Transcription
Welcome everybody to the July live session of Mad Masters. I hope you're all sitting somewhere sunny if you're in the UK, and if you're joining from other parts of the world, I hope you're also having good weather there.
Uh, so we've got a fair few questions that have already been submitted in advance. So I will go through those. I can see people already putting questions in the chat. Uh, so we will try and get through them all today as well. Um, so welcome Rory for another live session. Uh, and we'll kick things off straight away with a question we had in advance from Justin Cunningham, who said, "I work for an industrial 3D printing company, think planes, trains, automobiles. Any advice experience around marketing to manufacturers in the B2B space, which can be old school and see innovation as risky?"
"That's interesting. Um, one way of doing it uh is literally uh something that that often seems counterintuitive to marketing because in marketing your natural urge is to tell everybody everything and to point out how brilliantly new something is. Now, you could borrow from the BTOC space, which I think was craft mac and cheese, where they actually improved the ingredients and made it much healthier. And then they didn't tell anybody for six months. And uh, what they then did was that they said, "What we're going to do is we're going to change." And everybody goes, "I don't want you to change it. I like it the way it is." You know, if you've made it healthier, it's probably going to be less tasty. All the usual problems you'd have if you say this thing is new. People will automatically find a reason why they prefer the old one to the new one. And so what they did is they changed it, said absolutely nothing for six months, and said, "By the way, uh, just in case you're not sure whether you're going to like this, it's what you've been eating for the last six months." And nobody's complained. Okay."
"Nice."
"So, one thing is not to tell people things are 3D printed because I agree. I've, I've actually ordered things and I've, I've worked out that they're 3D printed. There's a device. It's a bit nerdy, but there's a device with four suction cups, which means if you've got a glass roof in your car, you can attach your Elon Starlink Mini uh to the roof of your, the inside of the roof of your car with four suction cups. And it suddenly occurred to me, actually, this thing's clearly been 3D printed, but it's not advertised on eBay or Amazon as if it's 3D printed. It's just sold as a product."
"Right."
"And so there may be an interesting way in which you actually give people 3D printed products and then say, "I bet you didn't realize this, but that was actually 3D printed."
"True."
"And so there could be something really interesting you do. Uh, in which case you actually, if people are frightened of change, you don't tell them the new thing's different until they've discovered it for themselves and liked it. And it's then a kind of proof point. Well, you can't, you can't be unhappy with this because you happily, you, you've been happily using it for the last six months. And that always struck me as a really, really interesting strategy, which is what you might call non-obvious."
"Um, but it, it is a really interesting point, which is, I mean, one, you know, one thing you, you know, one thing you can do is you can do old school direct mail where you actually send them something. Um, uh, you know, as, now that's an expensive thing to do. Uh, but there's some, I agree with you that the weird thing about 3D printing, it's like many technologies where the, the most extreme example of this, I think, is video calling, where the early, not very good manifestation of a technology serves to kind of discredit it long after it's actually become excellent. And..."
"I, I think, I think that happens with quite a few, quite a few technologies. You know, the early iterations, you know, by their nature, just aren't very, very good. You know, the first cars were very unreliable compared to horses. Okay. And what then happens is that because you have that natural tendency for people to, "We've always done it this way," as I said, habit and social copying. So, the other one is to, the other one there is that obviously with habit, what you're doing by saying you've actually been using 3D printed objects for several months but you didn't realize it, that's a way of basically hacking the habit problem. And the social proof problem, you can obviously just do by saying actually, by, by actually using that old B2B, um, that very old and established B2B, um, principle of the client list, you know, that you, you know, and, and by the way, I mean, I don't think there's anything irrational about being influenced by a client list, you know, uh, this person has, you've been good enough to work with British Airways or whatever it may be. British Airways patently do a lot of due diligence and checking over the quality of their suppliers. Therefore, if you are on the client list of this entity, then it is a reasonably good heuristic. Not that you're perfect, but it's pretty good proof that you're not bad."
"Nice. Very well put. Um, so next question we had, this is from John, uh, who's always very engaged with these sessions, I must say, and he has said, "How quickly will AI take over education, law, accounting, medicine, and in terms of improving outcomes from that sense?"
"There are two conflicting things here, which is what it can do and what we'll do with it."
"And I, there is a very good guy at WPP, uh, excellent guy called Roy Marley, who's one of their senior people, who has a thing which he calls the empathy spectrum. And he argues that with things that are low empathy, they're purely functional. In other words, you can define the outcome pretty much in advance and using, you know, fairly robust metrics. He argues that's where you automate. Keith says that in cases where you require quite a lot of human understanding, of which advertising is an obvious case in point, in marketing,"
"Uh, that's a case where you augment. And I actually have a theory, which is that there's going to be a whole business created around AI for people like me, actually. Okay? Now, if I look back and I look back on the time since 1994, well, actually since I first used the internet in 1988, '87, before there was a web. Okay. If I look back at the number of hours I've had to devote to basically mastering interfaces, downloading apps, installing things, there's a bit of me which goes, do I want to spend, you know, the next sort of 10, 15, 20 years of my active life learning all this all over again?"
"Or do I need a community of a hundred people who live in Fuerteventura and the Canary Islands who are really good at AI and really into kite surfing, okay? And people like me pay them a retainer to have what you might call is a human butler or concierge on steroids."
"Okay? In other words, I ring the guy up and go, or gender-neutral obviously, and say, "I'm, you know, I'm kind of interested in going on holiday in Greece in March next year. Could you just automate a system which sends me interesting places to get to stay and interesting places to go?" Okay? I don't have to learn the actual, you know, the prompt codes or whatever it may be. Is this a problem you can solve?"
"Now, so, so I would argue that there's, you know, there's this interesting argument. I think that's quite interesting. If it's something where the right answer is kind of um, what you might call definable in advance, then you automate. Okay. On the other hand, if, if it's something where actually a a degree of human fellow feeling is required in the performance of the duty, then you augment. Okay. And I think that's quite, that's quite the problem I have is that if we look at most of technology and how it's been deployed by commercial businesses, the IT industry knows something very simple, which is the easiest way to sell to a company. Well, there are two easy ways to sell to a company, actually. One of them is fear. If you don't do this, something disastrous will happen. And the other one is cost cutting."
"Which is, this will save you money, basically by decarbonizing your workforce. You can fire people in your call center and replace them with AI chatbot or whatever. Okay."
"M."
"And one thing that concerns me is that we have a technology which should probably be used 50% for cost cutting and 50% for making people, making customers' lives amazing, and it won't be deployed in that ratio. It'll be deployed in the ratio of about 95-5."
"Yeah."
"Okay. Okay. So the number of times I can think of where someone has deployed an online digital technology at extra cost to improve customer service, in other words, where there isn't a cost-cutting narrative involved, strikes me as quite small."
"You know, you know, if you think about it, you know, we've replaced physical direct mail with email. We, you know, cost-saving argument. We've replaced parking machines, pay and display with um, uh, with parking apps. Now, sometimes incidentally, those things to some degree benefit the consumer or the customer, but that's not the motivation or the justification that's used for implementing the technology. It's nearly always done on a kind of cost-cutting basis. And so I don't know, my fear is that we have here a technology which actually has the potential to make life wonderful for human beings, i.e., our customers and consumers in general, and the only ways in which it will be deployed is where it happens to save money."
"Yeah."
"And that's something that worries me. But what else would I think about? Now, there's a problem as well, by the way, with you mentioned law and you mentioned probably accountancy, law and management consulting. Okay. Which alongside the ad industry are three ways in which, um, uh, you, three businesses in which predominantly charge by the hour. Now, one, we've got to ask a big question here, which is, if you make the thing faster, and your customers expect it to be faster, and you're still charging by the hour, fundamentally, you're going to make less money, right?"
"Yeah."
"Okay."
"There's a bigger problem in management consultancy and law, I would argue, which is that I was talking to a guy at Blackstone uh last week, and he said, "We have this massive legal firm, and we spend a fortune with them. To be honest, I don't want the firm. I want the partner. But in order to get the partner and his high-level advice, the way I pay for the partner is not by saying to him, 'Here's 20,000 pounds for a day's work.' Okay? Instead, I give him a job. He then brings in 15 Oxbridge graduates who also do the job. Okay? Who do all the grunt work. They charge £2,000 a day. He ends up being actually only 20 to 30% billable. His staff are like 120% billable, and a large amount of the money gets filtered up to him, the partner. In this case, he's male. Okay? Because he, in this case, is the reason why they're with the firm, right? He's the decisive factor. Okay? But rather than paying for that person, they, the way it does, you leverage effectively your more junior staff because if you've got enough of them working. Now, if you, if the work, if the partner is still an essential person, but the junior staff are now automated, the entire business model of management consultancy and law, and, you know, to some extent possibly accounting, okay, is completely disrupted."
"Yeah."
"I would argue that's really true of an advertising agency, and the advertising agency holding companies have failed to understand this. Certainly true of, less true of media agencies perhaps, but certainly true of creative agencies, that you're with the agency because of six people. Okay? However, the way it works is that the six people may come up with a brilliant idea for which you don't even get charged at all. But the way in which they get remunerated for that is that you, is that you give them the grunt work and you allow them to make a fairly healthy 20% margin on the grunt work. Okay? Which pays for the existence of the David Abbott or the, you know, uh, you know, or, you know, you can just imagine. Okay. Now, if you take away the laborious stuff, okay, or, or the value of it goes down precipitously, the entire model of McKinsey or the entire model of Freshfields or Slaughter and May is fundamentally undermined."
"Yeah."
"And that strikes me, you know, in those particular fields, because of the way they pay, if they charge for outcomes, okay, this wouldn't be a problem. In fact, it'd be great because, you know, if you were Slaughter and May and you said, "Yeah, if we win the case, you give us five million quid." Then what you'd have is literally a guy who would do his brilliant, uh, delegate the grunt work to the AI, go off to a golf course, right? And come back, charge you five million quid. Now, admittedly, it's bad news for sort of middle-ranking lawyers, but it's fine for the business model of the firm."
"Yeah."
"If you're charging by the hour, if you're charging by the hour, I don't understand this. Right."
"You just come into the chat earlier. He said that's exactly what's happening in the accounting industry right now. There are no longer any junior roles, junior due to seniors using AI instead. As a result, there's a second problem, which is I would argue that an ad agency is actually, if we're being honest about how it really works, okay, is it's 10 brilliant people, 10 people who are on the cusp of being brilliant, um, a hundred people, another 10 of whom, but you can't identify who will also become brilliant in five years' time. Okay? Now, if you eliminate the bottom 80% of what's going on, uh, then potentially, you know, maybe what happens is that, you know, exactly what happens, which is Dave Dro writes a book or reinvents himself as a podcaster or whatever."
"But, but where's the feed, where's the feed line? Where, where's the inflow of future talent? Okay? Because most, you know, most partners in management consulting firms are really valuable because they did, you know, they did 25 years before the mast, you know, you know, you know, ahead of the mast. Okay."
"Yeah."
"Um, and so this absolute, you know, eradication of, uh, in, in some ways you can argue, well, you know, I mean, you know, maybe for, you know, a, a really 55-year-old, really senior M&A lawyer, they can just reinvent themselves as a sole, as a sort of sole trader. Maybe that's, maybe that's what has to happen. But for the overall business, this is a huge problem. And no one in advertising seems to have spotted this fact, which is, if we were paid on commission, right? And media commission, AI would be the best news the industry ever had, right? Okay. Okay."
"Oh, look, we need to rerun the American Express travelers check. This is me in 1991. We've got to rerun the American Express travelers checks. Automate media placement. Kaching! 1.5 million quid. Wow, we're off to the races. Okay? But we're not, we haven't been paid like that since 1992."
"Well, sorry to interrupt you there. If you want to watch the rest of this live stream, you're going to have to join the full edition of the Mad Masters course. Now, what is Mad Masters, you might ask? It's a course that blends behavioral science, creativity, and data. And it's all delivered to you by the one and only Rory Sutherland. Through monthly live sessions, on-demand Netflix-style modules, loads of bonus sessions, and even an in-person meetup at the biggest marketing event in London, there's 30 plus hours of content delivered to you exclusively by Rory. So, what are you waiting for? Pull the trigger and register now."