (03:17.937)

That's true too. Yeah, working with Greg Stuart for a while. Yeah.


Brett (03:21.105)

Since what, twenty twenty one? Or or was it way before that?


Vas (03:23.853)

No, way before that. Before the pandemic. Yeah, yeah, yeah. Well I was I was there for mobile. Yeah. Yeah. Yeah.


Krepsik (03:25.055)

No.


Brett (03:28.923)

Okay. Yeah, so you've been there from the from the very beginning. Right. So so today, Yeah, totally.


Rio Longacre (03:29.303)

Okay.


Very cool. Well we we love the we love the reports you guys are putting out by the way. It's excellent. I mean that's part of the I know Brett you're about to introduce is some of the great like research and reports they've been doing and and just so like we're stunked by that. And Brett, sorry I cut you off.


Vas (03:42.353)

Yeah.


Brett (03:43.634)

Yep, and so a report that just actually landed today, which is an interactive report online. I didn't see if there was a downloadable version, but it's a it's a very cool presentation that MMA Global did with Media Radar. Interestingly enough, I spawned that when I was at Media Radar a couple of years back and then it finally came to fruition, but it's evolved since my original ideas. I'm it's true.


Rio Longacre (04:06.261)

Is he trying to take credit for your ideas, Krebsic?


Krepsik (04:09.717)

One hundred percent.


Brett (04:10.637)

It's true, the whole media radar, yeah, it's definitely taking credit for at least part of that. and then Adelaide was another was another partner. and for those that don't know Adelaide, they have the attention, AU attention metrics, right? for the attention economy. and the the report's called How We Spend and What We Say, The State of Advertising Spin. It measured how brands spend and what they say across ten industries, six channels, twenty quarters, and roughly sixty six percent of US ads spend. So you came to this conclusion of something that you say helps predict or at least point to


Rio Longacre (04:20.471)

Tension metrics, right? Yep.


Brett (04:40.261)

industry messaging DNA, which I thought was a really interesting positioning. So we'll talk you know certainly about the findings there and how they apply to kind of the larger media and advertising ecosystem. so with all that said, thanks again Matt and Vaz for joining us again. This is gonna be a fun conversation. And did I miss anything colorful about your backgrounds? There's all sorts of colorful stories that that you'd like to tell the audience.


Vas (05:03.161)

I don't have a Zamboni story, so yeah, nothing to say, yeah, like man, it's like how


Rio Longacre (05:04.941)

Or a hundred mile race story, which in the green room chats Krupsick was telling us about his hundred mile races. That's pretty impressive.


Krepsik (05:13.885)

Yeah.


Vas (05:15.195)

Yeah. I mean I grew up in Greece so no Zamboni stories there, so not even snow. Not otherwise. Yeah.


Rio Longacre (05:20.897)

Not a lot of ice skating or ice rinks or no, no I'm sure.


Brett (05:27.431)

Sorria, take it away.


Rio Longacre (05:27.669)

All right. Yeah, so good. So yeah, so we we were excited about this, not only to follow up on the discussion we had at at during ad week, and we talked more about AI and some of the research you've done in AI and how it's being used and adopted and organizations are actually putting it into you know embedding it in tools and programs, and now employees are actually using it. So we thought a good follow-on on to that would be some of the recent research you've been doing to Brett referenced about creative, creative performance and creative


Measurement creative analytics. on this podcast, for those of you who are recurring listeners, we've talked about this to a certain extent recently. We had Alex Colmran VidMob. We had another Alex, Alex Maruka from Craftsman Plus. we talked a lot about the role of creative, how it's maybe kind of having a moment. w despite the fact that there is more AI stop and AI generated creative than ever. I think the role of creative and the role of tastes are becoming, I think you could argue, more important these days. So we've talked about creative data, AI production, testing, and


Even the risk that AI introduces by making everything kind of look alike and the same and introduces that Brett you mentioned it, that sameness, right? So today we really wanted to connect that discussion to competitive context and the research evidence that you've found with your recent report to talk about like what are brands saying, what are they doing, and what are they missing that


that your research revealed that customers probably want to hear. So let's maybe start with the first question. this is over to you, Krepsik. what is one creative decision marketers keep getting wrong even though they have more data than ever at their fingertips?


Krepsik (07:01.845)

They're trying to cram too much into a fifteen second, thirty second spot. I mean, we saw this a lot of in our research as well, but you know, we I think Vaz coined the term really well, like message complexity. Ads are just getting more and more dense and more and more complex and they're asking it to carry like the the foundation of the house, the full house and the entire like neighborhood.


Brett (07:08.732)

Yeah.


Brett (07:12.978)

Yeah.


Krepsik (07:26.565)

all on the back of like kind of one creative. And I think that idea of message complexity and message density is one of the biggest challenges that I see marketers kind of face today. They want to kind of capture attention, they want to have clarity, but oftentimes it's how do we pack as much messaging into that one ad as we can.


Brett (07:46.856)

Yeah, yeah. And the report seem to suggest that that that's increasing and it has increased over time and it's measurable and it might be it might be having downstream impacts on engagement and recall and things like that.


Vas (07:52.593)

That's true.


Krepsik (07:52.968)

Yeah.


Krepsik (07:57.081)

We've definitely seen a a consistent growth year over year over year in terms of message complexity. And it's it's been it was interesting just to see how much that's grown across a lot of formats. Now the one thing I would say that was interesting it was like a lot of your, you know, your video, your print, your radio ads, like your rich form media where you can tell a story, you can talk longer, you have more of that sight sound and motion, like the the bits that make advertising like so sexy.


those messages started getting more and more dense. The flip side of that in that research is that mobile ads actually started getting less dense. So when we got to that smaller ad format.


Rio Longacre (08:37.271)

So dense as in the the amount of messages, the the volume of like let's say words, information that's being passed in the message, or is it just the different the number of themes packed in a message? How are you defining that exactly?


Krepsik (08:48.084)

Yeah, I know l I'll let Vaz as the expert here define how we're thinking about message complexity.


Vas (08:52.527)

Yeah, so so the way that we defined it is, you know, obviously, you know, Media Raider has a ton of data in terms of creating almost like a a kind of a a a framework for for codifying creative and creative assets, right?


And you can see that from a messaging point of view, you know, do you have a kind of a brand kind of focus? Do you have a performance focus? Do you have a direct response focus? Do you have a product focus? And all of those things actually have more layers. So you can see that in a number of different ways. The way that we unpacked it is that we try to see the combination of brand and performance and product and and direct response messaging.


And what we did see is to your point that, you know, it is increasing. it is also very telling that in different industries, you have a much higher complexity, right? And I think that obviously has to do with with a number of things, with with with the the set of media choices that they make and with a set of, you know, like you know, other other product


Rio Longacre (09:54.293)

Is some of it like compliance or legal requirements too, do you think?


Vas (09:58.422)

Absolutely, right? I mean, you know, you do have to kind of you know follow some specific guidelines in some cases and you know you have to kind of obey to some of those rules. But the point is that you know we are seeing this complexity kind of going up or year over year over the last five years. I think this is a big wake up call because in a world especially when creative production becomes


I say easier, but at least you know you have a lot more kind of ways to to develop creative. Now you have to also feel the creative and and launch the creative and deploy the creative. But it is interesting to see that that trend going up.


Brett (10:38.289)

Yeah, well it and and you know, to like I was when we came up with this original idea, the idea of sort of dimensionalizing specific types of creative, and I think you put it well just to repeat it for the audience, because Media Radar has created sort of a a taxonomy for creative messaging, right? With with a ton of metadata and and Krebsik, you can you can certainly talk about about the the what you guys tag. I mean know tagged creative is tagged for roughly thirty thirty third level


Krepsik (10:38.558)

Yeah.


Vas (10:53.189)

Yeah, that was the word I w I was looking for.


Krepsik (10:55.987)

Ha ha.


Brett (11:04.775)

30, excuse me, messaging elements applied per industry, right? So you can start to break down things as as Vaz just said, like value, offering, you know, you know, pricing, product, brand type of messaging, direct response type of messaging. But to me, that stuff really allows you to dimensionalize and bucket. And what you're saying with mess message density or message complexity, well say complexity, excuse me, is that


that you're actually you're actually seeing more of that basically that metadata appearing within these ads. So they might be doing one or two things. A brand message plus a a price message, all within one sort of message packet, so to speak. Is that is that a good way to think about it?


Krepsik (11:41.012)

Yeah, and then you might add in a and then like a value message and then a product message. Like one of the things like to re to Rio's point earlier, there's certain industries where you say, Hey, if I'm talking about the pharmaceutical category, I expect that to have just because of that industry. Probably it Yeah, either there there there's a little bit of a higher view there, but you would be surprised to know that the number one category in term or industry vertical in terms of message complexity, restaurants and QSRs.


Rio Longacre (11:54.861)

You have to have all the adverse events. You have to have every single one listed out. Sure, yeah. Yeah, I there's no there's no avoiding it.


Krepsik (12:10.373)

So it's our good friends over at McDonald's and Taco Bell who are just saying, Hey, let's talk about the brand, let's talk about the products, let's talk about the value messaging, let's talk about the price point, let's talk about refreshers, you know. I think like one thing we've seen is like the we're in the middle of the refresher wars, which I didn't even know that was a thing. But everybody's got some


Brett (12:10.397)

Yeah.


Brett (12:16.989)

Yeah.


Vas (12:26.906)

No.


Rio Longacre (12:28.534)

I that i okay, I guess that is that a thing? What is that exactly?


Krepsik (12:32.349)

You know, honestly, it's this is where like I would say where marketers get it wrong with message complexity. Like there's so many things and I'm like, what is a dragonberry sparkling fruit drink with Red Bull? I don't know. But it's just like there's just like s so it's a thing. And it's like there's so


Vas (12:46.425)

It's a thing.


Rio Longacre (12:48.436)

It it seems like it's a product of a focus group. That's what it sounds like to me. That it's just like how could like like it's trying to appeal to everyone yet probably appealing to no one. I I I don't know. Like I I I'm not surprised you say that 'cause I think a lot of the ads, like like go to the Super Bowl commercials, most of them are terrible. They're either fart jokes or they're like overly complex, overly thought out. Like it's I I mean, I watched the super being an ad guy, I watch the Super Bowl I like I like the sport, but I d I watch it. I s I typically watch it for the ads. I just got the last few years have been terrible. I don't know.


Vas (12:49.029)

Yeah.


Vas (13:03.483)

Yeah.


Rio Longacre (13:17.504)

Maybe just getting to be a cranky old guy.


Krepsik (13:17.895)

Yeah. I you know, I think I don't th I mean maybe we're all getting to be cranky old guys, but like, yeah, I would say as I think about this world, just kind of like to kind of take a step back, I think there's there's this battle between attention attention and messaging, right? And so in many ways I think what marketers and what we've done well as an industry is like we're often chasing attention.


And we've got novel creative ways to capture attention. But then we've Liberty Biberty, right? But then once we once when what Yeah, but once we capture intention, it's then what's the message that's what's the follow-on takeaway? And I think that's where we probably are all letting ourselves down is like, wow, we've used our our


Brett (13:50.204)

Yeah. Liberty liberty, right? That's a perfect novel wave. Liberty Bibberty, right? You don't even realize what value that's bringing to me from an insurance perspective, but


Krepsik (14:13.019)

our taste to figure out attention, to grab attention to capture the moment. But then once we've grabbed that attention, we're we're falling short on what were we trying to communic what did we want people to take away beyond liber beyond liberty liberty and


Vas (14:25.071)

Yeah. Yeah. Yeah. It and I think you're I think you're right that that in some cases this may be a result of a choice. and in some cases it's the result of following the category convention. Right. So if you go back to when humor became a thing in insurance for instance, since we were talking about that, that was probably a choice. It was a choice because it's a boring category.


Rio Longacre (14:26.432)

Well but funny go ahead, Vass.


Vas (14:54.517)

And you just bring in and just do something, you you throw in a lizard or something, right? When everybody's doing the same, and if you look at the report that we published, you'll see that every single category in this two by two in terms of the complexity and the volume that you produce and the different kind of media, they have a very, very specific formulaic approach to marketing. And the question is why?


Like i is it because we don't want to be fired next year? It's just okay, you know, like everybody's using humor, I'm gonna use humor. Or is it a specific choice that we're making? I think the report that we published gets to that message of like are we making choices based on are we making decisions based on specific insight and hypothesis? Or are we basically kind of like taking last year's thing and we just say, okay, what is next?


You know, was the next kind of joke. Yeah. Yeah.


Rio Longacre (15:49.727)

I think they're overthinking it. I I think it's being overthought, right? This is funny. So I saw a tweet. This is by Tim Vanderhoek, CO ofiant, right? Who Tim, by the way, we'd love to get you on the pod. We've interviewed some other people. he's very outspoken. I I love I live his social game. It's strong. So he posted, you know, the Sydney Sweeney ad. Okay. It's funny, Mike. I got home one day a couple weeks ago and my wife was like really unhappy at this happy about this ad. So I looked at it and I said, Okay, actually, I think the ad's awesome. Like it's, you know, it's it's appealing to a certain demographic, like dudes who play sport


To do sports betting, right? It's basically her walking around her underwear. And apparently Tim is saying this ad outperformed all other ads by 25% when it went out. And he asks this question how many great ads die because someone might be offended? I just think people overthink it, they overfocus group it. They're they're too afraid of being controversial. They're too afraid of fr offending someone or being canceled. And they like, how many things can we pack in this ad to like the goji berry P fang group? And we want to, you know, the, but we can't do this because it's a


Vas (16:37.36)

Yeah.


Vas (16:44.195)

Real, that is a real problem. But it's a different problem. I think that is a real problem. I remember from my planning days, we would just pitch and pitch and pitch and then we just here's a great idea, and then from that idea that we win the pitch until the idea that will be launched, you would just the idea would be not recognizable, right? You would just basically water it down and water it down. Corrupt it by but so that's a different co completely different problem, right? But s if you go back into the Super Bowl


Rio Longacre (17:03.131)

got it. So the idea gets corrupted by this process. Okay, interesting.


Vas (17:12.465)

And if you look at some of these ads, it's not that they were not some of them were edgy, but they were not like so it's like but like what is the strategy? What it what like what are we trying to do? Are we just trying to kind of provoke yeah


Rio Longacre (17:21.824)

They're dumb, edgy. Like I I don't want to hear fart jokes. It's just r it's ridiculous. I don't know it suggests to me every other ad last year was like a fart or poop joke. I was like, Come on guys, this is like this like who you're a who are you advertising to? I don't know.


Vas (17:31.259)

Yeah. Yeah, we can do better.


Brett (17:35.228)

Yeah. So so so back to the report, there were there was you had one there was one other category which was sort of creative volume, right? So you looked at creative volume, you looked at message focus, which we talked about, which was sort of value product direct response, right? And then message complexity, which was when you have a bunch of those things happening in one specific ad. So did you see differences in terms of total creative volume by channel and by industry category and how did those things vary?


Vas (17:35.396)

Yeah.


Krepsik (18:00.378)

Yeah, yeah, we're so real quick, creative volume, just to define that for everybody, is the number of unique ads, unique variations. So how many different ways can you explain a car? and believe it not, there's a lot of different ways we can build a car ad.


Rio Longacre (18:17.974)

So so is variants of the same advertis in the same advertising campaign you might have different variants, or is it like like a different adge over a campaign?


Krepsik (18:23.736)

Yeah, so you could you can think about correct, like unique ads. So like even though like take pharmaceuticals, highly complex ads in terms of there's a lot of work we're asking for a pharmaceutical ad to do, but the the number the volume in the pharmaceutical category is actually pretty small. Like we're pretty consistent. It's typically everybody on a beach riding bikes and it's multicultural family.


Rio Longacre (18:43.73)

It's a multicultural family on a beach, right? Like smiling, pla f either doing a barbecue or playing volleyball or something. Okay.


Vas (18:45.209)

Smiling in the end, yeah, yeah. Yeah.


Krepsik (18:51.002)

Ye exact exactly. So it's the same creative, we just like new logo, different disclaimer.


Brett (18:57.17)

Yeah, yeah.


Vas (18:57.573)

Get your life back. Yeah.


Krepsik (18:59.024)

Yeah.


Brett (19:00.147)

Well and apparently those disclaimers don't have any negative impact on the perception or engagement or resonance of those ads with the t the target audiences that are trying to hit. It could be a target audience of ten million people that have a specific condition. And apparently through focus groups they say that those because for me they seem to be a very negative aspect of the ad where I'm like, My God, like 75% of this ad is conditions that seem worse than what the problems that's being solved with the medicine. Right? You know, I'm like, are you kidding me?


Vas (19:04.601)

Not anymore, of course, yeah.


Vas (19:10.629)

Yeah.


Rio Longacre (19:25.58)

No, but those diseases are pretty bad though. No, it's a thing yeah.


Brett (19:28.935)

But apparently that just completely is is is it's really not even heard by the focus groups, the people that are actually being targeted with the specific ads, which is interesting. But sorry.


Vas (19:35.875)

It it's not. Yeah, you're right. And and I did pharma advertising. Yeah.


Rio Longacre (19:37.932)

Yeah, if you like psoriasis, really bad psoriasis, it's it's debilitating, right? you'd love to be have it cured. Yeah. So speaking of pharma though, so in your report you had a nine X difference in creative volume between retail and pharma. I thought that was a really interesting call out. Like I've asked maybe you could t talk about that a little bit or or Matt.


Vas (19:43.341)

Exactly. Exactly.


Vas (19:55.442)

Right. I mean yeah, I mean it's a huge difference and you know, obviously for pharma, I mean we we we talked before earlier about the density, right? And you know, I mean it's it's easy to kind of think about like the constraints of pharma and the fact that you have to do TV and you know it's expensive and you have to pack a lot of things and all that and you know like and and then you end up with with with fewer creative pieces, because every single one of them as well


It's it's hard to produce, it's hard to kind of you know approve and all those different things, right? And on the other side of things, you have retail, right? Which obviously, you know, it's not a fair comparison. I mean, you have you can you can do stories about you know different categories, you can do stories about you know different kind of like your delivery, the e-commerce, this, that, right? So you do have a a a dramatically different model, but it's in a 9x difference, right? And and the question that


Rio Longacre (20:44.3)

That was remarkable.


Vas (20:46.109)

It's remarkable, right? So so the so the question that that that it provokes a number of different questions, right? So like what is the creative process that you have in every single one of these kind of environments, like you know, when you talk about retail and auto that have like they're on the high end, right? And then you have like probably pharma and so the low end, maybe finance as well is on the low end. And in a world now where we're looking at


you know, AI and and how that is gonna influence the whole process. And you can actually have creative intelligence in in terms of like, okay, what are we producing? What is the brief? How is it gonna kind of get us into producing better, more and better creative and and all those different things? How is that gonna change the creative process on this on those different sectors? I think is an important question to ask. You know, because you know you start with some constraints, but you know, other than pharma having those constraints and all that, I think some of those other kind of kind


categories don't necessarily need to follow those and I think you you can actually kind of rethink the creative process in in that way.


Brett (21:53.383)

Yep. So so what would you think what like and so Matt for you, like what would be a fair creative benchmark to compare like if you're if you're looking at a brand that's let's say a challenger brand, they're competing in a in a pr pretty busy, mature space with a bunch of incumbents. you know, I like how how should they be thinking about their advertising to stand out from kind of the category sameness?


Krepsik (22:16.134)

Yeah, I mean I I mean here's what I I'll answer that question in in a in a philosophical framing and then I'll kind of answer it more directly. I think and I've been thinking about this kind of lately a lot, even actually since our last kind of like live show when we talked a lot about AI in the creative space. And the way I've thought about it in our industry, especially in the world of creative, is that we we talk about the difference between like AI and I think about that as intelligence, right? I think AI in many ways


helps get access to lots and lots of information, right? It's it's a brute for it's I shouldn't a brute force. It's a qu it's a workhorse. It will do a lot of w heavy analysis for us. But I think about the difference between intelligence and wisdom. Right. And I think what AI is great at is intelligence. I think what we taught what I talk a lot about and I think about like taste and craft, that's wisdom. And wisdom in not in my mind is about having taking that intelligence


And and providing context for the situation, right? And I think that context for the situation is so important for a brand. So let's go back to your question of like a challenger brand. What I would say, and I think what we see from our research and some of the other research that MMA has done is like when you walk into a category or a vertical that has this certain I think Vaz says it very eloquently, like this fingerprint or pattern or conventional wisdom. What we find is like brands that break free of that.


Vas (23:22.449)

Yeah.


Vas (23:26.361)

Mm.


Brett (23:37.872)

Yeah. Yeah.


Krepsik (23:42.578)

Conventional wisdom, that challenge the norms, that do things differently, that understand both the context and of the category or that industry vertical and the situation that they're in, they're able to break three and do and do something different. Right. And I think that's what's so important for markers in the creative space is to realize that AI helps us with intelligence, but our job is to bring the wisdom.


And having that context and understanding that situation and say, how do we break free from that and challenge the the conventional norm? or what I think Vaz says really well, and maybe he he's gonna have some time on Broadway, they they defy gravity in in many, in many ways. And I think it's that defying gravity moment that's so important for those those challenger brands to like stand out.


In a category that may have either a lot of complexity or a lot of messaging volume, you have to learn how to defy gravity.


Rio Longacre (24:43.978)

I I so I would line in this report I really like that I think is related to that and Voss, I love your opinion on this is alignment or divergence is a valid choice, default is not. I kind of interpreted that as like that sameness that like that you're gonna get with with AI and LM's gonna spit out that kind of median. Like avoid that, make decisions. It was that was kind of my interpretation. Voss, love to hear like kind of like what you what that line means and how you're interpreting that.


Brett (24:44.391)

Yeah.


Brett (25:17.518)

we lost your audio.


Vas (25:22.83)

maybe.


Rio Longacre (25:24.768)

There you go.


Brett (25:24.966)

Yep, there you go. Yep.


Vas (25:26.081)

Yeah.


Rio Longacre (25:29.419)

Very good.


Vas (25:33.763)

So so to Matt's point earlier, you know, I mean there are some creative choices that we make that allows the f what what we say to kind of divide category convention.


Rio Longacre (25:34.86)

You're good.


Vas (25:52.013)

and the work that we did with with with the data that we have from Media Raider, actually, we went back and and we we saw some brands that defied gravity. And we did see that the brands that actually defied gravity meaning that they were relevant 20 years ago, and they're still relevant. these were the brands that actually made choices that they were different. They kept their messagings kind of very, very focused, but they also made choices that went against.


What the category was saying. Now you can say that you know there's a cause and effect, you don't know is correlation is causation. Are they big because they could defy gravity or they they defy gravity so they're big? I don't know that. Right? But the point though is that you have to think about two different types of creative, right? And maybe there's different rules. So let me kind of put a little thesis for you.


Rio Longacre (26:30.301)

Okay, that's good.


Vas (26:41.567)

You have your hard working, let's say performance creative, right? And I think you have to you you don't want to the same, but you wanna learn from your mistakes before, right? This is not about casting lighting in a bottle. You have to see what works. So you really need to build your creative intelligence here. And I think this is what Alex


Colbert will probably tell you about you know all the creative intelligence. What can you learn from your previous data of how your creative has worked? Now that takes you a long way. It's not gonna necessarily allow you to catch lighting in the bottle, right? Which is what you need to do on potentially your brand side of things, right? When you need to engage and when you need to stand for something, you need to kind of go beyond what you've done necessarily before. You need to build a story, you need to kind


Brett (27:11.473)

Yep.


Vas (27:32.253)

tell the story you need to kind of you know think outside the box right now how do you do that it's a process that it's not it's not it's it's a it's a part of like art and part science right and we know that you know the creative process to kind of get to cut sliding in a bottle it's not really kind of scientific right you have very very smart people and very very creative people that can do that. The question though is how can now intelligent get you to that can it


Right. When you get, for instance, data in terms of what your brand stands for and how your previous creative has worked, and then what your competitors are doing. These are all data that you need in order to at least kind of get to a thesis of like what what's next? I mean you need to have that as a starting point, but then you need to raise the ceiling yourself. So


Brett (28:22.64)

Yeah. But would you would you look outside the category? Would you encourage sort of a marketing team to say, Hey, looking at your competitive set is gonna is gonna kinda hold you within that sort of that limit, right, of of the category sort of habits and and legacy categ category thinking, arguably. Right. So would you encourage marketing teams to kind of


Vas (28:27.723)

Hundred thousand percent. Yes.


Brett (28:40.602)

Not only take analytics and data and insights from their own cohort advertising and understand what's working, what's not, so they can have a kind of a real feet on the on the on the ground understanding, but actually going to other categories, right? And to say what what's happening in other places, what's happening in the counterculture, what's happening in in in outside of just the advertising world? I mean, is that how these folks from a creative perspective really have to think about breaking the mold?


Vas (28:48.875)

Yeah. What is mo yeah, there's nothing more refreshing. Yeah. Yes.


Vas (29:06.592)

Yeah, I mean if if you ask me a hundred a hundred thousand percent. But I think I think there is there's things that you can do to bring culture in that are more strategic and there's things that are more


kind of creative. So let me kind of give you an example. If you have created your taxonomy of like, you know, these are the need states of my consumers, these are the set of benefits that I have, these are like the proof points that I have, these are the reasons to believe. These are like so you have like your your space, right? Then you just look at culture and then you just see like


Like what are my intersection points between this and culture, right? And I think this is the kind of the the place where at least you can start kind of finding ownable spaces where y the next creative idea will come from. So it's it's not just about basically saying, Well, let's think of a good idea, right? Intelligence will come into that and feed the next idea, right? And it and it and and and and your alpha is your data and your alpha is also your taste and and and the combination


Brett (29:57.992)

Yeah.


Brett (30:02.289)

Yep.


Vas (30:10.117)

of your of your taste and your data and your ability to measure and optimize and see incrementality is what is gonna give you a competitive advantage, I think.


Brett (30:19.505)

Yeah.


Rio Longacre (30:19.541)

So looking at categories that you you referenced a second ago, Brett, I mean I thought this was interesting in the report where there is a tr enormous difference, as we called out there's a difference in pharma and and and and restaurants, right? So I think look but looking even at QSR compared to automotive, right? I think there was a big difference there. I think QSR was at one point nine eight, message complexity and auto is one point zero seven. So Brett, you mentioned looking across categories. Is it w what do you think like


Vas (30:33.121)

Right.


Rio Longacre (30:48.479)

Krebs for you, like what do you think makes complexity work in one industry, not working in another industry? And does that make it more difficult for let's say a brand for one industry to look to another industry for insights? Or maybe should they be?


Krepsik (31:00.516)

You know, I think there's there's there's pros and cons of looking to other industries for insights. I mean, I think one example I I like to use a lot on this comparison is like what do like DraftKings, Chase Bank, Taco Bell, and Fabletics Hall have in common? Any guess? Kevin Hart.


Vas (31:15.745)

Hmm.


Brett (31:20.056)

d Yeah. It's either that or Snoop Dog, right? It's it's


Vas (31:20.823)

Ha ha ha.


Krepsik (31:22.862)

And so it


Well, I I actually think Shaq right now is still like hot ver


Rio Longacre (31:28.479)

Or Samuel L. Jackson, like he gets around his in all the ads too, right?


Brett (31:30.278)

Ha ha.


Krepsik (31:31.471)

I think Shaquille O'Neal and Joe Montana are still like ranked like one and two on the leaderboard of like most celebrity kind of representation and ads over the past kind of 25 years. But but so there's there's certain no, what I mean by is like to Rio's point, there's this cultural relevancy that we often brands will kind of cling on to to cut, or I shouldn't say cling on to, but find to kind of drive attention. And so there's there's some things you can learn about that from an attention standpoint to say, how do we kind of understand that?


Vas (31:31.957)

yeah.


Krepsik (32:01.164)

I also think from a messaging complexity perspective, I think automotive has done a really good job at simplifying complexity over the past few years that I do think like QSR and pharmaceutical brands can learn from. Because like for


Vas (32:09.737)

Hmm.


Brett (32:14.812)

Could probably learn from. Yeah.


Rio Longacre (32:18.187)

Except for that stupid farmers ad. I hated that. It was so


Krepsik (32:21.552)

Okay, that was a pretty that was that was a tough ad. But I agree with you there. But like they've gotten very good at making sure like the like it used to be in a in a car ad, you know, 10, 15 years ago. It's like we've got power seats, power steering, and they were talking about all the features and fun you guys remember these ads? They're like there was there was so many features and functionality of a car, and I'm like, I don't even know how a car works, what this thing is.


Brett (32:35.132)

Yeah.


Brett (32:38.475)

totally.


Rio Longacre (32:38.971)

yeah.


Brett (32:44.1)

Yeah, and now it's like this this Jeep Wrangler's gonna take you up the mountain. It's like the outdoor lifestyle choice, period.


Rio Longacre (32:44.501)

Yeah, no one cares.


Vas (32:45.121)

Yeah.


Krepsik (32:48.44)

And now it's like it's like correct. There it it is a lifestyle. It's very experiential. And you know where they took that from is they took that from fashion. They took that from luxury brands and said, what was that about? That was about experience. It's like I wanna be the person who does that or does this. And you see a lot of what's great about a lot of car ads now is they've been able to say, Hey, this car is this lifestyle or this car fits this need. And it becomes almost more experiential.


Rio Longacre (32:50.603)

It's a lifestyle. Yeah, yeah, that's g I like yeah.


Vas (32:58.519)

Mm.


Brett (33:04.049)

Yeah.


Krepsik (33:18.008)

And what you find is like those ads become more memorable. Because I don't have to figure out what like power steering really does, or do I need to make a choice on like power anti-lock brakes and


Vas (33:22.744)

Yeah.


Brett (33:29.168)

Yeah. It's it's it's this it's yeah, it's the Steve Jobs, you know, genius, right? Virgin is an airline and is an is an entertainment company, Red Bull, but Steve Jobs, I mean classic example that you know, probably have said ad nauseum in my in my life, but it's like yeah, he could have named the the the megabyte or gigabyte storage, the number of songs that you could put into his new iPod device, but he basically said, your record collection in your pocket


Rio Longacre (33:29.437)

Anti lock breaks or like all this, yeah.


Brett (33:55.194)

Right, imagine that for all of us that had like crates and crates and crates of LPs that we were moving around. Ten thousand songs in your pocket. So it was much less with they weren't the first people to come to market with an MP three. Right. MP3s were widely distributed across multiple companies like Sony, but they were the first ones to sort of really tie it into a broader lifestyle.


Vas (33:57.119)

Yeah. Right. Right.


Rio Longacre (33:59.06)

And you made it look nice, so you'd be proud to show it aro show it off too, yeah.


Vas (34:05.251)

Yeah.


Brett (34:16.56)

brand messaging play. And on the topic, Krepsik and N Vaz for both you guys, on the topic one thing that came out of this report of that was interesting of brand messaging is that it looks like there's that brand messaging, kind of upper funnel messaging, is kind of on the rise, right? And I'd love to hear your perspectives on that because you know we've seen obviously in the digital media ecosystems that we've all been involved in for decades, like almost this drive towards everything's about you know immediate impact, short termism, direct response.


Vas (34:32.077)

Yeah.


Rio Longacre (34:32.457)

I noticed that.


Rio Longacre (34:46.025)

Performance TV.


Brett (34:47.246)

Performance, performance, performance. it's a really positive thing to see that in actual research. So can you guys talk about that, the rise of brand messaging?


Vas (34:47.851)

Yeah.


Krepsik (34:54.955)

I actually would say I'm gonna give our industry some credit. I actually think that the rise on brand messaging is because we're learning to understand how to best utilize formats better. And so what we're finding is that on the major kind of brand formats where you can tell a story and tell a brand story, we see brand messaging on the rise. When you think about direct response marketing, we're actually using like direct response messaging, right? And we're just we're actually learning.


Vas (35:05.515)

Mm.


Brett (35:14.983)

Yeah.


Brett (35:20.338)

Yeah.


Krepsik (35:23.811)

how to use the format to do the work we want that format to do. And I think


Brett (35:28.08)

Yeah, in in the right type of channel and the right type of device and also with the right type of sort of cohort, meaning that person's at that hopefully pz that that path or that right point in time where that type of message is gonna actually help them engage with the brand in some way, right?


Vas (35:42.295)

Yeah.


Krepsik (35:42.883)

Y yeah.


Vas (35:43.255)

But there is another there's another point though, I think that those two things happen at the same time. You do see the rise in brand, which I think to Matt's point, it's yeah, I mean we see that trend and it's it is hopeful in that sense. And the MMA has done a ton of work to see how we can quantify the value of brand, right? Because brand has suffered traditionally because it's not on an equal footing with performance, right? You know, it's like, I got five percent lift in favorability. Well, can I take this to the bank? No, he can't. I mean he can't.


Brett (36:06.983)

Yeah.


Vas (36:13.269)

Now, because we have found a way to do that. So, but we're seeing this trend, but we're also seeing back in 25 minutes earlier the rise in complexity. So, what I think is happening is that there is a tendency to say, okay, well, let's do more brand, but at the same time, well, let's kind of add a little bit of direct response to the end because, well, I mean we have to do that, right? So it's not we're not completely out of the woods yet in that, right?


Rio Longacre (36:32.629)

Ha ha.


Put a QR code in the ends, yeah.


Brett (36:35.706)

Let's make it let's make it le let's make it less memorable by packing it full of stuff. Right?


Krepsik (36:36.483)

Verda. Well well I'm just imagining sitting in this pitch where it's like all right, we're gonna do this amazing brand ad and tell our story. Well, can we put some like direct response messaging on there? And like, well we gotta talk about this product we're launching too. And then like all of a sudden like what be


Vas (36:38.819)

Yeah. Yeah.


Vas (36:50.495)

Yeah. Just a little bit. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.


Brett (36:52.656)

Yeah. Bye now, get your free super chamois, right?


Rio Longacre (36:55.231)

Then the CO the CEO walks in and says, I want to do this because my my brother said this. You know, it's but but for those who d haven't read the report, i and this is I thought this is Brett, I I I thought that was one of the most interesting takeaways. So brand has gained six point one points and direct response has declined six point four points, which I thought was really interesting. And I I I guess I wasn't super surprised because I think we've overindexed on performance the last bunch of years and you know everything has gone performing, even C T V things that maybe aren't supposed to be really performance.


Krepsik (37:00.472)

Yeah, yeah.


Rio Longacre (37:24.841)

Or don't don't lend themselves well in terms of channels or formats for performance. They've s shoved QR codes into stuff that maybe they shouldn't have. So I it was very nice to see that, by the way. I I like that in a report a lot. I think it says a lot. And I think that's that's an interesting conclusion that I I don't disagree with Krebsik about like why people are doing that.


Brett (37:43.952)

Yeah. So brand versus performance, and I know Krebsic has spent a ton of his career in in the MMM space, the attribution space. Vaz, you've done a ton of work especially certainly at the MMA, with understanding MTA and and I call it sort of the attribution fallacy. And I think there's a lot of performance marketing, especially in the digital space, that misattributes actual con con conversion. Right? They're giving credit to down funnel, last click, all that sort of stuff. Where you look at it you know, you look at the numbers and oftentimes it's direct mail that really does the performing.


Vas (37:52.707)

True. Yeah.


Brett (38:13.561)

Right, you know the oldest of all channels, right, is actually there's there's a there's a kind of a measurable impact. How do you actually help brands sort of balance that brand versus performance conundrum and and realize the value of that that brand marketing's impact downstream on consumer behavior?


Krepsik (38:30.839)

Well real real quick on the direct mail front, I think there's there's a really good analogy and a really good story of I can't remember where this came from, but I just remember this from in my days in the attribution space. There was this these this this club promoter hired a bunch of college kids to go out and promote his club and they gave him a bunch of leaflets to say, Hey, come to the c come to the club and get a free drink or get in for two dollars off or something. And so and the and he paid


Vas (38:50.774)

Yeah, I know that, yeah.


Krepsik (38:59.029)

each kid based upon the number of redemptions that they had. And so a couple of these college students went out and like tried to get their cause the club promoters trying to get people to come to the club. Well this one kid basically like looked at a line of people trying to get in and passed them out to everybody in the


Brett (39:02.458)

Yeah.


Krepsik (39:14.199)

And so in many ways, like and he got tons of credit for it, but guess what? They are they none of it was incremental because they already were in line. So yes, you're correct. Just because I can I have a very clean view of measuring it and and guess what and like they're but they're already in line doesn't mean it really was incremental. And I think that's the fallacy of direct response marketing is that you already have like intenders who are giving you the signals that I'm going to buy.


Rio Longacre (39:14.261)

Smart.


Rio Longacre (39:18.335)

None of it was incremental, yeah.


Vas (39:22.635)

Yeah.


Brett (39:22.672)

Yes.


Vas (39:28.056)

Yeah.


Vas (39:31.532)

Yeah.


Brett (39:33.679)

Yeah.


Vas (39:34.007)

Right.


Krepsik (39:44.323)

And I do agree that marketing can help encourage to make sure like you convert the trip, you convert the transaction, and it does play an important role there. But when you already have a bunch of high intenders, how did you never stop to ask the question, like, how did we create these intenders that did not exist before? And that's the role that upper middle and upper frontal marketing plays. It's it's like you got


Brett (39:56.198)

Yeah.


Brett (40:02.16)

Yeah.


Vas (40:02.424)

Yeah.


Rio Longacre (40:05.173)

That w what you described a second ago was the retail media method.


Krepsik (40:09.174)

that is that is retail media one on one. It's like guess what?


Brett (40:11.885)

Yeah, yeah, don't get Krebsic started on retail media. We we should go down that path a little bit, yeah.


Rio Longacre (40:13.055)

Get it right when you're checking out, right?


Krepsik (40:16.492)

Yeah, I mean it's retail media. I mean, I I think there's a lot of value in retail media in this sense of like you know these are buyers, you can convert them, you and you can measure it. And there's a lovely story there where like you feel good of like, okay, I can trace the dollars, but nobody ever stops to ask the question, like, how did these people show up in the store in the first place? Or how did they land on like how did they land on the webpage?


Vas (40:16.739)

Yeah.


Brett (40:33.391)

Yeah, it's like it's like if you if you Yeah, it's like if you show up at like a if you s if you show up at like let's say a sporting goods store and you buy skis, there's a good likelihood that you're gonna come back to buy poles, to buy gloves, to buy ski pants, right? Th without any advertising. You don't need any advertising to push you to that.


Vas (40:38.135)

Yeah.


Krepsik (40:48.91)

Correct. So well, if you show up at a sporting good store, guess what? You're probably buying sporting goods. I don't think you're browsing. You know what I mean? you know, that's a high intent trip. You're high propensity or the moment you walked in the door, you're here for something. now we might be able to influence you. I can trade you up to a higher package. I can bundle different things in that basket. I think those are all great retail media execution elements. But let's not kid ourselves. You're already in the store, right?


Vas (40:49.858)

Yeah.


Rio Longacre (41:00.085)

You're a high propensity buyer, there's no doubt.


Brett (41:05.029)

Yeah, yeah.


Brett (41:18.255)

Yeah. Yeah. And you and you and you might be giving discounts to people that don't need discounts, right? It's the classic like mobile remember the lo the mobile location marketing claims, right? They're like, we got this person to come in to buy the twenty-five you know, the twenty-five percent off, you know, an extra large you know, Starbucks iced coffee. They're a Starbucks brand loyalist to begin with. They were going there at a specific time of day, a day part that suggested they're gonna buy from there anyway. Why are you just throwing money away?


Krepsik (41:19.19)

And so and so


Vas (41:23.33)

A hundred thousand percent. Yeah, yeah, yeah, yeah.


Brett (41:46.501)

To get them and pi and people will gain that system and then they take credit for the actual attribution, which


Vas (41:46.829)

Right.


Rio Longacre (41:50.124)

It's it's it's funny, Brett. actually did a project with a large like cosmetics company and I was working with our loyalty program, which had but had won awards. It's interesting. And I remember they could not tell me if their their concessions they were offering actually had actually paid for themselves and had incremental customer lifetime value over like whether they were just giving them away. It was I was an award-winning program, people loved it, but it was totally detached from there actually a CRM program. They couldn't tell you what CLV was, they couldn't under like


Vas (41:51.138)

Yeah.


Vas (42:12.364)

Yeah.


Vas (42:16.883)

Right. Right.


Rio Longacre (42:19.051)

So they were giving concessions that maybe they didn't need to give. So I mean I think it probably is working, but they couldn't tell me.


Vas (42:22.711)

Real I y you're right. At the at the end of the day, I think it's it's it's an incremental it's a measurement problem. And it has to do with being able to assess and validate incrementality, sort of obviously. But I wanna go back to what you said earlier, and I think it has to do with also the short horizon that you measure. So if your measurement is campaign plus look back window, right, then that means that


Krepsik (42:22.99)

Yeah.


Vas (42:51.963)

If something happens later, you're not gonna you're not gonna kind of measure the impact of that. Now, we did four studies that we called Brandus Performance at the MMA, and that was one of the main ideas to just understand what happens if you have individual level data. You connect brand brand KPIs and you know surveys and sales and ad exposure at the same level, and then you track not just at the end of the campaign but longer. Now


What this did is that it allows us to see that first of all, you know, we can measure the multiplier, and there is a multiplier. You know, things don't stop happening, you know, when the the campaign ended. And depending on the category, you know, this can be, you know, a two X or it can be a five X, as we saw. But also the brand favorability that you build during the campaign.


pays dividends because what we found is that a brand favorable if the person that you converted and you made that person like you that creates an annuity of conversions and that annuity of conversions for for a bank it could be like we found that it could be like a six X. Now that that is not gonna happen next Monday, right? Because you may not need an account next Monday. You may not need a loan next Monday but it might happen three months later, right?


Brett (44:04.334)

Yeah. Yeah. And that look back window probably is different by category, by industry category, depending on how frequently


Vas (44:10.195)

And it's gonna be look yeah, of course. Yeah, yeah. Yeah.


Krepsik (44:11.813)

yeah. I mean, I I'll I'll give you a really great example on this one. we I used to do a lot of work in the automotive industry, right? You know how often how often you buy a car in the US? Like roughly once every f what well, I think in the US we we're a little bit we've been incentivized to buy them faster with financing terms. but about once every once every three to four years. So let's let's keep the math easy so that way we're not doing public math right now on a podcast. So let's say once every four years, okay? So that means


Rio Longacre (44:22.068)

Seven years, something like that.


Vas (44:28.323)

Hmm.


Brett (44:28.546)

Once every three years?


Krepsik (44:39.179)

Roughly no more than 25% of the US population is buying a car that this year. This year. Do you know the average purchase cycle for a car? It's about a four-week window. Right? So that means if less than 25% is buying a car this year, that means roughly only 2% of the population is actively looking to purchase a car right now, while we speak. 2%.


Rio Longacre (44:43.882)

Is in the market.


Brett (44:46.734)

Yeah.


Brett (44:53.317)

Yeah.


Vas (45:03.607)

Right. Right now. Yeah. Right.


Krepsik (45:09.153)

But so the challenge that automotive has and why they think they've gotten better at branding is like the important thing is like to make sure like you're in the consideration set when you enter the purchase cycle. And then once you're in the purchase cycle, then it's like a different like mechanism to say, how do we choose to convert on our car package, value proposition, and dealership? But like you have this like challenge when you think about that view. It's


Brett (45:20.613)

Yeah.


Krepsik (45:33.354)

the the the long kind of like lag effect there can be wild in terms of like when you create that initial brand perception to when you actually enter the purchase cycle to say, I'm around ready to buy a car. These are the cars that are on my shopping list. And so getting all so oftentimes brand marketing is important to make sure you're on the list. So when I do go to to buy a car or when I go to change my bank account, which I do once every


Brett (45:48.27)

Yeah.


Krepsik (46:00.619)

Millennia? I don't even know the last time I changed my bank account. But like you it's been so long.


Brett (46:01.486)

Ha ha ha ha.


Rio Longacre (46:03.678)

Yeah. I don't remember, it's been so long, yeah.


Brett (46:04.688)

Yeah, yeah.


Vas (46:05.599)

Yeah. Yeah, right. Right.


Brett (46:08.536)

No, and that it met a c in that according to rule of marketing too. In B2B it's the same, it's the same exact phenomenon that ninety-five percent of the actual decisioning happens before they actually activate the sales buying process, like reaching out to a company, scheduling a demo, talking to a salesperson, going into, in the case of automotive, into a a dealership. You've already picked your your category the the the brands that you want to select, and then it's about that sort of next level experience. And so you've got to


Vas (46:21.005)

Yeah. Right.


Vas (46:33.047)

Right. Right.


Krepsik (46:33.527)

Yeah, you I mean you can still you can still win that last mile, that last kilometer, and that's that's the role that middle funnel, lower funnel marketing plays, because like, hey, I got you, you're in the consideration set, you're in the purchase cycle, you're in the store. Okay, if you're in that situation, that purchase c location, now like for the modern marketer, amazing tools to like win the last mile, the last kilometer, the last like ten feet. I think w in our industry we're we've gotten really good at the last mile and the last kilometer.


Vas (46:45.176)

Yeah.


Vas (46:54.978)

Yeah.


Brett (46:56.228)

Yeah.


Vas (47:00.427)

Yeah.


Krepsik (47:01.473)

I think what we have to get a lot better at to Vass's point is going back to our fundamentals of like how do you make sure you show up on the list? How do you make sure like you're even you even get a chance to win that last mile?


Vas (47:08.652)

Yeah.


And you made me think something now, Matt, because you know, the o that used to be


Krepsik (47:15.933)

That's the honestly, we can end the podcast now. I made Vass think about something. I feel like that is the most that's my career highlight.


Vas (47:21.636)

Well listen, it might not be a good idea. So so hang on for a second, it might be another good, but you made me think of something, right? So this used to be only limited to specific categories, meaning, like, you know, you're thinking about like you know long-term and all that, how do you get into the consideration set? Now, in the world of you know, moving from marketing to hearts and minds, to hearts and agents, right?


Brett (47:22.882)

You've stumped him!


Vas (47:47.329)

That may become the thing for other categories as well, right? So I think what we're saying, we're with thinking is like the role of brand and being in the consideration set, how is that gonna kind of influence your decision to buy, you know, something trivial, not just a car, right? When you basically ask Muse or whatever agent to do that, right? So it probably becomes


Brett (48:10.576)

You have your consumer agent actually yeah, fil filtering through all the noise on ba on your behalf, right? How does the agent on the on the on the sell on the buy side, the advertiser agent, actually break through that?


Rio Longacre (48:12.874)

Okay.


Vas (48:14.627)

Right. Right. But then you say why why didn't you put Oreos there or whatever? Why didn't you put that brand that I like? Right? And if you don't have right.


Rio Longacre (48:16.82)

Yeah,


Rio Longacre (48:20.936)

Well should we be c should we be offer


Rio Longacre (48:26.218)

But should we be advertising to agents? I guess this is a philosophical question. Love your thoughts on this. So, like, should we be advertising to agents? Can we advertise to agents? Agents, an agent theoretically is supposed to represent you and your interests exclusively, right? Agents should not be representing advertisers' interests, certainly. They should be representing yours. That's de th there is a legal definition of an agent, right? So, like, should we be? I mean, I guess this is a question I I fact I had an article, Brett. I don't know if you a chance to read it yet to three


Vas (48:30.147)

Yeah.


Vas (48:34.639)

Yeah.


Rio Longacre (48:54.482)

three rules of agentic AI. Talked about they talked about agentic commerce. I know some people believe agenti commerce will never happen. I think it is going to happen. It's already starting to happen, but I think


Brett (49:01.966)

Yeah. Well and name the three rules based on Isaac Asimov for the geeks in the in the crowd, right?


Rio Longacre (49:06.994)

Yeah, in fact they were based on that. And the rules are the rules, so the three rules of Gentic AI AI. Loyalty, an agent must represent its user faithfully with fiduciary style duty to put the user's interests ahead of those of its providers, advertisers, merchants, or any other third party. Rule number two is obedience. An agent must carry out its user's instructions accurately and completely, except where doing so would conflict with the first rule.


Vas (49:17.345)

Hmm.


Rio Longacre (49:30.974)

When such a conflict arises, it must disclose the conflict and seek clarification. You yes, these are like Asimov's rules, right?


Brett (49:36.526)

Yeah, and was this was it were these iRobot? Was this an iRobot or one of his his sort of artificial it is from iRobot. The three rules of robotics.


Vas (49:37.111)

Yeah.


Rio Longacre (49:40.053)

This was from iRoba. Yeah, these are the sh these are the three rules of robotics, right? That these are based on. And rule number three is continuity. This one is where it does very very d does differ from ESMO. So continuity. Agent must serve its authority and often a duty to act to serve its user, except where doing so would conflict with the first or second rules. It must always respect the user's right to pause, replace, modify, or even shut down and terminate itself. That's what's different from the eSmo rules.


Krepsik (49:41.613)

Yeah.


Vas (49:42.465)

Yeah. Yeah, right.


Brett (50:06.157)

That's pretty cool.


Krepsik (50:06.492)

I I I still go back to like the I like those those are pretty cool. I go back to like the the conundrum that I think about right now, which is to your point, Rio, like should we advertise to agents? There's a big gap between intelligence and wisdom. When you think about loyalty, the context window for most agents is really small. And so therefore it doesn't know the entire gestalt of your very technical term.


Brett (50:22.093)

Yeah. Yeah.


Krepsik (50:36.108)

of your enti of your of your entire life. It's a Yiddish term. but like the agent doesn't have that that full doesn't have that wisdom, right? And I think that's the challenge with when you think about advertising to agents. They only can get the intelligence that they see today and what they can index now, but they may not know what you experienced as a kid and how you loved a shiny red Ferrari.


Brett (50:36.143)

A Yiddish term.


Vas (50:39.203)

Mm-hmm. Mm-hmm.


Brett (50:47.236)

Yeah.


Vas (50:47.735)

Mm-hmm.


Krepsik (51:06.11)

Right. And like that's been your dream car. And then so how do you how how do you build that w how do you maintain loyalty when you don't have con when you don't have wisdom, you just have immediate intelligence?


Brett (51:06.692)

Yep.


Vas (51:07.723)

Actually Yeah, actually my


Vas (51:14.251)

Mm.


Brett (51:18.221)

Yeah, yeah. And we talked about a lot a lot on stage at Advertising Week, but it but it I think that's what's distinctly human judgment, human craft oriented. That that that intelligence, no matter how super intelligent it is, even if super intelligence is is more intelligent than all human beings on the planet Earth literally combined, that they're saying that's gonna


Vas (51:18.647)

Yeah, that's


Krepsik (51:26.443)

Yeah.


Brett (51:35.514)

be there in the next couple of years, right? That doesn't solve the problem of of the context problem, right? Like in and we keep talking about context and tech and data strategy and AI strategy, but it's still not human context, right? Human context is culture. It's stuff that's handed down from your parents, it's epigenetics, it's stuff that we experience through our entire lives that can't be replicated through just a super intelligent mind, basically. I think you could make that argument philosophically, right?


Rio Longacre (52:06.526)

Vaza, like you had some thoughts on this. Love to hear what you think.


Vas (52:06.683)

I yeah, I I mean I I don't I don't I don't know. It's it we're getting very philosophical, but I was reading this paper this morning that yeah, that it was basically saying the obvious of what we all know that you know if you if you ever kind of go to whatever your LLM of choice and ask what do you know about me, right? You will see that it knows a ton about you. Right? Absolutely, right? You can ask your IQ and actually it's funny because if you


Rio Longacre (52:13.214)

I love it.


Rio Longacre (52:30.878)

You could even ask it your IQ if you ask it if you prompt it the right way. It will it it will ans it will actually answer it.


Brett (52:34.989)

I'm kind of afraid to ask it what it knows about me.


Vas (52:37.141)

No, actually ask it because, you know, different I asked it on on two or three and you know, I got different different answers. One of them was like very like, it looks like you're on the top thirty two percent or something. And the other one says, I'm not gonna make a claim with that because you I'm gonna be sycophantic and that was the right answer, by the way. But


The paper was saying that, you know, like what we all know that there's a lot of memories and there's a lot of signals that we give away when we talk to our LLMs, right? So we as you know now not marketers but consumers, right? So and it was actually saying that there is a way for those models to actually, you know, systematically trace and collect that.


to create a a mental mind and and kind of a framework for for that specific person, right? now


when and to what extent this is gonna appear on your chat CPT ads, it remains to be seen. But it will definitely appear on your, you know, action of the agent at some point, right? Because the whole thing is designed to learn to understand you as much as possible, right? And and we all use them, you know, 24-7. So like eventually we will do that. So but the question then becomes back to my Mike to your point, like


all this wisdom that we get from from the consumers if like the agents from a marketing side of view of point of view how do we account for that and how do we still like kind of market to these people to this not people to these bots now to this agent


Brett (54:16.431)

Civilies agents.


Rio Longacre (54:17.246)

Okay, but the thing is agents here's where I struggle with it, right? So agents certainly need information. They need access, let's say if like I I like a certain shoe, that shoe is going and sell thirty percent. You know what? I probably want to know about it, right? 'Cause maybe I will want to buy it, or maybe I want to buy it instead of something else. Even so even if I'm not in the market, maybe I want to make it so I think agents need information. So I certainly think that they need access. Brands need to be able to communicate with agents, right? But like I don't know, is that advertising?


Vas (54:34.559)

No.


Vas (54:42.806)

Yeah.


Vas (54:46.325)

It will be and there's examples of I think time and some other publishers were trying to kind of put advertising for agents, right? Which is you know, I mean th they will experiment, you know, everybody's gonna experiment and advertising it means mm it's maybe about just putting text there for the agent to read, right? And you know, if that works, then it's gonna become monetized at some point, you know.


Rio Longacre (54:52.138)

Yeah, they were trying.


Rio Longacre (55:11.882)

But it seems more like a product feed to me than actual advertising. Because agents I mean, these agents are incredibly fast. They can rip through tens of thousands of websites really quickly. They have memory. They might not have the context grapsic for sure, but they can but their ability to t to consume information like much faster than humans is certainly a thing. Go ahead.


Vas (55:19.243)

Yeah.


Krepsik (55:22.827)

Yeah.


Vas (55:26.787)

Okay.


Krepsik (55:27.562)

Well, so but let me let me explain like the difference when like like I think this is a fun conversation. I think the difference between intelligence and wisdom, and I think it's a super important idea because like like the human brain, right? And let's use a let's use a mathematical AI term called it. You guys familiar with a petaflop? It's basically the a petaflop, all right? So basically it's like how many calculations per second can you do, right? And so you think about like thought calculations.


Brett (55:45.516)

A petaflop.


Vas (55:46.179)

Heard the words.


Rio Longacre (55:46.366)

Yes, I am.


Rio Longacre (55:53.31)

Pedaflops a lot of calculations.


Krepsik (55:56.266)

Okay. So the point one of the most widely known chips out there is like your your NVIDIA your Blackwell Ultra, right? And somebody can like search this up for me and verify it, but it's somewhere between ten to fifteen petaflops is how much compute it can do. Do you know the rough scientific estimation of the petaflops of a human human brain equivalence? Roughly fifty to a hundred petaflops.


Now it's not that the brain has the raw compute power that a AI agent has to kind of like brute force intelligence. Like AI is a is a workhorse and it's amazing that you need a lot of humans to replicate. But what AI can't replicate is the complexity, the breadth, the w context, and the situation that the human brain and like that's that's where I would say AI is great at volume.


way better than like any single human could be, but it's very poor at breadth in terms of situation and context. And it's a very simple when you think about it in terms of pedaflops, you're like, you got you're sitting there with a hundred pedaflops on your head and you look at your AI agent, it's like maybe running ten. But it can do a ton of deep it can do it could be your workhorse and do a bunch of research on it. Right. It's


Brett (57:15.694)

Yeah.


Krepsik (57:16.927)

And I think that's an important way as we think about AI coming into the space and how how you better best utilize it. It's great for that heavy, heavy research when I where I need a lot of like shovels digging ditches, right? I need somebody to go do a bunch of research. But ultimately it's like what the human brain is really conditioned for is like giving me given the information, I can understand the situation of circumstance. I have the wisdom to understand how do I think that information about that information in context to this current situation and this moment.


Brett (57:28.75)

Yeah.


Vas (57:29.517)

Mm. Yeah, yeah.


Brett (57:38.693)

Yeah.


Brett (57:45.189)

Yeah. Very complex environmental conditions that we've evolved to adapt and adjust to for survivability purposes, right? Think about all the nuances from weather and temperature to predators to other humans to that that the human brain evolved under those conditions, under con incredible stress. And that's that's part of the reason why. Right? You know, you gotta understand context at like a level that that you're saying that AI really can't. So at the end of the day, we are more intelligent than our AI.


but you have to define intelligence more broadly, is what you're saying. We have more wisdom than our AI versus kind of the brute force, right?


Vas (58:14.43)

Yeah.


Krepsik (58:17.343)

We're more we have more wisdom than our AI.


Vas (58:17.527)

Yeah.


Vas (58:21.225)

Yeah. But with great wisdom comes great responsibility. But the point here, right, is that that if we don't use this wisdom, because you know, like like we we said before earlier that you know we have to AI is raising the ceiling the the floor, so we have to use our taste or our wisdom to make


to ask different questions, to make different choices, right? If we take what the AI is giving us, then we're just basically like using five percent of this, you know, you know wisdom and we're not, you know, we're not really doing anything.


Brett (59:02.148)

Yeah, yeah.


Rio Longacre (59:02.322)

Yeah, I think that's an interesting point you just made about wisdom, the fact AI can create lots of things faster. Like one thing I think that we've touched on this a little bit is create creative fatigue, right? So I think the role of AI I think offers potential, right, to really address that because it's been so damn creative's been the bottleneck, right? Like making creative, getting it approved, especially in regul regulate regulated industries, actually putting it into the wild and getting it out pumped out, whether it's in you know ads or CRM.


emails, whatever, just getting that new fresh creative out's been very hard. So AI potentially offers the ability to solve for that, like br like remove that bottleneck. But I think at the same time, it can also introduces, you know, the AI slop, the lowering of the quality or the muddling of the message. Thoughts there?


Krepsik (59:48.853)

Yeah, I mean it's like it yeah, I can give you volume, but you gotta have intentionality. And I think that's that's the balance is like I love the volume and the speed with which I can produce more creative, but you gotta have that intelligence or or wisdom to kind of make that that intentionality you have with your messaging.


Brett (01:00:07.566)

Yeah. Well and the intentionality is how you break through and how you prevent sort of creative fatigue, right? 'Cause we're all as consumers, we're all being bombarded with what you know, d I don't have the numbers in front of me, but it's it's exponentially more in order of magnitude more advertising per day than we ever have been as a species in since the since the birth of advertising, right? And so there's you to you have to be the speaking of that relevance and intentionality to break through, it's not a versioning exercise. It's a differentiation exercise, right? And that


Vas (01:00:08.695)

Yeah.


Vas (01:00:23.359)

Yeah. Yeah.


Vas (01:00:31.768)

Yeah.


Krepsik (01:00:36.22)

It's it's a it's a memorability exercise. It's like, can I grab your attention, can I communicate a message and then do you remember it?


Brett (01:00:38.074)

Yeah.


Vas (01:00:42.751)

Right, right. Right. And and I think it's both actually though. Because we did a study about five years ago, six years ago I think now, a very large neuroscience study at the MA. About about nine hundred people, like with EG, with the the think caps, right? and one of the things that we found, and we were just trying to understand how cognition works in advertising, right? And how long does it take for you as you're scrolling down to just say,


Brett (01:00:42.905)

Yeah.


Brett (01:00:59.93)

Yeah.


Vas (01:01:12.342)

make a decision am I gonna keep scrolling or am I gonna keep going, right? Or I'm gonna slow down. It takes less than half a sec about half a second. Like between half a second and a second. Like in some like 0.6 seconds. This is crazy.


Brett (01:01:15.344)

Yeah.


Vas (01:01:26.276)

What that means is that, you know, the consumer the person is gonna make a decision whether to engage with your content in less than a second, let's say. Right? Now that varies by channel, you know, it's gonna be longer in desktop because, you know, whatever for for reasons and and and faster in mobile, which is like in your face and it's like it's gonna be you know, you're gonna have like video for instance is gonna be a lot more immersive and it's gonna immediately kind of create some sort of emotional response. But


You need to be thinking about like a first-second strategy. Like, so so you need to be thinking about like not just like like how do I tell a story overall, but how do I tell a story in an environment where people are gonna make a decision less than half a second, in a less than a second, right? And and I think this is where your intelligence and what


The machine is gonna tell you how it works, but also your taste needs to come together. And it's a it's a it's a tough one.


Brett (01:02:26.372)

Yep. Yeah.


Yeah, 'cause there's so many factors that go into that. There's behavioral factors, right? Where have you come from, potentially where you're going, there's contextual factor factors, there's time you know, time of day.


Vas (01:02:38.273)

Yeah. Yeah.


Brett (01:02:41.314)

Right? the all these things go into sort of decisioning like is this the the kind of magic moment that every CMO wants to hit? You know, that that those those two weeks of the year that you've got to hit two percent of the population to get your new ad, right, as for an automobile. Right? It's like, yeah, how do you and and machine learning I think helps with the raw compute of the complexity of information that it needs to analyze to be able to to get to that. Are we anywhere are we anywhere close to that though? I mean our brands


Vas (01:02:47.095)

Yeah. Right.


Vas (01:02:57.356)

Yeah.


Yeah.


Vas (01:03:04.395)

Right, right. And and remember


Brett (01:03:10.178)

comfortable with their ability to sort of predict like, hey, we know, you know, it kind of hints at at personalization and I'm not fully convinced that that's even a close to reality at this point, even in digital. I mean, what what what do you think?


Vas (01:03:22.134)

A personal


Vas (01:03:26.126)

With that format? Yeah. Yeah. Yeah.


Brett (01:03:27.37)

Krepsik, what do you You think we're you think we're close to that? I mean do you think CMOs are still searching for that? Like how do we actually get align all those factors so that we can hit people at exactly the moment that's relevant to them?


Krepsik (01:03:40.234)

I think that that would presuppose a level of efficiency in advertising that if I look at the data, we're still chasing share voice.


Brett (01:03:47.76)

Yeah, we really are. Right, let's be honest, right?


Krepsik (01:03:51.231)

Let's be honest, like it's still pretty much the share a voice game of like, you know, let's get that message out there. Now we do try to do share a voice in unique ways and different messaging and different formats, but it's like I think we're still like share a voice is still a pretty good predictor.


Vas (01:04:01.997)

Yeah.


Brett (01:04:07.588)

We're still we're still in the Wanamaker phase, so to speak. you know Yeah. Well and and there yeah, all the numbers you see say hey cor direct correlation between increase in share of voice and increase of market share within your category, right?


Rio Longacre (01:04:10.233)

Reach frequency, shar voice, all the old yell yeah.


Krepsik (01:04:12.232)

Three frequency share of Okay.


Rio Longacre (01:04:21.241)

Probably true. I mean it's there's things that probably haven't changed.


Krepsik (01:04:22.322)

It's you know, it is funny, it's like ever since I've I'd been in media and advertising, it's like the the one thing that's true is like share voice and share market have a really good correlation over time. You can buck that trend for like a week, a month, maybe a quarter, but over the long term, share voice and share market are incredibly correlated.


Brett (01:04:33.689)

Yeah.


Brett (01:04:43.396)

But back to that challenger brand comment we made earlier. How how if you're trying to enter a category, right, and you don't have the budget, you don't have the share of of of of voice sort of oomph behind you with with media dollars, how do you do it? How do you break through?


Rio Longacre (01:04:59.507)

Good advertising.


Krepsik (01:05:00.606)

Good advertising. I mean I mean I mean in the short term, good advertising can punch way above its weight. And I think that's the important


Brett (01:05:01.955)

Yeah


Brett (01:05:07.077)

Yeah.


Rio Longacre (01:05:07.177)

Yeah, but you think about liquid liquid death, like they sell freaking water, canned water for like the si the price of like beer. It's incredible. The people but it like that's great advertising, period.


Vas (01:05:08.707)

Yeah.


Brett (01:05:10.991)

Yeah.


Vas (01:05:11.936)

Right. Right. But but also how many things right, how many things find you on Instagram that you didn't even know existed or you didn't even know that you want it? It's like, my god, this is exactly what I did I didn't even know existed like a minute ago, and now it's like this is the best thing. Right?


Krepsik (01:05:15.048)

Water in in a recyclable aluminum can.


Rio Longacre (01:05:18.686)

You know.


Brett (01:05:21.838)

Yeah.


Brett (01:05:25.22)

Yeah.


Brett (01:05:31.406)

Yeah. And not not to be generic like with the Adelaide partnership with this with this report, but that sounds to me a little bit less share of voice because you're not winning a certain percentage of total exposure. You're actually breaking through with a much smaller percentage. So it's kind of like a share of intention or sh share of quality.


Krepsik (01:05:45.214)

Well well it's it's it was share of quality. So I think I think I I think the thing that Adelaide teaches us, I think the thing that like a lot of creative research teaches us is like in the short term, quality can compensate for like volume. Like if you a great ad can break through. High quality like impressions can break through.


Brett (01:06:00.58)

Yeah.


Brett (01:06:10.02)

Yeah.


Rio Longacre (01:06:10.173)

And I think that's where they're really like the the branding, the top the upper funnel. I mean, I think that's where you can make a huge difference, right? Because I mean if you like I I think you look at the reverse of that would be like Vas, you'd mentioned meta before. Yeah, I mean you can target the really, really well in that. Not as well as used to they hide a lot of the they they don't expose all the data points they used to. Used to be I remember like back in the day, I used to get ads based on your specific age. Do you need insurance in your this year? I mean, it was incre it was it were they really worked, but apparently, you know, it exp


Vas (01:06:29.475)

Of course.


Krepsik (01:06:34.339)

Ha ha.


Rio Longacre (01:06:38.579)

But certainly, but that'd be the opposite. But I think looking at like brand and upper funnel, I mean, for sure, that's where you can make such a difference. And you look, there's the Nike example, right? We talked about a lot in this pod about how they just neglected it for years and their share of voice completely went away. People no they they they exited the conversation completely.


Vas (01:06:41.269)

No. No.


Vas (01:06:55.882)

Yeah.


Krepsik (01:06:58.056)

Yeah.


Vas (01:06:58.424)

This I I th I think we we live in a world where both of those things happen at the same time. And I think a a big part of of your spend is still obeying the rules of y you know, like yesterday. And then you still Yeah, quality yeah.


Brett (01:06:58.745)

Yeah.


Brett (01:07:03.278)

Yeah.


Brett (01:07:12.227)

Yeah. C you're saying quality and quantity, right? R reach and effective reach.


Vas (01:07:16.088)

But then you also have it rates and effects. And then you also though have pockets where you see very clear signal to noise. We did a study, we've been doing a study what we call conservation for AI personalization, and and it's all about the delivery of like what is the DCO of in the world of AI, right? Can you basically take creative and dis and and find a way to match it with specific cohorts of people using not even like meta.


Science, like you know, metacontextual things, like you know, like time of the day and you know, IB under things. Now it turns out that you can, and we've done 29 studies or 30 studies that you know there's there's a lot of like good impact that we've seen, like double digit, sometimes triple digit kind of kind of impact. In in one case, it was like 272% lift because we do have an increment we measure incrementality. Now, the interesting thing to give you an example is that how


you see that there's signal to noise in some areas that you didn't expect that you will see. So the we did a study with Shell, it's a public case study, so I can speak to that. And they have like three concepts and you know with number of variations. One of them I was using a Formula One creative.


And we see the delivery like through like over the like the four weeks that it was playing, it was zip, zz, zz, zzp. So it was just really kind of nowhere and then with spiking. And nowhere and with spiking. And then we just see like where was it spiking? It was spiking during the weekends that the Formula One was playing. Now, obviously, the AI didn't have no idea that that was happening, but it was reacting in the way that people were engaging with the ad, right? And and then when the one one of the drivers that was actually featured on this ad went to


The hospital because he had an accident. That thing was moved. And we saw it too. Right. So I think like there are areas where you can see that there's, my god, like this thing works. I mean it doesn't work all the time, but it works, and then you just say, okay, maybe you know, and but then we still live in a world of like this.


Brett (01:09:01.987)

And you saw an extreme s yes, talk about contextual relevance, right? For


Vas (01:09:20.692)

And then okay, reach, frequency, you know, like more like so two worlds at the same time, which is which is which is interesting how you kind of balance both.


Brett (01:09:31.214)

Yeah. Yep. And it and it just seems like when if you've got you know, kind of unified data systems that allow for some level of let's say real time data feedback on all of these signals that are driving s changes in in engagement, attention, behavior, etcetera.


Vas (01:09:37.315)

Yeah.


Vas (01:09:49.9)

Right.


Brett (01:09:50.276)

Right, that's the stuff that's gotta be fully automated. And I think we're getting a little bit closer to that. So that you can make updates, you can increase reach or frequency, or specific cohort targeting changes, et cetera, or creative changes. Right. But a lot of that stuff has traditionally gone through a typical legacy process, right? We need a new brief, we've got to redevelop creative, right? We have some level of automation, but it's pretty limited. it seems like we're we're everybody's chasing this world where you've got kind of data congruence.


Vas (01:10:06.476)

Yeah.


Vas (01:10:13.164)

Yeah.


Vas (01:10:16.555)

Right. Yeah.


Brett (01:10:17.761)

and real time insight and you can make a whole bunch of changes both across Media Spend and Creative to to affect right?


Rio Longacre (01:10:23.815)

I I think people over overthink the personalization thing, Brett, just generally. I think that, you know, sure, personalization works, but I mean like why is meta advertising work so well? 'Cause they like it's like product th they'd be able to get the right product in front of the right type of person. I I don't think it's down to the individual. I mean, like I I bought I think about the stuff I bought from there. I bought a wallet from there. Okay, why did I buy the wallet? It's a wallet that like that it that probably dudes my age have a high likelihood in buying.


Vas (01:10:44.747)

Mm.


Rio Longacre (01:10:51.143)

Right. So they I don't think the targeting was any more accurate than that. Right. It's like a small little wallet that's expensive, that looks nice. It found me.


Brett (01:10:54.179)

Yeah.


Vas (01:10:55.671)

But it found you, Rio. It found you. I mean the shit sorry I mean that I bought like it's unbelievable what what finds me. And I said, it didn't just find me, it wasn't just the product, but it was the product and the benefit articulation and the fact that it was just like a straight line between my need state that I didn't know existed, with the benefit and the reason to believe, and boom, ready. Right?


Brett (01:10:57.261)

Yeah. Yeah.


Rio Longacre (01:11:17.575)

Yeah, I didn't know I needed a new wallet until I saw that damn ad.


Brett (01:11:18.349)

Yeah. Yeah. Yeah. Well no no but I Rio's making a point about but about but sort of one to one and sort of maybe the lie of one to one level personalization. That there's a cohort of people that can you can match to a cohort level that that that matches to that need state at that particular magic point in time in that context, right, that's good enough better than anything that we're c that we possibly would ever seen in advertising, right? Would you agree with that, yeah, Krebsick, would you think?


Krepsik (01:11:20.425)

To you solid here, man.


Vas (01:11:25.96)

Yeah, one to one maybe not. Yeah, you're right. Yeah. Yeah.


Vas (01:11:35.785)

Right. Right.


Yeah. Yeah. Yeah. Yeah. Right.


Krepsik (01:11:44.595)

I mean I mean here's what I'd say about like targeting and a lot of this. I think the power of like some of this personalization honestly is the fact that we know it's a real audience versus like in the open web advertising, like you know, actually it like how much of a are we really advertising to real people sometimes?


Brett (01:12:06.006)

Yeah, yeah. Yeah. No, you're right. It c it could be, you know, in in immin


Rio Longacre (01:12:08.361)

True.


Krepsik (01:12:09.872)

And so and so the the beauty the beauty of like when y Yeah, maybe it won't matter soon, like you think about the beauty of like authenticated platforms, you at least know the people you're advertising to are real for them real people.


Rio Longacre (01:12:11.421)

Maybe it will matter since


Brett (01:12:22.028)

Yep. Yeah, there's some yeah, and that's where beta really thrives. And that that was the episode I don't know if you've heard it, w if if people haven't the McKinley, Scott McKinley episode of True Set. And I mean his whole argument was like the entire data ecosystem, even today in twenty twenty six, is polluted. Regardless of what every identity vendor out there says, it is a super messy data space. and that's why you feel like you're you're receiving just bad right?


Krepsik (01:12:31.944)

yeah.


Krepsik (01:12:47.132)

Well well it's it's hard because a lot of the additional initial identity signals came from like I always say okay cupid and farmers only dot com. You know, that's where we got the the like cookie IP to hashed email address linkages. Like when we were all over first learning how to log on to like J Date and dating websites, you know.


Vas (01:12:59.91)

Mm.


Brett (01:13:04.514)

Yeah. Yeah. And then and then and then a bunch of database marketing companies that became sort of the identity spine of the modern internet, right?


Krepsik (01:13:12.296)

Yeah. I mean I I I think I think there's the best intentions. Like I think with the beauty of inter of digital advertising, it's like a really compelling rich media format. The challenge is like it's really hard to connect that to a real person. And the value of advertising on Meta, the value of advertising on YouTube, the value of can be advertising for B2B marketing on LinkedIn is like you kind of can guarantee its authenticated audiences back to your


Rio Longacre (01:13:34.001)

As long as as long as you don't check the off site or the audience extension box, right? And then you start blasting out to yeah.


Brett (01:13:34.093)

Yeah.


Krepsik (01:13:39.192)

As long as you don't check the audience extension box and off site box. But


Brett (01:13:41.358)

Yeah, and be sure b be sure you check that to turn that off, right? You don't want audio extensions outside of your walled garden relationships.


Vas (01:13:46.668)

Yeah.


Rio Longacre (01:13:46.781)

Yeah, no Yeah, 'cause it'll they'll flip it on automatically now, right? 'Cause they they want you they I mean the thing isn't yeah, I mean they run out of inventory, right? And like th and they they want more advertisers to spend more and there's only so many authenticated users, Krebsick to your point. So they will happily spend all your budget on off site, which is probably not humans for the most part, let's face it.


Vas (01:13:50.972)

All the time.


Brett (01:13:52.397)

Yeah.


Vas (01:13:54.945)

No.


Brett (01:14:05.667)

Yeah.


Krepsik (01:14:05.82)

Yeah, 'cause sometimes you look at these impression counts and you're like, Read'll be this one, Batman. There still is only three hundred million plus Americans. I do not know how we got to this impression volume.


Brett (01:14:12.778)

No I know Yeah. Yeah yeah. Your s yeah, your supply actually d it just doesn't exist. It's actually mathematically impossible for that supply, that ad inventory. Some people say it's infinite, but I I actually think you could probably calculate this out and say it's actually not infinite per se. Yeah.


Rio Longacre (01:14:28.146)

Well but Brett, that was Foo Doctor Foo's point. He was always saying C PMs are too low and the fact that, you know, they a lot of this supply is not real. Like it's


Krepsik (01:14:28.953)

I don't know.


Vas (01:14:31.299)

Yeah.


Brett (01:14:35.842)

Yeah. Yeah.


Krepsik (01:14:35.996)

Yeah. For the price point of CPMs, this can't be real.


Brett (01:14:42.796)

Yeah, yeah, and I think with CTV, and you I you've leaned a lot in there, Krebsic, for the you know, and I'm sure that that MMA as well, right? You know, th there's been a little bit of pollution in that ecosystem where people are chasing cheap CPMs and getting ads that are never served. There might be an impression triggered, that's n yeah, a JavaScript that's running in the background, so you're actually paying for it, but but people aren't being exposed. And so there's just a ton of garbage, so supposed CTV inventory. But it's typically the open web stuff, it's not the authenticated


Rio Longacre (01:15:00.048)

JavaScript running in the background. Yeah.


Krepsik (01:15:02.588)

Ciao.


Rio Longacre (01:15:12.242)

No real C T V. I mean it's not like you're not you're not buying Amazon Prime Video or Power Mount or Hulu and any of the yeah.


Brett (01:15:12.75)

Buying from Paramount, I'm buying from Netflix. Yeah. And so so Krebsic, b big theoretical question is that this is this points to the point of of of whether or not the open web, which be some people say it's already dead, there's it's just walking a long slow path to extinction, right? What do you think of the open web's ability to create some level of authenticated data moat? Is it a is it a premium versus long tail? I mean, how do you think that is gonna play out over time?


Krepsik (01:15:18.435)

Yeah.


Krepsik (01:15:43.438)

I think it's gonna go by the way of the classifieds. I think it's it's the it's gonna be we're gonna look back on it like like directory marketing used to be. Like directory marketing killer business in the nineties. Maybe in the nineties. I feel like I'm dating myself now. But but but okay. Like


Brett (01:15:47.106)

Yeah. Okay. Craig's list.


Vas (01:15:54.187)

Yeah.


Rio Longacre (01:16:02.802)

We were we were all working in the nineties, don't worry about it, yeah.


Vas (01:16:05.047)

Mm yeah. No shame here.


Krepsik (01:16:10.331)

But like I think I think in many ways it taught us a lot about new forms of advertising. But I think the future, candidly, at least this next chapter of media and advertising is going to be curated w audiences and walled gardens. I think it's like, can you own your supply chain? And I think if for those platforms and players that own a real curated audience, you know, you you've got your own world there and you can authenticate your users. And I think if I think about where the internet is at right now, it's like


Brett (01:16:22.563)

Yeah.


Krepsik (01:16:39.847)

You know, we're in that I used to say that internet like 1.0, 2.0, 3.0 phase. Like internet 1.0 was all about indexing information. It was all about access, candidly. Like that was the original internet. Was like, can we give you access to your information? And then internet two dot was super cool because it was all about, wait a second, we can build communities on the internet, right? It was the birth it w it it yeah, it was it was like


Brett (01:16:47.34)

Yep. Yep. Yep.


Brett (01:16:59.586)

Yes, it was the birth of social yeah, you've been listening to Shot Tobaccoala. This is you're laying some


Krepsik (01:17:07.099)

Back to like the idea of like we have communities, we have these different like platforms, we have MySpace, right? We have Facebook with like and it was cool, like you could find your people. And then Internet 3.0 in my mind is about like the person or the individual, right? And so in many ways, I think the internet is a great distribution vehicle, but for advertising, it's about the person. And so where the premium is, and I I kind of agree with Scott McKinley on this one. Like I think in the end of the day, it's like the area where advertising we see is the most effective.


Brett (01:17:09.379)

Yeah.


Krepsik (01:17:36.857)

isn't because we can do like like hyper targeting. It's because we just know these are just real people. It's an authenticated audience. And like it's it's that view that I think becomes the most value.


Brett (01:17:42.402)

Yeah.


Rio Longacre (01:17:43.101)

Yeah. By the way, I'm agreeing with you. Like the amount of times people spend on the open web is gonna collapse to close to zero. They're gonna spend their time in like in like like hopefully you know places like Signal and Noise or like they're gonna be going to places where we have our newsletter, we have our pods, we have our readers who come back every week, right? I think all of these all of these different services that that do that, I mean I think that that's where people are spending their time. They're gonna be spending their time in YouTube watching stuff. They're gonna they're gonna be going to these platforms where they where they have community, where they have content.


Where they they know where to go, right? And then and I think getting them getting them back, but the amount of time I mean, I I have so few tabs open anymore, right? I barely spend I'm not going I don't have fifteen, twenty tabs open 'cause you don't need to anymore.


Krepsik (01:18:24.433)

Yeah. You don't need to.


Brett (01:18:25.228)

Yeah. Authenticated content communities. I mean that that really is if this is sort of a a through line for the last probably ten episodes that we've done, Rio, including advertising week, it seems like that is incredibly important because you do. You find your your cohort, right? You find the type of co which is aligned to certain types of content in certain places. And as long as you're able to reach those people and they're authenticated as a as a brand, you should be able to do that at scale without all this unnecessary complexity of


Rio Longacre (01:18:34.343)

Yeah.


Krepsik (01:18:51.911)

Well, I I think I think that's always gonna be the hard part for the content side of the ecosystem is always like how do you build your audience and how do you kind of create scale? And that's hard. And I think that's that's can


Vas (01:18:58.659)

Mm-hmm.


Rio Longacre (01:19:01.722)

It's hard. It is hard to build an audience.


Brett (01:19:02.636)

Yeah. Yeah. I mean look at Net Netflix is is currently struggling with that. Despite their their good numbers, the streets still not you look at their growth numbers over the last quarter and their numbers are still not satisfying because of subscription demand has flatlined.


Krepsik (01:19:07.889)

Yeah.


Rio Longacre (01:19:14.248)

Yeah, but th they're but they're one hit away. They're one hit with like Zeek zeekgeisty hit away from being back, I think. They just had shitty content. Like it does there's so much bad you go to Netflix, it's like all bad content and like a couple things I wanna watch, right? Like I'll go to I'll watch Fauda, which is on bid really like is like I've been really into recently, but then Fouda's great. This this last season is just killer, right? But like so but I won't like but I won't like


Brett (01:19:30.883)

Yeah.


Brett (01:19:34.474)

It Fauda? That's a that is one of the better shows on television. Yeah.


Krepsik (01:19:36.007)

Yeah.


Rio Longacre (01:19:41.148)

I won't go to Netflix for a couple more months, right? And then until there's something else I want to watch, that I I don't know. I mean it's I so I I wouldn't count Netflix outbreak, but I I d I agree with you that you you know, it's a content game, you have to continually churn it out to keep people there, otherwise they disappear.


Brett (01:19:54.691)

Yeah. But you but you go to these platforms, these authenticated platforms, like Netflix is a perfect example, or YouTube, right? And the stuff that you see is hyperpersonalized to the stuff that you're interested. Mine my Netflix has got tons of stand-up comedy, but when I go to my son's c Netflix, not as much. He's not watching as much stand-up comedy as I am, right? So


Krepsik (01:20:10.897)

Yeah. Or you or you go you go you go to Vass's Instagram feed and it's highly curated for the things that he needs. Like there is there is a lot of truth to that curation of that experience of that content. Keeps bringing us back.


Vas (01:20:12.043)

I w I was full.


Vas (01:20:19.532)

Yeah.


Brett (01:20:22.262)

Yeah. Yeah, totally. Totally. Well, I hey I think this is this is good. I mean this is good for a good hour plus tw you know, hour and twenty minutes. I think we're


Rio Longacre (01:20:30.033)

Well but can't just one one final question for maybe for the two of them. So if like like one bit of advice that like if you have a CMO in front of you and she says what's the one bit of advice that from this report if you were to give to a CMO what would it be? Maybe Val start with you.


Vas (01:20:44.151)

Yeah, I I I think I think it's it's the finding that you need the intelligence and you need to well actually, let me change that because I think intelligence is more than data, right? And in I I I I I was about to say that you need the intelligence but you also need to know when to break out of it. but I think intelligence is something that's gonna be able to give you


an edge to make a right decision. Right. So you need that, whether you call it kind of data or you call it intelligence, you need to to have that as an input.


But you need to be able to crack your own egg, as as as they say, and and just in your own recipe and and you know, just just move away from from that and make choices that are intentional. especially now in a world of AI and the world of like, you know, regression to the mean, I think more important than ever to to do that. You know, invest in your own data, invest in your own in invest in intelligence, but also like, you know, invest in in, you know, taste and and


Human choice. Invest in creative talent in people, yes. Right.


Brett (01:21:48.599)

Which is in invest in creative talent, right? Like the right type of creative talent, people that have that experience are are gonna put you ab above and beyond what any AI is gonna do for you. Matt, what's your what's your perspective?


Rio Longacre (01:21:53.403)

Yeah. It's a good one.


Krepsik (01:22:02.426)

man, I I I I I'll I'll build on Vass's point 'cause I like it a lot. But I think intelligence sh should sharpen your judgment, but not replace your courage. Be courageous.


Brett (01:22:13.13)

Yeah. Yeah.


Rio Longacre (01:22:13.152)

I love that.


Vas (01:22:14.381)

Man, you like I you you yeah I like your point better. That was a Twitter that was a tweet right there. How do you say that again? Like damn it. That was good. Damn it. Yes, yeah.


Krepsik (01:22:18.286)

Ha ha.


Brett (01:22:23.096)

W yeah, we're gonna have our post production team just drop mic drop it, just it right at that line, right? Be courageous. But but you know, publicly traded companies, Matt, you know, oftentimes are under pressures as teams, as organizations, the markets, your board. and that pressure oftentimes is risk averse and


Rio Longacre (01:22:24.027)

Mic drop. Right.


Krepsik (01:22:25.99)

Yeah.


Rio Longacre (01:22:40.625)

Death by Focus Group.


Brett (01:22:41.91)

Yeah, death by focus group. And so being courageous, you know, going with the zeitgeist, going against the the the the you know the the current of your category or your competitors might be or or even your d you might be hitting a cohort that's the executive team's not gonna understand your messaging because it doesn't appeal to their cohort or their demographics, right? So being courageous is hard, but I still think it's it's the critical role of the creatives within these organizations to be


able to to say, hey, l just listen to us and trust us, at least test and learn in a real world environment and see how it works. Prove it out.


Vas (01:23:14.977)

Yeah. Right. Right. But also build a creative yeah.


Krepsik (01:23:18.458)

Yeah, I I had a I had a board meeting yesterday and like I came out of it and I I I kinda was thinking I was talking to a group of folks, I like I'd rather be courageous three hundred and sixty five days out of the year and be wrong half the time than like spend my time trying to be perfect once a month.


Vas (01:23:30.413)

Yeah.


Brett (01:23:35.712)

Yeah. Yep.


Vas (01:23:35.863)

Yeah. Yeah.


Krepsik (01:23:36.698)

Do you know what I mean? Like if I got three hundred and sixty five chances, like I'm gonna take that. Even though if I know I'm wrong half the time versus like being perfect twelve times. Like you can just get more done and get further. Yeah. Yeah.


Brett (01:23:43.425)

No, I know.


Yeah.


Rio Longacre (01:23:47.111)

Well I'd rather I'd rather I'd rather be controversial than invisible. That's that th those are words to live by.


Brett (01:23:51.202)

Yep, yep. It's like you've gotta take you've gotta take the reps and the right type of reps in order to hit the ball out of the park. And if you're not swinging because you're fearful right in the right way, you're just never gonna succeed, right? I think that's a really great way of looking at it.


Vas (01:24:03.137)

Right. But have the intelligence though, right? I mean that that like be be fearful be like you know, bold and and courageous when you know what you're doing, right? Right? Like woo


Krepsik (01:24:07.504)

Yeah.


Brett (01:24:14.348)

Yeah, and f and you're informed. You're inform you're


Rio Longacre (01:24:15.493)

Ha ha ha.


Yes, that is perfect.


Krepsik (01:24:19.686)

Don't don't go drive a car if you've never learned how to drive a car.


Vas (01:24:23.541)

Right, exactly. So so I think I think more today more than ever, I mean we need to build that intelligence spine, especially in the creative process, right? You know, your planning, your creative production, your you know, your delivery, your optimization, your measurement, they all need to kind of have a creative spine that makes you and a feedback loop that helps you make become better every you know, every time, you know, more and more like with every iteration.


Rio Longacre (01:24:23.736)

yeah.


Vas (01:24:51.963)

And then this is where you're gonna make your choices and you're gonna take your bets, right? Based on that, not just based on d do I f do I feel like it today. Yeah, as an input. Exactly. Yes.


Brett (01:24:58.87)

Yeah. Well based based on that as an as an input, right? So that you're informed in your decision making. And you know, but when you wanna break the the when you want to go somewhere that hasn't been gone before, oftentimes the past information's not gonna tell you a lot about what to do in the future. Right? well this was awesome. I mean we could go on for hours. thanks, Matt. thanks, Faz. It was this was awesome. and everybody that made it this far, visit us at www.deb dot signal noise.ai. subscribe to our newsletter.


Vas (01:25:08.885)

Yeah. Yeah. It's not gonna tell you that. Exactly. It's not gonna tell you that.


Rio Longacre (01:25:16.657)

It's great.


Brett (01:25:28.692)

comes out three days a week with just a couple pieces of content, articles as well as pods. And then, you know, obviously listen to us or watch us on all of your favorite platforms. And we'll see you next time. Thanks everybody.


Krepsik (01:25:38.865)

Sounds great. Thanks.


Vas (01:25:40.612)

Thank you guys.

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page