top of page
Search

The End of Marketing’s Age of Opinion: Greg Stuart on Measurement, Attribution, and Making Marketing More Predictable

  • 7 days ago
  • 56 min read







What if marketing’s biggest problem isn’t a lack of data… but a lack of discipline?

In this episode of Signal & Noise, we sit down with Greg Stuart, CEO of the Marketing + Media Alliance (MMA), for a deep, unfiltered conversation on why marketing still struggles to earn trust—and what it will take to fix it.

Greg has spent the last several years rebuilding MMA into a global force focused on one core mission: turning marketing from a field driven by opinions, proxies, and vendor narratives into a real profession grounded in science, evidence, and predictable outcomes. 

This conversation is a hard reset.

Because despite more dashboards, more tools, and more AI than ever, most marketing organizations still can’t answer the one question that matters: Is this actually making better decisions—and driving real business impact?

The “Age of Opinion” is Ending Greg argues that marketing has operated for decades without a codified body of knowledge—unlike finance, medicine, or engineering. And until that changes, trust from the C-suite (and especially the CFO) will remain fragile.

Measurement ≠ Better Decisions

  • Only ~30% of marketers trust their KPIs enough to use them for strategy

  • Only ~29% can trace decisions back to data

  • Just ~22% describe their analytics capabilities as robust 

The problem isn’t dashboards. It’s decision-making maturity.

The Attribution Illusion From last-click to MMM, Greg breaks down why most measurement frameworks still fall short—and why marketing continues to struggle to prove value in financial terms that CFOs actually believe.

AI Won’t Save Broken Foundations AI doesn’t create truth—it reflects patterns. If your assumptions, metrics, and operating model are weak, AI will simply scale those weaknesses faster.

Why Marketing Lacks Trust (and How to Fix It)

  • No standardized “science” of marketing

  • Overreliance on vendors and proxies

  • Weak linkage to financial outcomes

  • A discipline still driven too often by narrative over evidence

  • Marketing’s credibility problem is structural, not cosmetic

  • Measurement maturity is organizational—not just technological

  • CFO trust is the ultimate test of marketing effectiveness

  • AI is a force multiplier—but only if the fundamentals are sound

  • The future belongs to teams that move from reporting → experimentation → evidence → prediction

This isn’t another conversation about dashboards, tools, or tactics.

It’s about whether marketing can evolve into something more rigorous, more trusted, and more predictable—or whether it continues to operate as a function driven by opinion, intuition, and fragmented incentives.

If you’re a CMO, operator, or builder trying to navigate measurement, attribution, and AI… this episode will challenge how you think about all of it.

Watch the full episode and join the conversation.

🔑 What We Cover💡 Key Takeaways🎯 Why This Episode Matters Read the full transcript below

Greg (00:05)

Okay, good, should be all right, yeah.


Okay.


Brett House (00:09)

Hey,


welcome back everybody. It's Brett House at Signal Noise with my co-host Ryo Longacre back after a long trip to the South, to possible in Miami and New Orleans Jazz Fest. And we've got Greg Stewart we're thrilled to talk to. And for those that don't know who Greg Stewart is, he's the CEO of the Marketing and Media Alliance, MMA, formerly known as the Mobile Marketing Association. We'll talk about, and not to be mixed up with Ms. Martial Arts.


Rio (00:20)

So much fun.


Not to be mixed up with mixed martial arts, right?


Greg (00:35)

No, no, Only if you're trying to


only if you're a teenage son trying to date my daughters do I let you know that I'm with the that I'm the CEO of the MMA.


Rio (00:43)

They better watch out.


Brett House (00:45)

And as you probably got from that comment, Greg's got a really good sense of humor and I've always appreciated that about you. But you spent the seven years basically rebuilding, and I remember when the MMA first launched, I was back in the rich media space at a company called Unicast, company by DG Fast channel, and you guys were just getting started, and then there was like a bankruptcy moment, right? But you've turned this around.


Rio (01:06)

Or close to, was was actual bankruptcy


or close to it? But it's turned around since.


Brett House (01:11)

Yeah.


Greg (01:11)

Yeah,


yeah, we're all turned around. Yeah, we're all turned around. Jeez, before you guys start putting false information into the the the Digisphere here, I should step in. Yeah. No, listen, it was bankrupt and insolvent before I got here. Just to be really clear. Yes. No, I was brought in. Yeah. In fact, they'd had three CEOs in four years. That sounds like a board problem. Yeah. Yeah, that's a board problem. Not yet. Yeah. Yeah. Yeah. No, it's been over. been over well over a dozen years now. Yeah. Yeah. Yeah. Yeah.


Rio (01:14)

Okay.


Brett House (01:14)

It's been a long time!


Yes. Yeah. And it's now it's yeah.


Rio (01:31)

That's not good. Never good. Yeah.


Brett House (01:33)

Yeah, and you've been there for quite a while now. I mean, it's been seven years. Okay.


Greg (01:39)

Yeah, no, but you're right, Brad. It was a mess of a turnaround, you know, I mean, it's


Brett House (01:39)

Yeah. So it's 800 million. Yeah.


Rio (01:39)

Nice.


Greg (01:41)

safe, but same with the IB when I got to the IB where a million dollars revenue and 29 members and three of them hadn't paid their dues. I mean, that was ready to tank too. So I turned around the IB and then left there and then came over here to turn around the MMA. Yep.


Brett House (01:47)

Yeah.


And you guys


have what 800 member companies, something like 60 team members, you're in multi-country.


Greg (01:55)

Yeah, it's like 860


860 corporate members worldwide in 16 countries. And we're about 100 staff people at this point worldwide. Yeah, yeah, yeah, yeah, we become sort of big and legit at some level, I guess, right?


Brett House (02:07)

Yeah, yeah.


Yeah,


Rio (02:10)

Headquarters is in New York,


Brett House (02:10)

no, and I've heard


Rio (02:11)

right Greg? Greg headquarters.


Greg (02:12)

Headquarters in New York. Yeah,


I'm based out of New York and headquarters is here, right? Yes, yes, yes.


Brett House (02:15)

Yeah, and I've


Rio (02:15)

Nice.


Brett House (02:15)

heard nothing, honestly, the last couple of years, nothing but really, really good things. I haven't been able to go to your last events, but you're the events that you've been holding have sort of risen up to a level where I'm like, wow, everybody across agency, brand, vendor are all going, yeah, that was actually a really good substantive event. And I think you guys have prided yourself on sort of being kind of helping to architect the future, right? Talking about things like MTA and measurement and analytics and getting people like firing sharp on the...


on your great debate series. Like it's been thought leadership, but not at a surface level. You guys have gone really deep on a lot of these topics.


Greg (02:47)

Yeah.


Rio (02:51)

why


we ought to have you on this pod, Greg, so.


Greg (02:53)

Yeah,


yeah, well, that's good to hear. let's I'll give for the listener in case there's maybe those who don't. are a nonprofit industry body here to, you know, the fundamental thesis thing is that we think that we can advance marketers ability to create impact. That's the mission statement. But what that really means is that we think there's too many big unanswered questions in marketing and nobody's really leading with a charge to


I guess I put it like raise the stature and gravitas of CMOs. And we think the way to do this is having greater impact on business. Like I'll give you like one small example. You know, and I think this is the one you probably most know what most know us for because you guys worked on this back way back when. ⁓ But it's like, what's the value of brand relative to performance marketing?


Brett House (03:41)

Yeah.


Greg (03:42)

And what struck us, we looked at that issue about six, seven years ago now and said, well, that's a math problem. It's not a cold fusion problem. didn't need new science. You just needed to do the math, but math requires collecting the right amount of data and nobody had attempted to do that. So we created a launch. Nobody had tried to do it right. And you know, we'd all talk about brand. mean, I'm an ex brand guy too. I worked on P and G business for years as an agency guy, but it's like, we'd never answered the big question. So the MMA's whole thing is like,


Rio (03:55)

Certainly on the brand side, no one had tried to do it right.


Greg (04:08)

If we're going to be taking legitimacy, if we're going to be given legitimacy, if we're going to have trust is the big question, then we need to answer these questions in a way that the CFO of the board are going to appreciate. And so that's what we set to do with our Brands Performance Study to say, well, what is the value of brand long-term financially?


Brett House (04:23)

Yeah, no, we did a lot of that work, as you know, with Newstar, right? And we kept proving again and again with like 10, 15 years worth of data for big banks, primarily financial services, which are often bifurcated between brand advertising, that's doing your television advertising, and your direct response marketers that have their own budget, their own channels, direct mail being probably 78 % of that spend. And we kept saying, you've got to split the spend. It's got to be 50-50. You've got to put more money allocated to brand spend versus down funnel, which is where 80 % of the dollars were going.


Greg (04:25)

You did. Yes, totally.


Rio (04:52)

But when times get


tough, people always cut brand and they just invest in performance, right? Yeah.


Brett House (04:54)

Yeah. It was harder for them to prove, right?


Greg (04:58)

Well,


I tell you what we know now. since this is this is the end, listen, you know, there's a guy, there's some guys out of the UK names will come back to me here who did do some work around the long, long and short of it. That was the work that they did for 20 years. But they used kind of a funny data set. It was awards. And then they tried to supplement that. But it was an it was an imperfect data set. We set out to collect the perfect data set. So over 12 months, we tracked media at an individual basis, media exposure.


brand attitude, brand attitude shift, and actual purchase behavior. And nobody had tried to put those together. In fact, when we first came up with the methodology I showed to Google, they said, we love this, we don't think what you're trying to do will work. That was very funny. Yeah, yeah, yeah, yeah, yeah, yeah.


Rio (05:41)

they said that? That's crazy, yeah.


Brett House (05:42)

Yeah, it's sort of like you're


connecting top, mid, and lower funnel behavior.


Greg (05:46)

you're doing that you're trying to look at the whole thing and really what you're really looking for and I don't know are we recording video to the audience have some people might see okay I don't want to make it too hard for those who are still on audio but ⁓ basically what we did is it's kind of simple if you think you get perfect info and you have to collect the math and we've collected data now from almost a million people in the four studies we've done so massive amounts of data collection okay exactly


Brett House (05:52)

We are, we are, if you wanna.


Rio (05:53)

Yeah, we are.


Brett House (06:06)

That's a statistically significant sample.


Greg (06:09)

Well,


and the issue was is that you have to identify the same people here as you do nine to 12 months later. And that was what we defined as long term, I you could go longer term, but that was just further complicates the research. Okay, but then you have to have a big enough group here to begin and that you can still track them later and still collect data. But what we're doing is say if they shifted their brand attitudes here, then how much more did they buy nine to 12 months later or over the next nine to 12 months than the rest of the audiences? It's a perfect methodology. And we now know and by the way, the answer is 5x.


Rio (06:18)

this long term.


Greg (06:38)

So if you spend a, if you, and what are the campaigners now, if you generate a dollar revenue, you'll generate an additional $5 of revenue over the next nine to 12 months, and more probably later.


Rio (06:47)

That's massive ROI,


yeah.


Brett House (06:49)

specifically tied to the halo effect of the brand advertising?


Greg (06:52)

It's all directed to it's the direct effect because those are those are in essence RCTs. mean, you're you're really doing perfect randomized control testing and that so you know exactly what their change of behavior resulted in sales behavior.


Brett House (07:02)

Who's been exposed? Yeah.


Rio (07:03)

Greg,


based on that, remember that Nike case study a couple of years ago, not surprising at all. Remember they put everything into performance. They did it for several years, had immediate like net positive effect, right? Their revenues boosted because they were generating more immediate sales. But then over time, their sales started to tank. People forgot about the new products, right? People didn't even know the new products. All these new upstart brands.


Brett House (07:05)

Yeah.


Yeah, P and P and G did that with


Greg (07:25)

I...


Brett House (07:26)

Facebook, right?


Greg (07:27)

I... I...


Brett House (07:27)

They shifted a whole bunch of their money, like 70 % of their spend to Facebook early on and took it away from television and had a dramatic long-term impact and they had to switch back to their previous mix.


Greg (07:36)

What the thesis I


have and I've had other people articulate to all those not perfect information is that all those direct to consumer companies that were out there, there was such a hot thing here 10 years ago. I mean, many of those have now died and gone away. And I got to know those companies in the course of that. And exactly, well, or well, yeah, well, whatever that's what I don't know what the hell that means. But that was nutty.


Brett House (07:50)

Yeah, or been bought, yeah.


Rio (07:52)

Or got an AI like all birds, right?


I


don't think anyone does, right?


Greg (08:01)

Yeah, I don't


know. Yeah, I don't think they did. But they got a nice bump in the stock price to give him I guess currency to go do some but ⁓ the the issue is that they were so fixed. In fact, I would talk to him and say, Well, why don't you join the MMA? only like $25,000 ago, I'd rather spend the money into getting new customers. And I was like, Wow, that's really short sighted. Like, if you think well, I mean, I guess you maybe you've become a perfect transaction machine. And then many of those companies died because I think they didn't have enough brand. I've heard other people articulate that point of view. So


You almost got validation from that in addition to these other examples.


Brett House (08:32)

Yeah.


Rio (08:32)

Well, Greg,


would like, this would be really fun discussion. We're really stoked to have you on here. And you know, the theme for this, cause you know, this is a theme based or thesis based podcast here was that, and we loved it. This is based on your present, the presentation you did a possible last week in Miami that the Keith. Yep. The key theme here is marketers have more data, more dashboards, more AI than ever. Right. In fact, people are overloaded if anything, but the real issue is, is this actually resulting in like better judgment and better decisions? Your argument in the age of opinion.


Greg (08:45)

Yeah, yeah, I open possible every year. Yep.


Rio (09:02)

is that the ending gets it, it gets us something deeper, right? Than, attribution or analytics maturity. It's really a question about whether marketing can build a body of knowledge strong enough to survive platform noise, vendor incentives. And it's like, is this over, this confidence or maybe even overconfidence, can argue that AI ends up giving all of us, especially marketers. then underneath this, like, is the discipline strong? Is it weak? What is this automation doing for us? So we want to dig into this and,


look at it, like how is this impacting marketing? What does this mean about the age of opinion and what does this mean about the discipline itself at large? So with that, get this, let's go into some questions.


Greg (09:40)

Yeah, yeah. Well, here, let me, let me sort of, you know, I'll tee up the talk there. I'll tee up the, I'll tee up sort of based on what said. I opened, I opened the, I opened that with basically saying, if everybody in the C-suite of any major company is paid based on growth, and they are, and the CMO and marketer is allegedly the person at the table most responsible for growth, then why isn't she or he the most important person in the room? Why doesn't the rest of the C-suite love us?


Brett House (09:42)

Let's lay at the age of the possible thesis.


Rio (09:44)

Yeah, let's hear it.


Brett House (10:05)

Yeah, and they're often overlooked. They're replaced every


Rio (10:07)

If they're in the room.


Brett House (10:08)

two


years. Yeah, if they're in the room.


Greg (10:10)

Well, yeah, that's another issue, right.


And so, right. So that's my end all I think is and I think when I was trying to create there as a sort of a moment of self reflection is that we're really not value. And you know, I've had all these people talking on Dara, Dara Tresider, that opened up an event we did out in in Napa one year. And she said, you know, everybody in the C suite can do a marketing as a side hustle. It's like, you know, we've heard this. And we always kind of bemoan the fact that we


Rio (10:33)

Like devaluing


the discipline itself.


Brett House (10:36)

Yeah,


Greg (10:37)

Well, we value.


Brett House (10:37)

the profession, which is multifaceted, it's increasingly data driven. It's not easy.


Rio (10:40)

It's not easy. Yeah.


Greg (10:41)

Well, it's not, it's


not, but that's not my point, Rio, though, that's not the issue. The issue is that we don't know what the fuck we're doing.


Brett House (10:49)

Wait, who? The marketers or the, yeah.


Greg (10:50)

Marketers marketers marketers don't


know what we're doing. We do not know what we're doing And so you have to start with that moment of self-reflection say well, this is a reason for that. It's because nobody in marketing Was educated to be in marketing and there's no other profession like that


Brett House (11:06)

Yeah, they're the liberal arts majors, English majors, journalism majors, comms majors. Yeah, I always make fun of the comms majors, even though I've dabbled a little bit in the territory. But as those people that didn't know what they wanted to do with the rest of their lives in college.


Greg (11:09)

comms majors, it's a whole bunch of stuff that comes out. I'm not even sure what comms is or means.


Rio (11:20)

think philosophy, no, psychology is that way too.


Greg (11:21)

But the issue is that we're just not trained to be in this business. And what it means to be trained and then made to be a professional, which is really what you're going for, is lawyers, doctors, and accountants, everybody else has, is that you're trained to be in the profession. You understand, in our case, would be the underlying science of how marketing works. None of us have that. None of us are reeducated. None of us are recertified. mean, sure, there's a series of anecdotal information that gets passed along.


I'm not saying that experience is worthless and smart judgment. mean, you the people and you know, the people I work with on the board are incredibly smart people, but we don't understand the online science. So as a result of that, you know, there's a lack of predictability. And as a result of that, there's a lack of trust.


Brett House (12:01)

Yeah, like just, and there's some numbers, I think, to back up exactly what you're saying, right? And I think it comes from your State of the Marketing Analytics Report or Startup Analytics Report, your most recent report, right? Only 30%, this actually jumped out, only 30 % of marketers trust KPIs enough to use them for strategy. That's a problem. If you don't understand the data science, the methodologies underneath them, how this stuff is tracked, how you're going to trust it. So that to me, I think is, you can, you can overcome that. Only 20.


Greg (12:15)

Ouch.


Rio (12:24)

Well, it's not it's not


it's not a KPI then if you can't trust it right.


Brett House (12:28)

I know, seriously.


what do you, it's no wonder there's a lack of trust from the other C-suite members. Only 29 % can trace decisions back to data-supported rationale. 22 % described their analytics capabilities as robust. And Greg, you and I have spent a ton of time doing this stuff, but the genius words at ANA, celebrating analytics in marketing-oriented professions that are driven by data, it's sort of disconcerting to still hear those numbers, right? You think, and,


Greg (12:55)

Yeah, it's disappointing.


Brett House (12:57)

Yeah, and


so what do you think?


Rio (12:58)

Well, then you look at


average CMO turnover, It's like, or CMO tenure. It's like incredibly out of all the C-suite, it's most often turns over.


Greg (13:03)

It's low, right? It's low. It's low.


Brett House (13:07)

So what


is possible? What do they do? What are you counseling and coaching your CMO members? What was your recommendation?


Greg (13:13)

I mean, the big


thing for the MMA is that, you know, and listen, by the way, I don't have a degree, you know, over three decades of marketing, I don't have a degree in marketing either, mine was economics. So just to be fair to everybody, and to be fair to it's an incredibly complex profession is what I think we're not respecting. It's got both hard science of statistics and math and some of the other sort of neuroscience out there.


And then you've got sort of a bunch of soft sciences like sociology and psychology and much of it. So it's pretty complicated. Last year at possible I actually opened with a list of all the things you need to know. And it's it's almost impossible at some level. Okay. But that that said, some movement for us to be stronger than and to answer the big questions, which is what the MMA is all about. How do I and what's the value of personization? And how do you do it? Right? What's the value of brand long term? How do I have greater predictability in my analytics? You know, all the big questions.


how do I build a marketing org that drives sales and raises team net promoter score, which is other things we worked on. So MMA is not doing all the topics, but everything we do, we've been working on for years, even as much as a decade in one of those cases, to try to really help give the facts and science to marketers so that they make better decisions with more reliability.


Brett House (14:24)

You said team net promoter score. mean team, marketing team and how they're, how they're.


Greg (14:27)

So here,


let's play a little game. You ready for this? What do you think is the net promoter score for marketing department? So when you ask people, would you recommend your marketing department? And that's a specific question. What would you say the net promoter score? So now let's go, know, Apple's like a 70, right? know, most people are pretty happy at 30, 40%. Net promoter score is not bad. Utilities are like 10 to 15%. So what do you think the net promoter score is for people when you ask them about their marketing department?


Brett House (14:54)

From other from other teams. Yeah, it's got to a little bit 30 20 %


Rio (14:55)

It's gotta be low, yeah. 25.


Greg (14:57)

Pick a number.


negative two. Yep. So so and that's on average. So the it does go as high it does go as high as plus 50. And then of course, it goes as low as negative 50. So yeah, no, it's a it's a real problem. Now to be careful about that number is that it's not about markets about their marketing department, they could hate their CEO, they might not like their job. There's a bunch of factors that go in there. But it's a real dissatisfaction internally the profession. And I think it's partly because this lack of respect that we haven't gotten.


Brett House (15:01)

Negative two, they're not even on the scale.


Rio (15:05)

That's really bad,


Greg (15:27)

And the whole point of my talk, the whole point of that talk impossible to say, we need to take responsibility for that. You cannot continue to blame the C suite for not understanding you. You can't sort of say, ⁓ my mother doesn't get me. I don't know where you're going with the damn thing. You got to stop that. And we have to own the responsibility to learn the science of this business and get better at it and answer the big questions.


Rio (15:46)

Greg, it can't just


be about pedigree having degrees, right? It has to go much deeper than that. It's okay, sure, there's not, like most marketers don't have a degree, but the thing is it has to be the way they're communicating or not communicating to the other, to the rest of the C-suite, correct?


Greg (15:51)

Let I have the


Hey,


real, to be fair, listen, you my board is some of the biggest C-suite executives in the world, okay? Less than 25 % of them have a degree in marketing. And some of them have really quality MBA, so let's be fair to them, a good general manager, which is very valuable and helpful and why they do well. And like I said earlier, they're incredibly smart. And so part of what this business is, we don't have a codified body of knowledge.


Rio (16:10)

Sure.


And you can't teach taste either. think taste is a big part of it, right?


Greg (16:22)

We don't agree. Doctors don't run around and argue about on everything. I mean, sure, there's things that they debate, but for most health concerns, we've risen a bare minimum. That is what the Hippocratic oath was all about. We don't have it in marketing. We just don't agree. And nobody taught me when I was in the agency world.


Brett House (16:23)

Yeah, that's a good point.


Yeah, well, and yeah,


do question do we need it though and some of the best marketers? Yeah to be a profession you have to have a kind of a body of knowledge, right? But some of the best marketers that I've that I


Greg (16:41)

you got to have a codified body to be a profession. had to have a codified body knowledge because you have to know what to learn. You even talk, Hey, Hey, Brett,


you, talk to professors. They don't know what to teach. If you ask them what to teach in market, they're not clear on that. And they're still doing four P's. That doesn't help anybody.


Brett House (16:51)

Yeah. Yeah. ⁓ and I've taken


those courses, right? My son's going to a business for a business degree. was an MBA with kind of a international management plus entrepreneurship with a little marketing. And I took a bunch of marketing courses and honestly, they were the least helpful. It was the hardcore sort of HBS driven case study courses, the economics, macro and micro economics courses, right? Those were the ones where it was hard skills and I was really learning like the science and strategy of business versus like


Greg (17:05)

interesting.


Yep. Yep.


Brett House (17:20)

you know, brand versus, you know, that kind of stuff. The marketing just wasn't as helpful, I think, from a professional perspective.


Greg (17:26)

Yeah. But Brett, take your experience


from New Store. I you would have worked in I mean, I knew you weren't on the front lines of customer support there. you but you had heard some of your customers talk. Did you listen to what they said? And wow, that's a foundational fundamental strategy that we all agree on. mean, did you say they really understood the dynamics of how to grow that business? I think growing a business always feels like from working like a little bit of a bit of a crapshoot sometimes and a little bit of a ⁓ guesswork.


Brett House (17:33)

Yeah.


Yeah, and then I think that's the point of the age of opinion. And think the best marketers that I saw while there, and I talked and interviewed and met with a whole bunch of them, certainly had ⁓ a rigor.


⁓ around data analytics and they either had teams underneath them where they knew the questions to ask, they knew enough to be the conductor of that orchestra, they weren't necessarily data science driven people themselves, some of them were, but those always seem to be at GM and other brands, those seem to be at JP Morgan the strongest among.


Greg (18:23)

But, but, don't you feel like we,


that they trick themselves with some of that? Because how many people are still using click through as a measure? Let me be frank people. There is no relationship between click through and performance of the business anywhere, anyhow. It does not exist. Knock it off. And by the way, I wrote the clicks. I wrote the technical click standard back in 2004. So like, you know, I mean, we did that to define it. It's the worst mistake I've made is releasing that thing to the world, but like some issues.


Brett House (18:38)

Yeah, yeah.


Rio (18:40)

Probably never has been, right? Any research to back that up.


Brett House (18:47)

with the IAB.


Rio (18:50)

Yeah, but Greg, one thing I think makes marketers


different though from other roles. maybe I imagine this probably impacts the perception of marketers and how they interact with their peers in the C-suite is unlike other roles, a lot of what marketers do is actually done by vendors. that happens, I'm not suggesting that other, the CIO will, CTO will hire Deloitte's extenders to do a lot of their work. But I think for marketers, the fact that they're so dependent on vendors, agencies and other


Greg (19:07)

Yes, I agree.


Brett House (19:08)

⁓


Good point.


Rio (19:19)

to do everything from strategy to creative to media to measurement. I I've worked with some in some industries. You're like, what are these people even do? Are they just vendor managers? You kind of scratch your head. So I do you think that fuels this perception?


Greg (19:30)

Yeah. Hey, I'll put a


little bit differently. So I've had this conversation with probably one of the world's biggest CMOs and his company. We don't find people, we find people within marketing, too often in marketing partners, we see people who are good project managers. But they're not necessarily marketing people. They don't pride themselves on being marketing people. They're not there to study the art, the craft, the science.


of marketing their project managers. And we kind of that's why I give that the last year's talk was for the actually two years ago was we're too busy to do better was the title of that talk.


Brett House (20:06)

Yeah, it's


the kind of activity trap. Hey, we're putting out all this stuff. We've got versions and versions and versions and versions. we're hitting the airwaves with all of our stuff. People are like.


Greg (20:09)

Yes.


Rio (20:16)

But I argue people, they


Greg (20:16)

Yeah.


Rio (20:17)

end up spending their time in the things they like doing. So to your point, Greg, if somebody really is like at heart a PM, right? And then they end up in a marketing role, they're going to, well, guess what? This is going to be a very process oriented marketing org. That's all they're going to give a crap about. So I don't know that, but I imagine for the other people in the C-suite, like that's got to fuel their perception. Okay, what is this CMO doing? If they've got all these agencies, these other vendors doing all these things, what do they do? I mean, if they're not in the weeds actually,


I'm not saying that I CMO is going to be building campaigns, but they should at least know what's going on.


Greg (20:46)

It's not ⁓


that they're not in the weeds. That's not it. I think people are probably in the weeds. The issue is that they don't understand. Let's go after another one. So I say click through is dumb, yet it shows up in probably board presentations all the time. It doesn't have any metric to anything. It's not right. Or the whole thing you got around a couple years ago around duration, weight, and impression. So what a horrible cross-media measurement idea concept that was. And yet the industry believed in the bodies that were behind it believed that was the facts. It was just.


Brett House (21:09)

Yeah. This thing's floating off to the left while I'm reading an article,


right? I'm not paying any more attention to it, ⁓ despite the fact that it's been on the screen for a while, right? That's why that...


Greg (21:21)

Right? yeah, that's how


yeah, well, there's no relationship with time and advertising effectiveness, none whatsoever. We've looked at all that research. So it doesn't exist. So I don't think most people sort of understand that or one of my favorite ones is how many people still believe in broad reach basis strategies. broad reach is a terrible idea. is so unbelievably value destructive. And in a simple understanding the economics of that you just map it out. I happen to do it not hard to do here. But it's like, it's unbelievable how bad that is. And yet


Brett House (21:27)

Yeah.


Yeah.


Yeah.


Greg (21:49)

This is what a lot of markers and CMOs, some CMOs believe is just not right.


Brett House (21:53)

Yeah.


Well, and it's sort of gone from sort of the linear space, right? Where it was a broad reach vehicle. You could target, you know, at the television program level and it's sort of infiltrated the programmatic and digital media space. Yeah. And within what we've talked about with Dr. Fu and a whole bunch of other folks on this pod is that that push for reach, that same mentality has driven a lot of marketers to buy just bad inventory.


Greg (22:04)

The media forced us into broad base. Yeah, at one point. Yeah, originally, yeah.


Brett House (22:21)

fraudulent reach that doesn't really exist. And it's not, you there's a targeting strategy that's broken. They're going, they're paying less in CPM rates to reach as many people as possible. And then they don't understand why they're 0.001 engagement rates.


Greg (22:23)

Yes.


Are you getting into


how flaky or faulty the digital ecosystem might be? yeah, go real.


Rio (22:39)

Yeah, there's a of that. Well, but Greg, before you answer that too, I have a question


Brett House (22:39)

yeah. That's a big topic. Great. Yeah, absolutely.


Rio (22:43)

that piggybacks on what just said. like, I find heads of media, some of them, I don't seem to know, I'm not saying they don't know anything about media, but everything's outsourced. And you wonder like, much have they ever done this? Like, how much do they really know? And the complaint amongst you, people like Dr. Fugles, who will point to all this fraud, right? And then the complaint everyone always has is nothing will change until brands care.


until brands actually say, we're not putting up with this, all this, bad stuff you just pointed out, which we all acknowledge is there, it's 1 % or 50%, like, I don't know, but it's there, right? So is maybe they don't care because like they don't have the right background because they haven't done the work because they don't really understand. mean, what do you think is behind the fact that, cause I'll be honest with you. we, I don't, I think more brands should care and not many do.


Brett House (23:29)

What?


Greg (23:34)

Yeah, okay. So I'd like to suggest Rio, if I could, that that's what the MMA is all about. So I get it that they don't really necessarily have the time to go figure out all these, you know, some of this gets into real minutiae details. Some of it gets very technical, obviously, they might not have the expertise in house. But that's really what the MMA is trying to pick off as many of these things as possible that we can, like what's the value of brand long term? What's the value of getting a marketing org right? That means even that promoter score of teams, you know, which is one of the factors. We actually have figured out now how


Brett House (23:34)

Yeah.


Rio (23:49)

It's hard stuff, yeah.


Greg (24:04)

If you if you design the organization right way, you can actually improve sales like we can now tell you and measure like if you if you if what what what change in the world will improve sales, what would detract from sales. So there's just I think that what it is, it's a lack of knowledge. And I think there's also a big issue of like trust. And so you kind of represent there's a lot of vendors that come at you with a lot of solutions. It's very hard to sort through what some of those are if they're really meeting with your


Brett House (24:24)

Yeah.


Oh, totally Greg. And


they all sound the same and AI has made it worse because a lot of the AI vendors or vendors leverage. Yeah. Yeah. Yeah. Not only the AI tools, but those using AI to promote their tools, you start to get this kind of smoke and mirrors effect. And we've talked about a of it, a lot about this on the pod is that that CMOs or marketers need to have an inherent understanding exactly how this tech and the data underlying this tech works.


Rio (24:30)

Well, 15,000 more tech tools, right? And like another like 5,000 ad tech plus it's crazy. Yeah.


Greg (24:35)

Crazy. Yeah, and.


⁓ They should, they would be well served,


yes.


Brett House (24:55)

Like


that's where AI, if you're building with AI and it's something that I'm doing and O'Rio's doing, it really gets you into, it democratizes engineering to a degree. And five coding isn't technically engineering, but it does allow you to get within the weeds and it gives you just a much deeper inherent knowledge of what's happening beneath the surface. So you can call bullshit on that vendor that's promising A ⁓ and delivering B. You know, that's part of the problem is they don't understand the vendor relationships.


their involvements or the vendor tech and the capabilities, right? And it's all moving very fast. And so then they're relying on these vendors to drive reach, to drive.


Greg (25:31)

Well, I ask you guys


what you think, but you know, my sense sometimes is that we don't really have an orientation in market. It's funny. There's this funny duality that exists, which is, know, do we really believe in marketing or do we really believe in marketing? And my point in summer level is that I think some people don't believe that they have the impact they should. And so they don't orient to doing marketing improvement. Like, what are you doing to boost marketing productivity?


Brett House (25:46)

Yeah.


Greg (25:58)

And by the way, our own data or own numbers around this is just numbers, anybody could do this kind of math, is that if you can improve marketing productivity significantly make every dollar work a little bit harder, that you can change the stock price of a company, there's a seven, there's a point seven coefficient of profit to ⁓ market valuation. So improving market productivity, but I never hear any marketer talking those kinds of terms. I don't hear them say like, what do do that? I think they try to but I don't know that we really have a concerted effort to improve market productivity.


Brett House (26:21)

Hmm, even with AI and the... ⁓


Greg (26:27)

Part of it because we don't have the right metrics because there's a whole other project we're working on.


Brett House (26:30)

Yeah. Yeah. And for


me, it's, it's less, to your point, to that point earlier about activity, you know, and just adding more activity to the mix, which is easy to do. It's more about how do you actually get to better, better IP and better decisioning, right. And create, you know, and to me, the real use of AI is when an institution starts to build basically a brain, it's like a knowledge center, but it's not stored in your file system. It's not stored in your inboxes or your CRM, but it's, but it's a system that's powered by


Greg (26:35)

Yeah, yeah, exactly.


Brett House (26:56)

⁓ agent, agentic AI that stores all of that information and it compounds over time. Your competitive Intel, right? Your media Intel, all that stuff.


Greg (27:03)

Well, and I


think you use the key word there, Brett, too, that we're building knowledge. So sometimes one of the things I advocated for last year is that you need to run a series of experiments that validate thesis and out. It's not just data. Data is misleading because data changes situationally. But once you run a scientific experiment, and that's where I think we're missing is that we just don't have enough science background and know how to develop a hypothesis.


Brett House (27:20)

Yep.


Yeah, that's a good point.


Yeah.


Greg (27:31)

My favorite


one is I talked, I'll name names on this one, is I talked to Linda Lee, who was the CMO of Campbell's and she was on my podcast, right? Can I give a plug building better CMOs? There we go, I just did it. there we go, okay, fair enough. So, otherwise that's gonna get cut out. Okay, I'll see it. I see how this game gets played. But no, I talked to Linda. So Linda's the CMO of Campbell's. Linda's background was as a chemical engineer.


Rio (27:40)

Yeah, go for it. Yeah.


Brett House (27:40)

Yeah,


as long as you have us on your podcast.


In order to shill, need to, you know, we need some payment in return.


Rio (27:49)

Hahaha


Greg (27:58)

And she said she said the most interesting thing she said she goes, I see a lot of markers go, how big is my budget? And how big is the team I get when you give them a campaign to work on something to do? She goes, what we really need to do is say, what's the hypothesis that we have? And how do we validate or invalidate that hypothesis? That's what a scientific engineering name nine does.


Rio (27:58)

Interesting.


Brett House (28:08)

Yeah.


Yep.


Rio (28:18)

like that. I


guess relating to that, there was a, I remember seven or eight years ago, maybe 10 years ago, Gartner had that study about IT being outspent by the CMO. It never really happened. then there was another study a couple of years ago, actually, that was very interesting about the CMO needing to partner with the CFO, specifically to be able to look at the actual cost of things tied to outcomes, to be more financially driven and be a better steward of the budget they're given.


Brett House (28:29)

Yeah.


Rio (28:46)

I think that's a very interesting point and it ties to some of the metrics you mentioned, including, for example, attribution. So I love your thoughts on that, and like what you see going on there and why it's important. And do you agree with that thesis, number one, and like why is it important?


Greg (28:52)

Yeah.


Okay, my favorite board meeting was about three years ago. It was a global board, I'll even mention it. So this is the biggest of the boards that I have, biggest CMOs. And I'll never forget, in the middle of the board meeting, at one point they said, we need to know how to talk to the CFO. Now, I didn't say this out loud at the time, but my inside voice was, what the hell do you mean you don't know how to talk to the CFO? You're the CMO, they were cooperation, like, what does that mean? Like, I couldn't get it. And they raised a good, they raised a very good point. They said, well, we don't,


Rio (29:21)

Were they afraid of them or?


Greg (29:24)

talk in their terms and financial terms. don't, talk likes and clicks and all this other silly stuff, right? Okay. But the issue for us, right? The issue for us is that we need to be able to talk in terms of financials. And so, and they said they wanted, we need to standardize the way we do that. So CFOs believe it. So by the way, the MMA has developed that is a financial formula. We have developed it. We are in the process of now running, ⁓ capturing data to do that. We're finding no marketer has their data ready to be able to do this kind of analysis.


It's basically diminishing cash flow over time. And what it will do is it will tell the CFO that if she or he cuts the budget, marketing budget now, what will be the impact on the business later? We've never answered that question and we need to answer that question.


Brett House (30:04)

Yeah.


Rio (30:06)

So it's a great point. So


being able to say that, OK, you're going to cut my budget by 15%, you might think it's great. here's why. We're going to pay for this later.


Greg (30:12)

Well, here's a good example to


Brett House (30:13)

Yeah,


Greg (30:13)

that.


Brett House (30:13)

and then calculate that over time with the net present value application.


Greg (30:15)

Exactly.


Exactly. Right. Yeah. And listen, we may have to perfect and change the formula time we've it by academics and they said we're on the right path. So we think we're doing doing the right thing. But the fact that none of us have the data ready to do that is kind of funny. Or even that's an orientation from the CFO that marketing should be an investment, I think, is a radical concept to some of them. But we've got to talk their terms.


Brett House (30:36)

Yeah, and CMO needs to ask those questions to be able to dig into that. And I sort of think econometrics, having spent a lot of space at the MMM, media mix model, marketing mix modeling space, is such a critical, why doesn't that type of knowledge, speaking of you, Greg, you've got an economics major and you were a CMO at one point this year, you've been on the commercial side of things. Why doesn't that knowledge inculcate itself into the marketing organization? I think of what Mike Finnerty does at Newstern Mutinex.


Greg (30:39)

Yes.


Yep,


So I I think it comes back to this concept


of a codified body of knowledge that you need to know what is fact and science in fact I joke all the time in here if I can Just show that for those in video. They'll see I mean, this is the only codified body of knowledge that exists I've never once met a marker who read this book. It's called empirical generalization about market impact. It was compiled by Domak Hansen's at UCLA Did they see wait wait a little bit over here, is that right? There we go. It's okay. God is that work?


Brett House (31:03)

Yeah.


Yep. Here, push it over in front of face so that you can see that in video. Yeah, yeah, just go a little bit. Yeah, there we are. Empirical generalization


Rio (31:25)

There we go, yep.


Brett House (31:27)

about marketing impact. Who's that written by?


Greg (31:28)

Yeah. So it


was it was compiled by Dominic Hansen's in order to make it in that book, the study has to have it has to be a singular meta study that looks at 100 or 200 studies that came before on that topic and then crystallizes what's being said. Now, it in academics call it settled knowledge. And I talked about this in the age of opinion impossible, but it's about having settled on the stuff that we agreed to be factors point, it doesn't mean it won't change over time, we can get better and still continue to improve on that. But where we are now, what do we all agree on?


Brett House (31:43)

Yeah.


Greg (31:57)

And then you start from that basis. Like one of the things I pointed out, this is one of my favorite ones. What is the single greatest driver of financial performance for for in broadly market? I'm being very broad here in marketing terms. What do you think it is? Is it advertising? Is it loyalty? Is it a promotions? Is it I don't know. Anybody want to kind of make a guess? I'm by the way, I am setting you up here on this one.


Brett House (32:18)

Yeah, because it's a trick question. now now now like, what is it?


Greg (32:21)

It's a trick question, yeah. You know what it is? You know what is?


It has 67 times the impact that advertising does.


Brett House (32:28)

is this ⁓ distribution, right? Which ties to shopper. Is that shopper marketing kind of rolled into that?


Greg (32:30)

Distribution.


where the product shows up matters so much more. Yeah,


Rio (32:34)

That's interesting. Yeah.


Greg (32:37)

so you should spend all your time on increasing your ACV, getting your advanced distribution before you even worry about sort of other elements of the thing. Or the better...


Rio (32:46)

That probably applies


to B2B too, Greg, as well, because figuring out how you're actually going to get things to market, whether it's products or services, right? Like super important. People overlook that. Like, are you going to sell through channels? Are you going to sell through partners? Very important.


Brett House (32:58)

Yeah, and it comes down to whether you're a retailer or whatever, comes down to, that's your operations class in college. We get body of knowledge. That's your throughput operational efficiency. Can you deliver a good customer experience and deliver it at a speed and scale? Whether you're a dentist and it's about butts in seats, or you're running a restaurant and it's about flipping that table multiple times a night, those are all distribution use cases, right?


Greg (32:58)

Well, and then what's the value of those channels?


Yes.


But these are kind of


what the point here is though that there's series of science that we as marketers don't fundamentally have that we can incorporate into the thesis that we have that doesn't just evolve into opinion based storytelling, which is I think is why we kind of diminish ourselves. Back to my whole point, like listen, we can improve here. We just have to take the time to really put an expectation in learning the business and the science of this in ways that we just haven't. And in the context of it being really complicated, I want to take that off the table.


Brett House (33:33)

Yeah.


and


Yeah, not to sound trite,


but what do you think in, I've heard Michael Cassand said this at an event, at a Brave New Worlds event I had at New Star that I hosted at New Star. I don't think I got you to speak at that one. No, you spoke in the first one, but he basically said, whatever happened, and this is the Mad Men argument, and I'm watching that series belatedly now, but whatever happened to serendipity in advertising and then the trite comment of,


art, science, the combination of the two, which we might say is kind of bullshit. But whatever happened to serendipity in advertising, meaning when you're showing some delightful experience ⁓ for an advertiser to consumers that may not be in market, you might change their opinions. What role do you think that plays? Serendipity means like I wasn't looking for that. I wasn't in your target audience. You weren't reaching out to me specifically as a demographic or a profile. And yet you showed me something that I was


Greg (34:28)

What's what's serendipity mean though? What is that? Doesn't feel like a science.


Rio (34:38)

It's like a random happenstance.


Brett House (34:40)

Yeah,


random happens. I saw Rolex watch. I was in the market for Rolex watch.


Greg (34:41)

Yeah, I mean, I don't know. You're arguing


for what I'm arguing against, by the way. Yeah, so you're... Okay, good. Okay, just checking. Just checking. I want to make sure... No, I mean, that's kind of the issue. And listen, because we did some of this work together around the movable minerals concept, right? So listen, all targeting originally was based on demographics. We kind of touched on that with what reach-based media did. So we tried to improve it. We knew demographics didn't really work very well. I never thought it worked very well from where could tell. didn't seem that...


Brett House (34:47)

Yeah, I know. I'm not arguing for that. I'm asking for your perspective.


Yeah, yeah, they move on, though.


Yep.


Yeah.


Greg (35:09)

scientific, we moved to behavioral targeting the age of digital, we think the MMA now believes the answer around targeting salmutation should be based on neuroscience, because it's what the brain will be responsive to. So here's the thing, the brain makes a decision in three tenths of a second to pay attention to an ad or not. I don't give a shit how good the ad is. I don't give a shit if it makes you cry. I don't care if what you think sort of emotional response.


Rio (35:29)

So you're to decide, watch this or tune


out within three tenths of a second.


Greg (35:31)

The brain will say, I'm not


interested in that product. I'm out, I'm gone. And it makes the unconscious decision to ignore it. It has no imprint on the brain whatsoever. So why are you trying to target people who are not interested in and have no propensity or consideration to buy a Rolex ever? It doesn't make any sense. That's not good marketing. A concentration of those who have a potential to consider the brand, we think financially, which is this is the whole argument against respace planning.


Brett House (35:41)

Yeah.


Yeah.


Greg (35:59)

it economic makes it if you look at those numbers, I don't think you can deny it whatsoever. And it's all predicated on the concept, but it's back it's backed by neuroscience. That's the issue there that separates it. Now listen, to be fair, as you know, we're still validating the thesis under what conditions that true under what condition may not be true. I mean, there's still a lot more work to be done. But we think we're on the right path with that one.


Brett House (36:03)

Yeah.


Yeah.


Yeah.


Yeah, and I think the predictive capacity of AI, the compute ability of AI to be able to bring in a huge amount of for predictive analytics to understand that behavioral, contextual, all this stuff that it's absorbing, it's going to be able to more accurately predict reaction.


Greg (36:25)

Well, but okay, okay for analysis, okay.


Okay. Hey, have you ever, have you ever gone into an


LLM and typed like, Hey, what's the best marketing? Because by the way, I've done that and it gives you crap. And now it might, it might sound, it might sound good and it might reinforce, but all AI, all in LLM. I want to separate AI processing, which is where you went to, you know, the actual sort of LLMs that were all business that are so popular and helpful to us today. And listen, I'm a huge fan of AI. have another podcast on AI by the way, but


Brett House (36:44)

Yeah, I'm going to do that right now.


Rio (36:44)

Different crap each time too.


Brett House (36:54)

Yeah.


Greg (37:03)

Though that's all really valuable. But the issue is, is that if you go into an LLM and ask it for what's good, it's just going to repeat the opinions. You have to get back to knowing the science.


Brett House (37:11)

Yeah. Well, it repeats, it's been trained by


me, but yeah, it says the best marketing ⁓ is clarity plus proof in front of the right buyer at the right moment. Not the loudest campaign, not the clearest tagline or cleverest tagline, not the biggest content machine in the business. Yeah. So, so, so, yeah. So, so conclusion, well, the age of Mad Men and advertising is sort of the art is in an opinion.


Rio (37:24)

Well, there you go, Brett. There's our marketing strategy for the podcast right there. Right.


Greg (37:27)

There we go. You guys are


winners today. Great. Well done.


Rio (37:31)

So Greg, like...


Brett House (37:39)

That's kind of what that was based on. It's really over.


Greg (37:40)

But I


want to pick on that too. I keep interrupting. I want to pick on that too. listen, know, Rex and I co-founded, you know, Rex Briggs and I co-founded Multitouch Attribution many years ago. We were trying to figure out how to optimize media. That was the whole point of it. But here's what's interesting. It made us go in and look at individual ads to see how well they performed. In 47 % of the campaigns we looked at, of the 24, two dozen intense measurement campaigns we did, 47 of them failed on motivation or message.


Brett House (37:44)

Yeah.


Yep.


Yeah.


Greg (38:07)

In other words, the creative sucked, it didn't work. They didn't hit a motivation matter, they didn't get it. So I love to talk about the art and listen, I love the art and I've seen it be incredibly powerful when done right. But we don't even understand the science of how to create messaging. We have to be consistent about this. So we wonder why we're picked on and why we don't get it. And by the way, we think that answer also sits over in potentially neuroscience. We've not done as much work there on that. We'd like to explore that more. And that's a that's another very complicated.


Brett House (38:09)

Yeah.


Rio (38:10)

on half of them.


Brett House (38:20)

Yeah. Yeah. Yeah.


Rio (38:31)

And it.


Greg (38:35)

multidisciplinary area that we'd have to kind of figure out as an industry.


Rio (38:37)

I think


it's a good thing there, but it's important to test and have your tests be well thought and planned out and continue to test. Marketing is a science, but it is something you need to... I mean, it can be. I think a lot of people expect, we'll do this and it'll work or it doesn't work and we'd prove it. I mean, you're learning and you're proving and you might be disproving certain tactics, but that doesn't mean that like...


Greg (38:46)

There can be signs, yeah.


Rio (39:03)

This channel doesn't work, for example, just means maybe your creative socks. So, like, great. Let's talk about turn to thinking fast and slow. Like one of the best lines in the report is the need to think fast and slow closing the gap between data that is really immediately accessible and like provides robust ROI. Like, what do mean by this and how can marketers like, like, what does this split mean mean to them? Could you break this down a little bit for our listeners?


Greg (39:06)

Right. Yeah.


⁓ yeah.


Yeah, you know, I tell you what, that was actually compiled by Denise Champion on my team who's done phenomenal work around that. I don't have at my fingertips that information. So I know what you're referring to. It was actually I thought it was very good. I've been a long time so I've been exposed to it. So don't I don't think I'm your best answer around that one right now. But just that but the fast and slow is it's based on a pretty famous book that's out there now. And I can funny I referenced it last year when I


Brett House (39:42)

Yeah.


Greg (39:52)

possible and I'm just not remembering the name right now. System one thinking, right? And that system one. Yeah. Yeah, I don't know. I'll need somebody else to kind of walk me through that. don't I don't have it on my fingertips.


Rio (40:01)

You know, you get it just later. We can put it in the show notes. Not a big deal.


Brett House (40:01)

Yeah ⁓


Yeah, and so back to that question around where do you think MMM and MTA is going at this point? ⁓


Greg (40:10)

Well, but you know,


we're a little biased. We're very MTA focused and we think MMA is complicated and subpar. We just don't. Yeah, we don't think MMA is good. I think it gives you a night, maybe an okay and a long term view. But unless you get down to an identity graph into an individual identity level, we don't think you're really improving marketing. And we see that come up again and again. In fact, this financial formula that I mentioned earlier, it's currently called Project Romi will come up with a real name for this product or thinking here soon. But


Brett House (40:17)

MMM. Yep.


Greg (40:39)

that underlying, you have to operate the individual level. I'm just not a big fan of MMM and what it's really trying to do. That said, MTA is really damn hard. It's difficult to do. And we actually just did an analysis here a little while ago. It said, well, an MTA, what happens if you lose a singular channel data signal? What does that do to your analysis? It's not great. So we have a lot of work to do there, but we think that's where the answer lies.


Brett House (41:02)

Yeah, well, as long as it, you we worked in a lot of hybrid models, right? And you knew this at Newstar where you're bringing in exogenous factors that are not channel and not media related, that are driving consumer propensity, consumer outcomes.


Greg (41:06)

Yes, I know there's some of that. Yeah, yeah.


But are they improving marketing


productivity? At the end of the day, am I getting more for the dollar spent than otherwise? And I don't know that of every case I am around MMM because it just doesn't go to the fine level that you need. And yet, know, MTA's got its own issues.


Brett House (41:27)

Yeah, yeah, despite the claims of hypergranular MMM.


Rio (41:30)

But MMM is going


through, think the most changed I've seen in a long time, right? I mean, I'm seeing people apply AI to admit both for data ingestion, as well as like Brett, your point earlier, the analysis of that data and then turning it into more actionable insights. Because MMM used to be this thing, you do it once a year, you give it to, know, the CPG brand does it with their vendor, they give it to their agency, a year later, revisit it. like, not super effective in my opinion. Totally.


Greg (41:33)

I agree.


Brett House (41:52)

And it was very people driven too, right? Like


Mike Finnerty, we interviewed at Possible, is now the US president for Mutinex, which sounds like a cough syrup. ⁓ And what they're doing is he's basically replacing what was a human driven model, managed services, relatively low margin ⁓ of 200 people. He's an analytic partner, sort of does the same thing, right? They're very people driven. ⁓ And he's replacing those people with agents that do all, and each one of them have skills, aligning to a lot of data unification.


data, identity, resolution, and then pulling in every basically input you could possibly get from seasonality to every econometrics input that you could get into your broad model to be more predictive of what really is driving media impact. And to me,


Greg (42:39)

⁓


I love that as an orientation, absolutely. But again, going back to some of the core here I said earlier is that it's nice to have a good media optimization model, but if the ads don't work, then I don't think you've really accomplished the goal.


Brett House (42:50)

Yeah, yeah. And what you said before is it's basically a flip of a coin with that 47%, right? Like 50 % of creative, you just don't know.


Greg (42:57)

I mean, kind of feels like it.


Rio (42:59)

But that's just bad creative


messaging, the 47%. That's what that's referring to, correct? Just your message sucked, your ad sucked.


Greg (43:04)

Yes. Yeah, yeah, yeah. I mean, listen, we


said it was the book that Rex and I done way back when called What Sticks, which is basically said, there's three things you got to get right, you got to get a motivation that consumers care about, you got to get a message that they understand, and then you got to create, you know, media at right place, right time. And so the problem was that they were failing on the first two, which made the three invalid, third invalid, you spend a spend a dollar a dollar meeting and said, it's a waste of time. So that kind of fail rates got to stop.


Rio (43:26)

But does AI give us


the ability to create, like variants of creative, the creative was always a big logjam, right? Like that was always a problem because we talked about personalization. I've done, I'd love to see some of your studies. I know you put out some good ones about effective personalization and that it really works, right? When employed correctly, it does make a big difference in terms of like return on marketing investment. But the problem in creative was always


Greg (43:42)

We've done a lot of studies on personization. Yes.


Rio (43:55)

The creative, right? Like, okay, you want 50 very easy little gifts, but someone has to make those 50 creatives, get them approved by legal and compliance, especially in a regulated industry, then get them queued up in your, in your different platforms or activation. Really not easy, even given some marketing automation tools, does AI maybe change that and finally make it possible?


Greg (44:12)

Yeah, it certainly makes the production and the iteration of ads possible. Is that where you're going real? That's what you're kind of fixing on. Yeah, I totally agree. And I think it's becoming more possible. But even sort of making the decisions in high speed as is necessary in ad serving to get that to the right people and build cohorts around that, which is basically what our work did. We use machine learning, a couple of special techniques there in order to be able to find cohorts, then we could serve the right ad and see a boost in performance.


Rio (44:37)

For DCL, basically.


Greg (44:39)

It was DC, I'd


call it DCO on steroids because we didn't actually do, we didn't actually do generative AI for the ads and the fly. Although we could just, you know, no, no big brand was going to let us do that yet.


Rio (44:49)

Yeah, but I do think at some point, AI is getting better, brands are getting more comfortable using it, and tools are getting better that leverage AI.


Greg (44:51)

Yes.


Yeah, and Google's


got a solution and Metta's got a solution. So yeah, they're all behind that. Yeah, mean, it's part of the ads. But you don't have to be a rocket scientist to go figure out that probably personalization of ads makes a difference.


Brett House (45:00)

Yeah.


Rio (45:08)

Yeah, well, people will argue sometimes it doesn't work well, which I always thought was crazy, but fewer and fewer, I think.


Brett House (45:12)

Yeah.


Greg (45:13)

With DCO,


think DCO kind of damaged the space in some regards, it made it feel kind of stupid and dumb and not right. yet, you know, we've had to work with a lot of we've worked a lot of big CMOs on doing some of the lab experiments that we do around personalization. And it's hard to figure out it's hard to sort of, how do you hold state? Here's what it is. The CMO put it really well on a board made she said, Listen, she goes, we spent so much time convincing consumer or convincing marketers that it's a singular insight with a singular message to now ask them to do different.


And so how do you hold brand but then still personalize and get maybe a slightly different motivations or get a different messaging that sort of appeals? I think it's just an area we need to spend a lot more time. I do think the answer resides there though. And in fact, I asked an, no, go ahead. Yeah, go for it.


Brett House (45:53)

Yeah, well, but it's all


powered by good data governance and good data quality, because what is personalization if you've got underlying data that is completely missing the mark, right? Which is what a lot of brands are doing in terms of, you know, like we worked with GM, you know that at Newstar, they had 60 different data warehouses. So half the job of marketing was bringing this all together, unifying it, deduplicating it.


Greg (46:05)

It should make it better.


Yes.


Brett House (46:17)

⁓ resolving it down to the household and identity level and until you really, in device level, and until you do that, you're really not able to power a personalization engine of worth its weight in salt, right?


Greg (46:28)

Well,


I don't know. mean, the personalization that we did is we use just contextual signals. So we just looked at IP addresses. We looked at everything and then it would pick what would sort of kind of align people to a common interest. we saw perform. mean, here, you mentioned as I'll give the numbers to the audience is that we saw on average gains of plus one hundred and sixty percent. Now, a couple of those were studies in the US. They were all focused on digital metrics. So did they visit maybe these five pages that they go and fill out a lead form? Did they do something? It was a digital kind of dynamic. Yeah.


Rio (46:55)

This is all for paid media, correct?


Greg (46:57)

We've also done yeah, all paid media. We've also done studies now against can we improve brand effect with that or can we improve store traffic driving at the retail level and we've seen that the gains aren't as big because there you're you know, if you're going to go from brand awareness of 46 % to 60 % you know, it's only a 50 % gain. It's not as big as 106%. But when you see digital metrics, the gains were unbelievable to me. Nothing like nothing. I've never seen that those kind of gains ever and anything I've done. So I'm huge believer, I'm amazed huge believers.


Brett House (46:57)

Yeah.


Greg (47:23)

based on facts and evidence in our experiments of AI-driven personalization.


Brett House (47:28)

Yeah, so but you were doing that with a discrete finite data set, right? Whereas brands, what was the data set that...


Greg (47:33)

No, no, no, no, no, no, no, no, no, no, no, didn't use any none of that


use cookies. None of that was using cookies, we were just using signal.


Brett House (47:39)

Yeah, well, I'm not saying cookies. just saying what was


the underlying data set? And if a brand has fragmented data and is trying to train the AI on that.


Greg (47:45)

I, I, I,


I, yeah, no, that wasn't no, no, we were no, see, you're kind of back to the LLM kind of example, we were using machine learning. So it was one hot encoding with KModes clustering with the two techniques that were used in this in this particular approach. And that was able to build cohorts of people. And it looked at probably 1000 different signals time a day, it would look at location, it would look at you know, every everything you could pick up just from the from the from the data stream, whatever was available attached to that impression.


Brett House (48:10)

Yeah.


Greg (48:13)

And then it would build a cohort says, these people like this, this is the ad they like, and then we'll continue to reevaluate and test that. And it would just incrementalize itself really fast to high performance. It'd take like a couple of weeks and you'd be at the best performance you could get.


Brett House (48:16)

Yeah.


Rio (48:23)

Yeah. Well,


pay me a Greg. think one thing like brands have to marketers have to be aware of too is like, even when like, let's say you have, have the different types of creative, right? Like even understanding, okay, when you're targeting audiences, some audiences might be better defined than others, right? Some, you know, like, so you may think you're targeting like millennial restaurant restaurant lovers, but


Greg (48:39)

But see, didn't consider but


we weren't considering all that all that requires first either some variation of first part or third party data. We didn't do that. We just use signal that was available from from sort of what was out there that could be just done and it would find its own thing. And I will tell you, though, it did figure out some things like the one that was the funniest one was that we had a we had audio as a male female voice and the female voice performed 20 % better. So if you were regular, you just run the female voice move on. But we ran the male voice to the male voice, though, always performed better in southern states in Utah.


Rio (49:02)

Interesting.


Greg (49:09)

funny, patriarchal societies. And it always performed better everywhere in the country after 10 o'clock at night. So if


Brett House (49:10)

Yeah. Yeah.


Rio (49:17)

It is funny the


different way people act, right? I remember like, what are my first jobs in marketing from we used to do market research and we stuff. This is back when he used to call people, right? So I remember like the difference in experience, like calling up random people to ask them questions when you're calling, let's say in Alabama, where they talk to you all day, but maybe not answer the question. Whereas New York, they just tell you to f off and hang up. It was wild, right?


Brett House (49:35)

Yeah,


Greg (49:36)

you must have gotten me on the phone. ⁓ took a minute. It took a little listen. took a moment to kind of hear what they what I said there. Yeah, I was the one who told you to fuck off and you called.


Brett House (49:37)

yeah, yeah, so it's interesting and what your point Greg says to me is that is that you did you did a study?


Rio (49:39)

Yeah.


You're the one who up already.


Brett House (49:48)

Yeah. Yeah. what like what you're describing from that experiment,


Greg, it sounds to me like it's it's what I personalization, I think person's level, I think one to one, which has been a promise. Yeah. And you were actually able to drive much better performance with more generalized cohort, which is aggregate for specific cohorts of specific type, which which shows you just do that right.


Greg (49:58)

⁓ I totally agree with you. Yeah, exactly. Yeah.


Correct, correct, very good point, very good point.


Yes.


Brett House (50:13)

And then


you could potentially get to this promise of we're going to more effectively target you, not just a retargeting play, but based on your propensities.


Greg (50:18)

Correct. It's


100, well put, thank you. Better put than I said. exactly. Yeah, yeah, yeah. And by the way, I think what you just did is you just advanced the science that we would love people start to test and really work against. Like, let's bring in data on the customer. Let's see how we let the AI sort of figure out what's best. And it really was, I mean, it was most interesting to us about the experiment and all that we've done. Probably we've done 30 in total around the world at this point ⁓ is that it really does feel like magic.


The AI is making decisions and this is kind of back to part of your guys original point like it's making decisions we don't even see what's going on. We can tell we can tell it's sort of like you can't see the wind happen but you can measure what happened in its effect. So we can measure what happened but we don't know why it made the decision originally. Very cool.


Brett House (51:01)

And


the closer we can get to the understanding of the why, the better off as marketers we would be, right? Because we understand the science, we understand the data. Yeah. Experimentation is a huge. Yeah. So it's a call for more science and marketing. It's a call for potentially, I mean, you're saying build an academic body of researcher or understanding, and there's the Byron Sharps and those out there. But to me, it's more of people coming from different disciplines.


Greg (51:07)

maybe or just set up the right experiment and the right metric to measure against. That might get you there too.


Rio (51:11)

See if it works, right? If it works, just keep doing it,


Greg (51:27)

If that's right.


Brett House (51:30)

Right? Not coming from the creative liberal arts necessarily, English major, journalist background necessarily. It could be people coming from science backgrounds.


Greg (51:37)

Well, I'd like you to apply


science to that creative side and message development and message personalization to McConnell Appointment Mechan. I'd you to apply science over to the analytic side. And no, it's not in reach-based planning.


Brett House (51:41)

Yeah.


Rio (51:49)

Well, Greg, what about the argument


though that, I mean, like large language models and AI can do a lot of like mathematics and science and things that it opens up a lot of capabilities that, you know, someone didn't have. Like you can ask it to do very sophisticated things right now. So, I mean, I hear you like the need for marketers to be better informed.


Greg (51:58)

Agreed.


But when it comes to the


science, it's only using the collective opinion that it knows about now. AI is not expertise. It is not judgment. It just comes up with the next evolution of what it believes to be true based on what it was trained on. And the problem was trained on bad thesis and theories. Yes. Well, it's not even what might come in the future. If you kind of go back, those who hold the opinions who get


Brett House (52:14)

Yeah.


Yeah, it's predictive of what it knows, not of what might come in the future, right? Because it can't, it doesn't have the gift of foresight.


Rio (52:19)

Because of that average. ⁓


Greg (52:30)

ranked and rated in those systems or that you're trained off of don't have the underlying science. That's the problem. And so it would be nice if a nonprofit association might put together a good codified body of knowledge that might be helpful to everybody coming soon. Yes.


Brett House (52:34)

Yeah.


Yeah. Yeah, it's the reddits, the Wikipedia's, it's all that.


Rio (52:42)

Imagine that.


Brett House (52:46)

Yeah, yeah. Well, yeah,


but that's why if you build that institutional knowledge within a brand, right, that is collecting and amassing all of the information, every media engagement, every consumer, your lifetime value, everything that you know about every person you've ever touched or interacted with, and not to mention your internal employee base and executives, that's your IP. That's what gives a brand advantage in terms of how do we better represent our branded products to the consumers we're trying to reach, right?


Greg (52:52)

Yes. Yes. Yes.


I think that'll require a whole other episode to dissect.


Rio (53:18)

Yeah, and there's a whole argument about like who's going to own the agents. mean, in this world, the brands certainly would, right? Maybe using that body of knowledge, hive mind order to create them and use them.


Brett House (53:18)

Hahaha!


Greg (53:26)

I'm not so sure.


Yeah, I'm not so sure all that's been kind of completely resolved. Like, yeah, yeah, no, it hasn't been resolved. I this was last board meeting's discussion, by the way, on that very topic.


Brett House (53:30)

No, it hasn't been resolved, that's the push. There's a lot of people screaming. Yeah, like your marketing team doing


Rio (53:31)

It has not been resolved.


Brett House (53:37)

prompts is not AI. AI has to be institutional IP that's codified, that spans time, and it spans your executive team and all of your teams. And it collects and amasses and makes decisions based on all of this information.


Greg (53:53)

Yes,


that is the utopian view. And yes, I hope we will get there and the MMA is going to do what it to support that. Yes.


Rio (54:01)

Yep. But a lot


Brett House (54:02)

Yeah.


Rio (54:02)

of AI work that these consulting firms are doing now, funny. lot of it, Brett, is what you just described. Like literally teaching people had like open up the prompt and they use chat GPT. Like a lot of these big, like essential contracts to teach the government. Unfortunately, that's what I think we're getting. We're getting past that. We're actually now getting solutions built on top of AI, even if they're just LLM wrappers. But ⁓ I thought that is funny.


Brett House (54:24)

But we know you got a hard stop, Greg. I think we should do our quick hits and wrap this up, but we can always have you on another time. ⁓ So you start real.


Greg (54:26)

Okay, hit me with the quick hits.


Rio (54:27)

Quick kids? Time for quick kids.


All right.


All right, so I think I know what you're going to say, but I'd love to hear this. Like one marketing... Well, you're probably going to say bullshit, but one marketing metric the industry still values that they shouldn't or they should.


Greg (54:37)

It's bullshit. No, I'm sorry. I'm sorry. I went too fast.


Yeah, cl-


Yeah, no


brand, know the value of brand, but stop and click through. It's it's stupid, stupid, stupid. Yeah. I mean, all the research supports, even Google doesn't believe in last click attribution. They've come out and technically said that. I can't believe we're still talking about it. It's lazy. It's lazy. That one really irritates me. Yeah. Yeah.


Rio (54:53)

You're calling bullshit,


Brett House (54:55)

Yeah, last click attribution.


I can't believe we're still talking about this. It's ridiculous. mean, anybody that... It is lazy. It's intellectually lazy.


Rio (55:07)

It's inertia


too.


Brett House (55:08)

One idea marketers still believe in that the data doesn't support. mean, not clicked through. One idea marketers still believe in that the data just doesn't support.


Greg (55:13)

Wait, wait, wait, say again, one idea that...


Brett House (55:21)

That's a tough one.


Greg (55:22)

One idea that marketers still believe in, I mean, listen, I'm more concerned with trying to answer the big questions, what's the value of brand? How does Oregon impact what they do? Here's my favorite question is that I had a CMO call me, in fact, I'll name names, Frank Cross and Best Buy call me one time. He goes, Greg, said, goes, there a standard for quality marketing measurement? And the fact that that didn't exist until then, we wrote it six months later, was insane to me.


Brett House (55:52)

Yeah, Well, Greg, think you did. Yeah, yeah. And I think I'm going to answer the question for you. think reach, right? That's an idea people think raw reach versus the kind of movable middles concept.


Greg (55:52)

Insane. And by the way, the answer was incrementality on that, but go ahead.


reaches


so value destructive. I think it's the reason CMOs get fired.


Brett House (56:08)

Yeah, because they're spending so much money and they're not able to.


Rio (56:09)

Interesting.


Greg (56:10)

They spend so much


money in non-profitable customers, it's insane. Moveable Middles is what validated all that, Brad. You know that work. I mean, that was crazy. Yeah.


Brett House (56:17)

Yeah, yeah, non-profitable customers, that's an important point.


It's not just your lack of customers, it's non-profitable customers that would have bought your product anyways. Yeah.


Greg (56:23)

in non profitable customers. Yeah. Yeah. Or those


who would have bought anyway, you shouldn't be advertising those that tends to be really small, you know, but there's but you could have as many as 75 % of the population who will never buy your brand no matter what you tell them. They were just never going to do it. Stop spending money there. It's value destructive. It's wasted money. And they don't come in new the franchise as somebody likes to say, it's just not not not enough economic to make sense.


Rio (56:37)

So don't market to them.


It's funny you mentioned incrementality. So that was like the buzzword of marketing last year, mostly because of retail media, right? Like that's why it was popular. And all the platforms included these incrementality ⁓ dashboards in their platforms. I it was very interesting. OK, next one. Harder problem today. Trust in data or trust in marketing.


Greg (57:08)

Well, see, already I mean, another one of those are good. Trust in science. I want to change the question. Yeah, you got to trust science because science is a thesis and a hypothesis, test and evaluation that will last us down to time. So as the situation changes, that's a problem of data. Data is situational. And so you got to have science. Yes.


Brett House (57:13)

Trust in science.


Yeah, the end of the age of opinion is really the age of science. It's got to be, and it's really a call for almost re-skilling, right? And building knowledge around that as a CMO from an academic perspective.


Greg (57:28)

Yes, absolutely.


Rio (57:35)

Well, plus numbers, they can tell whatever you


want. What does that mark to end quote? Their lives, their damn lives and their statistics or something like that.


Greg (57:41)

Exactly,


Brett House (57:42)

Yeah. Well,


hey, Greg, thank you so much for joining us. I think we're going to have another conversation because this is fascinating. Yeah.


Greg (57:45)

Yeah, you guys are so much fun. knew this would be good. And I


knew you guys would be a little argumentative, which I really like.


Brett House (57:52)

It's part of the show and I think people appreciate the depth. We like to go deep and really uncover and interrogate a lot of these ideas.


Greg (57:55)

Yeah, it's like the great debates


that we did together. That stuff makes such a difference, I think. You got to hear a coin counterpoint.


Brett House (58:00)

Yeah, it's just so much smoke and mirrors out there,


so much noise, and that's why we call it Signal & Noise. Although there's a few other brands called Signal & Noise that we need to work around, but that's a brand marketing problem. Yeah, like in terms of people competing for the same name, and I'm like, well, we'll stand out eventually. ⁓ But hey, thanks everybody for joining the Mesa Spark. Greg Stewart, tell people how they can reach ya, and reach the MMA.


Greg (58:09)

Hmm. Hmm. Hmm. Okay.


Rio (58:10)

Yeah, but better put.


Greg (58:16)

Great, okay.


Rio (58:18)

It's great having you, Greg.


Greg (58:22)


if you want. I'm easy to find.


Brett House (58:26)

Yep. And for everybody listening, go to our website, www.siglennoise.ai and you can find us on YouTube, Apple Podcasts, and Spotify. And we will see you next time. Thanks everybody. Thanks, Greg.


Rio (58:36)

Thank you.


Greg (58:38)

Thank you.



Comments

Rated 0 out of 5 stars.
No ratings yet
Add a rating
bottom of page