AI Hooks for Meta Ads: Sora 2 and a Repeatable Process transcript

Today on the interference podcast, my COO and business partner in AJF Growth, Patrick Kadoo. You know him from building and selling supply razors in public. If you don't know him, maybe you know him from recently coming on the show and talking about how we're building a business together. Or maybe you don't know him and you're just going to like him. Anyway, Patrick and I are going to talk, we're going to start doing this, I think, every month where we're going to start talking about what we are seeing for early to mid 8 figure e-commerce businesses, the folks that are sort of in our core client range and what we're doing as an agency, how we're building to try to solve the kind of organizational problems and strategic problems that come up for agencies, for brands, for software companies, everybody in our space who's thinking about how to build and grow businesses and sort of what that ecosystem is looking like.

Today on the show, we're going to talk specifically about Sora 2, which chat GPT just released to the public. We're going to talk about the difficulty of building repeatable process more generally, just how hard it is to do that and why it's important and why we think for agencies and for brands it's really important. And then the resisting the proliferation of tools that is always a thing. Shiny object syndrome, the constant Slack, tap on your shoulder, endless notification world that we live in.

How do you resist that and actually build deep work? I have thoughts. Patrick has thoughts. Let's get into it. Let's not delay it. Let's get into it right now with Patrick Kadoo from Asia Growth. >> Hello, Patrick. >> Hello. Hello. I'm excited to talk about these things. >> I know. Well, some of this is like is like you and I probably need to talk about it. >> Yeah. >> So, we might as well record it and just >> Yeah, exactly.

None of this has been premeditated. I'm actually curious to see where this conversation goes. >> Me, too. Where do you want to start? You want to start on sore, too? We should say you're the COO. You you tend to I mean, you've got I don't know. you're the COO. You you are building out a lot of our internal processes, but you built and ran a brand as a CEO. People who don't know you, like you you understand how to think strategically about these sort of things, too.

Um I'm a little more in the work with our clients dayto-day. So, just give just to give uh people a sense of kind of where we're coming from here. Yeah. >> Um and and some of those things. So, um >> so Sora 2, let's talk about Sora 2. Uh >> so you have played with Sora 2. tell people like I mean I think people know but just like the very very quick like what it can do right now uh from an AI perspective like what what is the promise of it I don't I don't even know the total video length that's possible in sor 2 right now >> yeah I think it's like 10 seconds or something >> that's okay yeah that's what I thought and uh what what did you what are you seeing with like stuff you you've made cool stuff for a client the other day like really cool stuff >> uh and and some of the videos I've seen from store are like the most impressive AI videos I've seen the fastest Uh but yeah, what what do you think about it? >> I think and um I I think we should set the table um by saying AJF Growth doesn't use a ton of AI content in its advertising yet. >> Yeah. >> Um if we do, it's >> sort of not video content, but we we use AI a ton to write scripts. >> Yeah.

Yeah. Yeah. I mean production wise and like visuals. Now, we do use it for B-roll and for for graphics, and we're starting to use it more when it comes to like videos, uh, plugin in B-roll. Um, but, you know, we're not like we're not making like huge threeminut ads with with AI quite yet. Um, but we are always curious about we and I'll just say me, I'm I'm extremely curious as kind of the process guy that's got a whole team, you know, that's pumping out, you know, hundreds of ads each week for our brands.

Um it it's hard to you know there there's there's always some new tool being announced and released and you know it's one thing to go oh that's a cool tool like we can do like fancy shiny things with it. It's another thing to like make it a core part of your stack. Um and we haven't really decided we haven't really like jumped all in on a tool yet. So putting putting that on the table. One of the things that I think is really interesting about Sora 2 is um it represents I think the best kind of playground um for AI generation and and I say that because number one it's completely free and it's almost unlimited.

Um and then on top of that you're able to see what other people are making and interact with it and kind of play with it and remix it. So um so a couple things there. Uh, I have always been hesitant to play with AI video tools. Um, maybe because I'm cheap and I should stop being so cheap, but like every time you make a video in any one of these tools, you know, it's a certain number of dollars based on the tool, you know, and if I want to go just play and make hundreds of videos that, you know, that adds up pretty quick.

And if I want to ask my editors to go play and make hundreds of videos, that that adds up pretty quick. So Sora is really cool because I can literally just go make a bunch of prompts using our ad ad briefing tool, dump them into Sora, go do some other things, come back and see what happened. Um, and so it's a great learning tool. >> How long does it How long does it take for the video generation? >> Uh, so you can generate three concurrently and each one takes roughly I say a minute.

I haven't really >> Okay. All right. >> Yeah, it's pretty quick. It's pretty quick. So, you know, yesterday I did, you know, I spent maybe an hour playing with it and I probably made 30 or 40 of them. Um, and so there's that piece of it. The other piece we haven't implemented yet, but like I actually want you to get an account. I want other people to get an account because I've posted on my account all of the ad stuff that I've kind of been playing with.

And what you can do is now you can come in and you can kind of remix my stuff that I've made and you can go, "Oh, that's cool. I kind of like that. What if we tried it uh a different way?" So, >> so you can prompt edit the video. >> Yes. >> As opposed to just building from scratch. >> Yeah. So, think of it like Tik Tok. You're flipping through it. You can like, comment, and then, you know, prompt edit. Um, and so it's kind of it's kind of a fun way.

Uh, so not only that, but then you can prompt edit anybody else's stuff. So, there's like kind of a viral one going around now where a kid opens a Christmas present, you know, he's like, "Oh, wow. A COVID test." Or he's like, "Oh, wow." Oh, you know, uh, a bill from the IRS. And I was like, uh, so I turned it into a fruit cake for one of our clients. Um, and so I wouldn't have come up with that idea if it weren't for the fact that I was kind of flipping through the feed looking at some other stuff.

Um, so I I I view Sora more as like a brainstorming and uh like playground than Well, by the way, it's in my opinion the best video generation tool out there currently. Uh, easily. Um, so if you wanted to use it, and by the way, it's it's got APIs that plug into all kinds of things that people are already making sora videos through other tools as well, like arc ads and other stuff. But if you just want to start playing with some free video generation, like that's a great fast fun way to do it. >> My feel for so many of these tools that they're they're a lot of promise, but that that like there's been like not quite there yet for video generation.

Um, and I mean something really specific with that, which is that like it's not that they actually aren't there. is just that they require so much wrangling at the level of like incredible prompting and um and sort of reprompting and and so much work and then you run up against limits like you said and that makes the process actually uh pretty tedious and and so so I have resisted the like full embrace of a lot of this stuff yet because it has just felt like it's going to be there but like like it just feels like a potentially a big time suck in the short term for for relatively little output.

My my feeling watching you work with Sora too and what you've sent so far is that that just changed that that and maybe I was using the wrong models like I you know this is certainly not an area where I'm like going to be a thinking on the leading edge here but like I I think when I watch that it's the first time I've gone like these videos are really cool. They look and feel 100% real and great and uh and they're fun and interesting and I think that social element that you're describing is also really interesting because it just really does provide this kind of other element of what's possible.

So um so that that leads to all kinds of questions. Go. What were we going to say? >> Well, there's another piece like so I've sent you a lot of good videos, but you haven't seen the ones that I had to prompt and reprompt and reprompt and reprompt and like finally gave up because it just wouldn't. for example, it's really bad at like eating stuff and you know how to eat a fruit, it can't do it. Um, the other challenge is it works really well for generic stuff like eating a a passion fruit or something, but we have clients that have very, you know, specific products, you know, that can't, you know, you've seen some people on like X, you know, mock up like, hey, I've got this package of gummy bears that I've been able to get AI to do really well.

Well, we have some people that have products that just they cannot be there. there's just no way to to model them in AI. So, it really kind of cuts out a lot of, you know, you can't show the product. So, what can you do? Well, you could do some like fun street interview style stuff. You know, you could there are other ways to think about incorporating it. But, you know, one thing you've taught me to be really well good at is we never start making our ads from the execution details.

We always start from the message and then we see how the the B-roll or the assets can support the message. And for for now, you know, we we have never found ourselves going uh we can't really support this message with with our assets. Like we need new assets to support this message. And so that's why we've kind of not really felt the urge to go to AI yet. But I I want to start developing those muscles internally so that that we can start to do that more.

Taylor Holidayiday. So this is where my mind goes with this. Um, Taylor had tweeted a while ago now something that stuck out to me and I think stuck out to a lot of people in the world of AI, which is was something the effect of so much energy and time is being put in right now to to making AI look real, you know, like if AI video look real. And that that's the wrong use of AI. the right use of AI is to instead use AI to make things that are obviously fake but really fun and interesting, you know, or or whatever, however he phrased it, you know, but that was the idea was like don't try to make things that look real.

Instead, do things that are crazy. So, so sore too represents to me the possibility of maybe the easiest pathway to doing that. And I'm thinking about very specifically in this narrow lens of the hook. So, if you like go back to our to what you just referenced is like sort of our philosophy of advertising at AF Growth. It's like the message is the thing. The message is the core thing and that's what everything we're doing is in service to the message.

Now we can limit the scope of what we need Sora 2 to do for us which is like okay if we have a message we want and it's just a talking head explainer video or whatever. Um what we can do is we can say like okay if the message like for a client of ours is about reducing bloating as an experience and you know that's where the ad is going but you also know there's disproportionate value in creating a great hook in an ad that early engagement on an ad is a really strong signal to meta and to performance.

Then it's like okay how can we maybe make our prompting easier by we don't need 10 seconds of video we can get them to 3 to 5 seconds. we can use the hook to create things specifically that are creative and weird and funny and interesting that are just attention grabbing. That's the point of them. Then then we could uh then then then maybe we can sort of narrow again create a a more defined sandbox of saying like we're going to get good at this tool within this hyperspecific way and that might also help us develop the skills as a team to to prompt.

So we're just going to leave stuff on the table. Because one of the other things I feel in this whole conversation is this sort of constant pressure to embrace every new tool. And that sort of works against the actual operationalization of things. Instead, it's just to chase the next thing. It's like, well, okay, what if instead we built muscles in certain areas that way when we want to make at some point down the line a two-minute video that perfectly pro that perfectly works in all of our, you know, clients products, you know, when AI gets two years better or whatever, however long it takes AI to do that.

We've got the skill set built up by starting with three second hooks that are weird, you know, and that becomes the way we think about it in the writing process. Like I always even thought like can we build that? This is an idea I haven't shared with you yet, but but like can we get Prometheus, our co-writing tool, to actually like give us as one of the kinds of output that it gives us like the weird AI hook ad and it's like so it gives us our core storytelling frameworks which we use a lot, right?

And we talk about this a lot like there's these core kind of storytelling frameworks we use for our ads and we're building those and adding to the to the tool all the time so we have more and more of those to work from etc. That's just a slow and steady game. But what what if one of those core frameworks could be okay Prometheus help us get weird here basically and think about like what would be a crazy way visually to demonstrate the kind of thing you're doing.

Like one idea I had for example was like could we do an ad about bloating where as a woman is eating her stomach is just like expanding drastically like like absurdly you know and then the then the copy overlay is something like or look overlay is something like this is what a lot of people feel like when they're eating and nothing they can do seems to stop it and then and then you know you so it obviously be AI and obviously be fake but maybe it would be really attention grabbing and then you just go into the explainer ad format that we normally use but it becomes comes a way to grab attention and be interesting.

And I that to me is like the the the potential ne next thing to do that, but then we have to actually build we don't just want to like tell creative people go use Sora that like we do want to do them. We do we do want to do that at some point, but we actually want to also say like how can we build a repeatable process for working that in as like one of the options on the table for the kinds of ads that we make. You know, that's that's the way I'm thinking about it right now.

Literally every one of the full growth strategy clients that we have at AJF Growth uses Intelliggeems to do their CRO testing. And CRO is really the wrong word for Intelligjs. It is conversion rate optimization, but it's so much more. Intelligence is an incredible tool that I've I've come to see is just sort of a core part of every brand's tech stack. If you are building a DTOC business, uh where you can maximize not just the conversion rate of your traffic, but actually the profit of every click that comes through your business because Intelligence is built to help you test the biggest needle movers on your website that actually make a big impact on how your customers are interacting with your products, how much they're spending, and critically to measure that at the level of profit.

Because intelligence ties into your data on your site, your actual COGS data that you put into Shopify as well as shipping costs and all those things. And then every time you run a test on your site, you get an output based on how much actual profit you are getting per customer. So that means it holds variables like conversion rate, average order value, and unit costs and shipping cost and all those things holds all those constant as you run tests.

So you can see the most important output on your metrics. And critically, you can also do that while testing the most important things on your website. Certainly, you can run standard content tests and intelligence is very easy to set up for that kind of thing. If you want to run landing page A versus landing page B, that's a good use of Intelliggeems, go do that, right? But beyond that, you can do so many other things.

You can run price tests. I have one client who's constantly been testing his prices over a long period of time and finding where the trade-offs are between conversion rate, AOV, and profit using Intelligjam. free shipping thresholds, how much you're charging for shipping, sitewide offers, like should you just offer everybody 10% off if they sign up for your email list and and are new customers, or should you do a stack discount where it's 10, 20, 30% off depending on how much is on your cart, etc.

Those things have a really big impact on how much money you generate from the customers who come to your website and how they move through your website, through your store, through your products, through their experience with your brand. So, you should be testing them, not just guessing at them. And you should do it with Intell. You can also get 20% off your first three months by going to intellgeems.io and using the code Ferris 20.

F A R I S20. There's only one R in there and an A, not an E. Ferris 20. F A R20 to get 20% off your first three months. Go to intelliggeems.io to get started today. >> Yeah. Do you I think about it the exact same way, but then I I come to the roadblock of well I actually don't know that these clients would be interested in like we say we want to use these tools and do fun crazy things but like if you actually showed you know that client the AI video of a woman's you know stomach getting bloated like would would he would that client actually go for it? >> I think so.

I think I think I think I'm thinking through our client Russ right now. I think all of them would probably say yes and they might I mean they might say no but that's okay. That's that's the reality of design and edit work in general is that there's going to be brand standards that limit you from doing some things and I think brands should do that. You know, brands ought to have ought to have some things that they call there.

So yeah, I think a lot of them would. And I also think even aside from that, I think that the clients would just love to see us playing with stuff like that and getting creative and interesting. Even if they said, "Ah, that's like a bridge too far for us." We we think we we we applaud the effort. I think they would all experience that as good creative leading edge thinking, especially if we could replicate it, especially if it was like every week they were getting ads like that that were just different and interesting and weird and and things like that.

And I mean, honestly, the other thing here is like with manual bids, no, no kidding, like like you know, with Bid Cap and Target Ross, like the thing is if the ad doesn't work, it just won't spend. So, it's like there's there's relatively little risk. A lot of our clients have really embraced that idea that like you have a sort of strong incentive to throw things at the wall unless you're really really uncomfortable with the with the um with the brand kind of perception.

So, I don't know, maybe this maybe the stomach expanding would be uncomfortable in some way specifically, but you could come up with something else weird, you know? >> Sure. Well, with Sora, I think we're the closest we've ever been. And like I I think we're at a point where it makes sense to try to go implement what you just outlined, you know, because I think Sora gets us as gets us as close as we need to be. You know, there's going to be mistakes and things to work through and like, you know, we're going to have to have people that kind of start to become prompting experts, you know, on these models.

Um, but like you know, just an hour of playing with it yesterday, I realized, you know, I'm kind of an expert on prompting how to how to eat fruit, you know, it's like, okay, they don't it doesn't do this well. So, so what if I said it like this? And like the more you play with it, the more you kind of get the hang of how you need to prompt it and get it to do. So, like we could, you know, I feel I think what I'm saying is like, you know, we could probably go build build that, you know, muscle, start to build it, you know, essentially now.

And well, so this gets into some other things I want to talk about. But like this is where this is where like AJF growth was never going to work with me leading it because like when I tell you this is the vision that I have for how I think this could work. I'm thinking first of all, I'm thinking backwards to some ads that actually we looked at in the Philippines in January before you would come on and they we called them the meme hook ad and there were these ads that were like there were these videos that people had pulled that team members in the Philippines had pulled as like ad inspo that was like all of this like really funny stuff like like one of them is like a somebody is in a like what looks like a coffee shop picks up a big bag of trash to go take it out to the trash throws it over her shoulder and as she throws it over her shoulder, uh, she falls backwards onto the ground because the weight of the trash bag pulls her back.

And this is like a 100% real video, right? This is way before AI. Somebody used that and then they used that video as the in as the hook for like a piece of viral content or whatever. And they blended that video into uh into whatever story they were telling so that like as she fell down it, they sort of seamlessly edited into something else. And there's like a lot of really funny stuff like that. And one of the problems, right, with doing that for ads is that like that content is owned by somebody.

So you you you have to license it if you want to try and use it in an ad. So like using GIFs and memes a lot of times gets you into like legal questions and things like that. Well, with Sora, it's like, okay, now you can just just go rec just go recreate it, right? Um and so so anyway, so I love this idea of being able to do this. But the place that I get stopped is that and and again this is like this is this is why this partnership exists, right?

Is like I hear that and I'm like the idea of moving from that idea to creating a process for exploration and measurement and and understanding if this is even worth our time replication. And actually not only that, but now we already do have some established like creative strategy work. So we don't just have to like create the process for this. We actually have to create the process within an existing framework of process where where that is there.

So you say well we could go do that right now. And like my question for you and this is really me asking is can we how long does it take for us to to build that into our process? Like >> you know >> well as you're talking so hard for me. >> It's actually kind of crystallizing a little bit and I want to tell you what I think and I'm I'm interested in your feedback. So we build currently and I want to continue to hone our thinking and our process around it.

But we build everything around the angle which I want to actually start referring to internally as the message because I think that's a better word and it's the same word that Meta uses but let's just call it the message. We build everything around the message. We start with the message. That is the core. And um if we're do doing our job right, I was talking to our growth strategist the other day about this, one of our growth strategists, like we uh we are, you know, the the the list of messages we we use, we explore, we expand, we test per brand, like the you know, we should almost have a database of messages per brand.

Um and then, you know, then we we can let's say I've got a database of all the different messages we're we're testing or using concurrently. I can then go I can go task somebody as a full-time job or a part-time job to just go try to come up with funny, exciting, interesting hooks for those messages regardless of the contents of the ad. So, let's say the let's say the message is your kids are going to love this fruit more than candy, you know, because it's so sweet, you know, go make a bunch of hooks, you know.

So, I sent one the other day of like kids testing this fruit and talking about, wow, it's amazing. I don't I don't ever want to eat candy again. like you don't need to know the rest of the ad to go make that six-second hook. So, we could literally have somebody's job is just constantly cranking out hooks, you know, for specific messages and, you know, then going ah, you know, kind of hook menu. Ah, there's a good one.

I'm going to plug it in plug it into this ad. Um, I don't know how how does that strike you? >> I think you won't get it to a menu. Actually, what I think but I think you're I think you're thinking of the right idea. So one thing I believe about humans first of all we should say explore and expand is the way we think about creative in general which is like explore is our way of saying find new messages that work for new audiences.

So when you talk about the idea of reaching new audiences that you're not currently reaching a lot of times like all our things being equal and oversimplified version we call that explore. Expand is when you say like you do the thing that every advertiser has ever done which is you say like okay here's here's what's working for our for our customer and for our brand. go make more ads that communicate this same message.

And so that's the way we categorize our ads. Okay, that's those are the two main categories of ads that we make and very purposefully think about them in separate buckets and our whole process around them in separate buckets. So that's the way we think about it. So what I what I think is like even that framework like a thing I've thought for a long time is that that one of the mistakes people make with advertising and creative strategy is that they go and hire extremely creative people to go make ads.

And that sounds like a weird thing to say as a mistake, but what I really mean is like I I think that the thing that makes an ad work is the message. And that means that like the thing you really need is a clear thinker first and foremost. Like this is what this is like the core skill of a copywriter uh for ads is like somebody who can put together a logic a script that follows logically or a message that follows logically that makes sense logically to to do that.

Now, it doesn't have to actually be the form of like an essay or something like that, but it's just they have to have very clear ideas in their head of what the messages they're trying to communicate and then think about how different elements of communication which include visual elements. Um, but different elements of communication can effectively communicate that message and how you organize and and build hierarchy around the messaging.

That's that's super hard work. But if you uh if you start with a visually creative person, like a lot of times they're my experience with those kinds of people is that they're amazing and brilliant. and they do all kinds of really cool things, but they're not always typically I mean you don't think of the art people, you know, as like the like, you know, rigorous rigorous line by line thinkers. You think of them as people who go do crazy stuff, right?

And it's a it's a beautiful gift. So, it's not to come down, but anyway. So, but Sora 2 now opens up a new possibility where it's like or or just AI in general and this this way of thinking that you're talking about opens up a new possibility, which is like maybe this is the place for the creative person to go thrive. And that's actually really not me so much. I I don't know. But it's where it's like, hey, here's the message we want to communicate.

Get weird. Like you you want again like you now have Hollywood level CGI in your pocket essentially. You know, you can do anything you want. You're completely unlimited. If you want a monster crawling on the Empire State Building, you know, and fighting King Kong or Godzilla, I don't even remember. Like like you can do it, you know, like it's just like it's all available to you. And so maybe this is what we do is we set set up certain kinds of people in the organization to say hey we want you to be the people who are in charge of like here's the message here's the thing we're trying to get to come up with a really creative first three seconds that gets us into that message that serves the message still it has to serve the message that's the only constraint otherwise the world's your oyster you know that's the way that's that's kind of what I'm what I'm thinking about serious meta ads and other uh paid Paid marketers understand, paid traffic marketers understand the need for customized funnels as you reach different audiences.

We were talking about it in this episode, the idea that we are trying at different times to have different messages reaching different audiences with the same products or with the same brand or different products in your portfolio, which means the notion of having just one website that you send all of those customers to is sort of silly. What Vermont has come and done is made it very possible to create essentially bespoke customized experiences down to the level of different angles for your ads so that as you run your ads uh where you are where you're approaching different audiences with different messages, you can also create different landing page experiences way beyond the landing page itself.

So it starts certainly with landing pages and pre-sale pages and those kinds of things. But Vermont also helps you build quickly and easily and without being a burden without touching anything on the main website and screwing up the brand all those things. Uh you can go and build custom PDPs with custom thumbnail imagery which we know is really important in conversion rate optimization. Uh you you can build um custom upsells and custom cart experiences in all of those different custom funnels.

And so it allows you to build a really consistent and custom messaging for different things and critically to test it quickly and easily in Vermont where you get all the input and output that you need. I've seen great success with this with clients of mine who have used it and you should check it out for yourself as well. So go to vermontCommerce.com vermontCommerce.com AF to see if it's the right tool for you to build better, more custom funnels for the different audiences that you're reaching.

VermontCommerce.com AF. >> Yeah. And and just like to get down to brass tax, like the way I see it is we do a lot of like the format of our video ads, which is what we're talking about is like a long form explainer or like a UGC video, you know, and usually the opening scene is somebody talking to the camera, you know, kind of starting the the storytelling of the ad. Not always, but it almost doesn't matter what V, you know, let's let's say you go do the crazy three to five second hook. it almost in in some cases if the message is is coherent it almost doesn't matter what comes after that hook in in some ways as long as long as the message is is tied between the two.

Um so in in that way like you can basically kind of set your hook creative person off to just go do creative things without really regard to anything other than it's got to serve the message. So anyways >> Yeah. Yeah. No, I agree. And yeah and I so yeah finding the processoriented way to set set people up to do that I think is the question and it's like I you know I don't and I yeah coming up with yeah creative things I wonder if there's also a thing where like ad inspo becomes a bigger part of the equation because a lot of times my experience with when people post ad inspo stuff in different channels is like it's sort of unreachable. uh like it's like a lot of it is just like I yeah that sounds fine but I got to go get a client to pick up a camera and pay somebody and do all that stuff but or I got to go find these things but now it's like man what if you could prompt your way to like recreating exactly what you want you know um if it becomes maybe yeah more more part of it but I think we I think we're going to have to push our team to get weird and I think that's that's a big part of thing for me is like that's the that's the thing is like try to be weird I saw a video I saw sort of video the other day I think you might have even been the one to post it I don't I don't know but it's like somebody like looks like a high school football game or something like that.

Walks up to a probably 80-year-old woman in the stands. Did you see this? >> Oh, yeah. Yeah. Like >> with a t-shirt cannon, >> shoots her with a t-shirt right in the chest from point blank. >> That was so funny. >> That was that the red pill for me or the blue pill? That was the pill for me where I'm like, "Okay, we're there." Like, we are there. Like the only thing holding AI back is the audio has kind of a tell to it.

Like uh this the spoken audio. you can tell that it's AI, but other than that, like the video is just in in a lot of cases beyond recognizable. >> Well, if I was using that video in an ad, I would have cut off the last two seconds or 3 seconds and I would have cut off probably the first one second where he's walking up to her. So, it would have been right up to t-shirt cannon. either gotten the blast the old lady falls and then it would cut to like hey does this weird thing ever happen to you or whatever you know like I don't know exact what for whatever the product would be you know like um uh so but like cuz and that's part of the thing I like about this idea is that like those things that that are the tails on AI I think with a three to four to 5 second version of this stuff you could exactly you could edit them out and so it sort of creates it removes another barrier to using this stuff that's my feel right now is like so how do we do it okay Let's move.

Let's talk about process because this is like like when I look at this when I look at the when I I mean when I look at CTC who I think is just best in class at at a lot of things in the agency space. The thing I come back to all the time and you and I have talked about this is just that like at some point Taylor got serious about the idea that every thought he had about what works for brands needs to be drilled into at the level of clarity to his thinking.

So he forced himself to get really clear about the idea. Uh you know it's fine to say you should have a lean opex. Well, how lean, you know, and then how do you measure that and what counts as opex and how do you show that to somebody, you know, and okay, like those kinds of things, you know, like you don't you can't just say you run a lean opex there and how do you do it? How do you go get a lean whatever? So, there's just this endless thing.

Okay, great. Cohort forecast when, what time period, how do you deliver it? Okay, that means a certain daily spend gets gets forecasted out by campaign. That's years of work, right? Like, and so I'm not expecting us to get there overnight. But what is occurring to me as like I think about this idea of of Sora hooks and creative is like is like it just feels so difficult to me. But it also feels like the task of the agency life to say like you and I are sitting on this call and saying we believe this would sell products for our clients, right?

This this would help them grow businesses. Now we have to institute it in a way that's replicable. like it just feels so hard and now to you maybe it doesn't feel that hard but to me it feels like extremely hard. So I I mean where does your mind go with the ability for us to turn that into a process? >> It doesn't feel hard to me where where things start to uh I'll just say overwhelm me a little bit or not like this seems this is so simple to go figure this out but it's it's how does this tiny piece fit into the larger thing of all the different things we're doing.

So we've got a digital asset management system. So like these assets have to make it into there. We've got like a weekly cadence where we're identifying our priorities. We've got certain people doing different things. So I'm going, okay, who am I going to assign to this? How does it fit into our weekly cadence? How do we actually keep track of what they're producing and giving feedback on if it's good or not? So like, you know, the actual task of going and making hooks on store is no big deal, but how do we plug it into all of the pieces of what we're already doing?

But in short like I think it's just as simple as take take one take one start with one brand go you know it can be a gut feel it can be a go back through the the last 60 days of creative and go what are the top messages we're spending on over the last 60 to 90 days you know go identify the top three to 10 messages and then go put those in a database or just in a document and then go assign somebody and say your new job one day a week you're going to spend on or half a day a week you're going to spend on just making as many hooks you can um when it comes to these messages and here's kind of some general outline some general guidelines on how we think you should do it you know you can use Prometheus to come up with ideas you know this is kind of what we think good work is versus bad work and just send them off and go have them make you know 20 different hooks and then make it a part of your weekly review cadence where you go through those hooks and go that's a really good one you know Let's make sure that gets briefed next week.

Um, >> so you that's interesting. So you turn the hook in because the hook is downstream from the message, but then it's almost like you would go back upstream and like >> say, "Okay, now this is a really cool hook." >> Well, like >> don't hear me say that. Yeah, I think I think I did say that, but but that's not what I meant. Uh may maybe what I should have said is now we've got this whole library of hooks that we've got organized by message and next time an editor is working on an ad that has that message in it, he or she goes to that library and goes, "Ah, that's a great hook that seems to kind of go go in line with this this asset I'm making." Hm. >> Except I think then you're going to run into the thing and I think Meta is doing a lot of things to try to cut this off where like essentially you're going to have multiple hooks on one ad, you know, where it's like one body and all that.

That sounds like what you're describing and I think we're going to run into an issue there. >> Well, I don't know how to define >> I don't think so. Like let's say let's say we got the message just think of it like a folder in frame which is our digital assets management system but just think of Google Drive. You got a message, you got a message folder, and in that folder is like 20 hooks that somebody has been working on over the past few weeks.

And when I say hooks, they're like 5-second hook videos. Now, when so and so is editing this video that has that same message, he goes to that folder and he goes, "That one seems to work really well with what I'm editing. I'm going to use that over here." Now, if you want part of the process to go, okay, now he has to mark that one as used and it can't ever be used again. Fine. Uh if you want it to be able to be used again in the future, fine.

No big deal. But um you know he's basically kind of picking from a menu of of hooks to include in his in his asset. Does that make sense? Kind of what I'm saying. >> It does. What I'm actually wondering though is if we should do the downstream then back upstream approach. Like what I what I kind of wonder is like there's an interesting idea here where it's like like you it's almost like you start you give a creative person the message and then you say to them I want 10 hooks that could conceivably be related to this message, you know, and you you go make the hook in Sora with a prompt and then you you you also write a note next to the hook that says something like here's the way I imagine this and maybe this is what creative strategist is now I don't know this here's the way I imagine this connecting to the message uh so it's so I'll just use my example earlier it's a anti-bloating supplement and then you write like okay I had this idea the woman's stomach expands to an absurd length as as she's eating okay and here's how it connects the the the the bridge from the hook to the rest of the message to the rest of the ad is does it feel like this most times you eat.

For a lot of women it does or for a lot of women this is what it feels like most time you eat. Something like that. And now now you go back upstream which is like okay uh so now you have to go from there to connect out to what is still like a the same core message which is like this supplement helps you not bloat and and here's why right it's because of it's because of the probiotics and you know it's like and so it's like all the normal product details and and then you go do it again.

So, it's like, okay, the next one is another bloating one. Um, and it's like uh it's a view of the inside of somebody's stomach and you're you're it's like two humans in there looking around, you know? Oh, look at that. I can see it getting bigger. Something like that. And then it's the same thing, you know? So, um so yeah, that that's like the thought that I have is like is like could you do it that way or do we just need to tell people like go make an ad like don't do don't build a library.

Don't do don't do 20 hooks at a time. like just like go go make an ad. Yeah, >> I think what I think we're kind of saying the same thing, but what I'm saying is like a caveman version of what you're saying, which is which is literally like, okay, go make that bloat hook that you're talking about. You could literally take that hook and put it in front of any asset we've ever made about bloating and just have it be the first three seconds.

Like, it doesn't necessarily have to have a nice transition into the the rest of the ad. Now it can it can it might and it should but like you could literally go from that to talking head you know because most of our talking head stuff starts with something like is bloating a problem for you you know um do you experience bloat you know it's like doesn't necessarily have to have like a real cohesive tie-in now it can like in version two of this process but literally right now all I'm advocating is just go do something fun and hookie and then go straight into the ad. >> We should do this.

So here's what we should do. We should, this should be an actionable next step coming out of this podcast right now. Then let's report on it after we do it. Okay. I bet we can get this done within a month. >> The the concern I have about that idea is that it's going to is that Meta is going to see that ad and the one you built it off of as too similar. >> Yeah. And so >> to be clear, I'm not saying launch both of them.

I'm just saying >> I know, but if you have a winning ad from the past and you're just adding a new hook to it or something like that, >> I'm not saying we go back and do that. I'm saying for future ads that we make, we go use the hook and then it doesn't matter what comes next. >> Yeah. Yeah. Well, but so how maybe I understand. How is that different from the idea that you would attach it to the front of an existing ad? >> Well, I'm not saying go reuse old ads. >> Oh, you're saying just write an expand ad? >> Yes. >> Just write a new ad. >> As long as it's about bloating, >> doesn't matter. >> That's interesting.

Let's try it. Okay, that's still We can try that. We can try that. So, use the standard long form explainer. >> Yeah. And we should and we should have like the standard people do this. Yes, there's gonna be there's there's going to be some operational challenges here which is like okay now you have a new hook and that means you need a new script so you need somebody to go read the script or whatever. But you're right, you don't have to have that person go make anything related to that content.

The hook kind of stands alone as its own unit. >> Exactly. >> And I like you don't even have to have the hook generator talking to the the video editor. like Dino just goes and makes his ad and he leaves an extra 5 seconds at the beginning and then like you know eventually he's either at the very beginning or maybe at the end of his making the ad he goes oh I like that's a fun hook let's throw that in there let's see how it works um that make sense what I'm saying >> yeah I would have somebody brief it to Dino ahead of time or tell Dino this he has to do this for a couple of these ads right so you go film >> yeah ahead of time yeah the brief includes like this is part of the process that you've got to go pick a hook >> that we've pre-generated for you.

Um, but >> we Yeah, go ahead. >> The reason I'm scoping it like that is because this is by far the easiest way, I think, to kind of start to test and build that that that muscle a little bit. >> Okay. Yeah. And then, yeah, we should definitely go do this. We should go take a couple of winning ads and what we should do is we should give like three or four people in the organization. Should be me and you and it should be some of some of like Yeah, Dino is actually a great example.

This Dino is super creative. >> We could incentivize the best hooks sometime. Yeah, I like this. >> Best uh best hook rate. What what's the metric in Facebook that you know it's like >> No, it's just spend just spend spend spend. It's all >> Well, yeah. Yeah. But specifically when it comes to hook, but yeah, I I get your point. >> You could do hook rate, but hook the hook rate >> ratio metrics are are annoying. Anyway, um >> there's too many contextual factors, but the >> I think it should just be spent, but that that is that does create its own challenge.

But anyway, I actually think in that process somewhere, one thing we should do is this hook would be so different than the rest of the ad that we should actually take a couple of old winners and just throw another hook on the front as well as part of the process and see if Meta reads it as a new ad or not because like I know Meta is trying to read it as a different as as the same ad. But somewhere in that process, just go take our best performing bloat ad and just like go >> throw a hook on it. >> Put a throw throw a weirdo hook on the front. that flies against everything you've always said about iterations and uh expanding. >> Well, yeah, because meta is actively trying to keep you from doing this basically.

So, >> but you're curious just to test it, see what happens. >> It should be fast. Like you should be able to go and and those first three to five seconds will be so different that like it may it may create enough distinction that that it works. What I what I really resist is like change the headline, you know? That's like that's not different enough. But this is like if you start with fully different 3 to 5 seconds of the video, you know.

So, okay. All right. You you you you know what you're going to do here. We'll come back and report on it. >> I think so. >> We should see if we >> should see if we can get permission from a client to share it, too. >> Oh, that'd be fun. Yeah. The only thing >> we should we should try it with our newest client, our newest client, who you know who I'm talking about right now. >> I bet I bet he would be down to share. >> Oh, yeah.

Yeah. Yeah. That's definitely the best place to start. >> Yeah. And and the other the bloke client is also probably open to sharing, but I'm not sure about that. >> So, anyway. >> Yeah. Love it. Okay. >> Okay. Great. And then you have to tell me what you need me to do in this process. That's fun. Um, >> sign up for go sign up for Sora right now or not. >> Yeah. And subscribe to this podcast. Don't miss the episode where we come back.

We'll show the ad. Is if we can get any permission at all to show the ad, we we will show it. So, um, and especially if it works. So, yeah. Um, uh, I like that. I think that's that's the use of sore that I'm interested in. That That's great. Okay. >> That's why we're recording this podcast. Look at us, dude. Look at us. >> I know. Getting work done. >> Yeah. >> Yeah. It's good. Um, uh, repeatable process. Uh, let me say one more thing about repeatable process.

I've had a couple people email me recently about wanting to work with us, and I've had to sell tell them the same thing, which is that like we're probably looking at spring next year as the time when we actually start taking on new clients realistically. So, a couple people have been like, "Yep, just let me know when the time comes." Like, you know, obviously they may have a different agency by then and and I hope hope they find somebody great to work with and their business grows, and that's totally fine.

But um but as we're kind of looking at those things, I'm you know, you and I are like saying we're going to slow down to speed up here. The idea is like build really really good foundations so that we can um create repeatable results over and over. So as I've been talking to those people, one of the things I keep using this phrase that an agency is operationalized knowledge. That's what an agency is fundamentally at its core is what an agency is.

And so there's two elements of that. One is you need good knowledge. So for example, it's possible that these hooks don't perform at all. All sort hooks are bad and no matter how good you are at the process, uh people just read them as AI immediately or meta reads them as something and they just never work. And so you can build an incredible process for a really bad idea and it's it's not a very good use of your time.

Um on the other hand, there's this thing that happens a lot in these conversations and that has been defined a lot of the the shortcomings that I've had uh which is like and I've had a million conversations like this really cool idea. We really believe it'll work. We're going to try and do it. we're really going to try and then it just kind of withers and dies because there's no way to follow up on it or go create it or whatever.

And that's the operationalization problem. >> And the more that you and I work on this together, the more I see the like grind challenge of both of those elements, which is like first of all, discerning truth is difficult work. Um, so and it changes in the world of advertising, right? Like the UGC ad worked great for forever and then like ah does it even really work anymore? I don't know. uh like you know you have to have an assessment of like that kind of thing.

Um you know partnership ads is all the rage right now. what happens when that changes, you know, and and everybody's just like, I don't know, everybody's bought and paid for and you know, I don't know, whatever. I I'm just making that up. Like, I think they probably work great, but like uh so there's a hard work of doing that and then um and the macroeconomic situations changing all those things. But then there's also this really really hard work of turning those ideas even when you get to truth into prioritized operationalized uh like work that actually gets done.

Um, and so what I what I keep telling people who who reach out is like if you're going to find another agency, this is the thing to prioritize is like try to find people who you believe as best as you can tell have the best knowledge and they should at least have some level of conviction about that knowledge and they should be able to to state it themselves, but then they they should be able to show you that they can actually implement it.

Um, and so anyway, so this is just like a state of the union question, but like how are you feeling about our ability to do that these days? And sort of where are we good at it? Where are we at it? Where are we getting better at it? >> Yeah. Well, I love that uh metaphor and I actually use it a lot in my thinking and in talking to people and I think it it's also convenient that it it it describes you and I well, you know, I'm the operational, you're the knowledge.

Um and there's certainly overlap between those, but that's our core core focuses. Um and and we're both very happy to to have those be our our things. Um I would say where I kind of um I I think we're doing a really great job. Like you know if you think I I kind of started uh in Julyish and then mostly you know kind of August onwards. So like we've only been operationalizing things for you know a few months and like we've come a long way.

Um the t the tension that I feel you kind of alluded to which is there's this constant tension between okay I want to operationalize well actually I'll start here I want to move we want to move fast we want to break things we want to do great work we want to like go do exciting things we want to go like you know there's like all these shiny objects it's like hey we want to go do all these things so like I've got this and we've got clients who are going hey what are you doing with AI you know have you tried this have you looked into this so there's like this constant you know and then I posted on X the other day about like you know the Slack culture which is always on always everything's urgent you know so there's this constant tension between hey we got to go do good things fast and then there's there's this other side of well well in order to do a lot of good things you actually have to go slow um and then so like where I where I wind up is you know what's that what's that saying like go slow is slow is steady and steady is fast or something >> that's right >> do you know is that is that >> slow is that yeah slow steady steady fast.

Yeah, >> studies fast. So, like that's kind of where I land, which is like I want my team um to um not feel like everything is urgent. I want them to be able to understand what's important and then I want them to be able to know what they need to do to go do those important things. And we're getting better and better at that. But that's that's like a massive hill to climb, especially when one we're we're literally figuring it out every single day.

We're like, uh, okay, we've got like this hook conversation we've had, like I got to go, okay, it's a good idea. How how do we go operation? So, I'm always throwing like new operational things into the mix constantly. Um, and then two, once I know that once I throw it in there, it's going to change, you know, and so it's like the engineer in me wants to create, you know, a nice box, you know, and like hand it over and be like, it's done.

Um, but like it's never really done. It's always changing and it's always kind of getting bigger and like the the I kind of alluded to this in an earlier conversation, the the tendrils between things are always kind of multiplying. And so since I'm the one kind of holding it all together right now, that feels a little at times, you know, it's it feels a little overwhelming, but um you know, when I really step back and and look at the progress we've made and and where we are and the work we're producing, um I think we're doing a killer job and and I'm I'm excited about the prog where we are. >> I think that's right.

It was helpful for me to think back to like when you were at supply and getting agency services from people. >> Yeah. I think I think I think you would be really pleased to get our service level right now, right? Like >> Yes. >> I think I think right now you you like I'm putting words in your mouth there, but I just think >> there'd be stuff you'd see as shortcomings, but stuff that is also like this is really good. Um I feel that way certainly.

Um, >> I think that the clients also, um, >> I think they value the I think they've valued my input or they they have valued the kind of calm and uh, order that I've brought to a lot of the chaos because clients and I say this about myself like I'm it's funny when I work with clients, you and I have this conversation a lot. I'm like, "Yeah, that was me." like I did those things, you know, I was the one like, you know, constantly chasing shiny things and like, you know, there's never a situation where we're working with a client and we're like like what are you like I'm never like what are you talking about?

Like where did you get like I'm always like yeah I know exactly. Um, but my point is like I was a chaos machine as a client probably with the agencies I worked with and I always appreciated, you know, when agencies came in and brought kind of um order to to the chaos and maybe even forced me to to, you know, fit into kind of their systems. Um, even if I didn't always love them, you know, I at least I at least respected them and like understood that the output was good.

So, that's kind of my goal, being on the other side of the table with with our clients. >> Um, I think that's right. And yeah, and I I think you're right that clients have appreciated that. I also think that the um uh that that one of the challenges I feel in this scenario of sort of building process steadily and slowly and then resisting the chaos uh element of work is that this my theory right now is that the way to resist the chaos factory is is not primarily by setting bound boundaries around tools for example or like hey don't talk to us here you know we're off work at 5 don't text me whatever but it's actually to build such a sturdy uh process and sort of machine that it functions so effectively that clients actually don't want us to spend our time doing anything else and it become it's almost becomes an opportunity selection thing where where like we say you to the client like while it's possible we could go do this thing that you are asking us to do over here, we actually think if we do that that will take our time away from this and we're pretty confident that there'll be an ROI on this over here because we've done it a million times and we're also really good at it and we have repeatable process and all all these reasons.

We both we think it's the best use of our time and we've also invested the most in making it good. So, you know, those things together. So, so it's like it's like then that then it's like you put that in such a way where the client then believes we actually just want them spending time on the thing that they think it's best to spend time on because when they do that it creates a thing. So it's like so so then that will have a natural resistance to tool proliferation and to all these things is because we are then delivering our service and delivering our process so consistently and also like I think clients can feel I mean any partner in things can feel when there is like replicable thoughtfulness in what somebody you're working with is doing.

You can feel the professionalism in it. And so it's like it's like okay great you you that that will sort of set the client at ease that they don't need to tell us what to do you know and then and then just resist that. So what I have done a lot in the past is say like okay this is the thing that we are not going to do and this is the thing we are going to do and you kind of set up these rules and like I said boundaries for different things and I I just think that that is is actually sort of almost like almost like the the rules and the boundaries are the um are like the output uh actually of something where you've already inputed a bunch of other things like they're they're they're byproducts.

They're byproducts. That's the word I was looking for. they're byproducts of of a bigger thing. And that the more we build our thing to be repeatable and good and the more we have clients who who came in via the repeatable process as opposed to like clients who've been working with me for a while and experienced my sort of like being all over the place at times and some of that stuff, you know. Uh I think I think clients who are more used to that I think I think will will feel that. >> Yeah, that's just a long way of saying something else you always say which is basically performance solves all problems.

And and I would I would add performance and trust solves all problems. You know, it's like so I think of our newest client like you know, he's we don't have performance yet because he's brand new. Um but we have trust. Like we're walking into that relationship with trust. And so every time I suggest something or I'm like, "Hey, you know, this is how we're going to do such and such." You know, they're like, "Great, no problem." Like whatever you want to do, just tell us, you know, because there's such a level of trust there.

And there's also, I imagine, a sense of like I like I don't care how you do it. This is what I'm paying you to do. Like go figure it out, you know? It's like just tell me tell me what you need from me and you know I'm going to stay out of it, you know. And um I don't know if that's how that client particularly views it, but like they're certainly it's and it's precisely what you said. It's not because they're hands off.

It's not because they're they're being lazy about their business. It's the opposite. They they're just saying like there's limited resources. I have a very high amount of trust in you. I think you guys doing your thing will be the best thing for us, you know? Yeah. And that's that's a good point. Like and that that will sort of create >> the tears and now we have to go do it. One of the things I think about this whole thing is like like I just look at where we've come and I go like man I think in six months like we're going to be really really really good and then in six more months we're going to be even better.

It's like I feel like one of the things I feel really good about right now is that just like clients who stay with us over the long term. It's actually a hard thing to assess with with um with a with almost any vendor partner of any kind is like how much does their past reflect their future? Because I I think our past is going to be real different than our future in a lot of ways and and in that our future will be really positive.

It's not it's not just because I'm optimistic. It's because I'm looking at the work we're doing and >> going like, oh yeah, this is like we're just going to keep getting better at this stuff. >> Yeah. Yeah. >> Yeah. >> All right. Uh that's great. I would love people's comments and questions on this episode. Uh, and uh, and so just leave that in the comments. I I interact with and try to respond to all those. You should also email me at podcastfgrowth.com.

And like I said, do subscribe uh, if you want to see what happens with the stuff that we just talked about here. By the way, Patrick, um, did my audio change quality mid call? >> I don't think so. I didn't notice it, but it may have. >> So then it was bad the whole time. I just realized, so if you're watching or listening to this, I'm I think my mic had unplugged accidentally. So hopefully hopefully my audio is okay.

But if you notice it was a little worse than usual. It appears to have been a technical issue and mistake. >> Yeah, it seems fine to me. >> Yeah, I think it's fine. I think it's fine. But just uh just didn't know. So, all right. Cool. That's it. Um, anything else? >> That's it, man. Look forward to next one. >> All right. Thanks. Yep. It'll be fun. >> See you. >> Thanks so much for watching or listening to this episode with Patrick.

Uh, I really enjoy that conversation. I'm really curious to see what comes out of it. It's really fun to be working with somebody where we start talking about an idea like this and like it happens and I don't have to go figure out a process for it because Patrick is much better at it than I am and my team likes that way more too because then it's not me fumbling around telling them random things to do with no expectations and timelines and all that stuff.

So, uh so yes, I hope you enjoyed that episode. I think it's super fun. Uh please please please give me feedback on that. Like I said, podcastfrowth.com or just leave a comment. Uh I'd love to hear from you. Uh you can see everything we're doing and join that weight list if you'd like to at afgrowth.com where just ask for my input on what's going on in your business. You can do that at podcast at afgrowth.com as well.

Um and do uh subscribe as I said not only for those future episodes with Patrick but I do have another episode of Taylor Holiday coming up very soon. So you're not going to want to miss that. Got Jordan Menard coming up soon and talk about some things we disagree about. We were supposed to have that a little while ago had to get pushed back for a couple reasons but that's still coming. Um so a whole bunch of really good stuff.

I've got oh my goodness I've got um a business coming on an operator coming on who's built like a 200 something million dollar a year business I like a huge monster ridiculous I don't even know what the number is but gigantic business uh that's going to be great so come check it out uh by subscribing make sure you don't have to don't miss it you know what to do I'll see you next [Music]