Lydia Chan is joined by Sabrina Talma, CEO and Co-Founder of Human Made Machine, a creative intelligence platform that helps global brands test and refine creative work before it goes to market. With clients spanning Google, the NBA, and The North Face, Sabrina brings a rare, data-backed perspective on what actually resonates with audiences — and what quietly falls flat.

Sabrina shares why most brand campaigns miss the mark from the start, and how marketers are often too close to their own work to judge it objectively. She walks through a real campaign example where two nearly identical ads targeting Gen Z in the US and China landed in completely opposite ways — and what it revealed about the danger of treating global audiences as one. She also weighs in on whether higher production value actually drives better performance, and what regional differences her team is seeing in the data.

The conversation digs into AI-generated creative, including results from a real Puma head-to-head test comparing AI and human-made ads, and why AI tends to converge on average rather than differentiate. Learn why AI personas for creative testing are not ready for prime time, drawing on both her own experiments and a Stanford and Google DeepMind study on digital twins.

If you're a marketer, creative director, or brand strategist trying to understand what makes audiences truly connect with content, this episode is packed with insights you can bring back to the team.

 

• 03:22 - Marketers Are Terrible at Marking Their Own Homework: Why even the most experienced marketers consistently overestimate how well their creative will land and why testing against your true target audience is the only way to close that gap
  
  
• 09:37 - Short Form vs. Long Form: Why the industry is being pulled in two directions by platform-driven short form content and the growing case for deeper, more resonant brand storytelling, and why you actually need both
  
  
• 14:18 - AI vs. Human-Made Creative: What a real Puma head-to-head test revealed about why AI-generated ads tend to converge on average rather than differentiate, and the one scenario where AI creative actually works
  
  
• 23:26 - Why AI Personas Are Not Ready for Creative Testing: What happened when Sabrina's team tested AI personas against real humans, and what a Stanford and Google DeepMind digital twins study reveals about the ceiling of this technology
  
  
• 29:06 - Back to Basics: Why the foundational principles of effective advertising haven't changed since 1983, and what that means for how brands should be building creative in the age of AI

 

Olly (00:00)

Hello everyone, and welcome back to The Audience Connection. So, around the office, on set, and actually on this podcast, we spend a lot of time talking about the craft of storytelling, the instincts filmmakers and marketers rely on to create something that really connects with an audience. But today's guest brings a slightly different lens to that conversation because she spends her time testing whether those instincts are actually right.

Lydia (00:26)

Yeah, that's right, Olly. Maybe a touchy subject for our audience, right? So today we're joined by Sabrina Talma. She's the CEO and Co-Founder of Human Made Machine. They're a creative intelligence platform that helps brands test their creative with real audiences to understand what actually resonates before the campaign actually goes live. So Sabrina has a fascinating vantage point. She works with brands across industries, from tech companies like Google to organizations like the NBA and consumer brands like The North Face. So she gets to see at scale what kind of ideas connect and what quietly falls flat.

Olly (01:07)

Yeah, that's some real insights. And I think this conversation is going to be really interesting for our listeners because we all believe in our ideas when we're making something. We've all got behind something, whether it's a brand film, a campaign, or just a piece of social content. But Sabrina is across thousands of pieces of creative and seeing, okay, what does the audience actually connect with and what actually drives behavior?

Lydia (01:31)

Yeah, exactly. I mean, we get into topics like what testing reveals about audiences in different regions. Does higher production value actually mean more engagement, and how brands should approach the creative brief?

Olly (01:45)

And you guys close the conversation with some really big questions around AI, like whether AI-generated ads can actually outperform human-made ones and AI personas for testing. So definitely want to listen through to the end. Well, with that, let's get into it. Here's Lydia with Sabrina Talma.

---

Lydia (00:00)

Today we're joined by Sabrina Talma, CEO and Co-Founder of Human Made Machine, a creative intelligence platform that helps brands test and refine creative work to understand what truly resonates before it goes to market. I met Sabrina at DigiMarCon in Silicon Valley and really had to have her on the show because as storytellers, we always think we know best, right, in regards to creative and execution, but Sabrina has a front-row view on what actually works in the real world and what actually connects with audiences, which is the core of our show. So Sabrina, welcome to the show.

Sabrina (00:39)

Thank you Lydia and thank you for that lovely introduction as well.

Lydia (00:42)

Well, it's really great to have you. For our listeners, tell us a little bit about your background and how Human Made Machine came to be.

Sabrina (00:50)

So my background is split really between the media world and the creative world. So I started my career in media and I was focused on really all things about media planning and measurement. And in that world, the focus is really on the quality of where you're actually showing the content, who the audience is and whether or not you have that kind of combined media quality signals of what people are seeing and whether you have a true match to the target audience that you want as well. And back then, I guess I was in the world of what we called big data. It's now actually pretty small data in comparison to the data that we work with. But there was this really heavy focus on one side and we knew that with campaign performance, there was also the creative that had a huge impact on what we were seeing.

And the creative signals were relatively shallow and weak in terms of what we were looking at. So we were looking at things like clicks and engagement with the creative. And we didn't feel that really told the whole story of how they were contributing to the performance of the campaign. So Human Made Machine is all about understanding creative and unpacking, really kind of unpacking the performance of the creative. So we collect lots of structured data about creative and then we can provide a more unbiased view of how that creative is contributing overall to lift and specifically we focus on brand lift.

Lydia (02:09)

Yeah, no, that's amazing, right? You sit at the intersection of creative and data and you see what works before it hits the real world and the real market. So I definitely want to take advantage of that and maybe we'll uncover perhaps some of the biggest misconception, misunderstandings that brands have about audience connection. But before we get into that, in our previous conversations, you mentioned most creative or brand campaigns fail because the creative misses the mark, right? And, you know, sometimes brands get the foundation wrong, right? And everything then downstream is noise. How do we make sure that we get it right at the start? Is there- can you test the brief?

Sabrina (02:55)

Yeah, exactly. So if you imagine you have a brilliant idea, you have the consumer insight down and you know exactly the kind of message you want to deliver. I think the first biggest gap is us. So it's the gap between us and our audience and what we think will resonate with them. And that's very common. Marketers are terrible at marking their own homework. They say, you we've done a brilliant job. This is amazing work and this is exactly what we had.

Lydia (03:19)

Yeah

Sabrina (03:22)

what we had envisioned and whether or not that actually connects with the audience is something that you can get through testing because you're testing material against your true target audience rather than thinking about what you think and what we as marketers and even as experienced marketers, know, I look at ads all day, I look at the performance of the ads and I've been doing that for 10 years, but I still have ads that really surprised me and we try to work out what is it about this ad that really worked.

I think the most common gap that we see in terms of whether an ad will work or not is actually just the localization of that ad. So, you know, quite often I work with a lot of global brands and they might have specific messages that they try to deliver. And how you actually deliver that to different audiences is very distinct. So, for example, we were working with a global brand and their message is all about self-empowerment and they created two identical creatives, one for China and one for the US. And the only difference was, is they had done some localization. So the celebrity was localized for the Chinese market, the actual setting was localized as well. But otherwise, it was identical in terms of the concept itself. And what we found really interesting is it was a totally different interpretation between Gen Z in the US and Gen Z in China. And my team in APAC described it as, in the US, you're kind of looking at a mirror and in China, it was like looking through a window. And what they meant by that is the feedback from the US team was all about how relevant it was to them, whether they saw themselves in the person in the ad and they felt that that person represented them and that was really important to them. And with the Chinese audience, what we saw was something very distinct and they kind of looked at that person and whether or not they were aspirational to them. And so it was more like looking at them from a distance and thinking, is that something that I aspire to and want to be like? So relatability was actually much less important to them in terms of the messaging compared to that. And then there was a difference in the way they tried to interpret the message. So in the US, they have a lot of patients for interpretation. They try to kind of extract meaning out of the ad.

Lydia (05:29)

Yeah.

Sabrina (05:43)

Whereas with the Asian audience, it was a much more literal interpretation of what was happening in the ad. And so, I mean, we got to kind of the iteration of that. We were saying to them, you really need to give this Chinese audience much more literal cues of what's happening and you need to not shy away from this person being more glamorous and something that they aspire to rather than having the more relatable talent in the US market. So, you know, that...

that end goal for them was the same. It was still that message of self-empowerment, but the actual path to delivery was totally different when they actually had to think about the localization.

Lydia (06:18)

Yeah, no, that's that's really interesting. I think a lot of the brand marketers and filmmakers, directors on our show have have really mentioned the importance of cultural specificity. Right. And like localization as well. And I think, you know, having this almost thinking audiences are homogenous is is the wrong way to go. And really understanding that local market is really important. But I guess for.

Sabrina (06:33)

Exactly.

Lydia (06:47)

How does that, ⁓ that must be hard, right, to translate for a brand to sort of make sure that it's still aligned to their tone of voice and their brand, not brand values, but you know, their brand messaging. Do you see sort of a, sometimes that being like a difficult challenge to balance?

Sabrina (07:09)

Yeah, so I think in an example that I just gave, it was fairly universal that for Gen Z globally, authenticity and self-empowerment is important. But it's just actually the difference is the delivery. So having a slightly different delivery style for the audience is important. And they often reflect that, you know, that this looks like an American ad if we're working in APAC or Europe sometimes. So they can kind of feel if they haven't had that kind of local touch. And actually that filters all the way through the organization. Honestly, it's about the people on the creative agency side, whether they have the experience working in different markets and the talent that they use in the ads and the celebrities. So they notice all of those things, the styling of the people sometimes, like all of those kind of little touches help them feel that this was made for them and not something you've just kind of pushed across.

Lydia (08:06)

Yeah. No, absolutely. I mean, at Casual we work a lot with global brands and, you specifically we work a lot with a global hospitality brand. And we certainly, you know, you want to make sure that you're really working directly with the local team, whether that's APAC and EMEA because they have a better understanding of how those kind of local travelers and, you know, folks sort of engage with the type of content that you're putting out there. And it's hard when, say, the headquarters is like US centric, right?

Sabrina (08:39)

Yeah, yeah, exactly.

Lydia (08:43)

So you have, I know your clients span industries, right? From Google and tech to the NBA to consumer kind of brands like Northface. Are there any patterns that you see at how creative has changed over the past few years or even the past few months because of just the speed of you know, how things are changing nowadays.

Sabrina (09:08)

Do you mean in terms of the creative development or how they're actually made?

Lydia (09:11)

The sort of creative execution, right? Like just sort of the type of creative that you're seeing coming out of these, coming out of these brands, right? Whether that's maybe a investment in more direct kind of storytelling or short form content, or are we seeing, which we're seeing actually a bigger appetite now for longer form content and kind of deeper storytelling.

Sabrina (09:37)

Yeah, yeah, yeah, no, that's a super interesting question. And I think it's actually we're seeing it on both sides of things. So there's a pattern in the industry and it's actually driven on the media side by the platforms. So TikTok, for example, encourage advertisers that after you've shown two ads, they'll just swap it out and put something else there in front of someone. So it's short, it's fast and then it's done and then it's moved on.

And then we're also seeing patterns. I don't know if you heard, you know, Mark Zuckerberg talking about it, where you have platforms like Meta and what they're doing is you give them, you know, an ad or a brief and they say, we'll make the ad for you and we'll create lots of variations for lots of different people. And then on the other side of things, I think you have this kind of other group who are more committed to having core more resonant messaging that's kind of enabled by AI. So in my personal view, the first iteration of that, of actually, you know, constantly iterating different messages for different people is a kind of madness that we've seen in AdLand before, and I'm not sure we should be exploring it again, honestly. We used to talk about in the era of big data, lots of one-to-one messaging.

I think, you what we saw from that experiment at the time was you end up with a lot of disparate messaging. And really, if you look at really powerful brand marketing, there is this kind of consistency and this kind of golden thread of a narrative that goes through those specific brands. So if you're talking about Coke, would be like taste the feeling. And they always show the ads in these high emotional moments. If you look at something like CPG and Dove soap. You know, for fifty years, they've talked about it as a cream bar, restorative, like not drying out your skin. And if you allow an algorithm to create that messaging for you, there's several dangers. So one, it does kind of this hyper contextualize messaging for the individual, but it's not creating those continual brand associations. And I think for smaller advertisers, there's also a danger that what it does is it kind of converges on a message regardless of your brand.

And that message, for example, might just reinforce what your competitors are doing, and they're spending more money on that messaging as well. So, you know you're essentially not creating any kind of distinction. So I think there's a real danger in how that's used. I think there's still a use case for it, but I think we as advertisers need to be really cautious about how we actually use the technology like that. But the second approach I find actually really interesting and actually it solves some of the problems that we've had from creative testing.

Lydia (12:02)

Yeah.

Sabrina (12:25)

one of the biggest challenges we have is I could come to you and say, okay, Lydia, your video is great, but I think you should like reshoot this scene. And I think this would be more impactful. And you'd say, okay, there's nowhere I can make those kinds of changes, but I can change the branding for you and I can change the tagline or something in my edit. So we had a client and they did it kind of in reverse. What we worked with them on is they created the concepts with Gen.AI. And then we tested those. We gave them lots of feedback on like the elements of the ad that they liked, where to focus on, what to think about for the short form content versus the long form content. And then they shot it with a real person and a crew and changed the settings. And they made a super impactful ad in terms of its salience, its consideration of the product and the campaign really worked really well.

And I think that's a really kind of like nice approach where AI is speeding up the process, it's enabling the process, but it's actually still maintaining that kind of authenticity of human connection. think it was very interesting for us comparing the results of the GenAI creative and the other one, because it's not always that people pick up that generated by AI. They just pick up that it's slightly odd. There's something slightly like unnatural about it.

Lydia (13:42)

Yeah. Yes.

Sabrina (13:46)

And you could see that coming through in the feedback.

Lydia (13:51)

Oh interesting. So this is definitely something I had sort of in my list of questions later on, but we can definitely jump to it now. I, you know, I'm really interested and you kind of just mentioned this, but do you see, or have you seen a scenario where the AI product, the AI output actually performed better than the actual, maybe human production that occurred later?

Sabrina (14:18)

Yeah, so we've done quite a lot of testing of AI versus non-AI, not always as apples to apples comparison. So one we looked at was for Puma and they created one ad with AI, one without. And I think what was interesting about the AI generated ad there was that it was actually very realistic and so people didn't actually pick up that it was AI generated.

Lydia (14:25)

Yeah, yeah.

Sabrina (14:45)

But what they said were comments like, this looks like every other sports ad. And I think if you're taking a model, what you're doing is converging on something, which is on average, what does the running ad look like? And so, and they picked up on that essentially. But then it was actually, you know, in ways that it wasn't a bad ad, it still drove lift, it still kind of created, had enough branding, and those other elements. But the human-generated one was much more impactful. So I think from,

Lydia (14:49)

Mm-hmm.

Yeah.

Sabrina (15:14)

A longevity perspective and ideation, you still really need that human touch. But if you're kind of recreating a version of something that you've done before and you're happy with that, then maybe that's where you can use AI. But I think for lot of advertisers in the competitive world, you need like a bit more differentiation than what we're seeing with that. And the other area that we've seen ads which AI generated work quite well is when the content itself is not human.

So for example, the Coke polar bear, it's not very obvious to people that that's AI generated because we've been using CGI for a very long time. So doesn't even, we know it, but they don't really care, the audience, that that's the case. So if you're not trying to be very human with your AI concept, then actually I think that doesn't seem to be on people's radar at all. They just see that as an animated ad. So I think that's another decent application for it.

Lydia (16:11)

Yeah. No, I think the other side of it too, is that there's, think a, like a negativity that comes to mind when you tell someone something is AI generated, right? So, you know, not to pick on Coke specifically but, you know, having, knowing that that ad was AI generated, it sort of had a lot of backlash, right? On the internets. But, but you make a good point, right? It's like, well, we could have done this with, you know, our, sort of, traditional CGI approaches, right? But what, like why, right? Like as the sort of commissioner of the video, as the kind of creators when timelines are shortened, budgets can be shortened, and the product is just as good.

Sabrina (16:57)

Yeah, exactly. And sometimes Lydia, I think that backlash is actually in our bubble and not actually when you ask consumers about the ad as well. So if it's very obviously AI generated, then people will talk about the harms of AI that they perceive and things like that and their concerns and worries. But if it's not very obviously AI generated, then we as marketers might talk about how it doesn't do what we expect and isn't the quality that we expect. But for your average person, they actually don't care that much if they can't tell.

Lydia (17:33)

Yeah, no, I agree. I have I have this conversation sometimes with my creatives and, you know, directors and then DPs is that sometimes, you know, the audience doesn't necessarily put that much value in and how cinematic this is, right? This piece is or, you know, how high quality the production budget is. Have you, I guess, seen or what are your thoughts on that? Like, does higher budget mean, you know, better performance? Does- is UGC still a good way to go in some of these platforms? I'm a little bit UGC fatigued, but it could be because I'm a filmmaker, right?

Sabrina (18:12)

Yeah.

Yeah, so I'm going to be straight and say we don't have an answer to this yet. So I think when we've explored talent, there's a couple of things we've seen. The story and what you're showing is still the most important thing compared to the, you know, the quality and whether it's the more amateur footage or polished. I think actually we were talking about this the other day. And one thing that was really obvious was that we had a regional disagreement.

And so in the US and Europe, they were insisting that the unpolished content works better. And in APAC, they were saying, no, we disagree. We think that the polished content actually works a lot better.

Lydia (18:55)

Interesting. I would think that is that's flipped actually, but that's interesting.

Sabrina (18:58)

Yeah, so actually, we're actually just going to have a look at that. That's actually one of our upcoming projects. I'm going to get back to you on that one, Lydia. But for now, I'd say it's, you know, what we have seen is it's really more about the message though. What kind of celebrities, the talent, it's additive. It helps with things like whether an ad breaks through or is noticed, but it's not necessarily going to drive lifting consideration or salience of your brand if you don't have those other kind of foundational elements within the

Lydia (19:12)

Yeah.

Okay. Well, that's, that's really great to hear, as well, right? That, the story is still extremely important. we've, we've had

Sabrina (19:35)

That's right.

Lydia (19:37)

Dr. Paul Zak on the show. He is a neuroscientist out of Claremont University and he has this app called Immersion right and really testing how immersive certain experiences and I've been like films ads videos and sort of things like that. It could be events. It could be presentation you're watching right or even an interaction you're having. What he says is that you know stories really do peak immersion, right? And lodge into our memories then, right? When we view this content. And I think, you know, this hyper focus on short form, quick hitting content is perhaps not, you know, going to give you the kind of brand lift later on, right? Eight to twelve months from now.

Sabrina (20:21)

Yeah, I think the, you know, one thing that we definitely have a challenge with is just the fact that where the audience actually is. So, you know, it's undoubtedly the case that you're going to get the biggest reach from going on the platforms, which purely use short form content. You're much more likely to hit the Gen Z audience within those platforms as well and get those kind of that really broad reach. And I think the most important thing for us is actually to have a really varied media plan, because you're right in terms of that, like, deeper level of engagement, there's nothing like detailed long form content, which helps people understand what you do and you're offering. But in terms of getting the word out more generally, you can't ignore the short form content. So I think it's, it's kind of incumbent on us to just work with that a bit and kind of think, well, this is the attention that we have to work with. So we're still going to produce.

Lydia (21:04)

Yeah.

Sabrina (21:14)

quality content, but maybe we're not trying to do such a heavy lift with some of those channels. We're just trying to get the word out for those short form content and then we're building that deeper engagement with content elsewhere.

Lydia (21:24)

For sure. There has to be this understanding that you need both, right? Like you really need to invest in the whole funnel of content. But maybe if you're a smaller brand just starting out, you kind of need to invest in that bottom of the funnel, right? To get people sort of just getting touch points with your brand, maybe quickly buying your product so that then you can start building that community.

Sabrina (21:51)

Yeah, exactly. And I think also we'll see a shift again, because we're now seeing that content is so important to AEO. And so if you want to be found through an LLM, then it might be well worth investing in content and releasing FAQs and things like that. I think, yeah, I think we'll definitely see a shift in more long-form content, which I think is a good thing and it's more fun for us really.

Lydia (22:00)

Mm, yes.

Lydia (22:18)

I agree. So cheers to that one for sure. I want to touch on AI persona. So we talked about this briefly when we met at DigiMarCon. And on the show, we've been speaking to a few behavioral experts who have spent decades in marketing and the communication field as practitioners. And some have developed AI driven platforms that allow you to create audience profiles and be able to interact with them and test ideas, ask them questions, etcetera. So what are your thoughts on that? And are we close to being able to test creative and messaging with AI personas?

Sabrina (22:59)

Yeah, so I'll refer to a couple of types of personas because there's a few. And the first is we work with these companies called Panels and they're responsible for surveying lots of people. They'll survey for creative testing, brand tracking, general primary research, that kind of thing. So in practice, they build a panel and they have lots of information about these people and so they can build an agent, so a mirror kind of person of them.

Lydia (23:17)

Yeah.

Sabrina (23:26)

So we tested that head to head with real humans and we said, okay, if we show this agent and ad, this group of agents and ad, and we show this group of people an ad, like what would we see? I think we saw some things which were super concerning actually. And the first was how realistic the personas were in terms of some of the things that they said, but also these like hints of stereotypes. So for example, in the UK study, they all had like this kind of cockney accent as a Londoner for 20, 30 years of my life, actually, I found that quite offensive. And then we also saw that kind of common problem with AI of being overly positive. So for one brand, it scored 100 % for unprompted awareness of the brand, spontaneous awareness of the brand, which is unheard of. And so...

Lydia (23:57)

Yeah.

Sabrina (24:24)

What we saw that it did do is it kind of reflected the content. So it might, if the ad had a football theme, it would talk a lot about football, but in practice, that wasn't actually what people were talking about. It wasn't actually how they scored it as well. And we have looked at different external sources. So every AI, they kind of benchmark lots of different AI technologies. And I think they gave the accuracy for AI personas as something like 50%, which for me is really guesswork because if it can kind of describe the content for you quite well and people do talk about the content and it gets it right. But otherwise it gets it wrong. So that was a specific test on showing stimuli, creative stimuli and asking for a response for it. And we tried that for several different brands and markets and we really didn't think it was there right now to be able to support that.

Lydia (24:54)

Yeah.

Sabrina (25:16)

The other more promising persona that we've seen is a study with Stanford and I think Google DeepMind. And that was really interesting because what they did is they interviewed people for two hours. That in itself is problematic, Lydia, because after 15 minutes, people's interviews tend to deteriorate. But they did this kind of two hour interview. Exactly, yeah. And it was for, it's quite high sample, a thousand people over the US.

Lydia (25:34)

Yeah, yeah. It's an interrogation.

Sabrina (25:44)

With that experiment, they then gave the, they called them digital twins of the people, they gave their digital twins a fresh set of questions and they gave the real people the same fresh set of questions. And when it was about their kind of like attitudes to life, there was like an 85 % match between the responses between the twins and the others. But when it was about money, risk, the match rate was really poor. I think that's quite interesting because really we're often trying to sell something in advertising. So if you're asking someone whether they're going to purchase something, I think that also reflects the fact that it can maybe mirror an attitude and a view, but some of these more complex problem-solvings that you go through and whether you want to buy something or not or use something, it wasn't really picking up on. So I thought that was also a really, really interesting, very scientific experiment that we can kind of look at. I think that's quite useful for upfront ideation and insight about an audience, but maybe not for actually testing ideation.

Lydia (26:46)

Yeah, but that tracks right because I think as humans too you know, a lot of our decisions are, are subconscious, right? Like you mentioned that at the top of the conversation, likes and then views, it doesn't really indicate really our, sort of views or ideas around the brand, right. Or around the product. So, you know, it's just so complex the way that we think about risk and, and our behavior, that I, I don't think AI is, is there yet because perhaps, I know there's just like not enough, maybe data points or equation that kind of solves for that like you know what I mean?

Sabrina (27:27)

Yeah, ironically, there actually aren't because you talk about clicks and engagements, but in fact, that is what platforms like Meta and TikTok ultimately have when they decide whether your ad is good or bad. That's the most common data point that they have to lean on and to train their algorithms. And whenever you talk about kind of clicks and engagements, you're often talking about a more primed audience and...

Lydia (27:47)

Yeah.

Sabrina (27:53)

maybe one that's already engaged with your brand. So they're more likely to click on your ad because they already like your products. And so it's not necessarily reflective of a growth audience, which is really what we care about in marketing. How do we kind of actually push things out to actually reach a new audience and create growth through that. So I think that's really problematic in the fact that we talk about these algorithms having loads of data, but what is that data and who are you actually optimizing towards? and I think that's why we've always leaned on experiments where you have that counterfactual. There's already this many people who are aware of your brand, but with the ad there are this many people. So I think experiments are a good way to kind of balance out using algorithms, but still understanding the incremental impact.

Lydia (28:36)

Yeah, for sure. Well, AI personas guys approach with caution.

Sabrina (28:41)

Yeah, yes.

Lydia (28:43)

So Sabrina, to wrap up, where do you see the future of the creative industry? Will we be in a place where every single detail is tested and you're almost...

I don't know, maybe picking a whole creative apart and then putting it back together. Will it feel like a whole? Maybe I'm putting words in your mouth because again, I'm biased, but.

Sabrina (29:06)

Yeah, I think in terms of AI, what we see it's actually really good for is understanding failure. So if you've totally failed to brand your ad, if it doesn't have these kind of attentive features that actually capture your attention, which are fairly foundational to making something work, then AI is pretty good at detecting that and saying, okay, it's branding well for YouTube or for a longer format.

What it's not so good at is understanding the audience connection as well. And that's where we still really need those human inputs. And I think, you know, at the moment I'm reading Ogilvy on advertising and he talked about running kind of experiments and he talked about how first show the image because that will get the attention on your print ad when you're constructing it and then show the headline. And he says in the headline.

And that's what most people will read, not the rest of your article. So make sure you've got your brand name in the headline and you say the benefit that they're going to get from using your brand. So you get that kind of message in that headline. And then he says, talk about that kind of depth and detailed research about why your product is great in the body of your text, of your print ad.

So that book, Lydia, was written in 1983. I was born a year later. And if you think about the things that we're talking about, these like foundational elements of advertising haven't really, really changed over that time. But what changes are the audience, the localized audience, the brand that we're working with and the context of the media that we have today. And I think that's what we're doing is kind of translating all of those principles into the world that we're living in today.

Lydia (30:44)

Yeah, no, I completely agree. I was on a creative panel about a year ago and really I was just like, we gotta go back to the basics sometimes, right? Don't forget the foundations of what actually makes a good piece of content

Sabrina (30:54)

Exactly.

Lydia (31:00)

and let's start there.

Sabrina (31:01)

Yeah

Sabrina (31:02)

Yeah exactly we don't need to relearn these things.

Lydia (31:05)

Exactly. Well, all right, we'll end there, Sabrina. Thank you so much for joining the show.

Sabrina (31:10)

Thank you, Lydia. My pleasure.

---

00:34:08:18 - 00:34:17:07

Oliver

If this episode sparks something, curiosity, a new way of thinking, or something you're going to take back to the team, we'd love to hear about it.

00:34:17:08 - 00:34:24:23

Lydia

Absolutely. Make sure you're subscribed, leave us a rating, and drop us a comment. Tell us what stuck and what you want us to explore next. We want to know.

00:34:25:01 - 00:34:32:22

Oliver

This is the Audience Connection sponsored by Casual, the video partner for global brands trying to build trust with their audiences. We'll see you next time.