S4 Ep2: The Reach Delusion: How Marketers Are Wasting Millions & What You Need To Know

If you look at the paper, it doesn't sight barns right into years, none of the reviewers or professors analyzing this work and try to find flaw in it, asked how this relates, otherwise we would have to sight his work. He's never been relevant in this discussion. Welcome to that's what I call marking the podcast where you'll hear from the leading lights in the marking world and listen to their unique insights. Thanks so much for joining us again. If you're enjoying episodes of that's what I call marketing and you give us a rating or a review if we is incredibly helpful. So we love those five stars and of course if you subscribe wherever you're listening or watching you'll never miss an episode again. So what a great way to stay in touch with that's what I call marketing. Speaking of staying in touch with this, if you would like to partner with that's what I call marketing through our sponsorship opportunities, do get in touch. You can find out all about those on that's what I call marketing.com forward slash sponsor. So I'm to today's episode and today's guest came to my attention when I read an article on one of my favorite places, MI3 Australia and I was a tension grabbing headline that said he was coming out with research that says Byron Sharp and Aaron Berg Bass's marketing signs rules no longer hold and he is thousands of campaigns and million customer journeys as evidence with peer reviewed paper coming out very soon. So onto today's guest, it is professor Philly Baythalmus and he's associate professor of marketing at the University of Oxford side business school and deputy director of Oxford future of marketing initiative. He's recognized as the first marketing scholar to publish peer review studies on criminal marketplaces including the dark web. He's a graduate of the University of Florida animal sciences, University of Pittsburgh, MBA and PhD in marketing, he teaches marketing management and AI in marketing while contributing to innovative intellectual property initiatives at Oxford. He's bringing out a paper and it explores the effects of combining multiple advertising media channels and brand performance. I said it leverages huge numbers of data sets across multiple countries and identifies common media archetypes and evaluates their effectiveness on brand performance metrics. There's so much to talk about in today's episode. First professor Thomas's own peer reviewed work that is coming out very, very soon but also the implications it has on some of the things that we've all been reading and using for a number of years and I'm really interested to understand why Professor Thomas felt this was an area that needed a deeper examination and what his view really is on Byron Sharp and Aaron Bergbasso stay tuned for all that in today's episode. If you would like to reach and engage community of marketing leaders, get in touch with that's what I call marketing to discuss sponsorship opportunities. So much for joining me on that so I call marketing really delighted to have you here and kicking off the new year with brilliant episodes with you. We've all got to talk about it. But first of all, people who may not know you yet, you're a Brazilian living in Oxford and you've many degrees, you're working lots of different areas, you want to gain out of the science. Lots of experience in marketing. Can you just tell me about that path to get you where you are and that is Professor? Yeah, bit of insight. So first of all, thanks for having me, really appreciate the invitation and the chance to chat. So yeah, hopefully like everybody else, I have a very windy, any planned path to where I stand. I didn't even know that being a professor was a career choice, never occurred to me that people chose that, I'm a Brazilian born and moved to the US, as soon as about 13 with my family, a very classic Brazilian move, like didn't go very far, we went to Florida. A Brazilian economic collapse, my family, my dad got a job there and he was a marketer. So I always kind of grew up around, he used to tell me, you know, go into finance, don't go into marketing. That's who makes the choices, clearly it didn't listen. And I also wanted to be a doctor and yeah, so that's where like the hard sciences, Indian biochem, ended up turning into a degree in animal sciences, went into and like a side of econ and stuff, I was very lucky with my lack of planning and some, the brings that I was exposed early on, actually just a quick aside, like even in my undergrad in marketing at the University of Florida, a fantastic marketing hub for academics. At the time, I wouldn't have known, I was one of three thousand of his students or something at the time, but one of the editors in chief of the journal of consumer research was my professor for marketing 101. So I will remove some of my blame of ignoring advice and doing marketing by being influenced by these people. So yeah, went into industry, I worked in development, product development, food development, got my masters, got recruited into a PhD out of the masters because I think my head, Laura, I asked too many questions, you're clearly not made for that world. Once you get out of there and you come here, it's been that ever since I graduated from the University of Pittsburgh, which was a phenomenal experience, got my first job as a professor University of South Carolina and then got poached by University of Oxford from there and been here ever since. Amazing. And fast, one of the things that you look at, because we're going to talk about the picture that's out right now, but you look at criminal market cases, yeah, so that was early on in from 2013 onwards, and in part, that's why I got oxidized when I talked to the Dean's like that sort of research should be done here and do you work with the government? The simplest way to put it is we spend a lot of time talking about growth, like how do we make things work better? How do I remove friction? How do I get products in people's hands? Not every market is like that that you want to grow. If you think about unhealthy demand for bad products, human trafficking, you don't want to grow those markets, you want to do the opposite, right? So how do I create friction instead? How do I disrupt dissemination? How do I create mistrust? How do I change the information flow and the structure in the market as an outside agent so that it collapses on its own? And that was the work. It's literally all the stuff that we know and we do tweaked backwards. So yeah, it turns out like I've been able to publish a few things. There's a lot of things that can't publish because as you can imagine, as soon as you say, "Hey, this is how it works," they patch it and it doesn't work anymore. So there's things that go strange to governments, reach to various militaries to disrupt, but I think I've been able to help with cyber weapon markets closing shortly after some publications. The largest cocaine hub in Europe was closed. That's amazing. And that's why you move haste every seven days. That's why now I have a big beard tomorrow. I'm going to be skinny and no beard and one with a head on. I just got to keep moving. You'll be fine. You'll be fine. You'll be fine. I'm stuck in your website. And I love this line there, but it's as truly you can do no greater favor to your competitor as they underestimate the complexity and power of marketing. It's such a great articulation of our ability to do things in a really great way, but then also sometimes our inability to delve deep into marketing and understand the power of and understand the history of and all of those different elements. I love that. Prior to that, come from, I guess, probably through all your work. No, thank you for that. Yeah. It's trying to find a way to describe a pretty deep love for this resilience and also this attachment to complexity, which is a perpetual fight. You say complexity and every C-switch, I mean, eyes roll back. Nobody wants to hear that stuff is messy and complex. Reality of the thing is marketing is a relative competitive game. I am only going to achieve if the two of us are competing. I'm only going to achieve that which you allow me to. And that's a beautiful game, right? Like if you have a slightly competitive mindset, there is an aspect of I win you lose. It's from an earliest part of how I grew up in marketing in a Latin America world, where winning in marketing meant paychecks for your company, right? Like everybody's mortgage, depending on you doing your marketing and sell, right? So there was a responsibility to your people and company, but not in this. We're now in like your responsibilities to your shareholders. Yay. Great. Everybody loves money. But it comes from a deeper responsibility. It's kind of like that disrespect of what you are capable of doing, showing up with a strategy for Tic-Tac-Toe when everybody else is playing chess. Yeah. Right. And you're just like, oh, it's simple. You move this piece in the center and you always win. People are dancing around you and you don't notice. I like people that exploit the full power of this thing and win. I do like that, like that winner mentality, the go get a mentality that like I will kill my competition sort of mentality, really aggressive, competitive marketers, like they can't have a bad quarter, or they're just, you got fired and you're starting your new company. And they're like, how do I generate the man? How do I price correctly? That matters. So that's, that's what I like. I think like we have tremendous power to do all of those things. And I like to play with the weird dichotomies that we have in belief about marketing, right? So the, you go into a boardroom, like marketing doesn't work as a waste of money. I'm like, oh, cool. So you don't care about your privacy settings and you share your data. and you know, meta Facebook, whomever your big bad is at the moment, you don't care. It's like, oh no, they will shift every possible election. I believe Cambridge Analytica happens. So marketing can shift the world, but all you sell your product. And people maintain this in Congress ideas, and it's funny. We have that in marketing. It's a waste of money, but too powerful. Yeah, I keep them all as far away as the world, but they have the order, bros, what's been money on it? It's everyone gone, all the marketing, douche. You're going to lose your budget now, sorry, spend it or lose it. All you need is more X. Why I was told by somebody who would know that the CEO of well-known low-budget airline won't name, has claimed we could do, we could do that marketing. We don't need marketing. We still sell, you know, millions of seats. It's been done. This comes down to people not understanding the mechanism and thinking they're clever operators. Yeah. You don't understand the assets you have in their decay rates. So you can say, oh, I can stop advertising today and I'm going to sell as much tomorrow. Absolutely. Go for it. Come back to me in six months. Yeah. Come back in a year. Now, if everybody has also, remember the competitive aspect, right? If everybody else also exits the market, then you're going to lose at a different rate than if somebody jumps in and starts advertising in the excess of share of voice in that sense. The decay rate of brand elements, right? So this is an old study that I worked with some friends that I don't think have ever made into print. So this is one of those desk files. It happens a lot, unfortunately. But the decay rate of salience awareness is about a, here would be a point nine. So in six months, you lose about 10% of that ability to somebody record it. But you will lose over 60% of what you knew about that object or that person or that brand. Right. So I still say, oh, yeah, I've heard of them with, but you don't have so, and then you lose 10% over the next six month period. So now you're like that 81 of bays, but you're now at 20% of content image awareness or kind of consequence and connections. Now you're unsure of what opinion you had about that. Okay. Depending on your relationship, if somebody was really aggressive to you, you'd remember that aggression quite a while, but just brands that decays fast. So you will still have that name check, but no reason to buy or not buy, which is why eventually we get to brand repositioning as possible is because we've decayed all of the content. So the brand is hollow. I have heard of it, but I have no awareness of what's inside of it. If you know the mechanisms, then you can say, oh, if I need to cut my budget now, I'm going to take this out, but I need to come in heavier than, or it's everybody's going to cut. I'm actually going to go heavier now because I'm going to get abnormal returns to my marketing spend. The problem is people are playing games with systems. They don't understand. And that's yet, but that really goes right, right. People on that, like, it's interesting, if that's in your desk, mother of God, like, wow. What an algorithm there. It's easy to get your desk. Yeah. So, yeah, so this year, I am largely research focused doing less teaching to do a lot of the catch-up because the process, it's a lot of work to get something from Finnish study to published study. And I have arguably like the, I call an embarrassment of riches, like I'm really well privileged. I sit in the University of Oxford. I am surrounded by CMOs that support me in, like, here's data. Here's questions. Here's all this. Like, as a professor, this is a dream, but they're still sitting in the desk until a human writes and explains it. So, 2025 is the year of writing for Felipe. You come out with a gravely valuable paper, got these peer reviews after two years, and it came to my attention. I read MI3 in Australia, a lot of them share on me, but it came to my attention only then. And it was a tension grabbing headline as MI3 log to do. I'm going to read the headline. You're coming out with research, it says, "Birish, sharp, amber, bass, marketing science rules no longer hold, and you've a thousand campaigns, a million customer journey as evidence. Lots to start with and lots to unpack." I'd love to first talk about the tens of campaigns and median customer journeys, right? There's a lot of data. It's going to make it really robust, but am I right in saying there are the two projects or two parts of the project? Yeah. So, that headline, that was a great conversation with MI3 and South by Southwest presentation that was around that time. What I've shown in that presentation was a triangulation. So it was, here's various separate studies, and they converging to the story. I had as projects in those studies in different stages of development. With core of that, that has a thousand campaigns, which is data from Cantar and collaborations. The idea was a conversation with Cantar and Meta, and it started on just how do you, just essentially, can we explain cross-media better? Cross-effects, integrated marketing cons, complementary. Can we do better? And that was the start, and that's the paper that it's been kind of several names, and it's existed for a couple of years because it didn't peer review for that long. But it has, I'm going to talk a lot more about that specific paper, but it has, to me, a very rich set of results and powerful results that have been improved even through the review process, because we've been challenged and asked to prove even interesting things. Prove to us that industry doesn't already know this and use this, like, okay, we agree with the math. This is reasonable. It doesn't match what we know in the literature. So scientists haven't explained it this way, but does it have any value to anybody outside of academia? So even that became part of it, but it sets the ground. It asks some important questions, and this proves some lay beliefs and lay theory. So I think that's where it became spicy because it attacks the LinkedIn kind of belief ecosystem pretty strongly. But it doesn't start explaining why a lot of that stuff happens. Right. So it says, the things are different and you need to do things differently. It doesn't give a lot of the mechanisms of why and how and so on. So that's the next set of studies. I was like, okay, if we know these things are different and all channels are, as we say, like non-fungible. So you can't exchange a view on Facebook for a viewer TV. They're even on the same human being. They are doing completely different things. What does that mean? What is giving rise to those? What is that differential power across KPIs? And that's where the next study comes in. That second one is with WaveMaker. They have this ridiculously cool data set, which has a, it's like a 1.1 million customer journeys, think debriefs and surveys post purchase. Okay. Like what have you interacted? Well, like there's 72 or so different touch points that are measured and evaluated. And free existing brand priming and awareness, consideration sets. It's just like if you're a marketing nerd, oh my God, like the sheer joy of seeing that much, right, in the same place. And they reinforce each other because they come at the same question of media. You know, I'll use the wrong word here because it's the simplest word. But just synergies in the media ecosystem. It's not just synergy. There's more to it. But being able to speak to all of that, the reason for triangulating the reason for that is there will never be a single study that answers every question. Yes. Right. Or we have to do narrow things, prove those very narrow things thoroughly. And then we just do that. Lots of times. Right. Yeah. Because if you just write like, oh, I now have a theory of everything. And here's all media explained, it's going to be such a weak study that it's going to be almost impossible to prove. We need this sort of triangulation argument. That's where that came from. So one of the things that's kind of come out maybe is a narrative around your work. It's almost like an anti-Amber bus barge sharp sentiment, I don't think that was an act. Like that was the starting point. Like somebody didn't come to you and say, disprove anything that the bar and the institute have done. An outcome has been some of the principles that have been, you know, widely accepted. And as a result of the work from the institute and barge, have been maybe shown up to not be not be correct or how would you articulate that? Yeah. Honestly, I feel like I've been positioned as the, in a lot of sites like, oh, everything you're saying is just disproving by a sharp or doing this and that has never been the starting point. It's like such a weird motivation, right? Like, oh, like, I'm going to, my life is just to undo this book. It isn't. It's coming from a question that was unrelated to buying media channels. Even earlier, earlier kind of fight with barge sharp was, God, it was like around a contagion in London event or something that was presenting work on media complementarity and barge sharp made a comment of like, oh, media across effects are a myth and it's like, well, my work is not a myth. Here are the results. And then they put us against each other was like, oh, he said was, you know, mythology. I'm saying, here's. hard science. Oh, they're fighting and now it's spicy. And that's how it started. And the same thing here, like this started from a question from Meta from their side, you can imagine they're like, oh, how do we combine into media plans? They're always going to want to fight for some of that TV budget, right? Like you all know this. But they want to understand that they're placing the environment. So they're like, can you figure it out? I was like, okay. And that's part of our internal Oxford approach to identifying good projects is problems that are not solved in the science and problems that will have commercial value, right? So somebody cares, not just me, but it's not just consulting. The science itself doesn't have answers in one of our top journals. Then you have a killer project and everybody cares because I care because science moves forward, moves forward, and industry care is the same with this one. It started from that point that reviewer asked, can you show it to us that industry does not use media in the way you're describing? But I had to actually prove they were doing something different. And that's where I did this initial kind of exploratory analysis. So it's not a big part of the paper. It's like half a paragraph of the paper, like, like, like half a page of the paper in a second page. All right. Well, but if we go like my market share of this model is really, really small part of the whole thing, right, is to show that 80% of the thousand campaigns can be shown to be optimizing. Right. So one, we have to depart from the point that managers are not it is, right? They're not assigning things at random. They're making choices. They're spending millions of dollars, whatever it is to get these campaigns out. And these are all very large campaigns. We have an paper, right? You can imagine they're also paid to study the effect of their campaigns. Not every campaign does that. And they're all optimizing to reach. So I could show they have, it's not like their idiots. They have guidance. And there he is. And they're doing what they think is right. Yeah. They are optimizing. Then you look at what is the distribution of lift outcomes that we have or as the consequence of the campaigns. And it's normally distributed away from efficiency. So it's kind of like it's a random shot. Yeah. It's not clumped. If you thought if there was a high correlation of if I get reach, I get everything else, it should be equally optimized. Right. But it's not. There's a distribution. So it's, it seems to be disconnected. And that proved to this group that says, look, they are not doing it because they're not getting optimal outcomes. They are optimizing on a planning KPI, a planning coefficient, but not optimizing on outcome. Yeah. And yeah. So that's what people said. Like, oh, this goes against buyer sharps recommendations of reach as opposed to this or just salient as opposed to positioning. And it's like, actually, that is not doing that at all. And frankly, if you look at my paper, and again, this like, am I going against this person or does academia have something against buyer and sharps? There's no persecution. We're just interested in what works, what doesn't, and what the science has to say. If you look at the paper, it doesn't cite barns in two years, none of the reviewers or professors analyzing this work and trying to find flaw in it, asked how this relates. Otherwise, we would have to cite his work. And it makes very interesting conversations or headlines. In the on the record, I, I don't know, barn and sharps, we've never spoken. I don't have a basis to like or just like the man in the least. I have argued with some of his colleagues online. I did not have a very great experience in the argument. It wasn't very academic and it wasn't very founded in science. And I like academic fights. I like it as agreement. I think science is better when we challenge each other. And I thoroughly disagree with just about every single piece of insight and recommendations from that work. So like, in terms of, in terms of what it says and implications. So it's one of those weird things. Everything Andrew Ehrenberg said is correct. Everything Frank Bess said is correct. The work that has come from that of, therefore you do this, I completely disagree with that interpretation. And I think it's thoroughly flawed. It has nothing to do with the gentleman it has to do with the idea. Yeah. And to your point, it's driven a lot of organizations, big organizations, marketers to plan and buy that way. It's had an impact. And so I think when I wrote in my knees that are about this, is I think actually, it's encoded upon all of us to go, oh, here's something, here's something new. Actually, this is scientific to you peer reviewed. So we absolutely need to pay attention to your work. The thanks. I'd love to get into some of it because there's some really interesting study touched a bit about outcomes. And really, it was, it's basis outcomes. And he said, only one percentage of campaigns delivered substantial results, but most achieved mediocre result. That's kind of frightening. Yeah, it's, and again, that comes from that overall distribution that I mentioned, like that, that motivated the study and then motivated additional things. But that's worrisome because if you tell marketers, this is what you need to attend to. This is what you need to pay attention to. And you need to get more of this, they deliver, right? So if you say, I need reach, and if I'm from measuring region in all of these sorts of aspects, and I believe in thoroughly in the funnel, right? Like the funnel is my God. And if I get more top funnel, everything must follow through the funnel. Nothing falls out. Eventually, you know, my conditional probabilities means I'm going to get sales. So more has to be better. I can feel that argument, making sense in people's heads. When you're looking at this analysis, and I'll try to be precise about why I say 1% each one of these analysis was done as if it was a portfolio. Okay, so you're investing in media and getting returns from that media. But the media kind of interact with each other, that's kind of cool. And they meet different parts of your portfolio, which might be risk adjustments or whatever it might be. And let's assume there's an optimal portfolio construction, a certain percentage of stock that you're buying in different parts, that's going to get you a good trade off of the maximum amount of return, lift for us in this case, for the risk exposure that you have, a classic finance result is that, whoa, I can give you 60% lift. But, you know, there's going to be a high of volatility. You know, there's, there's, it's not all the time. I get that just on Tuesdays from, you know, the moon is full. Every other time I can actually get a zero return. But, you know, it's a moonshot. If your risk preference accepts that, you play that game. And that's rational. You're maximizing something with high risk. If you don't have that risk appetite, you move down the straight-off curve, right? This is called an efficient frontier. It's like you're still making an efficient allocation where you're accepting a lot less return, a lot less lift, but you kind of guarantee it at every single time. Right. You have tighter bands around your list or your performance outcome, but it's still efficient. So, what I try to see is, if I look at all the data that I have now, not modeled just how far, how close to these lines of efficiency are each of these thousand campaigns. What I had is then on reach, you say, maximize my reach and minimize my volatility of reach. 80 percent of all my data is 80 percent of the way to that curve. Right. Right. If I look at lift, 1 percent of my campaigns are 80 percent of the way to that efficient frontier. Right. And everybody's, so it means like there's always a different allocation that would have moved you closer to perfect. You are listening to that's what I call marketing. Yeah. Right. And if you looked at outcomes, what does it mean to be close to that edge, that barrier? You can imagine, on reach, you would have had a hundred percent reach, you can't go higher than a hundred percent. That's why you see what were the numbers like 85 to 95 percent reach for those maximizers. For the efficiency on outcomes, people were getting about 15 to 18 percent lift in their campaigns. Right. And again, just 18 as if you use all of the data, 15 is if you get rid of outliers, people that for some reason maybe lucked out. But you're still 15 to 18 percent depending on your beliefs of outliers are real or are not able to be replicated. The bulk of the people and the campaigns, multi-million dollar campaigns are in the bottom with that 1.75 percent lift. Right. Yeah. Right. So that's that's in that difference from 1 percent outcome to 18 percent. That's why I say like this is mediocre. Yeah. If you have 18 percent on the table. Yes. Yeah. And you're winning 1 percent. That's where it's just like that to me is a gut pun. Yes. Right. There's much, much better that you can do. And now it's both things and what this paper started going towards is, well, you need the scale of reach. But what are you scaling? Are you scaling one and a half? What if you could scale 10? What if you could scale 15 percent? Right. So it's that dual optimization of, it's not just that you get one, so there's no reach sufficiency. If I get reach, I get everything else. How do I plan for my goals? How do I scale my goals? And what is sort of the trade off that I can do so I can joint optimize both of those drives? Yeah. And that seems like the big thing because there is, certainly, like it does, create you, you are looking for reach and we all plan like we're planning for reach because, you know, we believe there's a couple of things that I think just to address here because there's internal naval gazing, there's also a few other components that might be worth addressing. Okay, so one is the agency clients, part of my privilege as an academic, the non-commercial neutral party is that get a lot of truths behind closed doors, people that, you know, something that a client will never, right? Like the brand is never going to tell you because you're the agency and the agency's never, never going to tell the client because the client's always right. I've seen and heard a lot of this like media planners have known this forever or we just tell our clients what they want to hear or we're just selling them what they want to write. There's a lot of that, right? So there's, there has been an over simplification of the media planning and buying process because some brands have bought into this all I need is this and that's what they're getting. So part of this is a rejection of that lay theory that this is all they need to buy and this is the only KPI they need to buy. So those two parties need to have a really good frank conversation. There needs to be an, you know, a little bit of a bravery of going against the CMO that says actually we have evidence to the contrary. How about we run an AB test and maximize reaching this geographic boundary we're going to run this lifted adjusted sort of thing a little better over here. So there's that and the other part is this evidence of we see like out of a thousand campaigns the vast majority of them are doing this wrong. So it's not like we found something that is not applicable is because nobody's doing it right. The reality of it, it's people are doing this wrong and it's reflected in every single conversation I see on media composition because you say, oh, why are we not going to, why would you stop advertising on Facebook and people go and why TikTok, right? So here's a nascent, it feels maybe over the past two years conversation that we have. We're leaving this, this channel in exchange for that channel. What is the advantage of TikTok? People say because the young people are in TikTok. What sort of argument is that? It's an audience construction, right? It's a targeting and audience construction. Who are you trying to reach? And no point was a part of discussion was what is the mechanism? What is the effect that that channel has on that individual regardless of who they might be? Right, there's a strong likelihood, for example, and I don't, here's a speculation. So let me put a red flag in warning inside of here, right? So this is unproven ideation. This is us jamming and having no basis for this whatsoever. But what if TikTok by virtue of how it's consumed and the velocity of the channel, the content on the channel, the formatting of the quick video content and the delivery is really good for brand recognition, but actually is really poor for recall or association, right? So you now know that maybe it does that because it's good for one type of memory formation, but it's really bad for different types. But meta-platforms by virtue of the multiplicity of content inside of it, like direct messages from people and your aunt is there or whatever it might be. And there's pictures about something that you did. There's contextual information that doesn't exist in the other channel about your life and it creates more associations. Right. So like you're actually getting completely different psychological pathways across different channels. But if you go to somebody that said, why did you choose channel one and channel two together? It is because of who was in that channel. Yes. And how it reaches. It doesn't matter who you ask. I had this conversation with professors. Why do we create multimedia campaigns? Why do we need multiple channels? Why don't you take all of your budget and just toss it on TV? Yeah, well, yeah. And most often they go to TV anymore. Yeah, that's it. So it goes back to who am I talking to? How do I maximize this? Nobody's talking about the fact that there's actual different functional benefits that the channel delivers. We're leveraging that to create a business outcome simultaneous with who we're trying to reach. So is it then like so take that then? Is it is that the starting point you're going looking at the platform functionality? Like so so we go actually we wanted to be on TV because we know that people are watching TV at least will have made that probably choice to watch TV. Right. They're not watching Netflix. They're watching you know, I'm a celebrity or love island or whatever it is. Even if they don't pay active attention to my ad and I've got good distinct brand assets and audio assets are a benefit from passive attention. Like it's like the starting point for us is I'm then after 18 to 24 years. Exactly. It's exactly that. And I think that's like if I tweak a little bit into this follow up work, which is what we're trying to do, I still don't have construction components. So things like the speed of the environment, algorithmic versus timeline based aspect ratio, length of videos like static versus video, all of those things play a role as well as what you described the utilization. What is the mood was that like, you know, an ad on cinema where you're captive and staring at a huge screen and there's only six ads before something that you intend to see versus an interruption based versus a print. Like all of those are going to have very different utility for people as well as audiences. But so this further work with wave maker with the million journeys shows that so we were able to estimate how easy or difficult people are to influence. We generate an individual level characteristic, which we called omega that captures how static or movable you are in a given category. And each channel has different power and a different likelihood to influence your journey at different levels of this quality. So some channels are actually particularly good at moving people that are low out omega, like they're just let's call like low energy people that are static stable and they have their preferences and they're not going to do. But actually TV and word of mouse have a good chance of shifting them. But other stuff won't right like online videos won't influencers won't like any of those. If you go towards the middle of people that have a bit more omega bit more energy kind of every channel starts to come in play. But the rank order of this channels are different by category. I would imagine yeah yeah yeah. Right. And then once you go to extreme omega people those that are like you know I jokingly call them the lemmings like whatever this stimulus is they do they're like oh that's a great idea let's do that like that doesn't matter because then every channel influences them right right so you think like everybody's like high-percentive and it could be that sort of thing and different categories are centered in different parts. So like personal care is very low omega values people are really sticky and really difficult to shift because they have it could be a time in category time out of category preferences built over a long time and connected with identity or how they present themselves so they're not going to change their shampoo face routine whatever my like like that kind of goes I clearly don't have the shampoo problem no my beard oil right but you go into cars it's the exact opposite right so people are inviting a lot of information when they're looking for a car and they're not very sticky so it's a very high risk journey because any input can move somebody to a different consideration set or a different brand alcohol tends to be a variety seeking category if somebody says oh have you tried this beer and you're like oh yeah let's try it you're now completely else and it took very little to me maybe a self-disclosing on my beer consumption but kind of behaviors and everybody else in the planet like no I only ever drink one beer but it's like that's it's centered in that but it's understanding the differential power of these channels in your context to get different behaviors one of the other things you kind of thought something a bit there was and you called us a scattergun approach so the like the answer isn't then go go everywhere because then we'll get well thanks because then we'll get people who may be thinking this like yeah so um maybe this is a nod to the intelligence of our media planner so let's let's recognize that the worst possible thing you can do is turn out every button yeah right so we part of this project and if you've heard me speak before no silver bullets there's no single answer that solves all our problems. I think it's a very sad marketing world when suddenly everything is reduced to a single variable. I think we're all out of jobs at that point. There's no single campaign construction that delivers every goal. Sometimes get close. But there's not one construction that just, again, is the silver bullet. If we want to play with fun kind of statements, there's no dominant strategy, right? There's no something that is better than all others. There is, however, what are called dominated strategies. There's always worse. Okay, right? So there's whatever the opposite of silver bullet is, is the one that always misfires. There's always a better bullet. There's always something you could have, whatever your goal was, there's always a better alternative than this, and that dominated strategy, the inverse silver bullet, the dumbest plate is to turn everything on. Right. Okay. Yeah. From moment to second, we do everything wrong. No, we're not turning everything wrong. No, no. The thing isn't see it happens. That's the scary thing is it's not like we all know this, but we see it actually as a, yeah, it does get deployed more often that you would hope considering the fact that there's always a better alternative even by just turning the thing off. And I think he could be from a, a variety of reasons, but a very common thing that could happen is simultaneous exploration, or I need to do everything. Sort of like, oh, you know, maybe I will use a little bit of budget here. I have left over there. If you are lucky to have left over budgets, let's live in that dream world first. Or, oh, but I need Pinterest for discovery. I need this for that. I need X for controversy. I need this for this. Now, suddenly you're everywhere. Yeah. And you are just not, you know, at all. Yeah. Yeah. Or, you know, let's go. It's closer to like if you're trying to be everything to all people's, you're nothing to nobody. That's pretty good equivalent sort of behavior is like this catagon spray and prey. You're always going to be outplayed by everybody else. If you're lucky that everybody else in your category is similarly incompetent and we're all doing the worst possible thing, then you're lucky. It's going to work because it's relative. So if a marketer and a mediator, like you said, listen to this and, you know, having lunch together, yeah, whatever magical world that that's happening, you know, listen to those chatty. Like what conversation they should be having now, right? Because they should be having a conversation. But like, is it throw at everything we've done? How do they start again? Like what's the, I don't feel like again, I don't feel like what do we do? Yeah. I hope it's not like you burn everything to the ground and start over conversation. Be careful what recommendation take. It's more of what recognize that I was able to prove a lot because I had a huge variety of companies that I observed over a huge number of years. This data set is frankly not available to companies or a lot of agencies like this is, this is rare stuff. So a company that just always does X would have never seen this open space, right? Because that exploration is not happening. So don't feel bad that I've been doing it wrong is just like everybody has kind of these blinders on and every channel, every media provider, every media owner has the same blinders. They know how their stuff works. They don't know how the other people's channels works or the combinations that work. So it's not like, shouldn't we, shouldn't we have known this? Like, no, we took decades to put this together. And if it wasn't for Kent, our measuring some of this stuff for decades, I never would have known. So no self-crisification needed, but a more honest conversation about what is working or not working relative to what they see kind of printed in the paper. I think a good departing. The papers are a very general set of statements about categories. We won't go into countries and conditions, but why are you using certain channels? Is it a inertia? Is it because your MMM that you've paid a quarter million this year, you told you to do that and did it give you the same plan as last year? Is it still just dominant by the channel you never turn off? Like, you start analyzing, like, why are you using it? Is it built on the basis of audience only and scale? Or is it built to drive a specific goal? Because if you're just saying it's purely reach, you're probably misplacing budget, as you're going to be optimizing to the wrong thing. The question is then if I'm going to build a portfolio of marketing interventions with a priority towards specific outcomes, some might be sales and direct kind of intent to buy. Others are going to be on acquisition, market entry. Others are going to be on association because of something your competitors did. That's a portfolio of activations and those activations are all going to have different media spend against them. That's kind of what are those combinations that you need to worry about? Those are different for different verticals, categories, industries. That is one of the more uncomfortable parts for several parties in this media ecosystem because this is where it's fun because I'm an academic and I don't have money. Yeah, it's just saying so a lot of media is mispriced because we're all buying TV as if it was worth the same to everybody else. And you know, maybe there's a channel somewhere where it's priced for the average channel, whatever, for the average category, for the average company, you can essentially buy a cheaply relative to the value that you generate. There's a huge amount of arbitrage right now in the media ecosystem until everybody figures out that not necessarily dynamic pricing, but you can adjust the price to outcome. But there's opportunity, like the amount of the normal returns you should be able to get in media on the back of this. Yeah. Because there's going to be expensive media that's under delivering and there's going to be cheap media that would ideally over the liver for your specific category. So there's probably a short run, sort of abnormal gains that you can get over the next year or two until the world kind of comes up. If they ever start adjusting pricing, there's probably two years of really strong and normal marketing returns that you can get. And because a lot of this stuff builds over time and it's sticky and you have long-term marketing assets built off of this, there's this proportional brand advantage that you can build off of this work. Because it's like for two years, you can outplay your competition by display book. Until it becomes price-adjusted and then you can do it, but you're going to have to pay for that. There's a window of opportunity right now for brands to really punish their competitors. Which sends really incredibly exciting and like all the media and it's bought on eyeballs. So that's going to be a shift, right? If it's backboard and I take comes, but they have to be able to measure that in some way. But again, maybe back to the practical application of this. I'm listening to you and going, I want to win. Crewing to be competitors in the next two years while they're still buying in the old methods. I have practically used a way to do that because again, I don't know if my media agency or any media agency know this yet. So where is the connection point? Let's play across two different kind of markets. People that might have strong agency support and those that don't have strong agency support. So I might or might not have gotten a lot of calls about how to implement this for a lot of media agencies and a lot of media owners. Yes. So it's coming, right? They're going to start embedding this into a lot of stuff they do. More insight coming from that. It's not like we're going to be helpless for a very long time. A big issue on the connection between academia and practice is productizing the stuff that I do because I'm supposed to be generating insights it. But I won't turn it into a dashboard. That translation is the painful bit. What we have in the paper is two things that can be incredibly helpful. A mapping of the archetypes and a mapping of how the archetypes leads to different outcomes. Archetypes are a combination of channels. So it's kind of a media plan. The media plan is described in probabilities. Somebody with this media plan is very likely to use outdoor billboards, right? So you can say like, oh, you know, if you don't use one, there's a high chance that you're not falling inside of this kind of plan. Oh, you can say, okay, I have a direct mapping of generally which archetypes or which media plans connect to a given outcome. You might want to see how far your plans are to those combinations of channels given that specific outcome. Looking at your history and what you've done, which ones or which kind of combination of archetypes you have been running and see if that matches to what we're saying, like, oh, that will drive this, but it won't drive that. We have a few points of adjustment for industry. Now there are more industries than we're able to print in a paper, right? So they necessarily have been generic. So like, I think one of the categories we have is like CPG. you're in a CPG world, fast-roving goods, like you might want to think of those terms. We have financial services, or if you're a service-based company, you can imagine that it behaves very similarly. So you can actually fine-tune the recommendation from the general to the specific and saying, oh, actually, this is one of those where I said the, there was close to a silver bullet, 'cause CPG all go towards one central kind of plan and across all KPIs, which is really weird, but there we are, right? So they have a single goal, a dominant strategy. You can fine-tune recommendations based on what you're trying to sell, which again is very useful. I think if you're a planner or a marketer because you're not just learning one way of doing things, you're learning a way of thinking through, if you change jobs tomorrow and you have to sell something else, how would you go about adjusting your media spend? It's not just whatever you learned and brought it here. It's like, oh, go all in on snap. Well, in your category emergency. And if you go on the SSRN page, which I might have linked or will make sure you have as well. So the paper is supposed to be available open science, but every other day I look at it's closed or open and it's random, but there's also a pre-print copy that has all the things in it. It's the same paper in an open science thing. And it has actually the computer code on how to run this whole thing. Right. Okay, well, that's all right. So it's the whole, like again, that's the whole thing of open science is the whole thing. Again, it requires data and it requires. So if you're a larger company and you have your own history, you can start recreating a lot of this stuff. It will require a bit more data science know-how on how to deal with even computer packages, the packages of the language is like Stan, it's, but it's there. Literally every meaningful piece of code is there if you wanted to recreate, to redo, redeploy. It's the full thing is kind of there. The only thing we can't give is the data, obviously, because that's proprietary to can't talk about. Probably the part of the thing was, I have no idea what this stuff is. All these numbers. So it's different people will hopefully get different things from the paper, you know, if you're one of the people that read the abstract and conclusions, thank you for doing that. Like that's, you're already engaging with content and getting an understanding and that's sufficient. If you went a little deeper and you're trying to understand that the rationale I'm taking and you avoided all of the Greek in between, still amazing. I'm glad you're trying to get into the theory and why things might be different. And there's, there's lots of little fights inside of the paper of me fighting against the scientific community of, in a friendly way, these are people that actually know in real life and, you know, have dinners with. It's just going and saying, this idea is wrong. Like one of those, like I alluded to earlier, I said synergy and in this paper I had to fight that the concept of synergy was wrong. Okay. It's because it's too simplistic as one explanation of why media works better together. Yeah, rice, rice, yeah, yeah, it is both cross effects and complementarity separately that build it. Okay. They were trying to combine cross effects and complementarity into one thing. Right. I had to argue they are two separate components in media planning where complementarity is, they do different jobs. Yeah. Right. So they affect different mechanisms. And they don't overlap in mechanisms. But a cross effect is that the presence of both of those channels in the same plan changed the power in which they do their job. So, you know, channel one does X, channel two does Y. There's no overlap between X and Y. They're complementary. Right. Okay. But that's interesting. Yeah. But which channel two is present and now makes channel one really, really strong at delivering X, that's a cross effect. Okay. Okay. So very, very interesting distinction to make. I would have thought I probably would have talked about the synergies and the layered effects of media. And you know, so that's, that's it. Yeah. Yeah. So this, again, that's that whole complexity. Like, so if you go and try to explain why a channel behaves in a certain way, how it functions, if you just say, oh, it's synergistic, you're again kind of, it's very easy to fall into hand-waysy sort of bits, right? But if you're trying to explain why it works sometimes and it doesn't, you can say, actually, they're both acting on the same psychological mechanism. There's zero complementarity. So their overlap is canceling some of the value they could have delivered and you have an empty slot where there needs to be another channel to do work. So you can become more precise about your planning and strategy. Again, I'm kind of obsessed with media functionality at the moment because of these papers. So in 2025, 26, probably 27 is going to be a lot of, how does this work? How do I build the differences? I don't have a lot of explanation as to what these mechanisms are. But I think from this work with wave maker, respect for human heterogeneity, we're all different and affected differently by different media, different times, different occasions, all of that matters. But still explaining construction so that we can go down to a very, very faring grain level of saying, I am in this channel, in this occasion, because I don't have another channel that gives me mechanism three for floor memory for me. Right, right. So if you understand down to that level, you don't need to go to that level. Yeah, that's that complexity. How far do you want to go? You can say, you know, I'm just going to buy more reach. That's fine. You were 100% allowed to do that. It is your choice of how you're going to do your marketing. But again, that to that original statement is, are you're going to lose to the person that understands it a bit more? Yeah. And that's again, I think to our starting point is, like this has to be read and explored by people. They can then decide to fundamentally disagree with it. And that is okay. You know what I mean? And you're enough buying, you're saying, I could apparently have them to you, by the way. And so I absolutely have to elevate your approach. There's so much more I could chat to you about because I think there's so many dynamics this, like even, you know, the growth of creativity and all that kind of stuff, you know, is big. And, you know, I mean, we'll have to maybe another day we'll come back. I could do this for hours. And one last question. I did read, I don't know if it's correct. Are you writing a book on marketing and AI? Yeah, so there's probably a few books now. I don't know how to bend with this year's gonna be. I had been commissioned to write a book on marketing interaction with AI. There has been slight delay in that. I was supposed to be done like a year and a half ago, but because again, my position of privilege and looking behind the doors of some AI providers who might talk to you, how do I put it? You know how conversation has changed about the timelines on AI? Yeah, that changed how the whole book was written. Right, yeah. Right, so when you go from, oh, you know, you start writing, LLMs are a thing sure, but there's no agents and there's no reasoning and there's, you know, I've said it to my class, arguments are being made. AI has been achieved internally. Like they're just giving timelines and so on, but we can argue over the definition. And that's what you're getting those different timelines. But the thing exists and it might not be consumer grade. So it's not like you're gonna get the AI updates on your cell phone. It changes a lot of not just how it do marketing. My biggest worries that nobody seems to recognize that it changes their marketplace first. Right, yeah, like everybody's where it's like, oh, so it's gonna change how it do my ads. No, it changes how people buy. Right, that's absolutely right. So like, I think that's a slightly bigger problem than like, oh, you know, my media planning is gonna be wrong. No, the economy is gonna be different. How people buy is gonna be different. How they search is gonna be different. How they consider is gonna be different. That changes the scope of the book as I was writing it. That's one, the marketing tech. I would need you to understand marketing mechanisms before we have their conversation. Right, because like this is marketing how it works. His story works under a high automated algorithmic kind of marketplace. And in my plans to actually then also do a pretty open book for essentially branding 101 and marketing 101. So since I've been asked, like, I'll do that. And then marketing 201, like I said, the operator, like if one is a book to explain mechanisms and then a second of a book on how to break mechanisms. And they bring that night on genuinely and I mean, because I think, yeah, I think there's gaps in the understanding and that's all. I'll have those, I know I've, I'll have those on my, there's a couple ways this is done on computer science that I actually really, really love on some of these fast moving things. If and when I do at least the other two, not the AI one, 'cause that was kind of commissioned and asked by my university, but the other two's, 'cause I just think it's kind of needed. My one of you and my students ask you like, oh, how do I get your inside your head like? So if and when those are ready, I'll probably just have them open source and available digital on my website I'd love you to come back and talk just about those when they're available. I'll keep an eye out and Thank you so much for taking the time. I mean genuinely appreciate it. It's and It's gonna have a big impact and I think again as we covered The big impact isn't aimed at anybody and you know, I you know, easy headline for me to say it is It's not and people have listened with her. It does come from a request and a lot of data peer-reviewed over two years and so I certainly encourage we'll put a link to the open source in the show notes and really encourage people to as you say even if it's two parts, contingent summary or delve into as much as they can and press your feet Thomas. Thank you so much Read your presets today. Thank you for having me. It was a joy to talk Well, I hope you enjoy that episode what a fascinating conversation with professor of Philip Thomas and One of things that really can't look so much to it and so much to the papers big implications for it and Really for me like I came across really strongly was This isn't about two people being pitched against each other. It's not a battle Professor Thomas as he said at the start he was asked to look into something and that scientific journey has led him to outcomes and Things that you know are against what I in in the market and so I Think all we can do and all all of you can do is Is take time to read even parts of the paper and But also like follow professor Thomas and you know see what else he he's gonna be on other podcasts I'm sure talking about this and and it's not again. This isn't ego for him. We talked about it actually at the end is you know This isn't him trying to become a star In any sense the world like genuinely when you spend we spoke both ends of the recording for for quite a sense of period and That's not what he cares about and so he genuinely cares about progressing What we're doing as a marketing community and so take it with that. I think when you listen review his work And again, everyone needs to make their own decisions based on what they read what they believe is right for their business I certainly will be taking it on board and seeing how I can Apply it and what's what's right for me. So look a lot to take in that episode was a long episode I really hope you've enjoyed it. I hope you do share it with people and with your Media agencies if you're a client with your clients if you're a media agencies and And let's start the conversation do leave comments. I love comments that will You know help us engage with this even further Listen, thanks so much for tuning in for now a less episode today And if you did enjoy it and would like to review it, I would really appreciate that and of course you can subscribe too That's what I call marketing and never miss an episode again That's it for me from this episode. Thank you so much for tuning in and we'll catch you the next time If you would like to reach and engage community of marketing leaders get in touch with that's what I call marketing to discuss sponsorship opportunities