The harsh reality of advertising's shift to 'churn and burn'

[Music] Hello and welcome to the Walk Podcast. My name is David Thurman and today we are talking about the fragmentation of creativity and what it means for brands and marketers. At the starting point for this podcast is a presentation I gave about the age of more with less and the need for marketers and an age of efficiency to find ways to drive more value. Now to help me illustrate this session I got hold of some absolutely fascinating new data from Creative Technology Company, Creative X. It showed just how big the problem of the fragmentation of creativity is within the digital space. To help me unpack that data I'm joined in this podcast by Anastasia Lang who is founder and CEO of Creative X. And we're going to go into this data in some detail in the course of this podcast. I'll bring on Anastasia in a moment but first, word from our sponsors. Good news. Decades of benchmarking insights and best in class case studies are now available at the push of a button. Lions Intelligence is a new operating system for creative marketing excellence. With state of the art AI technology and established marketing insights from the work, contagious IQ and walk. All in a single adaptive workspace that evolves with your thinking. Save hours of manual work and share your thinking with clarity and confidence. Lions Intelligence is where groundwork becomes great work inside strategy, creativity, execution, all in one place. To learn more, check out intelligence dot lions dot co. Anastasia, welcome to the podcast. Welcome back. I believe you. I'll welcome back. Yes. Yes. This is a second year running, I think it is. And I go, so I've set it up in the intro, but the background to this is an absolutely fascinating data. I think, you know, I've shown this to several people now and they've all been very excited and, you know, it's said, blown several people's minds. That's why we're going to get into here. But I think what it would be really useful to do is just set up because it's so important to what the data is. Just set up what creative X does like just explains people what they do and then we'll get into that research and a bit of depth. Yeah, I mean, very simply creative X is trying to bring a level of visibility into the part of marketing that has historically never been able to be measured at scale. And that's a creative. We all bang on about how important it is to advertising. And yet we measure our audience, we measure the keywords we use, we measure everything down to the time of day, the best time of day to show an ad. But when we get the right user, the right place, the right time, we have absolutely no idea what are the creative decisions we're making that are going to systematically be to better consumer response. And that's where creative X comes in. And so you're looking at very large sets of data on behalf of clients, large sets of assets and you're coding them in all sorts of ways. Yeah, that's right. I mean, a lot of traditional creative testing was doing small sample set analysis. And that's because the technology was in there, what we try and do is bring every single creative brand is running into our data set. So we're analyzing millions of creatives per client. I would say creative X probably has the largest platform, agnostic, agency, agnostic, creative data set from Fortune 500 advertisers. And what we're trying to do is take that creative, which is an unstructured data set, turn it into structured data so we can actually translate that into actionable insights, where a marketer can say, every time I make this type of creative decision, here systematically, what did this my performance. And I can then go and replicate that decision and feel pretty confident I'm going to get the same results. And just to be very clear, this isn't kind of classic pre-testing is that you're not sort of like testing emotional response. It's a different style of testing. You're kind of like almost like the auditors or the yeah. Yes, what we're doing is we're deconstructing the ad into all of its many ingredient parts, right? If the final creative is that dish, what are all the ingredients that went into baking that dish? How do you classify them? How do you understand how they come together and what flavor they make when you actually combine them into one. Some of that stuff is more for auditing purposes. Is my creative fit for platform? Am I using my distinctive brand assets? Am I featuring people in my ad who look like my audience? But you can also start to group and cluster that data into ways that get into more. Am I using a sustainability message in my creative? Because that's one of the key things I want to associate my brand. So think of them as a bunch of Lego pieces, right? That you can assemble to answer whatever type of question you want to know about your creative decision. This is a really interesting area at this moment because it feels like we're properly having more grown up conversations about medium creative coming back together. We've sort of separated these things out for decades in ways that don't always make sense other than to maybe polko. So it makes sense to some money trails, but it doesn't make sense for the industry. And if I've been presenting on the idea of more with less and how we operate in an age of efficiency and one of the ways to see creative in that world is as an exercise in medium efficiency. Better creative makes work harder. Absolutely. I think the insight that you had there is brilliant. I think people forget that creative and media are part of one coin and you cannot decouple one from the other because these two are the two critical elements of marketing effectiveness. And so when those two things start to pull in different directions, the entire operation falls. And I think as we go into this research, we're going to come back to the next. We'll tell you where it's breeding. Okay, so let's get into it. So just give us a bit of the background to research. So from my point of view, we started talking about this as we are preparing this sort of can session around more with less and you came to me this and these numbers. This is crazy town. This is it. But this I've got to dig into this. But what sort of encourage you to put this research together in the first place? Yeah, well, you were very prescient with the talk you were putting to do put together for for can. But the reality is the inspiration came from a lot of our clients out of big brands. And the reason you talk about the fact that their strategy is fewer, bigger, better. Right. I know for you, it's been fewer, bigger, longer. But it's essentially it's a similar concept. Right. And so and we saw this, this narrative being escalated to their shareholder presentations. It has become a company strategy. And the question we wanted to know is great. If that is a strategy, fewer, bigger, better. Our market here is actually executing that strategy when we look at their content operations. So we very simply wanted to answer whether or not the reality of what they were putting out was matching with the strategy that they were trying to put forth for their brands. Great. I'm just going to explain quickly what we mean by fewer, bigger, longer and you're, you let's just merge it into fewer, bigger, longer, better. That's that's that's the little basis cover them. But the the idea here and I talked about this in in the can session is that there is an emerging body of evidence or like best practice guidance. And it's coming from people like Rickson system one, but more broadly, I think there's a body of evidence that says. Marketers should be focusing their efforts on kind of having fewer ideas that doesn't necessarily mean fewer assets, but fewer ideas that they can then use their scarce resource to back with more money. So talking about fewer bets. But really sort of focusing working capital behind those. And by making things better by making you know, investing in the creativity upfront by focusing investment in those ideas, we then get to leave those. Those things are then good enough to leave in market longer, which means over time we're compounding the returns where yes, we may be putting more effort in upfront, but we're then able to leave things in market longer because it is better. And so what that in theory what that leads to is compounding revenues, but also things like pricing and profit effects. So that is a theory and that's the theory we talked through in my particular session. Can I just challenge? Of course you can because I don't think it's just a theory. I think in some ways it is a return to how things used to be to some extent. Because before content production and this concept of fewer was a naturally occurring phenomenon because of scarcity. Right. That is a very good point. But today we live in a nature of abundance. So you can do anything. You can do anything. And so we have to be reminded that just because you can do anything doesn't actually mean that you should. Right. So that's the setup. And what we wanted to know and what you have the data to show was whether this was actually playing out in actual, you know, real life brand advertising deliverables and, and you know, were people actually moving in this direction. So let's talk about the sample size and what you actually, you actually talk about. because this is important because it's quite a big study. - I love a big data set. (laughing) - Oh, it's on us about it. - Yes, so the sample size is big. We analyzed 633,000 ads to be precise. Over $2 billion of ads spend across six markets to the US and the five top markets in Europe. A combination of 143 brands. These are all Fortune 500 brands. So again, as you said in your speech, these are not your mom and pop pizza parlors. These are your big global brands. - These are blue chip advertisers. The folks are the most sophisticated creative operations. - With access to the agencies and the support. - The platform VIP treatment. And for them, we looked at two full years of data. So from 2023 to 2025, I would be remiss if I didn't mention here that if you think about this timeframe, the effect of AI has not kicked in yet. So we're looking at-- - Yes, that's interesting. This is actually before the AI-- - This is before the wave. I would say we're probably in that early part where we were starting to, the wave has not crossed the yet. And we looked at four platforms. So this is across Instagram, Facebook, TikTok and YouTube. - So that's the another important point. This is digital platforms only. So we aren't looking at TV and out of home, all those sorts of things. But just jumping ahead slightly, the reason those platforms are so important is that over those two years, spend on those platforms went up by, I think, it was 22%. So a map from this sample. So a time when budgets are largely flat to down, these are the platforms where money is shifting and presumably something somewhere else, whether it's another media channel, another part of the marketing operations is being defunded to fuel growth on this platform. - Correct, this is the rise of the continued rise of paid social platforms because that's where the users are. Right, I think TikTok is obviously a big sample of that. And I mean, we've been living this for the last decade as money from television and offline has started to shift to digital. And those platforms, yes, the overall revenue allocation towards those platforms has grown over the last two years. - And let's talk about, I guess, the sort of top line, the sort of top line data. So yeah, spend is going up, but volume's going up faster, isn't it? - Yes, so the first stat that we uncovered was that a content production is up by an average of 29%. Now, I personally hate averages because they obscure all man-erfs in. So when we actually look at the range of numbers within this, one of the things that we see is for a lot of advertisers, content production is going up by 300, 400, 500% in that timeframe. - Wow. - And that's in two years. - That's in two years. That's in two years before the full effect of Gen.A.I. And the first thing that every advertiser that I've shown their specific numbers to says to me when they see those numbers is they say, "Huh, my media budget has not gone up by the same amount." In fact, my media budget has gone down and their lies, the tension. - And so there we start getting into spend per ad, again, we're talking averages, but I mean, it stands to reason that if volume is rising faster and spend, then actually media budgets must be going down on a per-ad basis. - Absolutely. I think this is where, look, I didn't get very far in math in school. Didn't do very well in calculus. I will admit that public here. I think of that of this as a simple equation. You've got your numerator, which is the number of ads, and then you've got your denominator, which is your media spend. And these two numbers together tell you this is how much you're spending per ad. One of these numbers, the number of ads, is going up 29% on average, triple digits for many advertisers. Media budgets are now growing at the same rate. Many of them are going down. So in fact, what we found is that the average spend per ad went down by 15%. - 15%. - 15%. We are putting 15% less horsepower on average behind every single ad that we're putting out, and this makes sense, right? One of your, the top, the numerator is growing, but the denominator is seeing fixed, you're taking your pizza pie, and you're having to share with a lot more people. - Yeah, so like 15% drop in media spend per ad. If we go back to fewer, bigger, longer, better, as we've now re-criscened it, we are definitely not seeing fewer. And we're not seeing bigger because the spend per ad is coming down. Let's talk about longer because that one is, there's more than mixed picture there. So in terms of, because you measured duration as part of this, so it wasn't just about the spend behind each ad, it was how long each ad was in market. Talk to me about what you found about duration. - So when we looked at duration, from beginning to end, the average amount of time and ad was running was 36 days. We then decided to cut a bit more deeply into the data to see if there was any material difference between your lower funnel direct response ads, which we expected to run for a little bit shorter, versus your brand ads, which you would think would be market for longer. What we saw was that your direct response ads were in market for 32 days, and your brand ads, your upper funnel ads, were in market for only eight days longer now. - To 40 days. - 40 days. - 40 days. - So it only takes eight more days, David, to build a brand. - Build a brand. Yeah, I think we've solved it. - I think he's solved it. - So the long term is now an extra eight days. Okay. I mean, this is quite serious topic from point of brand and performance, because I think you're right. When I first saw that 36 days thing, I just assumed that these digital platforms, it's just gonna be lots and lots of promo stuff and other things. But when you cut it out in that way, and it's just 40 days, and you think, well, we've had people like Marl Richardson talk about leaving an ad in market for two years. And we seem to be, there seems to be a real divergence here between what we're seeing from some of the evidence, some of the big sort of meta analyses, and what's actually coming out of the real world data. And I guess the question here is, what's leading people to switch their ads so frequently? Is this a sort of the way these specific platforms work? Is it lack of confidence in the advertising from the advertisers? Is it, you know, are they delegating a decision to the algorithm? What do you think's going on here? - So on one hand, what we're seeing is that, increasingly, it's much easier to do more. We have the capabilities and the possibilities. Why wouldn't you? If you can do, why won't you? On the other hand, there's increasingly a narrative coming from the platform partners that you need more assets in order to play and win the algorithm. So I think that's the first tension. I think the most surprising thing for me when I looked at the data is that it shows a real lack of conviction and confidence in the creatives that we're making. So we might now have total content abundance, but we actually have very little creative confidence. The reason I find that so profoundly disappointing is actually as an industry, we have spent so much time on research and tools and other things that should make us somewhat less susceptible to as you very eloquently put it, having to play to the virality of the algorithms. And yet here we are again, basically now saying the algorithm is our new creative director. Let them decide what our consumer should see. - Let's get into the way we then cut that data. So I think that then sort of brings us back to exactly that point because what we've presented so far is like the top line, the overall average is, but the story for me got even more interesting when we dug down into the media budget itself. And we cut the data, although 633,000 ads, we cut them by the amount of media budget behind the ad. And we had one bucket that was under a thousand, another bucket that was like one to 10,000, another bucket that was 10 to 100 and a final bucket that was $100,000 plus. Remember what we talked about, our fewer, bigger, longer better. I don't know what I expected to see, but when we did the cut, but we've talked about it, I've shown it to other people. And the sheer volume of low budget work in that sample is to me quite astonishing. So just talk us through the numbers and then we'll talk about what it means. - So what the numbers say is that 93% of the ads that we put out there have less than $10,000 of spent put behind them. - And remind us this is Blutchip advertiser. - This is Blutchip. - This is not like, so we're cutting out all that SME stuff. - This is all Fortune 500 again. again, you would recognize every single brand on the list. So 93% of the ads had less than $10,000 behind them. 70% of the ads had less than $1,000. - Less than a thousand. - Less than $1,000. - I mean, for a Bluetooth advertising, that is essentially nothing, yes, that's-- - That is nothing. And again, because this is before the full emergency vi, I would venture that many of those ads that had less than $1,000 behind them probably cost more than $1,000. - Yeah, that was a good-- just in terms of people time, even if this is a variation of something, in my-- - Wow, okay. Sorry, we've got 93% under-- - Under 10,000 or 70% below 1,000. Now, I think there's a couple of ways you could sort of push back on that. - That's right. - So the first one would be, well, look, this is just the way these platforms work. We just have to fill the platforms with content and the algorithm will pick the best stuff. Did we find-- We did look at creative quality, didn't we? So there is a sense that the best, the most money, or the most money per ad, or the highest, the ads were the highest, spend behind it, was getting the best quality stuff, didn't we? - That's right. So what we saw was that when we looked at, when we look at the distribution of ads and media spend, what I wanted to understand is, to what extent are there other signals that we could have extracted from the ad before we put media spend behind them? To determine if this was an ad, we should have invested more into. And one of the things we saw is that the ads at the bottom end of the range, the ones that had about $1,000 spent behind them, had a quality score that was 20% lower than the ads at the very top of the range. And just explain what the creative ex-quality score metrics, that I think that's important is-- - Yes. So the creative ex-quality score is one element of a much more complicated formula of creative, effective as the creative ex-alts. And this just looks at, in this case, platform suitability. Are you making ads that are fit for platform? Actually, when we overly advertise our specific data into it, which has some creative quality, some of their own advertising formula for effectiveness, that delta between the lower spend ads and the higher spend ads in terms of their personal custom creative quality score gets even wider, which is all kind of a fancy way of saying, you have information at your fingertips. Before you start rolling the algorithmic dice, to know what you should invest in with more conviction, but for some reason, folks are still letting the algorithm decide. - So we're-- so just on that sort of brand specific things, that would be things like internal brand guidelines or-- - It could be. I think this-- - Or things that are best practices they've already found from previous, so it's right. - That's right. You know, the way I think about this-- the creative effectiveness formula that every advertiser is chasing is, we're all just trying to peel back the layers of the onion, one by one, right? So a lot of them have agreed on, hey, we need our ad to be fit for platform, because that's obvious, that's hygiene. But once you start to get some indicator that hey, there are some creative elements I can measure, I can see how the drive performance, you want to know, how do I develop something that's specific to me? One place where a lot of advertisers go to is distinctive branding, right? Everyone, for the most part, believes that, if the ad is its completely branded as mine, is more likely to sort of impact. That's one element. But fundamentally, it is about figuring out, what are some of the creative hypotheses they have, and how do we start to test them? So they can get more confidence in knowing if they make ads that have some of these components, do as are disproportionately more likely to perform for them. (upbeat music) There's no shortage of opinions in marketing right now. But the brands winning in today's market aren't adding to the noise. They're helping marketers navigate through it. Lines advisory partners with the world's leading brands, agencies, and media platforms to transform their point of view into game-changing thought leadership. That means evidence-based research, frameworks, and expert commentary all built on decades of expertise from walk, headlines, effe, and contagious. In today's market, it's not enough to just have an opinion. You need to be a credible authority. Take the next step at lines.co/sortletorship. (upbeat music) - Let's talk about the final sort of finding, which was around the amount of budget that's actually going into these low budget ads. Because I think the other sort of pushback and maybe this will change as AI is adopted more, but I like this sort of low volume stuff, it doesn't matter. It's just we can just create it with AI, just keep pumping the platforms full of it. And again, the algorithms will pick the winner and the stuff with the best quality schools will rise to the top. And indeed, we've had in the past year, people like Mwang Zuckaburg talking about infinite creative, this idea that, you know, it's constantly generating stuff. And that, I guess that kind of only works if you're spending a very small amount of your budget on those low, at the bottom end of the range. - Yeah, and we're thinking of like little test ads. - Yeah, so you're writing a little test then. - Yeah. - But we're not spending a lot of money, we're just testing what works and then we'll scale. The best stuff, and we'll scale the stuff that we really want then to have an impact. That's not what we found, was it? - No, not quite. So what we found is that these low budget. - So sub 10K, we said that as the. - Yeah. - And it was slightly arbitrary, but we set that as a sort of. - Yes, the sub 10K ads, which again, made up 93% of your ad volume, took up 30% of your media spend budget. - Yeah, that's quite a lot of money. - That's quite a lot of money. And the way that I think about it. - So we have $2 billion, we took it as a, it's like $600 billion. - Yeah, good mental math. - $600 million going into sub 10K, wow. - That's right. And again, because media budget is a finite limited resource, the at the expense of funding, the bigger, better ideas that are more like it have disproportionate impact on our consumers, we are putting that money into testing things. In some cases, we already have some signal that those things are more like not going to work. - Just like on that final point, because I think it's really important. The 100K plus group, so yes, look, it's a very small percent of our samples, in like 0.3, but then. - 0.3 exactly. - So if there are biggest swings, we're not thinking they many of them. - No. - But more importantly, we're not funding them that much. They took up about, I think, 18% of the budget, versus 30% of the sub 10K. So there's quite a big disparity there. This is it, and I think this is where we again, have to zoom out and look at our marketing effectiveness machine as a connected operation between the creative and content side and the media side. And what you have right now is that these two sides are fundamentally pushing in different directions, and they're going to break your machine. Even if you look at this and think about, and by the way, here, we haven't even included the cost of production, the cost of your team's attention to make those ads, but also, I think, more importantly, let's think about it from a consumer point of view. How many of those ads do we believe were in line with those two or three key big creative ideas or key communication point you'd want to land with your consumer when you set the strategy up in the first place? Increasing what we're finding is a lot of these ads are what we call disconnected ads. They are ads that are not related to the core messages that your team decided to land the beginning of the year. And to be very clear, we're not saying that all the money should just go into the top thing you should never test anything. Clearly, that's the opposite problem. But something seems a bit out of whack here in terms of the amount of money that's going into, as you say, things that could be tested before we did ever hit a platform with versus where in the theory, the fewer, the bigger, that stuff suffering, apparently because we're taking a lot of money and putting it into the testing set. That's exactly right. And I think this is, again, this is a real crisis of creative conviction, right? We do not trust ourselves as marketers based on this data alone to know and to feel confident that these are the ideas and the messages and the ads we want to land with our consumers. So we have entrusted the algorithm to be the almighty God who will make those decisions for us. Now, again, I am not here yet to say this is good and this is bad. But what I can say is when we've cut this data in advertiser-specific ways, every single person we've met with who's looked at this data has said, this is not how I would like my operation to run. And that's important, isn't it? Because if we looked at this and your clients or the people I've shared it with went, well, this is just the way. worked. You've misunderstood this is just what happens in this sort of new media landscape then maybe it would have retreated but actually everyone we're showing this to is actually quite concerned by what they see. I think they're quite concerned but I also think they feel perhaps a little bit vindicated or heard. I think because Jenny I and creators and all these modes of prolific content production they had such sexy demos and a lot of marketers have been under pressure to adopt AI and Jenny I was the first sexiest thing you could go on right no one was going to get fired for trying Jenny I and but yet behind closed doors what I hear a lot of market theory says look I don't need more that's not my problem I wasn't asking for more but it was difficult to go against the grain of the prevailing narrative even last year it can right this was this was the theme of the conversation and I think what this data actually does is it gives people conviction and again the line will be drawn differently across every organization to say we don't need more because we have some limited limited finite parts of our operation and we need to invest those really wisely and that brings us back to the whole more with less you know more with less age of efficiency but it does mean making really thoughtful decisions about where where the money is going yes but I'd like to operationalize that a little bit because I think you know we same word with less and you and I have dissected it right so to us we hear that and it means something very specific I think a lot of market tears might hear them be like cool I agree but like what do I do about this tomorrow and so my sense is there there are a couple of things that can be done the first one is how do you make sure that every ad you create actually ladders up to the big creative idea that you want to land right that you're not testing and putting out a lot of these the so if you're all putting out lots of volume it's still you need to ladder up right exactly yeah your your lots of littles have to map back to a big yeah the second thing based on this data I would say is I would recommend that marketers think about establishing media sufficiency threshold and again platform partners will help you actually if you go to them and say hey what's your recommended media sufficiency again it'll vary by region but I would if I were a betting person I would say none of them would say it's a thousand dollars right and so establishing and figuring out what is our level of media sufficiency which means that every ad we put out there has to reach at least this level because that is our minimum level of conviction and that's how I'm trying to put willing to put behind it the third thing and the last thing I would encourage people to do is again when I look at the SMAP it shows to me a real a real crisis of conviction and confidence that we know what's going to resonate so how do we continue to extract learnings from everything that we're doing so that we don't have to play let with our ads and can systematically know when we make these types of creatives we feel really confident getting them out there and putting the full weight of our budget behind them because they're going to perform with our consumers do you think this I guess is a so quite a broad brush question but do you think on these sort of platforms that because of the way they work because of the algorithmic nature that the gravitational pull is always towards small is it inevitable that as money goes in this kind of pattern replicates? That's an interesting question. I think the algorithms have been designed to play the law of numbers right and so I think there is something very tempting about taking something that has been historically very difficult for us to quantify and measure and say you don't have to do it the machine will do it for you. In this case yes that does lead itself to the law of our numbers where you will have a ton of ads at the top but this is also where I start to think about incentives. Everyone in this equation is a business and if you want to in more cases than not if you want to predict how someone's going to react the common advice is follow the money. If you can follow the money trail you can pretty much understand this is going to be the recommendation the individual and it's just true for you as it is for me and so this is where I think we is going to see have to figure out we are here to serve the brand and the brand only the brand what is best for them and their consumer and what is the least wasteful way of delivering it that does not decrease the odds of hitting those effectiveness highs right because this is also where you know will the algorithm be good at creating those you know those mass market hits is that that that element of virality something where we can only rely on the algorithm to do I don't know the answer to that but I think it's we're trying to solve. Oh agree. Let's end with word might of optimism which is that we did find one group of ads that did see an increase in duration over this period didn't we? We did we did so the good news is and I think this was an optimistic chat but I think the good news is is that we saw that ads that had over a hundred thousand of spend to put behind them. So that those big swings the big swings were in market for an average of 89 days. So that's 89 and that had it increased over the over the yeah so there are there are ads that have been kept in longer and that that duration is increasing over time which is which is good news. The bad news is just around very many often. There are very many of them. I think we we don't have a benchmark the difficult thing here is 89 days good or bad. Well no. I don't know. You know the we've heard that last year and this year that you know we should be thinking in terms of years not not months but that's right. You know three months sounds like a better minimum but I guess the point is if we're only spending 18% of our budget on those big swings then actually you know just think what more we could do without that that's our best quality creative based on the creative x score it's the stuff we're obviously you know we're really to double down on. Yeah so could we be spending more on those and pushing them more could we be spending more and I think this comes back to the the age old question that I don't think we've solved yet which is you know our our ads wearing out or have the barely worn in and the question I'd have here is our consumers getting tired of the ad whereas the algorithm getting tired of the ad because they're different CPCs or CPMs or whatever they can fetch for that ad and that's a that's an interesting opportunity for some some further research down there as always with these talks we end with we need some more research. A good place then a job never done the AI will not take our job David. I saw me home. Hey right. I miss Daisy thank you so much a few times it's been a pleasure having you on the podcast. Always good to be back. Thank you and it stays here and we will post details of all that research on to our podcast channels and of course it will be available on walk as well. Now if you like what you heard please do follow us on your podcasting platform of choice. If you really liked it please leave us a review. Until next time thanks for listening.