Les & Sarah’s big review of the year

32:15So another benefits of character ads, I suppose, that you kind of bypass all that stuff. Yeah, so you can do that with characters. But also if you have some kind of universality of a story or an emotion or a feeling or whatever. So I think there's a quote in the book. Was it you that said, or you were quoting, you don't have to be black to appreciate Shakespeare's Othello. You just have to understand jealousy. Yeah. And, you know, you don't have to be old. to appreciate the Amazon ad. You simply have to, you know, have, understand friendship and, you know, about longstanding friendships. There was one really interesting insight because we've been working on this feeling scene kind of series of studies for three years now, actually, looking at how different people feel represented in ads.

33:04And there's a lovely summary from it, which is when you see yourself in advertising, you feel more happiness, you feel stronger emotions, you're more likely to engage the brand right so there's a lovely headline story and we called it how diverse advertising unites a nation as well yeah and how actually everyone else also feels good seeing other people represented so lovely story great one group of people for whom that's not true older people so we suddenly did the study on older people going oh let's just let's see this happen again and then i got this funny call from the team going john we've got a problem i'm like what do you mean it disproves your law and i think that's how you represent older people You've got to get the emotions right.

33:43We had to call it Wise Up, which is actually the way to make older people feel good about themselves in advertising. Give them a lead role. Make sure it's funny. Show them as being smarter. Tap into the years of cultural references they've got. Make them useful to society. It's not like, oh, we see them sailing into the sunset. That's not it. They don't want that. So it's very interesting nuances, actually. But this is where Amazon got it right because they were leading, they were funny, it was a twist. You know, they had connection between them, all those things. Yeah, emotional connection. We talked, didn't we, about that, one of my other favourite ads, that Southern Comfort ad of Live Comfortably.

34:22I mean, not a classic representation of who you would want your drinker probably to be, someone with a pretty big beer gut kind of swaggering up a beach. But, you know, we all identify with that feeling of being totally happy in your own skin. And a brilliant example of casting. And it's just all about interesting casting, really, isn't it? This isn't tick box stuff, but just characters that leap out at you and you want to be with and you can identify with and tell stories brilliantly. That's what it's all about. You mentioned characters a little bit earlier, didn't you? And it actually brings us to our number one campaign nicely, which is Audi's Kevin the Carrot.

35:03So on top of the poll again, the only advertiser, apart from Coke, because they've been going 25 years, the only advertiser to get five years, five star in a row and continually get better and better each year, which is spectacular. But in a way, they've deployed that character fluent device, as Orlando tends to call it, doesn't it? Which has universal appeal again. So it's a lovely example of that. Yeah. And he tells emotional stories. We're not all carrots, but we can identify with that. And interesting, as you say, it started off as a three score, three stars, something like this. 3.6. That was going to be my question.

35:38Yeah. So has it sort of worn in over time? Popularity has grown. Two things have happened. So there's definitely wearing, but there's also, what they've done is they've kept the campaign idea, they've kept the construct the same. Yeah. But what they do brilliantly every single year is they play to sort of, you know, a slightly different story. Yeah. Willy Wonka this year. Yeah. You know, last year it was Christmas cow and so on. And then they also tell relevant jokes in there. So, you know, there's a few kind of, you know, panto-like, kind of this year's joke here. But what they do, actually, every one of those jokes is tested, right, to make sure that it lands, it gets the right response.

36:18With you? Yes. Yeah, yeah. So every second. So how do you test? So we start quite early in the year, actually, with a long animatic. So it might be usually about four minutes. So imagine the story, a black and white animatic, really stretched out. So voiceover, the jokes, kind of all done in an animatic style. And then what we work out as we see, does happiness and surprise change when that scene happens or the joke lands or the voiceovers? And literally we do that and we edit down to 60 to 90 based on the longer version. So it's thoroughly tested. Forensic, isn't it? Very forensic. That's interesting because I know I've heard people say in the past that humour is one of the things that's hardest to test with animatics.

37:02But I think when you're dealing with an ad which is effectively cartoon style, you're closer to the finished film. It would be much harder to test, say, Rowan Atkinson. An actor's comic timing. Delivery and all that stuff. That's very hard. But because it's animatic, it's almost like a cartoon, isn't it? so much easier. And it's quite voiceover heavy as well isn't it? It's not like you're relying on a lot of expressions and kind of expression humour. There is public, there's physical kind of knockabout humour. And the Jim Broadbent voiceover has become a fluent kind of voiceover even doesn't it?

37:37And that's a craft thing as well isn't it? That's craft, yeah. One of those things that you don't really I think it's probably an ad that's very easy to overlook how they make it look very easy don't they? But there's a huge amount of thinking that goes into that. Massive. And very layered as you say like a great panto where the adults laugh at some stuff and the kids don't really get it or whatever. You can watch it two or three times and get little different things. It's one of those ones that you do discover, oh, I didn't spot that joke about Willy Conker and all that stuff. Yeah, so I, you know, this is probably a terrible thing to admit.

38:04I don't see many apps because I, you know... Breaking news, ladies and gentlemen. Let's be there, doesn't watch ads. I don't really watch television. Double breaking news. Let's be there. He recommends television. Yeah. Yeah, so I'm quite insulated from a lot of actual ads unless I see them during the working day. I think this is the first time I've seen a Caravan and the Carrot ad. Is it? I think. Wow, yes. And so I think it had less effect on me. I thought it was nice. I thought it was charming, et cetera. You're a three-star level. Yeah, I think I'm at the three-star level. So I've seen it once, and that's the first time I've seen any of the campaign.

38:46so I think I'm at the three star level and you know you talk about jokes and I was thinking are there any jokes in there and I think oh there's a couple of things Did you get the big willy joke? You didn't smell that one Ah well there you go you have to watch it again Look out for the big willy joke I can see his bum or something his bum crack or something like with the plum or plum crack So I think that's an example of how Wherein works You begin to know and like the character and you're expecting certain things and you see repeated exposure you get the jokes but i'm not at that level yet well let's bring excel in uh or the ipa um ipa effectiveness gold was last year uh kevin uh the kevin campaign and what's lovely about audi is that you know that they call it in the case study discounting the other variables so it's a wonderful it's a wonderful example where they've been able to isolate the impact of advertising because we're all trying to do that but there's so much else going on at the same time but over the period in question Audi weren't opening stores their discounting policy was the same so they were able to go well actually all the other variables are roughly the same and they could look at the market share gains of Audi each Christmas and I mean you'll know this but Christmas is the one time where people are most open to changing the store they visit and usually Audi loses share on Christmas because they want to go up to M &S right yeah sure sure hand yeah and that that trade down has got smaller and smaller and smaller every year and then the trade into in new year where you go, I've got to tie in the belt.

40:14It's gone up and up and up. So it's just a wonderful, I think it's the best case study I've ever seen of advertising's effect on business results. I mean I'll be misquoting them but it's like a very big percentage of the number of people that visit Audi stores most of it happens in the 12 weeks after Christmas. They can attribute back to the campaign which is amazing. It's wonderful to see and the ROIs and stuff are just fantastic on it. And I can see how it works. I mean You look at that ad and, you know, admittedly, I've seen one ad in the whole campaign ever and immediately it makes Aldi feel Christmassy, which, you know, it wouldn't naturally do.

40:55Well, actually, you guys are to blame, actually. I don't know if you know this. I don't know if you know this, but the original brief was Beat John Lewis. Yeah. Because the whole idea around Aldi, everyone knows it's cheap. We know that, right? Yeah, yeah. What we have to do is become John Lewis at Christmas for people. And that was, see, in a way you can take some credit for it. And they were one of the first campaigns, I think, weren't they, at Christmas? I remember the first year that Kevin came back, it was like, ooh, you know, Kevin's back again. That's a little bit of a surprise. You don't normally see that.

41:26So they were kind of doing the consistency job and campaign before other people have started. The other thing I love about it is, from Marcus's point of view as well, is how it works everywhere. You know, you can buy the toy in store. Kevin pops up and has conversations on social media. It's on the side of the bus, not the bus, the truck when it delivers the store. You know, they really integrated it and very cleverly throughout, which is brilliant. So doing the long and the short, you know, it's kind of working in both senses. Yeah, good for them. The other thing, looking through the year, actually, so looking at non-Christmas ads as well for a moment, because we're also having a bit of a Christmas theme.

42:01There's also some consistency in fluent devices. So a little tip to my former employer, actually. So one of the top scoring ads this year was Ribena. Now, I've got a bit of inside intel on this because having worked there, but the Ribena brought back the Ribena berries. And I don't know if you've seen those, but the system when we tested it, the Ribena berries are one of the most recognisable kind of distinctive assets out there. And they got a 5.6 star. And again, it shows what another top tip I think for advertisers is to look back at what your brand owns as distinctive assets. I think we're too quick, aren't we, to sort of crumb up with something new and we don't use something else.

42:44Oh, absolutely. How long was it since they'd used the berries? Did you work on that? Well, I was there from 2016 to 2018. I remember this funny situation where we were making a new ad. and then i asked this casual question the team i mean i got in so much trouble for this and and i'm sure they're listening and we'll go john but um i said yeah yeah no let's let's do this new idea as long as it beats last year so while you're testing the new idea do you mind just testing what happened last year i've got this funny text message going oh good news bad news um good news last year was a 5.6 star the best ad that's ever been run in soft drinks in the history of ever sort of thing you know bad news is it's gonna be kind of hard to justify making a new one this year It was like, you know, that kind of thing.

43:28But it was a lesson for me, actually, because I didn't know, even though I was the marketing director, I didn't realize how good that asset was for the brand. I just didn't have the history and understanding. And so many, you know, marketing directors come and go. It's one of the greats, the new John Lewis positioning, I guess, has got this tension inherent in it, hasn't it, about letting your traditions grow. I mean, that is what you have to do as a manager of brand comms, isn't it? you have to decide what to keep, what you move on subtly and what gives howls of protest if you chuck it away, really.

44:03So it's about managing that, isn't it? Not chucking out your black currents because that's your strongest asset, but keeping it fresh every year. So I think everyone's getting better at that now, I think, aren't they? Rather than just defaulting to rip it up and let's do something new. This is what our friend Ritson did say to me the other day. He was like, is this the year we finally understood consistency? Yeah. It does feel like that, doesn't it? It does, yeah. The penny has dropped. I mean, we wrote about that. Yeah, we did. But we were way ahead of it. We were really struggling to think of any campaigns, weren't we?

44:33Gary Lineker was still going for walkers at the time. We had Felix Cat Food, which is our little cartoon Black and White Cat, which we worked on as an agent for 20-odd years. For 22 years. But we were really sort of scrabbling around trying to find anything else. There wasn't anything. Okay. So talking about consistency, you've got a good Felix IPA case study on the matter, haven't you? Yeah. Well, so we wrote this one in about 2002 or something like that. Consistency killed the cat. No, it wasn't that. Continuity. Continuity saved the cat. I can't remember. Something like that, yeah. And it shows the kind of the long roots of the problem.

45:14I think the business in general and then the marketing community and the advertising community off the back of that became obsessed with innovation and disruption. You know, it's all part of that mantra that, you know, comes out of Silicon Valley and all that sort of thing. Innovate or die. Yeah, exactly. and innovation is important and you know particularly when it comes to product and so forth but we bought that so much that we forgot about the other side, consistency and continuity and great great campaigns are about continuity and freshness you know, actually Kevin the Carrot is quite a good example of that in that you you know, you've got continuity but you've got something new each year, I gather.

46:09And in that Felix case study, we contrasted what Felix was doing, which was running a consistent campaign for 22 years, you know, the same little black and white cat, and often running the same ads. The very first one, the piano player one, I think they dug that out year after year. They bring it out every so often. versus Whiskers, which was at that point TBWA, I think. And they totally bought the mantra of disruption. And they were doing something new every time. And what we saw was that they were outspending us four to one and we were beating them. We went over that period from being the fifth or sixth biggest brand to beating them to become the number one brand, you know, on a quarter of their spend.

47:05Love it. Just with a little black and white cat frisking about to music. I love it. And that was a great example of that consistency breeds popularity rather than the other way around. Because people got more and more familiar with it, loved it more and more. We started doing tea towels, didn't we, and all the rest of it. So I found myself the other day talking to a bunch of marketers somewhere abroad, and I won't tell you what category, but as I was talking to them, it became really apparent that they totally bought into the idea that marketing was about innovation and purpose. You can imagine how I reacted to that.

47:43And I said, actually, clever marketing is when, you know, if you're a really clever marketing marketer, you can sell the same old stuff with no innovation whatsoever. You know, the real challenge of marketing is to sell more stuff when you've got nothing to say when you've got no new product when there's no you know killer fact in the ad but just to do you know you know to do the the same thing really well i know and we talk about marmite in the in the book where you know marmite was a brand that had no significant innovation for about a hundred years you know until marmite squeezy you know that was the one another good example and still grew Yeah, another good example of that is Guinness.

48:33So I don't know if you know, this year, Guinness became Britain's favourite pint. Again, one product, one idea, consistently executed. Really distinctive assets. Distinctive assets. Banging soundtracks. Soundtracks, yeah. Bit of animated stuff going on. And I suppose... Banjo beer advertising. Banjo, yes. You feel good. I mean, there are other ways of doing things. And actually, I suppose we've provided a big counter example with John Lewis in that, or on the surface it appears that we have, you know, because people have asked me over the years, you know, why don't you run the same John Lewis ad each Christmas and then go back.

49:13But I think what we did there was there was a consistent strategy, you know, the thoughtful gifting strategy, the idea that you're telling a story about an example of finding the perfect gift for someone you really care about. So that's, you know, the idea underpinning was the same. And the whole load of consistent stylistic things musically, you know, the palette of how the ads look and so on. But the other thing is what we did was to create, if you like, a ritual, you know. And again, a bit like the holidays are coming, you know, this thing of there would be a big reveal, the big ad that would make everyone cry.

49:55And so there was a kind of continuity and consistency in that, even though superficially there wasn't. Quite a hard one to do because to make it work, you've got to come up with an absolute banger every year. Yes, exactly. That's the pressure. I mean, well, unrelated to the ads, but the thing that used to make me laugh about the John Lewis was the guy that had the John Lewis Twitter handle. Yes. Well, the American. Yeah, in Canada or something. Wasn't he an American politician or something? But he used to get flooded with people talking about him for like a week when the ads came out. He won't have that problem anymore, will he?

50:43Now, actually, just before we go on to the next topic, two interesting stats for you. we tested the very first TV ad ever to air on TV back in the early 1950s I think 1965 it was a it was a toothpaste yeah yeah you got it it has to be 65 because that's when commercial oh no hang on is that I've got that wrong haven't I 65 commercial TV yeah it was 50s or 60s I can't quite yeah I can't remember the date but it was for a toothpaste plummy British a proper BBC news kind of accent scored over three star isn't that amazing now i know there's probably nostalgia and it just you know there'll be a little bit intriguing but it does go to show that would put it in the top 15 percent what percentage of database would have been around at that time not many no no not many exactly it's black and white i know you tested some of john webster's ads didn't you that the martians were up there yeah pretty well up the top when they probably Probably the best example, Hovis Boy on a Bike, 46 years.

51:49Well, we tested it several years ago, so it's now 50 years old, that ad, but was five star. I think this sort of slightly leads us into a point that we were talking about before we started recording, which is to interpret results like that, you've got to put them side by side with comparables. so for example hovis boy on a bike i'm sure part of the emotional response from that will be from people who remember it you know there'll be a nostalgia thing and not all the people in your group will be people who remember it but they and and similarly you know you show a black and white ad from the 60s it will have an emotional impact simply in the same way that seeing something really old yeah you know does so in order to assess the emotional impact of ads you've got to put them alongside things that are in some way similar so hovis boy on the bike versus cadbury's everyone's a fruit and nut case for example um and um like categories so this is a very very good point and i'm glad you raised it because i wanted to ask you about this so um i've pulled off the average star rating of the categories this year and ranked them, there is a big difference.

53:06Because I know we were looking at the top ads this year. Sarah, I know you and I were having a chat about this. Lots of sugar in there. Well, there is. There is. If I read out the top three in a number three Magnum, a new Magnum variant, number two, the advert for the British Bake Off, number one, Cadbury Dairy Milk, glass and a half, right? And let me tell you the top three on average in at number three, baking. So boy like. Number two, ice cream. Number one, chocolate. So the best ads mirror the most emotive category. Yes. You know, you see. Actually, I've done a little bit of work on chocolate advertising, not for a long time, but earlier in my career.

53:48And one thing, so econometrically, one of the things I noticed is with chocolate, anything works. You know? Yeah. It's really easy to sell chocolate. It's the same with ice cream. I would have an ice cream. It's very impulse, isn't it? Just have to remind you that it's there. Literally, it's called impulse in the category. Yeah, particularly impulse chocolate because you go into a corner shop. How many variants are there to choose from? A million. You only have to say chocolate and people literally start salivating. I probably will myself in a second. So you could sell chocolate by just going, you know, aero.

54:22and that's it you just say the word aero immediately sales go up you know so it's not hard to get a response I thought for example Bake Off there'll be first one you know people think like cakes but secondly people love Great British Bake Off and so again you could probably run a test of the programme yeah you could run a still that said Bake Off next week you know people go oh I love that great ad So there's sort of like built-in feeling to the category that… There is. And this is really important. We're coming to sort of… But just before we talk about why that's important, because there's a very important point that comes up all the time when I'm talking to advertisers about this.

55:04Before I do, here's the harder question to both of you. What do you think the lowest performing category is? And I'll give you some leeway here. So any of the bottom three, what do you reckon? How about ball bearings? Ball bearings, okay. Probably not a lot of that in the database. I was going to say pile creed. There's probably not a lot of that either. So ball bearing category would be what? So industrial products. Industrial products. What do you think, Sarah? I say that simply because I've been talking to a ball bearing manufacturer. Breaking news again on this podcast. Well, it gets you some effective ball bearing advertising next year.

55:43Sarah, what do you reckon? Bottom performing category. Cars? Cars are an interesting one. they're statistically exactly average. And the interesting thing with cars is all advertisers are hugging the average line. So there's like zero difference between the best and the worst. It's like the most stereotypically average. In fact, I think cars have been using AI for 10 years. Longer than that. Because if you use AI for cars, you get the same thing again. I was trying to think. I do think there is a problem with finance. They're low. Actually, what I was thinking particularly was things like life assurance and pensions.

56:19um because it's that's close because it's quite it's a again i've done a bit of work in that area it's both boring and frightening yeah you know thinking about you know whether or not you're going to starve to death in retirement no one wants to think about you know what it's one it's one of the only so one of the one of the only awkward moments i've had doing the job i do is because usually when you measure you test with the national population and then you test with your target audience the score goes up right it's obvious right that's why we do targeting right and I was working with a major pension provider and they got a one star.

56:54I'm in that awkward position. I was like, I'm in quite a lot of the time. But don't worry, we'll test it with your customer audience. I'm sure it's going to go up. Oh, how wrong I was. It went down. Yes, it reminded you of your inadequacy. They did not want to be reminded of, you know. And so it's quite hard to do good work in that category, but it is possible. I mean, you know, for the life you don't yet know. That was Ally Dunbar, wasn't it? Yeah. I think. And, you know, they managed to, you know, they had music, humour. Yeah. Paul Whitehouse stuff, wasn't there? Yeah. Or Aviva. Actually, interesting.

57:29So humour's quite good at diffusing stuff. It's a serious case for humour. Yeah. I think more generally in society, one of the things that humour can do is allow you to to go to touch areas that are sensitive you know push boundaries say truths that you couldn't say explicitly in a po-faced way and so on i think that's quite a good example it's never a truer word than spoken in jest yes yeah yeah exactly exactly you got you got close you got very close i think i think financial service is a mid mid uh mid to high one star the bottom well i should go So the third least popular category is gyms.

58:12The second most least popular category is slimming aids. And the least popular category is sport betting. Sports betting. It's interesting the juxtaposition with the eating and not staying healthy. Yes, it is, isn't it? I hope the government's not listening to this, by the way. Because we might end up with a bit of... So sports betting presumably is that there'll be a certain amount of antipathy to... It's probably regulation. because actually one interesting thing is actually we work in the US with a major manufacturer of health products I'll say, medicines and in their category they actually by law have to state the negative side effects that come with the category.

58:52Now if you think about sports betting and their category average is 1.2, it's the lowest on our entire database globally. Now of course with sport betting you have to have all the terms and conditions and you have to have the gambler wear and that sort of thing so it's I'm pretty sure it'd be government regulation that's driven that down, which is a point of it, is actually to suppress the impact of advertising. So interestingly, again, going back to finance, financial advertising is heavily regulated as well. And I remember working again on stuff to do with life insurance and pensions. I remember that the company I was working with years ago, they had more people in the compliance department than in the marketing department by a factor of something like four or something like that.

59:32And the ads were so heavily regulated that effectively you couldn't say anything, which was tremendous because it meant that they ran ads with no messages in, which really worked. There was no option but to do emotional advertising. And it worked. I love it. Actually, one of my, isn't there, there was a Millwood Brown study, wasn't there, where they found the more messages you have, the less gets remembered. Yeah. It's like the same thing, isn't it? Yeah. As you say, the more you get remembered. And the more you crank the rational bit up, the more the emotional bit goes down. Exactly. Yeah, yeah, yeah.

1:00:06Real treasonous. So, you know, people often sort of say, oh, you know, well, what's the optimum number of messages in an ad? And you go, yeah, yeah. One. One. No, the answer's... Nil. Nil, exactly. Now, I'm glad we talked about this because one of the things I have to explain to people, when Orlando built the kind of system on methodology, inspired very much by your approach, IPA database again, looking at... And what he did is he looked at the relationship between ESOV and market share changes in the following year. And then he worked out the emotional response and created, I think he called it the kind of creative amplifier sort of thing.

1:00:39But what's really interesting, and we have to explain this to people, because obviously we do the five star system and everyone kind of celebrates that because it's simple and intuitive. But actually the prediction is actually related to your performance within your category context. So what he found is that your ability to outperform your category average creative, exactly as you do with ESOV, is like a creative weighted ESOV kind of measure. So if you're in ice cream, you better be doing five-star advertising. Equally, if you're in sport, betting, slimming or gyms, a two-star is going to take a look at that.

1:01:14So category context is actually, so when you get into the data, the contextual thing is critical. And I suppose the other thing is that in order to really outperform your category, you probably want to be doing something that's a little bit different. And that's one of the reasons why I think there will never be any hard and fast rules about what is effective in creative terms. Because one of the rules is to break the rules. Yeah. So, you know, for years I've been thinking, you know, the John Lewis thing, we're not going to keep getting away with this forever because other people will copy us.

1:02:00And, you know, there's more and more of that kind of advertising around. Sooner or later, if there's if a formula works, other people will copy it and then it will stop working. Yeah. Well, and we're kind of seeing that at Christmas, aren't we, with the, you know, the system on testing. Yeah. You know, you can kind of see it coming through. that when you were talking about actually the lack of a distinctive asset per se for John Lewis and when we talk about handwriting didn't we in Tony and that is the stuff that's just so widely copied now isn't it? Yeah, yeah. We've had it to ourselves for so long.

1:02:32Where are babies? I was just musing on your list. They must be up there fairly high. Yeah, very, very high. The couple of categories actually that will probably round out the top five or six pet care. So we always joke about putting a dog in your house. But often when you looking at you go geez there's a very average dog ad in the top 20 kind of on our database you know so again with that you need to do something more interesting with the dog than just the dog so that's one um and babies of course i mean in fact the super bowl winning ad was it this year or last year was huggies um the only super bowl ad to get a five star was i mean it was brilliantly funny brilliantly cute you know it was very very well done only super bowl had to get a five star it's Rare, actually, yeah.

1:03:13Every year there's one or two. Well, here's an interesting thing we should all celebrate, by the way, as UK advertisers. The average Super Bowl ad is, and we test about 80, and Christmas is about 80 UK ads. If you say Christmas is our Super Bowl, right? Same sample size. Average star rating is 2.7 for Super Bowl. Average star rating is 4.3 for Christmas in the UK. So the standard is demonstrably different. However, to turn your own point against you, I mean, Christmas advertising, if you think of it as a category, is going to be, you know, Christmas is an emotional time and then the products are going to be, you know, indulgent food and stuff like that.

1:03:58I mean, we used to say Christmas is half a star. Probably now it's a star, yeah, roughly speaking. Very good point. The category contest is enormous. Now, I asked our friend Ritson to pop some questions over for both of you. So we'll give these a go. He's obviously in his Excel spreadsheet, actually, as well, because we've got some quite good, quite techie ones here. Anyway, he wrote a column recently on the power of synthetic data. Very interesting, actually. So I don't know if you caught this story, but he created this perceptual map of car brands and then revealed that he'd done it on ChatGPT.

1:04:34And then he went and tested it with a real human survey responses. and there was a 90 % overlap between what the computer was able to generate, scanning the internet, versus what people actually felt about these car brands. Now, 90 % feels like within the realms of statistical significance and, you know, pretty... I mean, most people would take 90 % as a... If I can get 90 % of it right and save time and money, I will. I guess the question for you two is, like, that being the case, what else can you do with, you know, with AI to get shortcuts in? Lots to say about that. I mean, it was a great article.

1:05:10I was on holiday when it came out and I was quoting. We were waiting for a bus in Mallorca. And I was quoting bits of it to my family. And my wife was going, shut up. You're on holiday. Yeah, you're on holiday. Les doesn't take holidays. And it was classic Ritson. Poor wife. Classic Ritson because it was, on the one hand, there was some very powerful and interesting ideas in it. Plus also the best description of programmatic advertising that I've ever read, which was describing it as a black box full of turds and spiders. It was brilliant. But yeah, I mean, I think I've been messing around with ChatGPT, you know, for a while myself.

1:05:54And it's powerful and interesting. And I think as an industry, we probably kind of got the wrong end of the stick in that I think the really powerful uses for at the moment are actually in the research area rather than the creative area. Creative that's generated by AI, I think once you've seen it a couple of times, you recognize it pretty much instantly. But it's a great way of speeding up research. I actually so I'm slightly biased because I know the people behind the synthetic data stuff uh they are they are great guys and and they um they they will they're doing interesting things but I think the other thing that you learn from messing around with AI is you have to be really cautious with it because it sometimes just completely fucks up and it will tell lies and it will make things up i mean somebody once said it's a bit like it's like having a really clever um intern who's completely unscrupulous you know who will go off you know just go and bring me data on that and you don't know whether they've made it up or not so so you do really need to double check everything but the but the speed with which it can come up with that yeah stuff that i mean you know a quick way of testing hypotheses putting putting out failures for things what areas do you think it'll touch i mean can it help with my media mix for example or which which audience i should target those kind of questions i don't know i mean i think i think we should all be playing with it and finding out but the point is that that um i don't think you can use it well unless you already know quite a lot of stuff because you don't know whether or not to trust what it's telling you I think it can go badly wrong.

1:07:41As you said, I've had a bit of experience of it when it's already in pre-testing. We're working with one client who used sort of AI pre-test to save time and money. And yeah, you can get really bummed steers from it. I mean, one where it wasn't understood that there was a celebrity in this ad because that couldn't be kind of detected by the methodology. It completely changed the response to the whole ad if you didn't know that that was this certain person. So proceed with care. I can give you the data on that. So we've been using AI to compete with ourselves. Can AI outpredict the system one approach?

1:08:19The correlation between the system one prediction and market share change over 12 months is 0.66. That's been tested 264 times in different studies because we continually update our validations. the correlation between ai trained on our data so this is not this is ai with the biggest advantage you can imagine because it's using our own data correlate whatever 0.33 so it's not there yet it's not there yet but interesting on creative another thing i'll chuck in as well so we helped with a study looking at could uh so we gave some art students a project to with a fictitious brand to come up with some posters.

1:08:59And then we gave AI the same brief, almost identical. So one group, the students got 1.9 star, the AI got 1.83. There was one cell within this experiment where we had to ban the result and they actually came out top. The reason we had to ban it was they used both. So they tested with AI. They then used their judgment to improve it and they retested it with AI. And that is the smart way forward. And they got 2.8. So I think that was actually the story. It's another both-ism, isn't it? Yeah, well, exactly. And this is exactly what I find with messing around with ChatGPT, is that I think if you just used it and say, well, that's the answer, you get a lot.

1:09:43It can do some things quite well, but you get a lot of stuff that was not right or just very bland or very generic. But if it becomes a dialogue between a smart person like myself and this tool, then you can do something a lot better. And I mean, the really clever thing about it is the speed because that means that you can explore research avenues that you just wouldn't have the time to explore. A little example, a trivial example, I'm reading Trollops, The Way We Live Now at the moment. At one point in the book, one of the characters says something which is about America being bigger than Britain.

1:10:28And I thought, oh, okay, so at what point did the US population pass the population of the UK? Was it around about this time, 1870s? That's quite a hard thing to answer with Google. And even with Google, you wouldn't bother. But I asked ChatGPT and it came up with an answer. And immediately I thought there's something wrong about this answer. The figures didn't stack up. so had a conversation over the course of about a minute or two and then eventually got ChatGPT to admit that you know, made a mistake, I misspoke about, and then eventually got to an answer that was satisfactory but all of that stuff kind of like, I would not have bothered to do that if I didn't have a thing on my phone that allowed me to do that Was that in the bus queue as well?

1:11:15No, I was lying in bed reading my book I'm sorry, I'm a nerd but But you have to watch it. You do. I've actually done an experiment to prove this, a rather funny one. So I had this very negative review about the podcast earlier in the year, right? So this guy filmed a seven-minute video doing a takedown of the episode I did with Bob Hoffman and basically going, oh, he's not uncensored. In fact, he called me tyrannically English bullshit. it was tyrannically polite english bullshit anyway that was how he kind of summarized the review and it was a properly you know properly i in a way it was quite funny actually so i you know when you get this situation you think well what do i do do i get offended do i just treat it as banter and i decided to treat it as banter and just enter into the spirit of this guy and he'd copy me into everything and that kind of thing anyway i i messaged him on late too and said look with respect there are plenty of podcasts out there to pick on and this is the first time he'd done a review like why mine and he said well i asked chat gpt for the world's number one marketing podcast and you came up right and i said that's interesting because i'm not technically the world's number one marketing podcast but as a joke when i launched the podcast you said it's the world's number there you go yeah yeah you see so you know chat gpt is learning the only problem is when i tried to recreate the result it obviously learned itself in the time between then and now and i couldn't recreate the result and got a lot of other things so um i don't even even if you know this, but at DDB, we're working on an AI, a Lesbot.

1:12:54A Lesbot? Whoa. Which is trained on, you know, all the stuff. Yeah. Peter and I. What you would say. And Hamlots Plan's in there as well. But does this mean we're going to go to conferences and Lesbot will be taking a slot? And so I was messing about with it and I said, okay, I'm just asking it some questions. One of the questions I asked was, do ads wear out? And it gave me a long considered answer about the evidence that ads wear out. They do, yeah, blah, blah, blah, blah. And I said, you know, didn't we write in our book how not to plan that ads don't generally wear out? And it sort of went, oh, sorry, I misspoke.

1:13:31You know? Yes, ads don't generally wear out. You're having to train your own boss. Well, this is an example. It is already trained on stuff we've written. but all the same it you know it needed some you know to be kept on the leash as it were i see um my view is that ai will do will help low skilled people to become much more skilled and high skilled people to become much more skilled the people in the middle i think will be the ones who are in in some danger all right so mr rixon had another question for you les uh But ESOV, given that some data out there is walled and it's difficult to get to, how reliable is ESOV now?

1:14:18Or how easy is it to get to a proper ESOV calculation? Well, it's a valid concern. And certainly Peter and I have been expecting that share of voice would stop working. It would stop predicting market share. for years we thought this you know as as the proportion of ad spend that is audited goes down you'd think that this stuff would no longer work but it still does work so i'm literally um you were asking me earlier on what you're going to do and when you know after this i've got two things to do this afternoon the third one is we're finishing off a share of voice model uh and it still predicts market share right now.

1:15:03Why does it do that? I think the answer is two things. The cynical answer would be that some of the unaudited media that people spend a lot of money on don't have much influence on long-term market share because they're short-term tactical stuff. The other one is that the unaudited share of voice tends to be proportional to the audited share of voice. So you're only measuring the tip of the iceberg. Um, but you know, if you've got 20 % of the audited ad spend, the chances are you'll probably have about 20 % of the unaudited ad spend because people have similar ratios of the ratios. So I think somewhere between, you know, that's, that's a kind of like one's an indicator of the other to the other stuff doesn't work.

1:15:56Yeah, it makes sense. Yeah, that follows. His third question for you was, if I ran three MMM models, would I get three different answers? That's a very interesting question. It would be an interesting experiment to do. There was a famous experiment done in the late 80s, early 90s, the ARF copy validation project. You know that one? So they did that with pre-testing. So they got 23 different pre-testing companies to all test the same ads to see where they, yeah, you're getting nervous now. And the result, we didn't take part in this. They left out the correct answer. Well, it's relevant to you because what they did was they compared, they didn't just compare whether or not they came out with different answers, but they compared the results against econometric measurements of effectiveness.

1:16:54And the overall conclusion was that pre-testing as a whole was no better than tossing a coin at predicting effectiveness. But if you looked within that, some methodologies were better than others, and all the ones that had any predictive power were based on liking of the ads. Anyway, that wasn't the question. But I reckon if you've got different econometricians to all tackle the same problem, they might come up with different answers for two reasons. One is they might choose different data sets. You know, one econometrician might think to include rainfall in the variables you test and one might not.

1:17:35And the second thing is, you know, the actual methodology of how you build a model has more craft in it than one might, than a non-expert might understand. One would hope to get around this, what companies should be doing a lot more of is actually forecasting, you know, and using forecasts to test models that way. you know if if a model's good it should have predictive power well that i mean going back to the sov thing that's actually probably the reason i joined system one is that they were able to evidence in every single case that applying a creative weight on top of esov was more predictive in every case that we've ever looked at so it it has all the always it tells you what you knew to begin with that the creative matters but still good to have the evidence but the important thing from our point of view is that the prediction did correlate did improve the correlation to uh in market performance and the other point is and i've sort of made the same point myself several times is is that creative is really really important but also media is really really important and the you know basically weight of spend and share a voice is really really important and you know you know there's no point in having a great ad that no one sees you know your quote about cathedrals in the building cathedrals in the desert and brand size is very very important brand size is very important as well so you you've got to you know it and really great advertising you know is is is popular advertising it adds that most of the time it's ads that have you know that people like makes them feel something makes them you know emotions feelings associations all that stuff at scale um and that's what we get wrong a lot of the times we we think it's all about just winning a little creative award for something that nobody sees well listen it brings me up quite nicely actually i wanted to round off right with with with uh the best performing adam and eve ad on our database because it joins it joins our conversation up in a really beautiful way today as we're recording this is the first day of cop uh we talked about purpose we talked about advertising and um the most successful adam Eve added this year on our database is Great Western Railways.

1:20:00Yay! Which I don't know if you knew that because it's a lovely example. And also reason to mention it as well. We've just included it in our Greenprint report, which we've launched at System 1, about how advertisers can help deliver sustainable messages in the right way. And what I love about this ad is brilliantly is it's funny. It's an Enid Blyton story, Famous Five. You know, it does all the things that great advertising does. It's a campaign. It's a campaign. Yeah, yeah, yeah. It's basically a fantastic ad that delivers a message about taking the train rather than taking a car. And we talked about how to do purpose.

1:20:34We're doing purpose with humour and storytelling and making great ad is what we should be doing. We might be seeing a lot more of getting people to do the right thing for the wrong reasons, almost, which is sort of where purpose, I think, went wrong a lot as well. It was sort of preaching at people. But this is a really good example of that, isn't it? It kind of slides under the door, that kind of environmental message. and people just see a lovely story and something that reminds them of their own holidays and it's all very nostalgic. So very great. It's great to hear. I didn't know that. There's no carbon messaging.

1:21:04No. There's no kind of real messaging at all apart from a big advert. And we can relate to the insight of getting stuck in traffic and being late and diversions and then people on the train have a beautiful time and get there early. And to the extent that it does the sustainability carbon stuff, It doesn't do it explicitly, but there's, you know, quite a lot of clouds of choking smoke in there, isn't there? It's implicit. It's implicit. Yeah, very nicely done. It doesn't preach, as you say. Yeah, no, so brilliant. So it's lovely to show that we can make a difference and do some purpose in a funny way.

1:21:39Exactly, yeah. And, again, you know, the purpose is baked into the product, isn't it? You know, inherently. Yeah. Back to your point before, yeah. 100%. So as you look ahead to this year, because we're now in 2014, 2024, I should say 2024 as this goes out. Oh, I'm still in 2014, don't worry. Exactly, yeah. There's lots to be said for that. What do you think we're going to be talking about this time next year? I mean, there's a big case for AI. We talked about how that could change how we do research. There's a case for is this the year that sustainability finally actually would do something about?

1:22:13I don't know. It feels like we've been talking about it for a while. Yeah, I think that's right. And probably more focus on doing that through products, packaging, pricing even, more than promotion, which is where it's been sort of focused arguably too much and in the wrong way up to now. So I think a lot more marketing on those sorts of basics rather than some of the things that have been diverting people's attentions over the last year. So a lot of climbing back down the benefit ladder into products and pricing and new products and that kind of stuff. that's a good point because if this year or last year was the year of consistency maybe this is the year we get the basics right you know because there's been a lot of distraction in the last few years with global events it'd be nice to think we get on and deliver the return for marketing it'll be the year where we don't talk about what's new at the end of the year it'll be much more about just getting the basics right I don't know about what we'll be talking about but the advertising industry I'm sure will be going up its own ass about something irrelevant that's one thing that doesn't change because that's what they always do One consistent thing about the marketing industry.

1:23:16You know, as an industry, we love something shiny and new. I mean, we were all talking about blockchain, weren't we? Yep. And NFTs. And NFTs. How much of that is, you know, apart from, you know, actually, have you seen the McDonald's Crypto Bros ad? No. It's a poster that says something like, you know, Crypto Bros, we're hiring. because they're all going out of business or going to jail that's very good but yeah i mean we we forget this i think is how often that as an industry we become obsessed with things which turn out to be you know absolutely ephemeral uh like do you remember you put went the point back in the uh late 90s when we thought that wap was going to be the the big technology it was going to change everything you know does anyone remember what WAP is anymore no it does it does still exist but well it brings us I think full circle back to Sarah your point at the beginning which is you're not the customer so if there's one thing we do this year it's getting in touch with what our customers actually want to do rather than what we want to hang on about all year yeah yeah yeah less wanging less wanging more and with that thank you on that Bob Shill.

1:24:36Oh, he's in his and Peter. Brilliant. Sarah, Les, thank you very much. It's been a blast. Thank you, Orlando. Thank you. Thanks for listening and watching the Uncensored CMO. Hope you loved my conversation with Les Burnett and Sarah Carter. I certainly had loads of fun doing it. If you'd like to never miss the episode again, please do hit the subscribe button wherever you get your podcasts. And if you're watching this on YouTube, hit subscribe there too. If you want to get in contact with me, I am over at X, or formerly known as Twitter, at Uncensored CMO. And you can find me on LinkedIn under my own name, John Evans.

1:25:09Thanks for listening and watching. I'll see you next time.