Phil Barden is the Managing Director of DECODE and the author of “Decoded: The Science Behind Why We Buy” which is a classic text about the behavioral science of marketing. Before going out on his own, he held leadership roles at T-Mobile, First Choice Holidays, Diageo and Unilever. As his profile says, “25 years client-side marketing + 10 years decision science = ‘Decoded.”
Okey-dokey. All right. Well, Phil, again, I really appreciate you accepting my invitation to come on here and talk to me. I'm not sure if you know this, but I start all of my conversations and even my interviews at work, actually, with a question that I borrowed from a friend of mine. She's an oral historian, and she helps people tell their stories. And she's got this beautiful question that I stole because it's really big and beautiful.
But because it's so big and beautiful, I kind of over-explain it. So before I ask it, I want you to know that you're in total control, and you can answer or not answer any way that you want to. And the question is: where do you come from?
Oh, wow. This sounds like I'm on a therapist’s couch. Where do I come from? You mean sort of literally or metaphorically?
However you want to take it.
OK. OK. So, where do I come from? The origin story. Very, very interesting. So, literally, I'm from the UK, born and brought up in the south of England. My parents were both teachers, and my mother exposed me to creativity. She was a music teacher, and I think that probably stood me in good stead for wanting to get into marketing. My father was a science teacher, which is where my worlds have collided in my most recent career. So I think that's probably where I come from.
Yeah, that's quite remarkable to have the science and the creativity. Can you tell me a story about the creativity? I listened to another interview where you talked about the role of creativity growing up. What was it like?
Hmm. Yeah, well, my mother insisted that all her children learn a musical instrument, and she exposed us to many various types and genres of music. She had a very eclectic taste, which I think is admirable because we were brought up not only with classical music, opera, and ballet, but also musical theatre, pop, and rock bands.
She was a huge Queen fan, for example. So she would listen to the entire album of *A Night at the Opera*, and then right after that, she'd put on the *Peer Gynt Suite* or some Rachmaninoff or something. So we were all taken from quite an early age to operas, to ballet, to the Proms, classical music performances, and also musical theatre.
So, Gilbert and Sullivan, a sort of light opera, and then into rock opera even. She loved Andrew Lloyd Webber's work—*Jesus Christ Superstar* and *Joseph and the Amazing Technicolor Dreamcoat*. She was the musical director for all the musical productions at her school.
And I grew up learning the flute, my brother learned the violin, and my sister learned the piano. So we all were instilled with a love of music from an early age. And I think that exposure to the arts in general, because I then joined a youth theatre group as well, and school friends who participated in that—we played in a band together.
So we were always into the creative side of making music. And one of the guys from school, who I'm still friends with, said, when we were at school, "I'm going to go into advertising," because he and I used to write stuff for a school magazine and sketches for a school revue show and things like that. So the creative process really appealed to me.
And when I first learned about marketing, and that was through the husband of a babysitter actually, who worked for a marketing company, I had no idea what it was. But when I learned that it meant that he got involved in creating advertisements and packaging designs and promotional materials, I thought, "This sounds amazing," because it was very tangible. And I liked the idea of creating something tangible, but it also spoke to that creativity, that spark and inventiveness that I found so exciting.
Yeah. And did you—I mean, how old were you when you discovered advertising? Was that—were you that young when the idea of working in advertising arrived?
Yeah, at school, teenage years. So this friend and I used to buy *Campaign* magazine and devour it and learn about all the greats of the advertising world, not just in London but worldwide, and the big agencies. He also had a friend of a friend who was a copywriter.
So we got to hang out with him a bit, and it seemed like a very glamorous lifestyle. So when I did my degree course, it was what was called a "thin sandwich" course. You spent six months at college and then six months working to get practical experience.
And you had to be sponsored by a business to do this course. So I got sponsored by a company called United Biscuits, which is now Pladis, a global company. And my placements with them included manufacturing and then the sales force.
And then the last placement was in the marketing department. And for me, this was like a dream come true. Because I'd heard about all this stuff, I'd been on the periphery of it.
But then to be in a marketing department, and actively involved in doing some of this stuff, talking about packaging and promotions and pricing and product changes and innovation—that was wonderful. So I knew that really cemented for me that I wanted a career in marketing.
Yeah. When did you first encounter the idea of the brand or the concept of the brand as an asset, or something to be managed or built or created?
I think it was after that placement when I joined Unilever. And I think it was the schooling and upbringing that I got there—that brands are the lifeblood of a business, that they create current and future revenues. And it was always this idea of brands having an intangible nature. A brand was a feeling, a brand was something in your head.
It was more than just the physical product and what it did, and the way it performed. It had some other essence to it that was always quite mysterious, but very exciting and intoxicating. So I think it was—I mean, obviously, we're all exposed to brands as we grow up, right, and adverts, etc., etc.
And, you know, kids from quite an early age become brand-conscious and brand-aware and very quickly decide that some clothing brands are cool and others aren't. And they wouldn't be seen dead in certain clothes, you know, from quite an early age, which is interesting. But then actually working with brands and having to make decisions that were not just in isolation, like, if we change the price of this, there's a financial calculation.
But over and above that, what does that do to the perception of the brand? Or if we change the packaging design or the materials used in packaging, we change the feel of it, the texture of it, what does that say? What meaning does that give?
So the idea that the brand was not just the physical components, but there was something else around it, this intangible equity, was something that was certainly cemented during my industrial placements.
Yeah. And you spent a long time managing brands, right? I mean, at high levels before, excuse me, the sort of the arrival of—you use the phrase "decision science." So I'm wondering, when did you first encounter decision science? And what did that do? How did that change the way that you operated?
Yeah, I guess that's the question.
Yeah. Well, I spent over 16 years at Unilever. And then I went to Diageo, which was also a real brand-centric company. And then I moved to T-Mobile. So I was invited by a former Unilever colleague. And it was a very exciting invitation because he said to me, "Listen, you and I know that businesses have to be based on consumer demand.
And we have brands that are running through our veins. We take them for granted. But I've joined a business which is technology-led, it's supply-led. And the role of marketing is very different."
It's about sort of chucking stuff out there and seeing what sticks. But the problem is that all competitors in this market operate on pretty much a level playing field. There's one technological advance, and then very quickly thereafter, everyone has the same thing.
Everyone has the same handsets. So it's not just a question of who gets 4G or 5G first or who's got the fastest mobile internet or who's got the latest iPhone. Everybody does something very similar. So the only way to differentiate is through the brand.
So that's what I joined my former Unilever colleague to do. And slowly but surely, we managed to turn this ship. And it was when I was at T-Mobile that I first got exposed to decision science.
And decision science is a catch-all phrase that we use deliberately to include different fields of academic study and science. So a wide range—from neuroscience through the different flavors of psychology, cognitive, social, evolutionary, but also semiotics and cultural anthropology, and then more recently, what's become popularly known as behavioral economics. And I was exposed to this because I faced a dilemma. I commissioned a very expensive piece of research in 12 European countries.
And it was research with which I was familiar from Unilever and Diageo. But the data we were getting back just didn’t seem to make sense. And I didn’t, honestly, I didn’t know what to do.
It was a huge personal risk because I’d convinced the business to spend a lot of money—I mean, a high six-figure sum—on this research. And yeah, my own personal brand was in grave danger at this point. And someone said, “
You should have a talk with these guys from Decode.
They’ve got a very interesting perspective on what we do.” And I met them, and the two founders—one of them is a neuroscientist and the other is a psychologist. And I showed them the data, and I showed them the methodology that had led to and created the data.
And they immediately critiqued it, using frameworks and language and views that I’d never heard before. But what they said and the questions they asked seemed to make intuitive sense. And I commissioned them on a small piece of research to kind of replicate the approach but using their approach in one of the countries.
And it came back with far more useful data that intuitively made sense. So I took a very bold decision to cancel the original research and to commission Decode to run their approach, which was an interesting and novel one for me. I’d never heard of this implicit testing.
I’d never heard of so-called System One and System Two at the time. It turns out, actually, some of the guys I now work with were at Unilever at the same time as me, and they were using implicit testing in their product testing and sensory testing because they had PhDs in psychology. And it was obvious to them that you had to use this particular method if you wanted to validly measure certain aspects of brand performance.
And implicit testing grew out of social psychology because the psychologists knew there’s this big thing called a say-do gap. You know, what people say is not necessarily what they do. And particularly when you are asking for information about which people cannot say anything or will not say anything.
So, you know, racial, gender, stereotyping, and biases are where all this started. But now applying it to brands and understanding what brands meant at an implicit level, how communication works at an implicit level, this was a complete revelation to me. And the upshot of it was that we commissioned this research through Decode.
We used their approach and their model to create a new brand proposition and brief an agency. And the first manifestation of the brand relaunch in the UK was so successful. It grew sales by nearly 50%.
It actually doubled footfall into retail stores within 48 hours. I mean, I’ve never seen such a dramatic cause and effect before. And it’s got 40 million YouTube views or so, which is still pretty famous.
We didn’t have Instagram or TikTok in those days. It was Facebook and YouTube. So, it had 72 Facebook groups set up on the back of it.
And it was this unique idea of a flash mob, and that was the creative leap that brought what we wanted to bring to life. And it was so successful. I went back to the founders at Decode and said, “Nobody can believe this.
This is incredible. The company’s never seen results like this.” And these guys kind of shrugged and said, “Why are you surprised?
Because you know what we’ve encoded in that ad are motivators of behavior. So why are you surprised that it works?” And that for me was just like a huge lightbulb moment when I realized the sheer power of leveraging what these guys knew.
Because they were working with 150 years worth of scientific and academic study into human behavior. And ultimately, marketing is about behavior change. We want people to buy our brand, talk about our brand, share stuff, buy more, switch, whatever it might be.
It’s behavior and behavior change. And learning, as I did through this experience, that leveraging what the different fields of science know about behavior change has dramatic and direct commercial impact. And that’s what got me into decision science.
It was so exciting and so fundamental to marketing that I ended up deciding to quit my 25-year client-side career and join these guys because I just really wanted to be part of bringing decision science to marketing.
Yeah, what an amazing story. It's so clear that you’re right there on the front edge of this paradigm shift, right? I mean, does it feel that way, that there was a before and an after, clearly, in your story? But how would you describe the way marketing worked before? What’s the understanding that’s out there for people who don’t understand what decision science is?
Yeah, it worked. It worked. And it probably still does to some models and approaches that have evolved over many years that are simply either wrong or incomplete.
For example, I was schooled that to get behavior change, you need to change someone’s attitude—that people behave in a certain way because they hold certain attitudes. So what you needed to do was create an intervention, which typically was an advertising campaign that would get people to change their attitudes. And as a direct result of that, they would then change their behavior.
And what decision science tells us is that’s not the case, or it is the case, but in very few examples. What normally happens is that attitudes form following the behavior. It’s like a post hoc rationalization because we need to, in order to stay sane, we need our attitudes to conform to our behavior. So the relationship is not causal in one direction—it actually happens more post hoc.
So that was one example that, you know, we shouldn’t just go chasing attitudinal change and measuring attitudes as the be-all and end-all. And the other was the so-called AIDA model—the Attention, Interest, Desire, Action—as a sequence of events, and that you needed to get each one nailed in turn. And then, learning from behavioral science that it doesn’t work in a sequence like that, that actually, you can get attention because people are interested, rather than having to get attention first in order to create desire and interest.
And so understanding more about the mental processes that exist in the brain was fundamental to me but also very challenging. And being honest, very uncomfortable because, having grown up with certain paradigms, it’s hard to shape them. And it hurts. It’s effortful to change, change our minds.
And that even is described in behavioral science, it’s the so-called Semmelweis reflex, because change is inherently risky. And that’s why we have a bias towards a status quo. It’s why change management, per se, is very difficult in any business because we stick to what we know, because it’s comforting, it’s familiar, and it’s safe.
And once you rock that boat, it takes effort to change, to rethink or reconfigure your mental model of how the world works. And I think it was only because I had that personal experience of living through the T-Mobile relaunch and seeing the impact that helped me reshape my mental models. But it’s no surprise that, you know, when I first set up business in the UK with no clients but with this utter conviction that this was a very powerful tool, and I went to see my former colleagues from Unilever, and I would tell them about decision science and tell them the T-Mobile story.
And they were fascinated, absolutely fascinated. But they kind of gazed blankly at me and said, "Well, that’s amazing. But anyway, back to the day job," because it was too challenging.
It really, really hurts when you’re told that what you’ve been doing for years might not be the best way of doing something.
Yeah. Yeah. What are the implications? I mean, I’m a researcher, right, and a brand person, but what are the implications for how a team learns? You know, if you really digest the implications of this, what does it do methodologically in terms of how you go about building a brand or just understanding what’s motivating your customers?
I think you need to have a model that is rooted in how the brain works. And that’s what first attracted me to the Decode approach because what they said to me was, "Look, people buy brands, whether they’re physical products or services, to meet a job to be done," as we’ve come to know it popularly. And those jobs to be done are a mixture of things like functionality.
So, you know, I’m thirsty, I need refreshment, or I want broadband, I want fast connectivity. So it’s very, very important that you deliver those. But that’s not enough in a competitive world because there are many brands and choices that could satisfy that functional job to be done.
And that’s where there’s this other level, which is a more implicit level of social, emotional, psychological goals that people seek to achieve by using a product or service. And it’s the expected fit between a choice and its ability, its instrumentality, to meet a goal that drives valuation in the brain. Because if we’re faced with many choices, you know, going to a supermarket, and there are many brands that could meet your particular functional needs, but how do we actually make a choice?
That valuation is driven by the expectation of best fit. And that expectation itself is driven by the associations that we learn that the brand has. And that can be through personal experience.
But it can also be expectations and associations that are built through things like advertising, or word of mouth, or the signals that a brand sends through its packaging or its other activities and touchpoints. So they are all forming associations in the brain and the so-called System One associative network. And it’s those associations that we use to assess instrumentality when we’re faced with a choice.
So once you understand that is how human beings make decisions, then the question is, how do we measure associations, both at a category level and a brand level? And how do we, if our brand is deficient in associations versus what it needs to be, how do we strengthen them? Or maybe how do we weaken some associations if they’re kind of not helping the cause as well?
But once you understand that and it all fits together, you strengthen association, you get a better-perceived fit with the goal or job to be done, then you’re going to get higher valuation in the brain. That also leads to higher mental availability. So when that job to be done exists, which can be driven by occasion or context, situation, then which brand comes to mind first?
So it all loops back together, it all fits congruently as a model and as an approach. And then the choice of research method falls out of that. What construct are we trying to measure? Is it a System One response, or is it a System Two reflection or evaluation? You know, that’s why people, I think, often get confused and conflate concepts of emotion, for example, because it’s perfectly valid to ask people about, how did you feel about this ad you’ve just watched? What were your likes and your dislikes?
Those are emotion for sure, but the method you use elicits and evokes reflective mental processes, which is System Two. Right. So it’s perfectly valid to ask explicit questions or give people a like-at-scale or a rating scale for that type of thing.
But on the other hand, if you are trying to measure automatic associations or automatic emotional response, then you need a different method. You need an implicit method of some sort, whether that’s a biometric method to measure a physiological response or whether it’s an implicit testing method to measure associations. It all depends on the construct you want to measure.
Yeah. And I reached out to you because there was a post on LinkedIn in which you really drew this distinction between emotion and motivation, which feels like it’s really blurry. Those words are used kind of interchangeably in a lot of situations, but I loved how you pulled them apart. What is the difference, the functional difference? What is the distinction between emotion and motivations?
Yeah. It’s a confusing one for marketers. And the reason is, I think, because it goes back to this dichotomous view that we’ve had for many, many years in marketing and advertising, a split between emotional and rational.
And that has been conflated with System One and System Two. So people think now that emotion is System One and rational is System Two. There are a couple of things we need to tease apart here.
And we’ll come back to motivation in a moment. The brain doesn’t work on a basis of emotional versus rational. It works with automatic processes and reflective processes. And System One are automatic processes and System Two are reflective. So when we learn something, like I mentioned earlier about my mother insisting we learn musical instruments, you know, when you start learning a musical instrument or language or to walk or to drive, it’s really difficult. Lots of cognitive effort going into that.
And that’s a System Two task to learn those things. As they become automated and implicit, they pass into System One. So we no longer have to think about walking or driving. They are System One, but they’re not emotional. Walking is not emotional. So you can’t just pigeonhole System One as emotion and System Two as rational.
So that’s one thing that needs clarifying. The other thing is emotion and motivation, because people’s popular view is that emotion drives action. There was a very interesting meta-study by Professor Roy Baumeister at Florida State University. Baumeister and his colleagues examined over 4,000 published papers that purported to show a link between emotion and action—specifically, that emotions drove action.
And their review of 4,000 papers found, in fact, about 1% of those studies—about 40 papers—actually did show a link, a causal link with emotion driving action. And the vast majority of those were what Baumeister classified as extreme cases.
And what he meant by that was someone gets in your face and is really threatening and angry, and you hit them, right? You get fear and anger building up in you. And as a result, you hit them, or someone cuts you off in traffic and in frustration, you honk your horn at them. That’s an extreme case. But if emotion drove action, our visits to the supermarket would just be a rollercoaster of emotions. And they’re not, of course, they’re not, right?
So what actually does happen? And this is where, again, when we come back to science and examine what’s been studied, what drives motivation is so-called goal achievement. And one of the very senior neuroscientists at Stanford actually said, "Goals are the system units of human functioning, whether we’re aware of it or not."
Achieving a goal is what triggers motivation. And a goal, as I mentioned before, like a job to be done, can be functional, right? My body is, I’m thirsty, I’m in need of liquid refreshment.
So I drink something—that is a goal that I have. Or I need comforting, or I want to do something different, something adrenaline-filled, a bit rebellious, a bit risky. Or I want actually to—I have a goal that is about displaying self-esteem and superiority and whatever.
So the choices that I make to meet those goals will be very different. Motivation and emotion—because everyone says, "Oh, they share the same Latin roots." Well, yes and no. Motivation comes from the Latin verb "movere," which means to move. And emotion comes from "emovere," which means to excite. The "e" bit comes from, is similar to the concept of "out."
So it’s like moving out, so exciting. So they share some roots, but there is an important distinction between them. And when you talk to the scientists and academics about the role of emotion, they’re very clear.
I mean, Carver and Scheier, who are world-renowned psychologists, say we experience emotions whenever the likelihood of achieving a goal changes. And to give you an example, so if I’m playing tennis competitively, and so my goal is to win, and I am winning, I feel good. I feel powerful. I feel proud. But if I’m playing to win and I’m losing, I feel angry or frustrated or sad. So we experience them when the likelihood of achieving the goal changes.
They also act, and this is what Baumeister says as well, they also act as feedback to us, as a gauge of the extent to which we’re achieving our goals. And they then get linked with goal achievement to help us learn for the future. So imagine, for example, your friends say, you know, you’re going to meet a bunch of friends at the weekend and all go out for dinner.
And the task of choosing the restaurant falls to you. And you want to feel good about your choice. And you want to maybe exhibit your knowledge or your connoisseurship. So you choose a certain restaurant. Now, if you have a great time there and a great meal, and it’s good food and good wine, a good ambiance, you’ve achieved your goal. And your friends will maybe congratulate you.
"Oh, great. You know, thanks for choosing this. It’s really, really nice."
If you have a poor experience, then you miss your goal, right? And you feel frustrated and not good about yourself because you’ve let your friends down and it reflects badly on you as well as an individual. So that emotional feedback helps us learn for the future.
We will choose restaurant A and not restaurant B in the future. Or, to win the game of tennis, we do this particular tactic and not that one. So that’s how emotions work—entwined with motivation as a gauge and as a feedback mechanism. But the underlying motivation is driven by this goal achievement.
What’s my current state now? You know, am I thirsty? Therefore, a goal is triggered to quench my thirst. Or I’m going to choose a friend to go out with for an evening, and I want to have a great time.
So my choice of friend would be different to if I’ve got a problem that needs to be shared and solved. I might choose a different friend who’s a good listener and someone who’s great with advice, but they’re not necessarily the life and soul of the party. It’s a bit like, which brand do I choose depending on the job to be done, and hence, what my motivation is going to be.
Is it fair to say that goals and goal—the language of goals, motivations, jobs to be done—that’s all the same territory? That’s sort of the beating heart of a brand, basically?
Yeah, yeah, I think so. And the other aspect of goals and jobs to be done, which are core—they’re central. But the thing to remember, because human behavior is dynamic, is that the goal or the job to be done can change depending on context.
So when I first started working with the guys from DECODE, they said to me, there’s a very simple equation in psychology: behavior is the product of the person times the situation. And I didn’t quite understand it at the time. But then they said, well, imagine you’re choosing an alcoholic beverage to drink, and you’re on your own at home, and you’re sitting down to watch a movie.
Yeah, you choose a certain beverage. But imagine now you’re on holiday with your friends or family, or imagine the temperature’s high or low, you’re in a work situation, you’re with colleagues or clients or whatever it might be. You change any one of those contextual variables, and it can directly influence your choice of beverage. The one constant in that is you as an individual—you don’t change. But because the situation changes, then our behavior changes.
And that’s the dynamism of human behavior and how jobs to be done have to be looked at in the context of occasion as well.
So this also is the foundation of that idea. I think this is you quoting that people buy categories first. This is to the degree that categories are defined by goals.
Yeah, yes, absolutely. And this—because if, like, going back to my example about I’m thirsty, I need something to refresh me, I need liquid—that’s a category decision. Or I need broadband, or I need car insurance—that’s a category decision.
So what motivates us at the category level is very important to understand first, because that then creates the context in which the brands operate. And this is all quantifiable as well. This is foundational research that we’ve been doing for many, many years, because once you have quantified those drivers at a category level, you can then profile the brands in the category.
And that enables you to determine and define relevance. So which of the category drivers does your brand own? And also distinctiveness—the degree to which your brand owns the drivers uniquely, or it may share the drivers with competitors as well.
And those are definable, and they are quantifiable as well, which is what makes it very exciting. And that’s exactly the same research Decode did with me as their client on the T-Mobile brand. But understanding the category is important because you can measure brand associations, and you can say, well, these are strong, and these are weak.
Yeah, and is that good? Is that bad? Is that relevant or not?
I don’t know, unless I’ve got a context against which to judge it.
How has this territory, the landscape of decision science, changed? I mean, I think the book was 2013. It’s been a long time. What’s the state of things now?
And of decision science or behavioral science per se? I mean, what I observe now from when, you know, Decode started 17 years ago, and I think at the time probably was unique in its offering. Now, there are many more vendors around.
There are many more undergraduate and postgraduate courses available around the world in decision science and its various component fields. And I see even client-side, you know, there are chief behavioral officers, even there are new roles springing up. And certainly, you know, what I observe on LinkedIn and through marketing and advertising media, the language that’s being used now is much more embracing of behavioral science.
And there have been some great proponents of it. I mean, people like Rory Sutherland at Ogilvy, you know, he has set up the Ogilvy behavioral science practice. He’s authored a book, *Alchemy*, and he has really led a fabulous charge for the whole industry worldwide, on why it makes sense to look at so-called irrational behavior.
So yeah, I think it’s generally more accepted. I wouldn’t say it’s mainstream yet. In some businesses more so than others, some have embraced it for quite a long time, but others are still like, "Well, we’ll stick to what we’ve always been doing."
So yeah, we’re still—I think it’s still on a journey.
Yeah, I had a flashback. So in the late 90s, I got my first job at a brand consultancy, and we would do groups with deep projective exercises and free association. I had never experienced any of this stuff before.
Our client was Clorox, which was like the only bleach brand out there. And in the exercises, these sort of heavy bleach users would—they kept generating, you know, we would ask if it were an animal, if it were a day of the week, and stuff like this. And the imagery was so amazing.
And it was like a doorway, right, to goals, the way you’re talking. Like I remember the first woman in one of the groups said she saw a snow leopard. And the snow leopard was—she imagined a beautiful snow leopard with some cubs, like this very maternal thing.
But if she went anywhere near it, it would rip her face off, is what she said. And it was just amazing, that little image that told so much, at least to me, and to us, to the goals, the motivations, right, of bleach, the role that played for her as a mother in a house. And also how, because she kept bleach in the garage, you know what I mean?
She didn’t get anywhere in the house. And so I wonder, I feel like we’ve come at this truth of human goals and how it motivates behavior, drives behavior. But I came at it through this qualitative imaginative exercise.
And you came at it the way you came at it. And I’m just wondering, how do you feel about free association, qualitative? And also as an addendum here, because I became a qualitative geek, I don’t—did you know Roy Langmaid?
Yes.
So I became like a total fan of his. I actually had a phone call with him in the pandemic. And I remember he wrote very critically of behavioral science. He says, no matter—there’s some quote, he said, read all you want about behavioral economics, you will not find people there.
Oh, interesting. Oh, that is interesting.
Yes. And so I wonder—that’s my thought. And I guess my question is, with the frameworks, and with the implicit association tests, which seem very dehumanizing in a way, but maybe I don’t know what I’m talking about. But what’s the role of qualitative in understanding the goals of your customer?
Absolutely. Well, when I—again, when I first started with DECODE, I was asking them about the implicit quant methods because I’d never come across them before. I was only familiar with qual or more traditional quant, shall we call it.
And they said, well, look, there are two ways to access the implicit mind and system one. You can do the projective stuff. You can do it on a therapist’s couch, you need to have real skilled operators doing that.
But it can be done. But the trick is to distract or exhaust system two, because otherwise it will get in the way, right? It acts like the policeman or the gatekeeper on what we self-report.
So you need a skilled moderator to be able to do that. But you can do it. The only issue then is it’s not scalable.
It’s not quantifiable, which is where the implicit quant testing comes into its own, I think. But you mentioned, I mean, the behavioral economics thing from Roy’s is interesting because I always like to say that economics lacked the person, because economics itself was just the pure theory and lacked the reality of human behavior. And that’s where the human came in.
But coming back to your thing about the snow leopard, it reminded me of our previous conversation about emotional and rational, because people often say, well, bleach, it’s functional, right? It cleans, it kills germs. That’s what it does.
But when you listen to somebody with this metaphor about the snow leopard, and it’s the same that we’ve done it with cleaning products, you get really rich, evocative imagery about protection, and caring for other people, and reassurance, even pride, and a sense of efficiency and expertise and control. Those are amazingly powerful emotional responses to something that otherwise might be dismissed as a purely functional product with rational messaging. And that’s why this emotional-rational dichotomy is just completely false.
And it’s also very unhelpful because there’s nothing that’s not emotional, you know, and the Clorox is a great example of that.
There’s a few different ways I wanted to go at this point. Well, I know George Lakoff—I don’t know if you’ve ever read him—he always—I can never find the quote, but he calls reason imaginative. He has a way of saying imaginative reason, which I feel like bridges the gap between the—allows all of us into that idea about how we’re thinking.
The way that you describe the relationship between system two and system one, between, what is it, reasoned or reflective?
Reflective, yeah.
It struck me maybe just because of the use of the word automated, that sort of AI is kind of a next level of—it’s almost a linear progression, these things that we’re learning. And we said when we outsource something to generative AI, we’re sort of automating something. It’s like a further refinement of automation.
But I thought because you were there when decision science and behavioral psychology and behavioral economics kind of arrived, we’re now at a point where generative AI is probably putting—what was the—it’s making people feel really uncomfortable probably about how things were and how they’re going to be. What is your thinking on the value of it and the implications of AI on marketing?
Well, I mean, I confess that with Decode, we’re spoiled because one of the co-founders actually did a double PhD at Caltech. So Caltech, the number one academic institution worldwide, right? So he did one PhD in neuroscience. Concurrently, he did a second PhD in AI, and he co-authored a book called *Understanding Intelligence*, which is still a reference work in academia. And 20 years ago, he was teaching robots to learn. So he then kind of parked it and said, "I’m going to focus on Decode and the whole decision science bit, but I have a vision for AI."
But computing power doesn’t allow it yet. But now it does. Especially the last few years with the development of processing power in the chips, it’s enabled his vision. So we have actually set up a sister company three years ago, which has built a suite of apps that predict and optimize creative effectiveness. Now, this is based on—this is predictive AI, it’s not generative AI.
But what he’s done is to take everything we know about mental processing—so attention, through perception, through memory, emotional response, semantics and semiotic associations.
And he’s trained these—our AI tools with human data. So we’ve got, for example, 15,000 hours of eye-tracking data, 5 million eye-tracking, static images, 450 million images, and the corresponding human tagging with words. So if I showed you a sunflower, you’d tell me, well, it’s yellow, it’s a sunflower, it’s got seeds, you make oil, blah, blah.
But then at a more semantic, conceptual level, you would say, well, it’s outdoors, it’s the sun, it’s warmth, it’s nature, it’s positivity, it’s uplifting, things like that. So we’ve trained the AI with that. So now if you showed it an image of a sunflower, it would learn, it has learned what humans associate with it.
So we’ve got this suite of tools now that mirror—they’re like a digital brain. They mirror all of these processes. So we can now upload a static or a video piece of creative, whether it’s a bit of point-of-sale material or a pack design or a YouTube ad or whatever it might be.
And the AI will give us some metrics that predict its creative effectiveness. And we’ve got principles behind it that help—if you’re not getting the metrics you need, how you optimize. So then, because of the nature of the AI, of course, you can then iterate.
So you can chuck 100 ads at it and pick the best three, and then do some evolution of those, and then test them again. And it’s done in minutes. And it costs pennies.
It’s ridiculous how effective it is. And this is being used globally now by about 500 brands in about 25 languages. And our biggest client is using it to test every one of their social media videos now.
So that, I think, is a beautiful marriage of the behavioral science, decision science principles, and understanding of human behavior and then using AI deep learning models to then predict how a human would respond.
Yeah, that’s amazing. Have you had any interactions with the synthetic users that kind of—that approach?
No, no, I haven’t. No, we’re not using synthetic data at all. I’m not sure if this guy, Chris, our co-founder, has examined it at all.
I know he’s starting to think now about the relationship and the loop between predictive AI and then generative AI, because in theory, that could be very interesting. You know, you could generate and then predict. And if it’s not optimized, then regenerate and predict, etc.
He’s very—Chris is always very clear to say, you know, this doesn’t replace humans. You still need the human to write the right prompts, for example, for the generative AI. And we always talk about it as augmented intelligence rather than artificial.
So it’s got, you know, it sits alongside. We’ve had examples where clients have—and to give you, this is a real example, and I think it’s very helpful. A client has an internal design team that produces all of the visuals for all of their brand comms, and they’re doing, they’re producing lots every day.
They had a change in their brand positioning, and their CEO said, "Guys, I’m not seeing a lot of difference in comms now, even though we’ve changed the brand positioning." So they asked us in to talk to them, and we sat the design team down, and we said, "What we really need to do first is to define what is on brand, what sort of visuals are on brand and what sort of visuals are off brand." Because only by juxtaposing those will we really be able to tease out this difference.
And what it led to was a really deep, sometimes heated, but very useful debate amongst the design team to really distill the essence of what they meant as on brand. Then once we’d got that, we were then able to select about 100—no, I think it was about 200 or 300—visuals that reflected it and train the AI. So the AI was trained with what equals on brand, and similarly, what equals off brand.
So the next time they got a visual, they just stick it through the AI and it gives them a traffic light. Is it on or off brand? If it’s neither, if it’s an amber score, go back to the team to debate.
But the team actually then upskilled. So by having the AI alongside them, they were then able to have much more useful and efficient conversations about the choice of visuals. And it cut—it sped up decision-making by a factor of about 10 inside that business.
So I think that’s a nice example of where the AI sits like an assistant, you know, it is augmenting human intelligence, not replacing it.
Yeah. Well, I want to thank you so much. I mean, it was a very kind and generous thing for you to do, to accept this invitation. I really enjoyed the conversation. So thank you so much. I really appreciate it.
Likewise. It’s a pleasure. Thanks, Peter.
Phil Barden on Science & Behavior