What's Working in Meta Ads Right Now with Alex Cooper | AI Animation, Research & Winning Formats

0:00There are so many brands that are making more ads than ever, but their performance is worse than ever because it's volume without intent. The sales calls I'm having is like, we're making more ads and we're getting worse results. Yeah, Alex, it's a chat, all things creative. It feels like the feed has got a lot more hyper relevant over the last 12 months. It's just thinking about how we can really make it feel like an experience that this person actually experiences on their organic feed. And I think a lot of brands don't do that. It's like you've got statics, UGC, high prod, and then AI almost as like another vehicle completely.

0:32But what exactly does AI videographer do? How can you almost build a world around the brand through that mechanism, creating characters, creating personas within that format? We are going hard on Yapparad. I mean, Yapparad's an AI animation. Incredibly confident in both of those formats. How are your team approaching iterations differently now? I think that's probably one of my favorite use cases for AI in strategy, is pattern recognition at scale. But I almost think the pendulum has swung too far where it means that people aren't doing the scrappy manual research.

1:04Welcome back to another episode of DC Diaries. We've got our second ever repeat guest today. Wow. Second ever. Quite the privilege. Only Matt Kelly has been previous for just how amazingly valuable his episodes are to us and the audience. Yeah, Alex, here to chat all things creative. I think it was what seven months ago when you were last on with us I think it was more than that it was last summer was it? yeah I know time flies things I mean a lot has changed yeah a lot has changed in the last four months let's just call out the elephant in the room in the spare chair Lucas has unfortunately been done by the train strikes and is rapidly line biking his way over to join the episode so expect him to just appear out of nowhere like magic in maybe five to ten minutes or we can just AI him yeah just ai generating like maybe as a character yeah playmation lucas um but yeah i i agree lots has changed um since we last spoke in in all things creative ai i guess what's changed for yourself as well since since you were last on launching a sass yeah yeah we launched in uh in january um and and went viral it's been uh very interesting going from like agency side to uh to software side but yeah there's a lot of transferable lessons as well which i'm i'm sure we'll get into yeah keen to dive into that before we get back into it let me ask you one question how much did you spend on meta ads last month for a lot of you listening it's probably quite a lot but how much of it actually came back to you if the answer is nothing you're going to want to keep listening incard is a financial platform built specifically for modern digital businesses the thing most people will fixate on is the fact that you can get uncapped cash back on all of your expenses.

2:48Because unlike a lot of these programs where the categories are pretty much useless to all of us, Incard gives you 2 % cash back on things that actually make up your cost base. Things like SaaS, ad spend, travel, and lots more. The things that you're just gonna pay for anyway to allow your business to run and grow. Money back on your existing outgoings. And if you're anything like most of our clients, that's gonna be a lot and it's gonna add up fast. But it's not just cash back. Corporate cards with multi-currency accounts, a proper connected banking experience, and direct integrations with things like Shopify, Xero, Stripe, and lots more.

3:24All in one place. And if you've ever had 15 tabs open, working from about three different bank accounts, and a spreadsheet that nobody's touched in three weeks, just trying to get a full picture of where your money actually is, then that's the problem that this solves. One connected view, your whole financial experience in one place. Brands like Asio Beauty and Hello Paris are already scaling with this. So if you're spending serious money on daily expenses and getting absolutely nothing back on it right now, you're going to want to look into this. Back to the episode.

3:59Kicking off, creative. Everybody loves this. It continues to be the most talked about topic. I feel like brands' competency level has definitely improved since we probably did the last episode where it still felt early in many ways, in many areas. If someone watching had buried their head in the sand for the 12 months since we were last on the episode, what do you think is like the biggest shifts we've seen in that time? Yeah, there's a lot. From a macro level, I guess I would start by saying that it feels like the feed has got a lot more hyper-relevant over the last 12 months, And it's no longer like message trying to serve you, you know, things that I think that everyone will like, but just understanding exactly what each person likes.

4:48And I guess this plays into why there's now more of a focus on personas and avatars in our messaging. And that's something that we've tried to do a lot of over the last few months. I feel like there's been a lot of development in AI execution. I'm sure we can get on to AI animation shortly or, you know, actually AI realistic. But last time we spoke, I mean, AI animation ads weren't really a thing. And, you know, the AI models that are out there were, you know, they were decent, but they're not anywhere near the level they are today. I don't think people were using Higgsfield last time we spoke.

5:27No. I think that wasn't really a thing. I think it was using specific individual models rather than one centralized tool. Right. And, you know, 12 months ago, you had people on Twitter saying they were getting a lot of performance from AI video. We were getting like sporadic performance, but now it's genuinely like, you know, competing, sometimes outperforming real human creators with our AI content. Yeah, I think it's like another layer. It's like you've got statics, UGC, high prod, and then AI almost. It's like another vehicle completely. um we've definitely seen the same we've obviously brought in some ai videographers who are literally just there creating concepts all day which has been really interesting when you sit like alongside high production studio like the cost difference the savings for us the impacts for clients so yeah definitely seeing that i'm sure we'll dive into some specifics a lot more ai slot uh can you can you understand then because i've heard you guys talk about this concept of the ai videographer before we have something that i think is similar but what exactly does the ai videographer do so we've kind of we were trying to think about how do we bring ai into our production process in a like compartmentalized way so a way that worked with our existing process so you've got creative strat account manager managing research concepting strategy scripting briefing and we just like tagged that videographer onto the end of that.

6:54So when you get to the end of that, you may have, you may previously had 30 % statics, 30 % UGC, 30 % high prod. Now it's like another, maybe we'll separate out 15 % and that'll get routed to that person. And they'll then, spending all day within Higgsfield and various tools, creating a process that can generate a high volume of those assets at pace for a client. One way we're doing that quite efficiently is for example if we're doing those claymations is like creating a character that we then use across many different scenes because it's far more efficient than and if we were to do um individual like concept generation um starting to do a lot of like ai podcasts concepts ai green screen concepts so still relatively early but it felt like the most balanced approach to bringing it in without having to like really shift the process that we already had in place yeah this is this is really interesting because i was speaking to someone on a consultant call the other day about this and and they were uh like we've done the same thing at ad crate we've actually got two of those ai videographers ai video as is what everyone call them and uh this i was speaking to a brand that does it spends like two mil a month and they um they have currently i think it was like seven or eight video editors um and uh they were going through the process of training uh their video editors on on how to use Higgs field or how to generate ai visuals i think it's a better approach to go and hire a specialist uh or specialists who are focused just on ai content whether that be realistic ai or um or ai animations because it is a different a slightly different skill being able to look at a script or a storyboard and understand how to paint that picture visually versus just like take clips that already exist and and edit them together yeah i agree it's like fitting b-roll generating b-roll or scenes of a broader concept that'll be lucas you'll pick it up yeah if you're um if you're listening and you want to see a drowned rat on screen move from spotify to youtube right about now yeah lucas has arrived and he's he's had a disaster um soaking wet yeah coming back to that point i think i definitely agree i think generating like b-roll um generating elements of a concept is different to matching and and bringing a written script to life visually you you your video on higgs field where you broke down that hume health ad where it's like all about visceral fat and you see like that generating within someone's skeleton i just think is a best-in-class example of a brand that's done that really well yeah it's not easy to to do something like that i'm not i'm not gonna say you've got it or you don't when it comes to like these ai video editors but like you see a ai video edit or an animation edit done by a regular editor versus one who is like trained up or knows how to use hicksfield knows how to paint the picture visually of what the script is saying and like you can feel the difference when you watch ads like that so going back through to point i do think it's better to hire specialists and try and train like your whole suite of video writers on how to use higgs field or um generate air video yeah you've obviously been closer to that that like process and the way we're building that out what have been the learnings from having someone so like hyper focused on on that even if we are a little we are relatively early into the process but yeah yeah i like i i'm kind of viewing it through first principles so like how do you how do you approach production and you project like the because i agree like the our editors similarly to yourself they they have the capabilities to do ai and they do to some degree but i think what it's it's almost like how can you when you start working at scale does it make sense to like split those specialisms we're still going through that that's my hypothesis that you may be able to be able to increase efficiencies by splitting those two parts of the process off So you have obviously the AI videographer that essentially is like what you would do with a creative strategy when they hand a script to a production team.

11:11They would go and shoot that content, build a rough timeline. Then once that had been signed off, that then goes to the editors who can then obviously polish it and do what needs to be done from that perspective. I'd say like the biggest thing that I've learned through the process is like setting constraints to strategists. and i think there's two that's twofold right because you don't want to like you don't want to kind of put a strategist in the box where they can't do or you know experiment with creative ideas but we found especially with like ai animation where it's like you know it's much more time consuming than like an ai podcast as an example if you have a strategist that's putting through like a two minute animated ad with like 50 scenes in terms of like higgs field tokens and then days that will take our ai videographer it could be like two to three days just just just to get that rough assembly cut so we've tried to set some constraints initially that our strategist can work within so we can really optimize the process and then from that point we can look at then how we scale it further so that is just based on like length of script below one minute and then the number of scenes that you could work within but then in terms of that you can then do anything within that constraint so yeah yeah i think that there's also like slightly more nuance in in who it gets dished out to at least ad create because there are different types of ai animation ad right so like anyone can like anyone can go and prompt in hicksfield for like i'm a tomato and i do this and i'm those ads like this the pixar style uh like explain right they're super simple to make but if it's something more detailed like the Hume Health example where you're actually going into the skeleton inside the body or the claymations where you've got to weave multiple scenes together and tell that story that's the kind of thing that you'd want like someone more specialist to do 100 % and I think we've seen that like we said at the start that's become a whole new layer to an ad account strategy hasn't it that just wasn't present when we were last chatting and when you were last on what else do you think's changed in terms of like macro trends like what's working right now from a creative perspective.

13:24Yeah, one more thing I want to say on that actually. Sorry. No, you're good. Go off schedule. But like when, I was speaking to someone else about this the other day, like when these AI animations first came, like came up in the organic feed and people started sharing them on Twitter, I was like, oh, you know, just another one of these formats that is going to be here today. Then everyone's going to rinse them and they're going to start to not work as well inside the ad accounts. but i do think this is something different at least for the in terms of the animations um and i saw a good uh piece of content on excellent thing from camillo the ai uh the animations guy yeah um and he was talking about how like you know when we are when we were kids like why did we learn through cartoons because like we ingest stories and messages better through those type of visuals than often through a human telling those stories.

14:20So I originally thought it was like something that was, you know, just kind of here for a short time and it's going to be a trend that's going to go. But I do think this is a format like, you know, it's going to be here for a long time. And it is, like you said, almost like a new division of ads that we should make in. I quite like the idea of like being quite, I think brands should be quite intentional with it. Like thinking of it as like you say, like if it's going to be something that's that's present for a long period of time how can you almost build a world around the brand through that mechanism like you've obviously mentioned creating characters creating personas within that format it's also working so well organically like there's some crazy youtube and instagram and tiktok accounts just like ai animation i feel like that's like a really good and we've always talked about this we've talked about it together you talk about a lot as well on x and you know organic is like an incredible source of inspiration And normally what you find with any of these trends that have worked on paid, like they, you find these like TikTok pages or, or, you know, Instagram, Instagram pages that all they do are churning at scale, like one type of format.

15:26So I saw one, which was like the, yeah, the, the, the fruit one or where it's a different vegetable talking about the benefits of that vegetable as the vegetable, but they were like churning them at scale and getting millions of views. And that was before it had ever been used as a format from an ad perspective. so I feel like you know really again it's like the obvious thing but that is where you're going to find where these trends start before they then translate into paid yeah same with stuff like Pinterest as well finding loads of interest in our teams finding so many things on their like I remember this part of it where it's like upcoming trends surfaced for you you mentioned right at the start um that view of Facebook becoming better or meta becoming better at like persona level delivery or just like like i think of it through the level lens of like fragment further fragmentation of audience within within the algorithm so it's gone from like macro persona almost like demographic delivery to like a micro persona it now feels like um type of messaging psychological zone bias i feel like that like the end state of that is the concept of like soulmate theory like can it get to the point where it has the ability to deliver on an individual level by asset and your ad library inventory becomes more impactful on an average level because it's able to to find those people um yeah i just what are your thoughts on that it definitely feels like that's that's the direction of travel on yeah i mean we were speaking about this before um before the pod like this everything that's happening right now and has happened over the last few months has made the stuff that we've been talking about for the last couple of years not like a nice to have like the stuff like persona analysis going deep on journey mapping creative consumption analysis like these things aren't nice to have anymore like they're absolute necessaries if you want to make uh ads uh that convert and i see like a lot of brands when i do these consulting calls we'll see everything that's happening and go like oh we need we need to pump more volume we need to make more ads than you were saying before the pod as well it's like there are so many brands that are making more ads than ever but their performance is worse than ever because it's volume without intent yeah and i just think that there's like a lot of brands out there that think they're making 100 ads a week when they're only making you know 10 20 ads five different ways yeah you know 100 the sales calls i'm having is like we're making more ads and we're getting worse results or we feel like we're in like a creative echo chamber because they're just, they're almost in this black box of the same ideas.

18:07They think like their brand, they think through the lens of their brand. They're not got any like diversity of thinking within the process. And that's just, they're struggling to break out of that as well. Yeah, we've started doing this exercise with brands. We've built this report that essentially like scans the ad account, downloads the videos and transcribes them and essentially will like bucket into, um you know psychological messaging zones um and it will show where the gaps are and usually what people will find is like they are they are speaking to one of those quadrants which is usually very fear focused versus actually you know different type of psychological messaging that people might resonate more with which is actually more positive up uplifting or talking more so about the after state than sitting in the problem and like by going through that analysis like it was so clear how many of our accounts were actually just like missing completely you know just not speaking to certain emotional zones um so i think that is a really like interesting exercise as well when it comes that's like a layer below persona and obviously you don't want to over complicate things but we've seen that be like a massive unlock not only being super specific on the persona but then really ensuring that we are like diversifying the way that we speak to different people yeah i think that's that's probably one of my favorite use cases for ai in in strategy is pattern recognition at scale looking all the ads uh inside your ad account picking whatever the uh whatever the category is whether it be persona whether it be uh awareness level whether it be you know emotional tone and then just go and like map out all of our creative across this matrix and then look for where the underserved personas or whatever it is, maybe, because you can find way bigger unlocks through doing that than, you know, just setting up a chat GPT image, a workflow that's going to churn out, you know, 250 ads a week.

20:03Yeah, it's that gap analysis. I think the real strategic thought that then is a layer above that is the prioritization framework you apply to that. Where do you start when you've got that view of like, here's the gaps based on where the upside is, based on the ease, impact, confidence of an action. And I think people are getting very caught up with that as well, like chasing, doing like all of your personas rather than just doing one if you're at a lower scale really well until it hits the point. And yeah, I think that's a really high level strategist is capable of making those decisions really well.

20:39Yeah, this is a great conversation. This is why I like, I get passionate about this because I get on a lot of consulting calls and it feels like it's very easy for like if i'm working an agency i'm working at a brand i'm a creative strategist i log in on a monday morning it's just blank slate i know by friday afternoon i have to make 30 ads or 40 ads or whatever that number is and you just sprint sprint sprint towards that that end of week and then monday comes the next week and then you know you're back to square one so i what i'm saying is i see a lot of strategies do that and then not uh you're not actually do the work to to retro on on the sprint and they did and go oh actually what was my hypothesis for this week uh you know did i prove or disprove that there are any open loops that we need to close um and doing that work so you can uh like start to be more intentional about the testing that you're doing yeah i think like as ai begins to commoditize execution more and more and actually strategists have more time unlocked as a result of that i think one of the one of the hardest but most important parts of that role does become prioritization and like knowing how how to prioritize what um and then as a result what that translates to in terms of strategy your point on learnings from output to output a week to week or whatever that time frame is is really poignant we had a we had one of our strategists on last week chris to do like a day in the life and we talked on that episode of how like this there is a lot of complexity in that in that week or two whatever that batch is there's so many variables so i feel like it's very easy to just get to the end and be like i've learned nothing i haven't i've i'm not compounding forward any learning i've just got loads of data yeah um i think that's just that in itself is a skill it takes restraint and like um quality of thought and intention um because otherwise and it may almost require you to move a little bit slower in some ways and do less like less variation with the like trust and belief that the process will compound there's another sentiment i'm that i'm feeling at the moment uh and i've heard from strategists that they've got more data than ever but they feel less aware than ever or like they got less information than ever um so yeah i i agree for us like we like we have our strategy time block uh in their calendar so it's like you know our we do like increments of sprints so we have like small ones on weekly basis and then monthly and quarterly and the level of retro is dictated by how frequent the bigger ones are the monthly or quarterly um but i mean it's it's it's difficult to to know what to prioritize like when i'm doing it i have a like a stupidly simple process but it's it's mostly based on vibes if i'm doing persona analysis for example i might like just literally make a spreadsheet and say like here are the different personas and i'll just have like what percentage of revenue they account for cac ltv and strategic prioritization i'd like rank them out of 10 and then like the one with the highest weighted score i go okay that's where i'm going to allocate my creative between you know 60 number one 30 to 10 number three then that will what dictate that we will dictate is like the next sprint i feel like that's what what you just said there like it because you said it's like vibes and it i feel like that's what an experienced strategist is when they've just been in the game long enough and it's it's it's just it's like subconscious isn't it really it's like that intuition just from being just having the experience through different types of scenarios is to actually know how to prioritize because you're right like there's you it's so easy to overthink it and i think strategists have a complexity bias especially because of how many data points there are now so it's like how do you actually bring it back to the basics and simplify the process while still ensuring that you are driving that prioritization 100 why i think that's really important is and this is another thing i'm speaking to brands that are suffering with this issue as well is like i think the intuition is even more important when when like looking forward because it's easy to pull ltv data like who we are acquiring now what our current customer looks like and validate that with a lot of data points from the research but it's hard to see like a low volume of data that points you to a completely new person to advertise towards and have conviction that that's the right place to go so example would be like creatine gym bro obvious creatine menopausal women 18 months ago not very obvious find like a few things on reddit and be like shit this is like a massive opportunity so you go there you don't go further and further into the current persona we audited one of the biggest supplement brands on meta and they were spending a lot and they were making hundreds of ads a day and they were basically just advertising to one persona and going deeper and deeper into this persona but that and i think that's just like confirmation bias in data yeah 100 it's interesting you guys like have a have a similar approach i was speaking to someone the other day about like this is this is also true in copywriting because we're going to talk about i'm sure like you guys did a great episode recently on on codifying uh your intelligence and like you can build skills and build processes to to make better script writers or make better copy for uh headlines but then like i really sat down and thought about uh you know when i'm writing a script why like why do you write like this line after the previous line like why why does it flow in this way i don't know about you guys like i for the most part it's just vibes and a lot of the top copywriters i speak to is like yeah i don't really i mean there are some frameworks and processes that you can apply to certain parts of it but there is no reason just feels right to write this line after the next time or change this word to this word yeah it's reps isn't it it's like when you've done anything enough like riding a bike like you don't think about the next pedal um forward it just happens i think this is interesting as well that what you were just saying about you know strategists even though they've got so much more information now that they actually feel almost more confused at times and i think that brings us on nicely to this idea of uh that we kind of speak about quite a lot where we are obviously very bullish and ai forward in everything we do as are you but i almost think at times the pendulum has swung too far where it means that people aren't doing the scrappy manual research that actually means that they are truly absorbing the information it's like the equivalent of you know if your report is just auto generated for you and you're just reading the report versus you know an element of manual input to that report it's the it's the equivalent of like reading a whole book back to front versus just reading like the back of the back of the book so i think there's i'd love to hear your opinion on that balance between what we do automate through from a research standpoint and what research we do that is more scrappy that isn't it's scalable necessarily but it can extract insights that maybe ai wouldn't yeah this is this is a really interesting conversation because like we do a lot of our uh research with um with ai and a lot of the pattern recognition like looking through uh you know richer data like customer reviews push push service uh support tickets etc um but again it's so dependent on like the context that you feed it and the the thing would like when you're i'm sure you guys have all been like you're deep in like eight claw chats and you just want to bang your head against the wall and it's like uh regardless of uh whether ai is right or wrong it will give you the answer confidently yeah so it's like it's so difficult to understand uh you know what is the correct answer versus what uh just needs more context to give you the correct answer and i agree you want to make sure that you're doing a combination of that plus um manual research so i tend to lean on ai more for the the stuff with rich data anything that is harder to scrape or more obscure sources and we'll go with a lot of manual research so for example comments like comment mining TikTok comments, YouTube comments, Facebook groups, just places where it's hard for AI to be reliable.

29:21And I think you can really pick up a lot of sentiment in comments because, at least from what I've seen, like a lot of Facebook groups, Reddit seems to be more, like, it feels like it's more AI generated than it previously was. But like going into the TikTok comments, Facebook groups and comments, you can find a lot in there. So we do a lot of that. sometimes i found that ai tends to struggle with uh web scraping um so we do a lot of that uh manually as well but in general it's just it's hard to find the balance um but that and then creative consumption analysis so it's going to organic feed and getting a feel for what type of content people are consuming what does it look like what kind of creators what's the lighting etc yeah before we get back into it let me ask you one question how much did you spend on meta ads last month.

30:12For a lot of you listening, it's probably quite a lot. But how much of it actually came back to you? If the answer is nothing, you're going to want to keep listening. Incard is a financial platform built specifically for modern digital businesses. The thing most people will fixate on is the fact that you can get uncapped cash back on all of your expenses. Because unlike a lot of these programs where the categories are pretty much useless to all of us, Incard gives you 2 % cash back on things that actually make up your cost base. Things like SaaS, ad spend, travel, and lots more. The things that you're just going to pay for anyway to allow your business to run and grow.

30:48Money back on your existing outgoings. And if you're anything like most of our clients, that's going to be a lot and it's going to add up fast. But it's not just cash back. Corporate cards with multi-currency accounts, a proper connected banking experience, and direct integrations with things like Shopify, Xero, Stripe, and lots more, all in one place. And if you've ever had 15 tabs open, working from about three different bank accounts, and a spreadsheet that nobody's touched in three weeks, just trying to get a full picture of where your money actually is, then that's the problem that this solves.

31:22One connected view, your whole financial experience in one place. Brands like ASIO Beauty and Hello Paris are already scaling with this. So if you're spending serious money on daily expenses and getting absolutely nothing back on it right now you're going to want to look into this back to the episode i think it's like as especially if it's you're a strategist that hasn't worked you know you're still learning about the brand yourself and the the personas that you're targeting and um you know that whole market i think there's there's there's a lot of value and like for example start started working with a new brand recently and you know I wasn't involved in the process but I was just quite interested in uh the pain points in which it was like you know looking to solve so I was and I haven't done like manual reddit research in a long time like because we scrape a lot of that now and you know we're able to to do a lot of that uh through automation but the value that I got just like going through and manually searching through Reddit and just actually just reading the whole threads, like stuff that wouldn't have been picked up necessarily because it didn't have the upvotes, but just seeing how people are interacting was actually just like a really valuable exercise that maybe people wouldn't do.

32:40And it's not something you do all the time necessarily, but especially to just really get, do that deep research initially to understand the brand. It's like a really valuable task. And then to the point a minute ago, like that's how you improve your taste or like you know that's how i improve your intuition so that when you're copyrighting you have the right vibe because you've been like deep in the manual research examples on reddit is i've just done the same over the weekend for a brand project i'm working on and it's like it scraped reddit and then it obviously served all this data on the product but not the context in which the product was mentioned so then going deeper and really read because it was often when when i went for a few threads it was like the storytelling that had led to the product or the problem being surfaced gave you so many ideas for like narrative driven problem aware ads that wouldn't have been provided with like just the raw extracts um so yeah i think it's really interesting how are you um obviously with meta's creative diversity uh updates um last year how How are your team like approaching iterations differently now?

33:48Like, do you see them still as valuable as they were before that update? Or is it something that you've kind of evolved and changed as a result? Yeah, I don't necessarily agree with some of the sentiment online that, you know, iterations aren't valuable. They don't carry weight anymore. I think they absolutely do. They just need to be meaningful. Yeah, exactly. So whereas, you know, before we might have looked at an ad that's working really well and changed, like, I don't know, the text overlay on the hook or changed a clip in the middle of the ad um or added a couple of clips somewhere in the middle of the ad now we really try and change the uh dna of the ad and make it feel like i mean still the same uh concept but a different uh feel of the ad and we'll do um less iterations but try and make them like more different of a feel than before we've got a lot of brands that literally like some of their best performers right now are you know ads that have been iterated on so i don't think that it's something that that should not be done i think that there's a people are slightly over indexed on like you know don't do iterations because of andromeda um they just need to be meaningful yeah i still think small iterations are worth doing if the concept's new because you try to like when you're trying to find a message that works or you say you're iterating on a headline in a static and you're you're early in the journey of exploring like a pain point or a persona or an angle and you can still get value from the same image with 15 different headlines as a signal for what to do next so yeah but i think when you've found a concept that works that's when it comes to that point of like really needing to to drive some significant differentiation yeah 100 yeah i completely agree with your point i think that's exactly what we've done we've just we've sent obviously not to the to the point but like halved the amount of iterations we might do dependent on the brand and pushed more of that volume towards more net new variations but then as you say with that other half being way more intentional about what you actually are manipulating within that creative and why we have a brand that we work with i don't think this is valuable or that anyone should copy this but i just think it's such a really interesting case study we work with a brand and we took them on board when they were doing sub a million dollars ARR and now I think they're at I'm gonna say like 40 or 45 and there that we found one um like winning script super early on that that scaled them up uh and they've literally just like taken that script and or taken that concept and tried to remake it in as many ways as possible so I turn it into AR animation turn it into authority figure turn it into a podcast ad turn into a founder ad um and like most of their account's been carried to 40 45 mil off the back of that one concept again people listening to this should definitely not do that um but i at least thought it was a very interesting uh you know example of why uh you know we we can all say things on here but it's just like you know you just got to see what works inside your ad account it makes sense if you found like a really good problem aware ad message and angle that appeals to a pretty broad tam makes sense that just iterating on the way you deliver that would unlock more and more people who have different content preferences over time yeah just making sure i think the caveat to that is like you have to have the the audience profile i guess like the size of audience to go after i i rate that strategy but it's like part of a strategy isn't it as you as you're saying like it's yeah doing that whilst also doing everything else but i i i also one thing i've seen work really well on on the reverse is once you found like a really great winning script or framework use that almost exact script but then adjust it to fit a brand new persona that you're looking to target so the messaging only slightly changes but do the exact same framework a slight change of the script and obviously the hooks because the because obviously it will be targeting a different persona and then rerun it and we my my brand like i we did this with one framework that was just just working really well and every time we tested a new persona against that same framework with a just slight tweak of the script to align with that persona like it became a top spender so like it's a good way of you know once you really have those solidified frameworks testing new things whilst keeping the variables the same that you know have worked previously saw a linkedin post and this is missing me switching gears a little bit i saw a linkedin post from yourself that i definitely agree with we had a conversation last night lucas about this but like the concept that you're winning ads are either those like super ugly um native pieces of content that kind of naturally blend into the organic feed or they're like hyper optimized elite direct response storytelling through the lens of like a podcast ad with like classic three second increments really well scripted um i think that's been a a shift since we since we last spoke as well um certainly around those like native yapper ads like some of these really organic feel organic styles that have sort of come to the forefront and deliver a lot more performance yeah i mean i i agree well i mean it was my tweet i do i do agree with my tweet uh hey um obama meme where he puts the medal on himself

39:26uh yeah it's interesting because you look at a lot of um a lot of the like creative that gets put out if you actually scroll the organic fee for five minutes and then you look at most brands creative like there is a big difference or like what most brands think are ugly ads sorry there's a big difference in how it feels as a piece of content whether that be how it's shot or like how edited it is or like whether there is a a soundtrack or not um i feel like most brands hear like oh make ugly ads and they try and go do to do some version of that but then they still go and like you know over edit it or put captions on the whole ad or have you know light in somewhere that maybe doesn't need to be light in um and uh yeah as you guys said like a lot of the ads that are working today are less scripted for lack of a better term yapper ads i just feel that a lot of brands getting caught in the middle of like trying to make a direct response you know like hollow socks dr squatch pros style of b-roll mashup and like something that's just someone picking up a camera and yapping they have something in the middle and it doesn't really hit it's not really authentic enough to feel like a yapper it's not really good storytelling enough to feel like a b-roll mashup like direct response ad and uh that's where ads go to that and i believe will continue to die um as you know content gets more and more ai i think that these yapper ads are going to get more and more uh influential because people are going to crave that authenticity they're going to crave listening to actual people rather than people reading from a script so it's just thinking about how not how we can make our ads more ugly but how we can really make it feel like an experience that this person who we're talking to actually experiences on their organic feed and i think a lot of brands don't do that 100 i think another trend in the ugc is that similar to that is like more hooks and and concepts that sort of start with a in an environment that's very native to that consumer that maybe isn't the house like gyms um i mentioned one where the guy was carrying his daughter down the road um someone like third part person filming themselves cooking just that like bringing stepping into the world of the customer from a visual perspective um i think yeah like that concept of like authenticity looks very different for different demos as well like say a 50 what 50 year old woman would see as authentic and what someone at the age of like 21 would see is authentic is very different.

42:06And it's completely different per persona. Like if you're a luxury brand, you shouldn't hear this and go, oh, I'm going to try and make something as ugly as possible. For you, like authentic and relatable, maybe someone sitting there holding their phone in business class on a flight because that's what appeals to your avatars. so yeah it's just like really working out exactly what your prospects are seeing on their feed and how to recreate an experience like that I guess with like how UGC has changed over the last 12 months and you know if we rewind three four years ago to the fact that you literally just take a a very proven script framework direct response and every creator you just send them like that exact script as we've said with like these yapra ads now like it requires a completely different approach obviously with partnership ads as well there's there's there's more of like a push on really tapping into creators through that lens how are you in the agency approaching ugc differently in terms of that process yeah uh we we are going hard on yapra i mean yapra ads and animations like super bullish yapras super bullish animations that is like not necessarily most or maybe most of the ads that we put out uh now they're just incredibly confident in both of those formats um our yapra ads today are uh mostly uh not fully scripted um i think uh a like underrated or not as much talked about part of like what makes a yapra is the actual creator team component of it and making sure you're sourcing the right creators because your brief, your idea, your like hooks that you dictate can be as good as possible.

43:56If you don't have the right creator paired with it, then it doesn't matter. You can have the best strategy in the world. So for example, I was on a consulting call a few months back with a pretty big brand who are struggling with their yap rads. And I was like, okay, how many creators are you testing this across and when they say well we've got a team of 10 creators and it's like yeah you need to test a lot more than that to find actual top creators that are good at yeah it's tough like it looks easy but it's a really difficult skill to get on camera and talk authentically about a product uh especially if you don't actually use the product yeah so you know they end up i i uh told them to go and hire a creator source, which is something that we did at AdCrate.

44:46We now got four people in our creator team and we just brought in two extra people from the Philippines whose full-time job is to scroll TikTok and Instagram and scroll different platforms and recruit these top yappers. We test them and we have hundreds of these in our pool being tested at a time. And then the best ones get kind of graduated into our retainers that we use often. But it's easy. It's not easy, it's it's one part is coming up with the idea and then kind of the putting the brief together but then also making sure that you've got enough top quality creases which i don't think a lot of brands spend a lot of time doing how are you from a if a brand's listening to this and they're like okay we've got the person create a coordinator or whoever's in the i've just brought them in so we've got that person that's scraping um what do what advice do you have for them in terms of like for their strategist briefing what goes into that brief versus like a fully fledged this is what the script looks like yeah um i don't know how how you guys operate but like i i am a little more laissez-faire with with strats on on how we get them to brief so we do have brief templates but i more so say like you put whatever you want down on that brief i don't care if you use a template if you record a loom, if you make a drawing on a notepad and you scan it in there, you just do whatever you can to best get across your creative vision to the person who's reading this brief.

46:16In most cases, when it comes to Yapper ads, we have hooks dictated. So we say like, here are five different versions of hooks you can run. We'll put talking points and then an InSpo reference. if it's like a really good yapo we trust and work with on retainer then you can be a little bit more you know you can let them go and let them have freedom we have a few creators who we can literally if they know the product really well you can just give them like full creative freedom say i don't want to put any constraints on you um but for the most part we are uh giving them direction with hooks and then just let them go off on uh on different talking points nice yeah i love that yes it's so true isn't it it's like that idea of just giving the prompts and showing them an example of what the best in class m result looks like and then they will be able to fill the gaps as a result of that like it's again that thing of you know not over engineering it because if you over engineer and you've got this huge brief then they don't know where to focus on and it just as a result the performance is likely just not going to be authentic yeah and then the other thing that i'd say on that is also that having a great relationship with your creators will improve like how good the outcome of that brief is.

47:32So this brand that I was talking to, like before they didn't do anything, but now they've got a whole like creator program where every other month or quarter, they fly them into the office and they do like training sessions with them. We've done a lot of training sessions, like virtually with our top creators before and to get them to have more buy-in than just like another agency or another brand they're working with and really feel like I'm a part of this rather than just I'm someone who's, or read now a bunch of ad briefs for them it's a really good idea i think um i think great creators are in short supply there's a lot of creators but there's not very many good ones so i think you've gotta you gotta work hard to retain same on tiktok shop actually i spent quite a bit of time over the last few months helping brands learning tiktok shop and it's like a lot of a lot of that platform is just mass affiliate gifting but the brands that win are the ones that retain high quality creators for a long period of time because they're getting slammed with offers and get them to make 30 videos a month rather than one it'll just transform the business um two final things on like creative trends because i want to jump into some ai bits and parker and parker two other things that i think you mentioned positive you mentioned positive framing i think that's going to be a big trend like macro trend over the next like 12 to 18 months i feel like the algorithm is just scaremongering and fear and like shame i feel like and i think that's a symptom of when it being the easiest thing to go after and when you when we do that analysis on pretty much every ad account it's the only angle that's saturating i think that that is driving and from a macro perspective like general culture it's like desire for more positivity more positive frame i think we're going to see that play out in in the creative world what do you think that looks like like in terms of ads yes it's that's that's the difficulty and i think again it's similar to when you look at like satire there's a reason why people don't do it as much and one of those reasons is because it's harder to execute on because naturally as humans like especially through social media it's controversy sells and like you know really agitating problem cells but I think it's more about sitting from a script standpoint like how do you sit more in the after state and you know the the kind of or the ideal self versus actual self so it's like that that ideal self is really kind of going into this is how it feels to get to to reach the other side and you're kind of it's more through that that empowerment lens versus actually doing that more also from a more negative fear-mongering lens.

50:16Yeah, I agree to some extent. I don't know. I think it's like diversity though, isn't it? It's not like that's what you're doing. It's like that is part of your strategy. Yeah. I think it's harder to do. Less brands will crack it. The ones that can do it and do it well will cut through massively because they'll serve a completely different message to most of the market. Yeah. I think satire ads can somewhat sit in that because they are funny and positive rather than shame and negative. Probably why when you look at motion reports that relative to the percentage of the algorithm that is satire, the percentage of spend allocation is so much greater proportionally.

51:02I'm not sure how it's going to show up for, I'm sure we'll be able to maybe in 12 months chat through it as something, but I just feel like that's that that is going to be a macro trend um and something that people want uh over time another thing I've been I was doing some work with for a brand this morning is like this is just a bit of a random point is like I feel like a lot of brands have this complexity bias in the way they speak to customers and like you should tone down the reading age of like almost every concept you make to like people as if you're speaking to someone who's like 14 15 i feel like again it's been too close to a problem and speaking like you obviously know it very well the customer doesn't going really like science led rather than like problem led i don't know how to how to phrase that any better but yeah good example of that we i won't name the brand but we we have a brand we started kind of going down this route of like heavy expert driven ads which do work i'm not saying that but then almost that messaging translated then from a consumer standpoint which then wasn't working as a result and i think we were almost speaking in a way that because you're so close to the problem as you say you think that the consumer understands that like you do but then actually what we realized worked was like dumbing it down like you would with a five-year-old it's like visual metaphors so like pouring visually showing like the difference of of something through like the amount of water in this jug versus this one or like if it was a gut health issue like just pouring the water into a jug with holes in it and it's like leaking out you know so it's it's it's like by kind of creating those visual metaphors and dumbing it down to something a five-year-old would understand it's like easier to cut through yeah pairs with ai ads that human health example yeah it's like when you say something like visceral fat in an ad and you know what that is don't assume that the whole audience explain it for another sentence bridge it and then you and i even and so yeah it's easy to take like if you've got if you think you've got this just whack a script in claude and ask it to write it for younger reading age yeah that is a great use case for the ai visualizing and we sometimes i think think that you know because we will watch an ad 20 times yeah but like when the person is actually seeing the ad in their feed they're watching the first half a second and then like 20 minutes into their instagram scrolling session they're looking for every reason not to uh look at your ads they see that first split second if you if you can't make it immediately clear why it's relevant to them and and how it can help them um and just in a really simple way they're just going to keep scrolling but we look at our ads and try and make them perfect through the lens of like oh i've watched this through 20 times so of course it makes sense to me versus like the first impression i think i was again i think it's this carries across the full funnel i've i really have this belief that um brands aren't generally customer centric enough and like they're everybody in the business as well like they think too much they're too close to the problem therefore they create too much complexity because they're thinking about it every day not appreciating that you need to write every piece of creative and you need to design every pdp for someone who's never seen this before and just internalizing that as a thought process has a massive impact on how you actually approach every part of marketing i think just more customer centricity yeah also kind of uh uh different but related that's another reason why uh research sources that are pre-purchased can be very interesting to study like ad comments tiktok comments youtube comments because you're getting uh sentiment from people before they buy and like their actual mind states then uh versus like only looking at the um customer reviews or post-purchase surveys that might uh give you information on someone after they buy but like understanding what their objections concerns are before can help you talk to some people 100 so we'll get on to parker shortly which i'm really excited to talk about um firstly i'd love to know like how you are using ai to optimize your own life you know that could be around work specifically but let's let's kind of shift gears away from creative strategy for a moment i'd love to know some of the things you're personally doing there yeah uh i've been digging way too deep into claude coden i've got bags under my eyes that's because i've partially because i've just got back from traveling but also claude coden um there's like a few fun things i've been playing around with i don't know if this is like you know i'm going too much off the scope of this pod here uh but uh for example my uh my family are like really into football or soccer um and we will like every like big tournament like a euro's or world cup we'll do like predictions but almost like a march madness uh style like where everyone predicts what every score is going to be and um and i can get points for how many like predictions you get right uh so i just vibe coded a an app where you can everyone on their phone you can just type in the scores and it will also update based on like some api that i found online that's got all the world cup scores so for this upcoming world cup we've got a whole like app for our uh family predictions i think is a pretty fun uh i saw something the other day on twitter i don't know you guys saw it you see the the guy who turned his um his run it gamified his running you see that with his previous time but he was running away from his previous yeah yeah yeah i thought that was so cool and he can like pick up coins as he's running i can see like his previous runs I mean that's so cool I want to build one of those for like weightlifting in the gym so it'll check your form like through the glasses and it'll tell you you know sometimes it tells you on the on the machine like if you've got a full range of motion in I want to build something like that for like if you're yeah Ben should know anything so is that through the meta Ray-Bans yeah right interesting and like in real time it'll could like check your form have you got a pair of those yeah are they good I mean I don't use them as much as I thought I would yeah they look really cool when you see like someone upload a story on them and you get them and you just don't use it that's what put me off so it just feels like like an ipad you forget it and you just never use it yeah yeah yeah one of my friends actually in um cape town is like a proper claude code nerd claude code nerd and i respect it so much he's like automated his entire apartment so he's got like a little dashboard that he's built on his ipad next to his front door whenever he uh walks into a room the the temperature like goes to the perfect temperature that it should be uh he has red lights turn on at the time that like they're for his optimal bedtime so he can optimize his circadian rhythm uh the curtains open at the right time when he's got the right amount of sleep and it's all controls this little dashboard he built on his ipad for through american psychos oh mate it's i mean definitely not worth the time that it took him to make uh but incredibly cool i've seen people on i saw a guy on twitter who'd who'd like set up a camera wired it into hermy's agent and then it was just all it was doing was remind him to drink water it's like you haven't drank in like 20 minutes drink water just like into his for his agent um it's funny i i haven't got explored much claude code for personal use i'm just you've made me think that i need to actually have a maybe sit down and think about that especially after my disaster today i feel like maybe that's where I need to be optimizing more.

58:51Schedule and location. Yeah, that could be the move. If you were to look back at like 12 months ago to now, like what tools, other than Cloud Code, I feel that's an obvious one, are you using more and what are you using less, would you say? Example for me is obviously, I don't think I've logged into ChatGPT nearly a year. I cancel my subscription. Yeah, which is wild to say. Like if you were to say that a few years ago. If you were to sit on the previous podcast and say that, I just wonder if we would have just been mental. you thought it was like armageddon or something higgs field obviously that wasn't a thing which is equally as crazy it was we probably was a thing but it wasn't like used much by i don't know about you guys i'd say my tech stack and our team's tech stack is actually compressed and it's pretty simple yeah when it comes to ai because i mean like the uh like most of it most of our work is generally genuinely uninclued or like uh we've also got our uh like I think you guys call it the CSOS.

59:48We've got like AI OS that we've built through Claude Code. There's just basically houses everything about create financials, goals, KPIs. We've got an ad building platform inside there that has like the Higgs field canvas style, like ad builder and it pulls in different APIs. Outside of that, Higgs field, Parker. Brilliant segue. Yeah. Brilliant segue. and then I'm looking now Whisper I'm still a big fan of Whisper I've seen people come out lately and say that they're not massive like it's played out for them and it's played out a little bit for me the iOS update I think sorry the iOS update it was great until they switched the iOS update then something about the newer iOS for me it'll just often I'll do like a three minute voice note and it'll just delete itself really?

1:00:37oh that's the worst I've not had that eight minutes into a context dump and it was just like it's been slow for me but I've not had that happen before. I love the videos on Twitter of the offices where they've got like the foot pedal for the mic. That's a great idea. Really? Oh man, that's a genius. They don't have to say they just type in the foot pedal the way into it. Oh man, I love that. That's so good. That's really good. I love that. Outside of that, I mean there are tools that we use a little bit you know, for some concepts. Reddit answer I think is good but like outside of that in terms of AI tech stack I mean there's not much that I can see here.

1:01:14yeah i mean we're we're the same to be honest um i think yeah similarly to yourself we've kind of built our own custom operating system for the agency and for create a growth strategy um but yeah very similar setup to be honest isn't it crazy how like this time last year i remember we did a course last summer right i think like we were just about to go into it we're just in the middle of it as i as we did this podcast last time oh yeah and like the biggest request by far in that course we wanted to do the whole course about context engineering because you know we saw what was gonna happen in 12 months time um but like the biggest request was uh nann and gum loop workflows i know everyone everyone was like please show us the workflows you guys are building i'm like so true well you know it's punished that's crazy i'd literally forgot nann was even a thing i use it a little bit for um some like some automations but barely barely compared to previous they have touched in yeah yeah it's just not needed um so that takes us on to parker which is your ai creative creative strategy software that i'm super excited to talk a bit about um i'd love to just like delve into how you found one launching the product and that shift from being like an agency owner to uh you know building and and driving a sass company like they must be completely different and things so talk to that a little bit you mentioned the timing to me when it's interesting to reflect on when you launched it literally right before yeah well i think it was either the week of or the week before called code and it's like mate yeah it is tough we didn't realize at the time like how significant that would be yeah i guess uh like if i could give myself advice like my prior self advice uh on like how to build a software in 2026 is to to build it for a world where your product is like, instead of rigid blocks, is like intelligent clay that can be molded into everyone else's workflow.

1:03:18Like MCPs and APIs so it can plug in. Because, I mean, I think I tweeted this a few months ago, but just the more I work in Claude, the more I just think that Eric is going to be done inside of Claude. It's like slope, not intercept. Build for where we're going to be. Yeah. Because it takes six months to build anything. So it's like, how do you build for the things that aren't going to change rather than the things that are changing? really hard to predict yeah everything software will become a foundation that is as you say like molded to what you need it for um like for example i could imagine the same thing with with like slack you know how can you mold slack in a way that fits your organization specifically that isn't in just a block as as it is right now um but yeah that's super interesting yeah it's it's difficult because we built for a world uh before that and then we realized pretty quickly oh like the way that we do it like the way that everything works is is now completely flip on its head and i know it feels like right now only you know not only but like there are agencies and brands that are like ahead of the game who are building their own ai os and it feels like it's a you know to people who are you know brand owners and or or teams who just don't have the time to do this it feels like oh wow i can never do anything like that like that's just gonna be for the you know elite agencies or elite brand operators but i don't think it's crazy at all to say that every business will have a version of of that um in the not so distant future so in that world where um where everyone has their own ai os for whatever their business is like how like what's your body prop how do you plug into that how do you give them something they can't get uh by building something on cloud code or codex or whatever it is so yeah it's just been different to think about instead of you know the outcome of the agency which is you as a convert um to build something that has like actual defensible ip yeah i mean like i think we're we're we are i think we're further off that than we think just in terms of everyone being in a position to be able to build their own custom operating systems because i feel like we are you know if you look at like d2c twitter as an example like that is a very small subset of individuals that are you know at the forefront of the ai revolution whereas actually got remember there's like 80 90 percent of the rest of the market that it's going to yes they're using things like chat gpt and probably claw just in in the actual llm interface but that next level up i i feel like there is for a period of time at at least for the next few years, going to be a strong opportunity still with, you know, very specific softwares without having to build that custom customization into it.

1:06:07Yeah. Sorry, I think the value is like building software as a service, as a layer on top. So I think the, I think traditional SaaS is value over time is going to trend towards zero, but SaaS and service is going to converge. And you're going to sell an outcome to a business and you're going to deliver that through a combination of time and tech. And that's the direction of travel that we're going to move into. Whether that's like, and you can see a lot of traditional softwares already move into this and staffing up massive customer success teams, which where they didn't have that previously. And it's just the agency SaaS conversions, isn't it?

1:06:54And I think the final piece is the billing change. Like how do you bill in that new reality where execution commoditized? I think that's just like you bill on outcome, you bill on growth, you bill on upside and value creation. Which is, I mean, for an agency, super exciting. Really exciting. Super exciting. If you're a good agency. Well, yeah, if you're not, it's going to be tough. Like, yeah, if we are still, however long it is away, once, two years before, uh like everyone has that type of software like if you a system built into their business then like if you're an agency today who can then go and build something and then build reasoning traces and compounding context into that that all your team can can build off of for a year two years you're gonna be very far ahead of people who are not uh building for this world yeah i think this i can't remember the stats but it's like the percentage of people who are using over a thousand who are doing over a thousand prompts a year on AI is just like less than 0.1%, I think, or something like that.

1:07:59I butchered that, but it was, you get the gist. Don't quote that data. I also think the point on DTC market is like, I had three calls in the last probably four weeks with brands that had no idea what Andromeda is. It's like, even in the creative landscape, there's still people who are just the adoption we're very far along the adoption curve I think relative to many businesses you look on Twitter and you feel like you're always behind but then in the grand scheme of things so final question what's next for Alex Cooper? good question we're going to continue to scale the agency make it a lot more operationally efficient like we're going to continue building our own version of the ai os and um think about what we can spin off on the back of that parker as well just gonna i mean keep on making it like the productized version of that internal agency system that we have um we're in the process of uh doing a massive like a context revamp of it so just trying to make it like a as good as a a really smart senior creative strategist um and make that accessible for all brands so yeah just the agency and the software are the two big projects i've got i'm real shout out to your co-founder at the agency by the way he's killing it at the moment i've been like seeing his content popping up a lot he's ali is ali is a machine he's built most of the claw code stuff that we can tell he's like so deep he literally will he had he messaged me the other week saying uh his wife got angry at him because he was clawing at 3am in the morning and she was like come to bed and he was like nah I'm opting so he's deep in it I learned a lot of the stuff that we built in Clawco from him yeah nice well thank you so much for coming on I think yeah it's been amazing having you sorry I missed the start we'll get you on again Q4 wrap up I reckon I'll be down for it anytime you guys want to do it where can people find you?

1:10:08X is probably best I am most active on there I'm also on LinkedIn YouTube although I don't post as much as I should but if you just search Alex Golf Cooper on X G-O-U-G-H I think Cooper links in the description we'll also put a parka offer link which is a great product by the way so definitely definitely try it out for all of you guys yeah who are trying to make more and better ads which should probably be everybody watching and we'll we'll put a link in the in the description check it out yeah thank you for coming on thanks for having me guys cheers bro cheers