Rory Sutherland on the Fat Tail of Marketing

[00:00:00] Rory Sutherland: That was best interview I've ever had, actually. Best podcast one of the best podcast interviews I've ever had.

[00:00:05] Phillip Jackson: Oh, thank you so much.

[00:00:05] Rory Sutherland: Absolutely fantastic.

[00:00:06] Phillip Jackson: I... You had a little ––

[00:00:09] Rory Sutherland: But it's kinda right, isn't it? It's a framing problem. It's a paradigm problem, literally.

[00:00:12] Phillip Jackson: You had a little T.S. Eliot in there, a little Marshall McLuhan.

[00:00:15] Rory Sutherland: Yeah. We're... This, it's all there.

[00:00:19] Phillip Jackson: If 10% of your work could produce 130% of your revenue gains, would you still be obsessing over dashboards? This is Rory Sutherland, and he says that every marketer is chasing certainty and that hunt for certainty is killing creativity.

[00:02:03] Phillip Jackson: Let's bring it back to life. Rory Sutherland is a legend. He's the chairman of Ogilvey, and he is the poet of persuasion. He's also championing this new idea called the "fat tail," where small moves are creating giant gains for brands. And he should know, because we met in London and we talked about it at Klaviyo London, where, among other things, he was ripping a fat vape, and he was telling me how in this search for certainty, we are forgetting that it's creativity that powers the future. And, how AI and agency are emerging at a time where brands are reduced to nothing more than a chat interface with your customer. I'm Philip Jackson. This is Future Commerce. Stick around because this conversation might just change how you value marketing.

[00:02:55] {Intro Music}

[00:02:55] Phillip Jackson: We have one or two of our own concepts that I'd love for you to, maybe, riff off of, if you're so keen. One is something that we call the "multiplayer brand." And I believe that the world is becoming more multiplayer. You see how brands, I think, are less in control of their own narrative in that.

[00:03:22] Rory Sutherland: A lot of categories have been what you might call artisan, haven't they? So in other words, you know, in everything from bread to gin, you've effectively seen this massive kind of diversification of many, many smaller brands.

[00:03:38] Phillip Jackson: Sure. I would even say that maybe, in terms of... Where the consumer has more control en masse of the direction of a brand than those in leadership of a brand. For instance, the backlash towards the most recent Apple ad campaign. The direction of a brand in general, I think, is more turning towards the ability for the crowd to steer. And now with generative AI tools that brands are looking to build for consumers to be able to dream, for instance, Nike just created this AIR Imagination tool for you to be able to design your own shoe. I think we're starting to see small glimpses of, rather than consumers using off the shelf tools like ChatGPT or MidJourney to dream what a brand's output could look like, we start to see in the future––I'm talking five, ten years from now––brands giving consumers the tools for consumers to co-create with.

[00:04:44] Rory Sutherland: I mean, there's an interesting possibility, which––

[00:04:50] Phillip Jackson: Possibility.

[00:04:51] Rory Sutherland: ––that the whole direction of marketing changes. By which I mean, that that in a weird kind of way, the to use a metaphor... Okay, it's rather like reversing the polarity of an advertising agency, which is previously a company employed an agency to find customers. And you may actually see a world where, effectively, consumers appoint agents to find suppliers. So you could actually see a complete reversal of polarity in the whole way in which commerce works. You could also, I think, see something which could be disruptive much sooner. I mean, you're already beginning to see things like that, smatterings like that, and things like the housing market, where traditionally it was the seller who appointed the agent. And now it may be the buyer who appoints an agent to find a property, which, by the way, isn't even necessarily on the market. Right? Now, the other thing that might be a more rapid change, which is interface change, is always disruptive because if you change the interface within which people choose and act, you fundamentally change behavior. I mean, quite significantly. You know, I mean... It's still largely said Amazon is an interface business as much as it's a logistics and retail business.

[00:06:09] Phillip Jackson: Sure.

[00:06:09] Rory Sutherland: Okay. You know, it's an interface, which people are very comfortable using. It is currently the default, whether it's the app or the website to which you would go if you wanted to buy something, you know, which you weren't sure how to find. What happens if––and this is a big if––let's say voice or some other new AI mode of interaction simply becomes the preferred mode of interaction and replaces, let's say, typing. Okay? That's actually quite significant. Because, it's a chance for what you might call entrenched advantages of habit and behavior suddenly to change very dramatically. So I'm gonna give an example. I mean, we can date this all the way back to the '19... Well, actually, 1980s, I guess, when direct brands started to appear.

[00:07:02] Phillip Jackson: Sure.

[00:07:02] Rory Sutherland: Where, in 1986––and I know this because I was doing it at the time––if you wanted to buy a flight, you went to the high street and you went into something called a travel agent. And you talk to them about what sort of flight you might want to buy, and then you bought it. Okay. And then easyJet, for example, or in The UK, US rather, it might have been JetBlue or Southwest came along and effectively, you no longer went to that trouble.

[00:07:26] Phillip Jackson: Disintermediated the...

[00:07:28] Rory Sutherland: Disintermediated the whole thing. But as well as disintermediating it, which is kind of the economic side of the story, you fundamentally changed the way that people thought about––

[00:07:36] Phillip Jackson: The interface. Right.

[00:07:37] Rory Sutherland: ––and, and chose...

[00:07:39] Phillip Jackson: But they gave a better experience as well to the end that...

[00:07:41] Rory Sutherland: Yeah. All kinds of reasons. It was open twenty four hours a day. You didn't have to leave home. I mean, there are enormous number of benefits to all this. You know, you could act on impulse. And, I don't know what the future of a decision making interface is going to look like. So, you know, one hunch might be that it may be an avatar you talk to, that in many ways replicates a travel agent.

[00:08:10] Phillip Jackson: Mhmm.

[00:08:10] Rory Sutherland: Okay? Where you say, you know, "would it be any cheaper if I went on Wednesday?" Or, you know, you know, "are there any routes which route via Spitsbergen?" Okay. It could be a mixture of voice and video. It could be a mixture of voice and photography. It might, by the way, not even be immediate... In that, you may say that it's the constraints of the current interface which force people to make a decision right now. What about an interface that says, I'm interested in going on holiday somewhere warm in January, February or March 2026. Can you just feed me interesting ideas for the next six weeks? And I may help you refine them. I may express opinions on your suggestions. Now, there are purists who I think get it wrong. They suggest that the way this is going to work is find me the perfect toaster and buy it.

[00:09:04] Phillip Jackson: Correct. Right.

[00:09:06] Rory Sutherland: I don't think humans can make decisions without the aid of comparison. So to me, a more likely result is, "find me five really good toasters."

[00:09:16] Phillip Jackson: Right.

[00:09:17] Rory Sutherland: And let me explain their strengths and weaknesses and let me choose between them.

[00:09:24] Phillip Jackson: Which is how every consumer reports has always worked.

[00:09:28] Rory Sutherland: It's always worked. There are economists who would argue that if the machine can optimize for the consumer, why would you bother giving the consumer a choice? I would argue that's a ludicrously overambitious aim on the grounds that, one, I have no faith that I've made a good decision.

[00:09:44] Phillip Jackson: There's a fundamental distrust, right?

[00:09:46] Rory Sutherland: And also, fundamentally, the value of something is comparative. You know, there are all kinds of interesting reasons there.

[00:09:54] Phillip Jackson: You remember—–

[00:09:55] Rory Sutherland: You know, undoubtedly, you will perhaps use, just as we might appoint a buying agent if we're looking for property, you might well appoint a toaster agent. Now, one other intermediate possibility, by the way, which is worth exploring, is that the interface in part remains human, but it's a human superpowered by AI. Now, at the age of 59––which I am––if I look back on the number of hours I spent of my adult life, since 1994, effectively perfecting the use of interfaces, learning new interfaces, apps, new operating systems, everything else. It is an extraordinary proportion of my life as a learning curve. And there's a bit of me which goes, "do I really want to spend, you know, five of the remaining years of my life continually having to master the use of AI prompts? An alternative, and I only mean this speculatively, would be there's a concierge service of humans who are really good at AI and who like kite surfing. And they go and live in Fuerteventura or the Portuguese coast or whatever it may be, okay? Well, they spend an awful lot of time kite surfing. But when the weather's bad or it's dark or they don't feel like kite surfing, they can earn a pretty good hourly wage by sorting out people's problems with spectacular efficiency through their superior use of AI. Now, we shouldn't reject this theory, although most people who believe that life is an efficiency optimization game would regard any use of humans as suboptimal. They'd say that that has to be solved by technology.

[00:11:42] Rory Sutherland: But I can see certain people using this. I mean, what's one of the most unbelievably robust forms of employment in The United States and The UK? Essentially, it used to be called domestic service. It's now called cleaning, you know, or childcare. In the same way, it's not ridiculous to me, famously, Simon Cowell doesn't have a mobile phone. He has people around him with mobile phones who relay the information back to the extent that it's actually relevant and useful. I don't think what I'm hypothesizing there is completely impossible, that you may actually... If you think about it, let's take a personal shopper highly alert to fashion and brands. Okay? Who is empowered by a comparable skill in AI that matches their taste, refinement and general knowledge in our sector. You could actually create a new form of human employment. You know, if you could appoint someone who could automate the process of being really, really useful in terms of serving him information about holiday destinations based, on a growing and deepening knowledge of my own tastes and preferences. That's not an impossible interface either, by the way, because it's rendered economically feasible by the fact that this person who, in an old-fashioned world, as a concierge service, they might be able to serve 30 people, and here, they might be able to serve 300. You know, there's a very, very big difference between paying a personal shopper £30 a month and £300.

[00:13:22] Phillip Jackson: So, let me ask you a question then. So you made this really great point, I believe, about having this world where we feel like we're more in control because we're given a set of options. We have maybe a set of agents or people that are powered by agents around us that are giving us more control. That makes me feel like, you know, there's an existential question around a set of free will. Like, I have been given a choice, right? I have a choice because I have free will. I have a choice. I don't want to be given just one choice that something has optimized for me.

[00:13:56] Rory Sutherland: Autonomy is one of the things that we prize for its own sake. If you take away our feelings of autonomy, we feel fundamentally powerless and lapse into a state of learned helplessness.

[00:14:07] Phillip Jackson: Which I think... So, then, this comes down to something that we talk and we've studied quite a bit about, is that... There's been a homogeneity cycle in the way that user experience and user interface design has emerged. Right? And so we've––

[00:14:23] Rory Sutherland: Which makes sense in a way.

[00:14:24] Phillip Jackson: It's utilitarian. Right?

[00:14:26] Rory Sutherland: Right. A familiar interface is system one.

[00:14:29] Phillip Jackson: For sure.

[00:14:29] Rory Sutherland: And an unfamiliar interface requires cognitive effort. It's a bit like that fact that when you're driving down a familiar road, your brain switches off, and you suddenly realize you've been driving for five miles without any conscious awareness of what was on the road. Okay? Generally, we don't do that when we're in an unfamiliar city.

[00:14:47] Phillip Jackson: Sure. And yet we're novelty seeking. Right?

[00:14:49] Rory Sutherland: I know.

[00:14:50] Phillip Jackson: Right.

[00:14:50] Rory Sutherland: The fundamental conundrum of the heart.

[00:14:52] Phillip Jackson: It is. And what I think is really interesting is that we're also now gravitating towards the most essential interface that can ever exist, which is a chatbot. It's chatbots. And I think there's an essentiality to this new experience that, I think, there might be a––you're talking about existential questions––that a brand might have a challenge around how do you have a more branded experience in a user experience that is fundamentally brand-less. Right? So, we've never had as much control of brand in this world. And we also are leading ourselves into a world where we have a lot less control over the brand experience.

[00:15:40] Rory Sutherland: I think large components of what constitutes the value of a brand will still remain. I'm not one of these, sort of, Scott Galloway doomsayers who argues that the whole thing is dead. And we'll simply use other heuristic. You know, we'll use ratings, reviews or AI recommendations to make our purchase decisions. I think there are a lot of reasons why brands work, not least because it's reputational skin in the game, and that you can trust a famous manufacturer or supplier slightly more than you can someone you've never heard of, for the simple reason that they have a lot invested in their reputation, and therefore, they have more to lose by disappointing customers. And they're more––by the way, they're also more exposed to scandal or opprobrium or anything else. I also think we use brands actually as handy heuristics for making decisions. In other words, they help us reduce choice, to an extent. "I don't want to buy one of those because, you know, I'm not planning to spend a huge stack of money on this. It's not that important to me." Versus, you know, "this brand is exactly the brand that is effectively appropriate for my price point," for example, "or appropriate for the degree of interest I have in a category," whatever it may be. I think there are... I think brand value is actually multifarious, And it's completely wrong to treat it as if it's actually a single homogeneous entity, "The Brand," because actually it serves all kinds of purposes. I mean, famous will always be valuable, and that's partly because people will come and find you rather than you having to find them. And people will respond to things that are famous in a way that is generally more trusting and requires less anxiety, than, you know, something that's totally obscure and unheard of.

[00:18:44] Rory Sutherland: So, you know, simply the value of fame and familiarity is a human constant. I don't think that's gonna change. Nonetheless, the nature of the way in which we build brands, and the way in which consumers use them, it's going to change. It really has a lot. I mean, you know, it used to give you a stranglehold over distribution. In other words, if I'm only going to stock three things, I'm going to stock the three most famous things in the category. Yeah, it doesn't really apply online in a case of limitless SKUs. We've seen also categories which used to be dominated by one or two brands that have kind of gone artisan and fragmented. So, yeah, I mean––by the way, this change is actually...it's always been...the pace of change may be accelerating. And also, the simultaneity of the change may be surprising as well. Previously, I think a lot of technological change, a lot of behavioral change took a long time to permeate through the population. And now, of course, this is why interfaces are so weird. You know, you can change your whole interface for everybody simultaneously in a digital world. That was never true of a physical shop. A refit was something that basically spread out and took years. But I mean...there are, by the way, when you talk about the existential question and also the philosophical question, there is a concern I always have about technology, which is the extent to which a lot of technology arrives as an option and ends up as an obligation.

[00:20:25] Phillip Jackson: Say that again.

[00:20:27] Rory Sutherland: Okay. So, one thing we've got to be alert to, is that...This is a bit of a banal example, but I think it's quite fitting. Let's take the parking app, which you'll all be familiar with. Okay, when it first arrived, it was an alternative to a pan display machine where you had to walk 100 yards, put some coins in or a credit card, detach a sticker, affix it to your windscreen, and generally fart around. Okay? The parking app was a wonderful addition to this, just as the ticket machine was a wonderful addition to the manned ticket office at a railway station. You know, if I know what I want, if there's a queue at the manned ticket office, I can actually bypass it and use this slightly more con convenient alternative. And at that stage, it's all fine and dandy. Until, of course, people discover that parking apps require less maintenance and are vastly less costly than machines, and...suddenly, what you find is the machines have been taken away. And the app's the only alternative. Mhmm. And then it becomes more stupid, still, because, you know, there are elderly people. In his nineties, my father never had a smartphone. He couldn't have––well, probably by text–– he would have found it almost impossible to park anywhere. Parking apps start to be used in underground car parks where there's absolutely no signal.

[00:21:49] Rory Sutherland: Parking apps have an absurd system where, combined with automatic license plate recognition, you will get fined if you pay for your parking six minutes after you arrive. Now, that would never have happened in a human-mediated environment. Right? So you start getting this degree of automation which is oppressive because it's no longer there as an alternative. It's suddenly an imposition. And you could argue the same, by the way, of the smartphone and indeed the mobile phone. Yeah. When you were one of the few people who have a mobile phone, it was great because nobody expected you to have one. You could do things that other people couldn't do. It gave you comparative advantage. Okay? When everybody has a mobile phone and it is assumed that you have one, suddenly, the balance goes from what you might call optionality to obligation. Mhmm. And I think this happens with a lot of technologies, and I think we need to be alert to it because something that arrives as an I I would I would take an example. Okay?

[00:22:49] Rory Sutherland: McDonald's, I believe, in The UK, is starting to actually open outlets which are screen only. In other words, there is no option of going to a human being and actually saying, "I'd like a Big Mac and fries." Or, as you might sometimes do, which a machine can't do, "what have you got that's ready right now, because I'm in a hurry?"

[00:23:09] Phillip Jackson: Right.

[00:23:09] Rory Sutherland: Okay? I'm not that bothered about what I eat. I'm bothered about how quickly I can eat it. "Okay. I'll have those two." Now, now, suddenly, what you have is that the cost savings of the new interface are visible and quantifiable and immediately measurable. The opportunity cost is invisible because you can't measure the things that people aren't doing or buying because you're no longer making it possible for them to do them. And I, you know, generally, the lesson of retail is to be multichannel, isn't it? The more ways you sell, the more you sell. Probably, the same lesson applies to payment, by the way. The more ways you offer people to pay, the more people buy. Alright? And I would say a third example would be around delivery in ecommerce, which is one of the most heinous things I think e commerce players do is they offer you no choice of who the delivery provider is. Right. Now, if I'm paying for delivery, a lot of people have a very strong preference for USPS, UPS, FedEx, DHL, every whatever it may be. A lot of people also have a stronger version to two or three of those entities. Let them choose. Why does that not happen? I would argue you would sell much more if you made it absolutely explicit at the beginning of the buying process that you will get to choose who delivers and when. Why doesn't it happen? Because that decision doesn't rest with marketing, doesn't rest with people who are looking for growth, it rests with procurement. Sure. And they say, want to funnel everybody through this one provider with whom we've got a deal. So one, I can justify my own existence to pointing to the cost savings we manage to negotiate. Okay? And two, we'll get the maximum rebate for which the procurement person who made the decision gets the credit.

[00:24:58] Phillip Jackson: And I would––

[00:24:59] Rory Sutherland: I would argue this is abominable behavior. I'm paying for delivery. Right? A lot of people will not place an order because of the fear that they will suddenly then receive an email saying, "your X is being delivered by Y," who are known to be particularly useless in their area. And it varies by individual.

[00:26:40] Phillip Jackson: And, Roy, having worked in that industry, too, for 20 years...I would have endless conversations with people in how to measure holdout tests and how to implement the testing of such a hypothesis, too, with––

[00:27:54] Rory Sutherland: Do you think I'm right, by the way?

[00:27:56] Phillip Jackson: I think you're right.

[00:27:57] Rory Sutherland: And, by the way, have to measure repeat purchase. You'd have to...it's not a short term cost saving.

[00:28:04] Phillip Jackson: The question shouldn't even really be, I think, is...Nobody in the process would even be asking what's best for the customer. The challenge here is, in our industry, there's a really strong desire path about going to the self-service of the chat. Right? And they're creating their own experiences. And that's how we started this conversation, where we just kind––

[00:28:39] Rory Sutherland: But the feedback of the chat can be remarkably interesting because it'll reveal, rather, as the call center does--

[00:28:45] Phillip Jackson: Right.

[00:28:45] Rory Sutherland: --the problems consumers have that they can't solve for themselves online. Okay. So it's the only feedback loop for where you're falling short. I mean, very clever...

[00:28:55] Phillip Jackson: I mean like the AI chat, right?

[00:28:56] Rory Sutherland: An AI chat, I mean, it's all, it's still chat.

[00:29:00] Phillip Jackson: Right.

[00:29:01] Rory Sutherland: At its best level.

[00:29:02] Phillip Jackson: Right.

[00:29:02] Rory Sutherland: And by the way, it should also allow you...It should allow you the option to request a human. Because otherwise, you could get stuck in an endless loop of mutual incomprehension.

[00:29:15] Phillip Jackson: Yeah.

[00:29:16] Rory Sutherland: If your query or your problem was slightly out of the ordinary and outside its training set. Okay? You have to give people the option of the human. Just as I think of McDonald's, I'm an enthusiastic adopter of on screen ordering. I hate queuing. Okay? But there will be occasions for me, and there will be people of different temperaments who want a conversation. And what we're failing to realize, unless you have some sort of free chat mechanism, is that there's a whole form of customer feedback, which the online world never captures because you never anticipated the problem in the first place. I mean, Amazon, interestingly, I think one of the most admirable things they do is the call me back button. If you've got a problem...

[00:30:01] Phillip Jackson: Right.

[00:30:02] Rory Sutherland: You're logged in, you identify the item with which you had a problem. You identify the nature of the problem. You hit call me back, and about three seconds later, your phone rings. You don't have to go through any of that identity stuff, Okay? Because they know who you are already, and they know what your problem is. You get straight into the gist of the issue. The most peculiar thing to me is that I don't know a single other business that's actually copied that. And they say to me when I recommend it, they say, "oh, we can't do the cost-benefit analysis." And I go, "that's already been done. Amazon tests everything." Okay? Amazon wouldn't have kept this thing alive for seven years unless they had some evidence of knowing that it worked. Just steal it. But no, no, no. The burden of proof of marketing is far too... Marketing is in a very, very unfortunate position, okay, where it's held accountable for every single unit of cost that it incurs, but it is completely inadequately credited for the value it creates. Okay? So we have this strange world of accountability, which doesn't mean overall accountability, in the sense of that a hedge fund, or a VC, or a pharmaceutical research entity would understand it, which is, "overall, do we have more breakthrough successes than the cost of activity?"

[00:31:21] Phillip Jackson: All of which, by the way, are measured in timescales of years and decades.

[00:31:24] Rory Sutherland: Timescales of years. Right.

[00:31:26] Phillip Jackson: Not in quarters.

[00:31:28] Rory Sutherland: Yeah. Okay. So brands are built over years. Why do we measure them in quarters? Right. Is not me, but someone else said it. Brilliant point.

[00:31:36] Rory Sutherland: Marketing is fat-tailed. Okay? This is Nassim Taleb who says this, probably the largest expert on fat-tailed distributions, by which I mean 10% of what you do delivers, in some cases, 130% of the value. But you don't know what the 10% is in advance. That's a process of constant iterative experimentation. Now, we have a world in marketing where every time you do something that fails, it's chalked up against you as a cost. However, when you do something that succeeds, at the most, you can lay credit to its effects over a three-month period. Now I know of marketing ideas, and I've had maybe one or two myself, okay, which have delivered, I'm not exaggerating this, at a cost of 25,000, maybe £50,000 value of an incremental––this is in UX typically. Okay. £10,000,000 a year of incremental high margin revenue in perpetuity. Well, you know, that's probably worth, I don't know, you know, the value of that is 50,000,000, a 100,000,000. You get paid £25,000 for it. Okay? But after year one, after which it's been used to justify none of that future revenue, which amounts to––

[00:32:47] Phillip Jackson: Is tied back to––

[00:32:48] Rory Sutherland: ––is tied back to the marketing activity. Okay. So it's a bit like, imagine what publishing would be like if you only were given royalties on the first five days of book sales. Well, first of all, no one would write a serious book.

[00:33:03] Phillip Jackson: Right.

[00:33:04] Rory Sutherland: Secondly, the only books people would write would be clickbait books. You know, sort of, "Why JFK Was A Pedophile." Right? Which would sell massively, and then everybody would realize it was rubbish and then never buy it again. Okay? Right.

[00:33:16] Phillip Jackson: Sure. {Laughter}.

[00:33:16] Rory Sutherland: Now, so fundamentally, if you have this massive asymmetry in terms of cost attribution and value attribution, it's a massive distortion on the practice of marketing. Never mind agencies, which suffer from the same thing in microcosm. I mean, there are ideas that Ogilvy has had which have made a million––sorry, a billion dollars for clients. And, they would acknowledge that, by the way. We had an idea that was undoubtedly, it's still going, by the way. The client is still using the idea. Okay, I'm not going to say what it is. And it has undoubtedly contributed to $1,000,000,000 in sales and massive brand-valuing changes and so forth. We did the calculation to work out how much money we made effectively from generating that idea. It's about $200,000 Now, if you have a case where what has happened is finance, and to some extent tech, and very deterministic uncertainty-averse people in consulting are trying to create a world where every quantum of effort and expenditure is matched to a quantum of revenue...Business does not work like that. It's a systemic thing, Okay, where 10% of what you do covers 180% of the other 90%. {Laughter} Okay, it is not a pointillist world where every unit of expenditure is attributed to by a unit of value. It's not like that. Some moments are priceless. Large swathes of time are useless. So, trying to take attribution down to that incredibly narrow granular level is basically a false god. The reason marketing is fat-tailed, by the way, is because marketing exists in the real world, and the real world is fat-tailed. Your life is fat-tailed. If you look at the things that really matter, I don't know if you're married, I don't know if you bought a house. I don't know how you got your first job. But all of those things were probably slightly freaky occurrences, which could very easily not have happened. They weren't the product of a long term plan, a cost-benefit analysis. They weren't the product of any of those kind of mental processes which are beloved of the kind of finance people, or procurement, or anybody else. And often, by the way, you made those decisions on the basis of things that weren't really numerical metrics. Now, you know, I assume you didn't choose your partner on, you know, on a spreadsheet.

[00:35:53] Phillip Jackson: Right. Yeah. We didn't sit down and...

[00:35:54] Rory Sutherland: No. Yeah. But Charles Darwin kind of did.

[00:35:58] Phillip Jackson: {Laughter}

[00:35:58] Rory Sutherland: He actually did an upside-downside comparison chart of why he should get married. (His ultimate decision was totally emotional, of course. But it was, you know, it was merely an attempt.) And, by the way, going through those emotions is healthy. Right? But, allowing the process to determine the outcome at the exclusion of anything that can't be effectively reduced to a number or a metric, that's a terrible way to make decisions.

[00:36:29] Phillip Jackson: And that's where the hypothesis then here in the fat tail is that, it's the––maybe there's an allegory––the best time to plant a shade tree is twenty years ago, and the second-best time is today, right? So, we have to, to be able to best prepare for a future in marketing, we have to be making decisions that pay off beyond the horizon of this quarter's payoff.

[00:36:56] Rory Sutherland: Yeah. And yet if we're not allowed to lay claim to any of the revenue line as the product of past marketing activity...

[00:37:07] Phillip Jackson: So, you're making a call for a change in the way we do attribution?

[00:37:11] Rory Sutherland: Fundamentally, it's a total change in the way we look at attribution, which is to stop making it within an arbitrary and limited time frame. Now, what we're driven by is the whole process of internal accounting in a business, which is itself completely arbitrary, including the time frame under which we think, why a year? It's calendar year. It's a completely arbitrary setting.

[00:37:37] Phillip Jackson: Yeah.

[00:37:37] Rory Sutherland: Or, indeed, the quarter. Okay? And, so, the need for the need to pretend that everything is thin-tailed and deterministic––which it isn't––is effectively destroying opportunities to create value, which are fat-tailed and probabilistic.

[00:37:56] Phillip Jackson: I couldn't agree more.

[00:37:59] Rory Sutherland: And this is––by the way, I said this to Scott Galloway last week. I said this concerns me a bit as a marketer, for entirely selfish reasons, and because I think that marketing can occasionally solve absolutely enormous business problems...But if the same mentality is now being applied to R&D and experimentation in general within the firm, that's a double tragedy. Now, Peter Drucker said, since the purpose of business is to find a customer profitably (and keep them, I would add), the only two things that add value are marketing and innovation. Everything else is a cost. Now we almost have a world in which everything else is treated as extraordinarily valuable, whether it's HR or finance or whatever, but marketing and innovation are looked at as a cost. So, it's complete perversion and reversal of the way in which Drucker looked at a business. And most businesses have this mentality because they're very comfortable with internal incremental optimization. Okay? Because it's deterministic, predictable, and kind of makes sense in advance. Okay? They literally have this mentality that value is created in the factory. Value isn't created in the factory. It's created in the mind of the consumer. Okay? It's and it's also created in the laboratory, I would argue. You know, there are two ways. I would say marketing and innovation are the same thing, or they're two sides of the same coin. Okay? There are two ways you can create new value in a marketplace.

[00:39:23] Rory Sutherland: You can either find out what people want and work out a really clever way to make it, or you can work out what you can make and find a clever way to make people want it.

[00:39:30] Phillip Jackson: Mmm.

[00:39:30] Rory Sutherland: And the value created is indistinguishable regardless of the direction of travel. By the way, getting the maths wrong, okay, that's what led to the financial crisis of 2008. They thought it was a thin-tailed distribution of probability where you could effectively discount the possibility of highly significant events. And they got that wrong, and it half destroyed the world economy. Now similarly, we're destroying the innovation economy by demanding a level of attribution that is simply too short term, too specific, too granular. Great, great line of this, by the way. Henry Ford, he got some consultants in, and they said, you seem to have a pretty well-run shot, mister Ford, but we don't quite get one thing. There's a guy down the end of the corridor who seems to sit around all day with his feet on the desk, but you pay him sort of $70,000 a year. And Henry Ford replied, yeah. That's absolutely right, he said. But the thing is, two years ago, that man had an idea that saved me $5,000,000, and at the time, his feet were exactly where they are now. Now that's an appreciation of uneven value creation. Okay? At the moment, what we're trying to do is pretend that a business is like Newtonian physics, okay, with a sense of proportionality and unchanging laws. The laws of psychology, you can rewrite them yourself. Okay? Laws of physics, not so easy.

[00:40:55] Phillip Jackson: We're tending to rely on AI a lot, and especially as we're starting to look more at technology for more automation and for more marketing creative. A lot of the conversations that we're having, especially in

[00:41:09] Rory Sutherland: This is excessive, by the way, which is not to say that AI is unimportant. Right. And I call this the sourdough effect. Okay. I'm really, really delighted that sourdough bread exists, but the extent to which it dominates the artisan bread category at the expense of cornbread, soda bread, and lots of other forms of interesting baked goods. Of course, as a Brit, I'd have to include crumpets and muffins and that. Okay. It's simply ridiculous. Sure. Okay. And in the same way, Zoom and Teams, the adoption of video conferencing is really, really important. Okay. That could conceivably allow you to completely restructure the way your company works, where your employees work, what hours they work, the terms of employment. It creates a spot market in human talent. Okay. Yeah. The fact that you can actually gain from someone's expertise without needing them to be co located with you is extraordinarily important.

[00:42:02] Phillip Jackson: Overreliance on it, it creates it creates an unbalanced We're very unfair,

[00:42:07] Rory Sutherland: By the way, because most people go, oh, I don't think it's as good as a face to face meeting. And I go, 90% of the Zoom calls you have would never have taken place.

[00:42:13] Phillip Jackson: They never would have happened.

[00:42:14] Rory Sutherland: For sure. It's a completely false comparison.

[00:42:17] Phillip Jackson: Correct. And I guess this is my question. When we're looking at AI, AI is programmed with a specific set of biases, right? And it's programmed with a specific set of rules and assumptions. That's, you know, a corpus of knowledge.

[00:42:35] Rory Sutherland: There's a danger that it's effectively and neither optimistic nor pessimistic. I'm baffled. There's a danger that what it's doing is kind of enshrining groupthink.

[00:42:46] Phillip Jackson: Sure.

[00:42:47] Rory Sutherland: It's taking groupthink and effectively engraving it.

[00:42:50] Phillip Jackson: Do you think that the thin-tailed mentality is the corpus of knowledge that's programmed by default into AI?

[00:42:58] Rory Sutherland: There are areas where...so Roger L. Martin, who fantastic business writer, argues that the three ways you can really produce creative strategy breakthroughs, I would add humor to the list, by the way, are analogy, anomaly, and resolving a trade off. Okay. Now what I want to know, I know that AI can't create a decent joke. So there are clearly there are clearly failings in its mental capability.

[00:43:27] Phillip Jackson: Okay.

[00:43:30] Rory Sutherland: At least there isn't a recorded instance of AI generating a joke that humans find funny. Maybe that'll change. I can see AI being quite useful at the analogy job, which is what are things we could learn from other industries in a different sector that could perhaps be applied to the business we're in ourselves? In other words, if you take the Uber map, how could you do that with train travel? Well, the answer might be better passenger information. Sure. Okay. I can I can see that there are functions which it could do actually to a level which I would deem imaginative in that I will actually look at something, know, I never would have thought of that? Yeah. I don't know what it's like in anomaly detection, and that presumably depends how you actually brief the thing in the first place. Okay? And I don't know how good it is at resolving trade offs. In other words, finding a way which says these two things have always been treated as as in opposition to each other. Actually, with a bit of ingenuity, you could actually have a win win. I'll give you an example of this. It suddenly occurred to me. Okay. If someone came to me from a body that wants to encourage British people to go to the cinema, and we have a long conversation. And, of course, we talk about the competition that cinema faces because you have streaming services, Netflix. In many cases, films now appear on streaming services at a cost while they're still at the cinema. And we're talking in this automatic idea that the streaming services are in opposition to the cinemas. And then suddenly, I had a weird thought, which is why doesn't Netflix sell cinema tickets? Okay. Now let me explain why. Okay? Because they're not actually in competition. Okay? Unless you frame it very narrowly because, one, if Netflix earned a couple of pounds for every three cinema tickets it sold, let's say, it would be incremental revenue on top of its subscription revenue.

[00:45:32] Rory Sutherland: If it discounted those subscription services to its customers, it would be another reason to subscribe to Netflix, which I could use to justify my subscription. But also, nothing is better for Netflix than that I and my family spend an evening at the cinema Mhmm. Because we're not watching anything on which they have to pay streaming royalties on. So in fact, Netflix would be ecstatic for three different reasons if it sold me a cinema ticket and I went and watched the cinema instead. Okay? Now what I mean is that there are sometimes occasions when what you might call what Roger Martin calls integrative thinking, design thinking, can actually take something which is universally assumed by everybody to be an either or and turn it into a both and. Okay. That's fundamentally what a lot of creative work does. It takes something that seems to be a trade-off. We're number two, so we try harder for Avis. Oh, dear. The problem is we're not as big as Hertz. I know what we can do is we can reframe being number two, so we turn it into a source of perceived advantage. Okay? Good things come to those who wait for Guinness. Oh, Guinness takes ages to pour. We'll actually turn that into a benefit. And you can do that, by way, in psychology more easily than you can in physics or chemistry. You can do it in physics. I mean, you know, in chemistry. I mean, a catalyst is to some extent, you know, a resolution of a problem. You know? It's equivalent to a kind of reframing, if

[00:47:00] Phillip Jackson: You like.

[00:47:01] Rory Sutherland: These two things don't react together very well, but in the presence of vanadium, they miraculously do. Okay? You know, you can do it in physics, which is, oh, dear. We need we've only got this much energy, and we need to boil a liter of water, and it's not enough. What do we do? We'll actually take the water up to 20,000 feet, and the boiling point of water is lower, and therefore, it's problem solved. Okay? You can do it in physics. Okay? You can perform this. But in psychology, you can do it in so many more ways. And I don't think we spend enough money on what I call psychological r and d. Simple as that. And that's partly because marketing, which should be effectively an r and d function, it should be part of the r and d function, has been weirdly placed as this weird cost center, which is nonsense. I mean, how can bringing in a new customer be framed as a cost? But if you work in finance, apparently, it's completely natural for you to think of it like that.

[00:48:02] Phillip Jackson: I think that this is the conversation to be having right now. I think that this is precisely the stage to be having it on. I'm so excited that you're the person to be bringing it to Klaviyo, and I'm honored that you would bring it to the future commerce audience. Thank you so much for joining us.

[00:48:19] Rory Sutherland: Absolute pleasure.

[00:48:20] Phillip Jackson: Thank you very much. Rory Sutherland.

[00:48:22] Rory Sutherland: What a pleasure. Thank you very much indeed.

[00:48:23] Phillip Jackson: Appreciate you.