How AI Is Rebuilding Every DTC Creative Team

But the act of concepting is basically to try to find the like higher level themes that are worth pursuing, that are worth creating five or six or 10 briefs around. And so it really is around like what themes are working, what themes are not working, and how do I explore ideas at a concept level so that I can basically go and create like a few versions of that concept. And so that was like by far, by far the most common use case. But there's one paradigm shift that they make is that if you can get an agent to go and do five minutes of work for you, great, then try to do 10, then try to do 15.

Keep trying to push that until you have an agent doing work for you for a longer period of time. Hooks are so much deeper than this like tacky pattern interrupt that you might do in the first three seconds. It's almost like the hook is like the DNA for the whole thing. And if the team gets good at hooks, they've got it all figured out.

Like everything basically becomes downstream of being really good at hooks. Maybe it's not like this month, next month, maybe not this year, maybe not next year. But I think we're getting to the point where that's the only job left is being an allocator of AI tokens. And if you're not doing that job, I don't think there is a job.

And now let's take a listen to the Scalability School podcast. Welcome to another episode of the Scalability School podcast, everyone's favorite podcast. Thanks to all the tens of listeners we have. Super stoked to have you.

We're very popular. I've heard we're the number one podcast in marketing in Poland, Google told me recently. So thank you to all the polls out there that are listening to this bad boy and implementing it as quickly as possible. Today, legend of the industry.

Legend to me, who's also a great friend. And I tried to email you recently and it came up with your shoelace email. And I think if you have Reza's shoelace email, that's like a sign that you're like a true friend. You've known him for that long.

And I was thinking and I look back and the first time I emailed you, do you want to take a date on the year, a guess on the year, Reza? Must have been like 2017 or something like that. 2016, 2017. 2017.

First time you and I traded emails. 2011. No way. Wow.

That's crazy. Yeah. 2011. And which is insane and also shows that we're old.

So there you go. So we're glad to have Reza on today. We're talking about AI and creative. There's tons to get into here.

Super stoked to talk about it. Everybody's always excited on the podcast, of course. And Brad has a story. Brad actually has something he'd like to admit to you, Reza.

This is the confessions part of the podcast. But by the way, this is not a sponsored episode by emotion or anything. We just really like love Reza and you guys are doing a ton of crazy stuff. And I think I've been listening and hearing you talk about this a lot in terms of AI and creative teams, what it means, you know, what AI, where we're headed and how people can really like make sure they're thinking about this the right way in their organization.

And so that's really what we're going to cover today. I know it sounds very general, but I promise tactic stuff is following from this. But first, Brad's, this is Brad's confession. That was beautiful.

You should sing. You need to sing more often. Well, it's funny that you started with your story because when I do, when I type in Reza's email, I also get the shoelace. And I originally started paying for motion in 2021.

And I was the idiot that after three months was like, I just don't know how we're going to use this. Like I can't, I just like ads manager. We stopped using it. And yeah, like long story short, it's we use creative analytics every single day, all day and everything that we do.

It's just like embarrassing that it took me like, well, that was that was 2021. We first started using motion. After a couple months, we went off. And in the last two years, I literally like can't.

It's like we just were perpetually involved in these dashboards. And so I'm embarrassed, which leads me to an like, I guess, like a starting point of the question. Am I alone? Or is there something that changed?

Because like, I think we're going to we're going to go through this journey of we'll eventually get to AI in this conversation. But before AI, there was just like getting people to open up creative analytics. And so was there a was there a change that like started to get people to care about creative analytics that you you saw? I feel like we've seen multiple like like something around Andromeda last year.

I know this wasn't necessarily a thing that tipped you over the edge, but like for whatever reason, towards the end of last year, everyone just started talking about Andromeda. I think a lot of it was a little bit exaggerated, like doom and gloom stuff to like YouTube videos and whatnot. But like, I think that that had a really big impact on people taking the concepts that I think we've been talking about for a very long time seriously, which are basically like the the classic motion POV is kind of crazy that like all this stuff has changed and so much has happened in the last few years.

Like the POV is still exactly as it was when we were first chatting about like why we're building motion, which was that there was a gap between the people that are like thinking about revenue and data and numbers and the people that are like producing the creative. So we noticed that those two disciplines were so far apart. And I think towards when we started working on motion 2020 2021, we saw a really big shift from basically like audience targeting being a really big part of media buying to like all of a sudden kind of felt like a bit overnight.

The revenue lever started to just basically be creative and the algorithm was going to be creative and the algorithm was going to handle targeting and figure out who to put it in front of. And so our view at the time, and I think it just becomes more and more true every year is that those two disciplines really do need to merge into one, like the people thinking about driving revenue and analyzing data, they really got to get their creative caps on. And then the people that are producing creative, if they want to level up, if they don't just want to be like order takers that are just kind of producing stuff, then eventually, even like pre AI, we always had the view that if you're a creative person that just produces what other people tell you to produce, you're not creating a ton of value.

Like as soon as you can articulate to someone, hey, go and make me this, executing that work is actually not that useful or valuable really. And so as a creative person, it's like, how do you rise up to a higher level of abstraction, create more value? And the best way to do that is to like, get your data caps on and like think about revenue and the impact of your work, which was like, a really empowering story to creative people. And I think we saw people come from both sides of it.

But in particular, creative people are like, you know, I actually think they, they will win in in whatever is about to happen, because to have like, taste and judgment and like kind of an eye for good quality, creative is a lot more difficult to teach than it is to be like, hey, by the way, look at what happens with revenue and data and so on, like that is easier to wrap your head around. And so our view is like, could we build products to really help bridge the gap between those two worlds?

And it just feels like every year since we started. The message of that idea reaches like a wider and wider and wider appeal, the the stat that recently just like starting to blow my mind, you guys know, we've been creating like, content and do a ton around like virtual events since the early days of motion. I think, Andrew, you're on the first virtual event we did. I think it was you.

Nick Sharma and one other person, I don't remember who it was, but like the first like marketing idea we had for motion was like, hey, there's a lot going on. Like we got to do like long form stuff. Like, let's get a bunch of smart people together for an hour and do like a live event for an hour. People are still talking about it, you know.

So I wasn't invited. He said smart people. Yeah, that's right. That's why you weren't there.

Smart people. Yeah. We did a creative trends event at the end of last year and it got 26,000 registrants. Just insane.

Like I think that's kind of been our marker for like how wide appeal these ideas are is like how many people show up to our events. So something's going on. I think I think the idea that this collapsing of roles I think is not just true in our industry, but like it seems like in many, many industries where we're entering an era where if people, the individuals who can juggle what seem to be like opposing skill sets and bring them into one are just so valuable. So extremely valuable.

So extremely valuable. And now like with the help of AI, like you can really extend the things that you're maybe not naturally as good at to basically be these more like polymath multi skill set individuals. And I think we're going to see a lot more of that. And just the fun, just the fun time to be doing anything.

Honestly, I think it's just such an exciting time to be living through. I mean, so, okay, so you have so right now let's think about like, let's say I have a brand, let's say it's an eight figure brand and we sell water bottles or whatever. We go through and have hired maybe a couple creative strategists. We have people that are creative strategists.

We have a designer and we have a couple of designers and we have a video editor. This seems to be like a typical setup now. Okay. I'm like maybe creative director if you're lucky.

Right. Or you're thinking about this, you know, in a bigger picture. What do you think is advice that you would give to organizations that have this model? And like, how is this going to evolve even in the next six months?

Because so much, I mean, already it's gone from like briefing to, you know, like people were putting together briefs, which obviously is still happening to a degree. And then it's like you go out and create it. And I think there's still some of that that you're going out and, you know, you're creating the UGC or you're getting UGC back and all this. But it seems like a lot of what these people they're doing now is they're like project managing to a degree with a creative idea and looking at performance as well.

So they have to have paid social performance. Like I'm doing a search, I'm doing three searches now for creative strategists. And I sent candidates to each of these people and every one of them wrote me back and was like, what's their paid social experience? And I was like, yes, let me tell you about that.

Like, I'm sorry, I didn't mention that. And they all want that. They want to have that because that's where they're headed. So how does this team evolve?

Like, how do people build the right way in your opinion? Yeah, it's an interesting question. It's also, I think, a question of timelines. I think there's so much that can be said about where we're headed with AI.

And it's hard to say, like, how much of it is going to be applicable next week, next month, next quarter, this year. Like, I think that's probably the piece to, like, whenever you hear any AI advice, it's trying to figure out, like, you know, is this something that, like, is going to be years in the making? Or, like, what do I actually do tomorrow or next week? I think the more useful, tactical, tomorrow, next week, next month type of ideas is probably less around reimagining the way org charts work.

Right? I think that will happen. I think we're going to have to reimagine org charts, but I don't think that is, like, the most important thing to do right now as, like, a business leader is to start reimagining org charts. The way that I think about it, and I've been trying to, like, start to instill this in our team, is that I think everyone's job at this point, I do think this is, like, this should be a next week thing.

It's going to sound like a later this year thing, but I do believe this should be, like, a right now problem. Is that every individual, no matter what job role they have, I think the right way to think about our jobs right now is as allocators of AI tokens. And what I mean, the reason that's an interesting framing is that step one, there is a question of, like, what is my AI token consumption right now? Like, I'm not being paid by the AI labs to say this.

And by the way, you can use, like, a lot of tokens on some of these, like, subscription plans. It doesn't necessarily translate to burning lots of money. But the first question is, like, am I doing stuff that translate into, like, AI token consumption? And that's the first place to start.

If me or other people on the team are still not, like, deploying AI tokens for some body of work, that's wrong. Like, I think we should be doing something that involves, like, using up AI tokens. Then the next question is, like, is my return on AI token positive? Like, am I doing useful work?

Am I doing valuable work? Am I, like, am I creating more value with these AI tokens that I'm spending versus whatever they're costing me? Which I think it's very hard to do poorly on that trade. Like, I think if you're deploying AI tokens to do any kind of work that you find is interesting, chances are it's going to be really valuable.

But I think that is a good way to start thinking about things in this exact moment. The problem, the reason why people don't do it as much as they should be doing right now is that in order to get value out of AI, it takes some investment. It takes some time investment. You've got to learn a bit of vibe coding.

You probably need to, like, do things you've not tried before. Like, the barrier to entry isn't, like, that's probably the hardest thing. My sense is that we have extremely capable AI models that have not been diffused into the economy. But it is, but they are real.

And, like, you see it in the coding industry and the way that it's, like, completely transformed that industry right now. Like, it's not a theory for that industry. Like, coding has forever changed. It is not the same job as it was even, like, 12 weeks ago.

And I think that is going to be true for every other industry as well. So anyways, long answer to the question of, like, I don't think people should be, like, shuffling around their org charts right now. But every individual in whatever role they are in should ask the question of, like, if I was forced to spend $1,000 or $2,000 in some kind of AI credit consumption over the next few weeks, what might I do with that? Even if you don't want to give your employees free reign to go and burn a few thousand dollars, which I think is money well spent.

But even if not, like, at least get them to, like, run that as a hypothetical. Like, if you had to spend $2,000 in AI credits, like, what would you do? And I think the answer to that question is basically the answer to what people should be doing right now. It's like, force yourself to think about enlarge blocks of AI tokens and then think about what's the most meaningful work you can do with that.

We can talk about certain examples. I can give you one that I did the other day that was quite expensive in terms of, like, token processing, but I found it very useful. Yeah, I would love to hear the example. Okay.

One quick sidebar. Sure. I've been trying this entire, like, last couple of weeks to just, like, force everybody to, like, go into Cloud Coork. And just, like, whenever you're about to do something, go there first.

And, like, I know that sounds like you're not outsourcing all of your thinking and decision making to this, but it's like you're using it as a sparring partner to, like, help you get closer to the right decision. And I just, like, I just keep getting random HTML visualizations from the team. It's like, okay, like, this makes me, this helps me understand this better. And it helps me, like, think of the process to even get to this in the first place.

So I've been seeing that show up a little bit more frequently with our team, just trying to, like, push that. And it's been fun to see. But tell me how you burned a bunch of money. Okay.

So our AI agent, which, by the way, I think this is the year of, last year was the year of overhyped products because AI wasn't really there. And this year is the year of, like, underhyped products that really work. We have launched an AI agent called Runeth. We've not really talked about it.

If you log into Motion, you will see it. Our goal is to actually not make a splash about it, but we just kind of release the agent into our little universe. And we kind of want to take the opposite approach of really big, splashy launches. But the thing is completely taken off.

When we see our usage metrics on Runeth, it's just like a vertical line in terms of the amount of usage that people are doing on the agent. And I can't remember the time period I looked at it. It may have been in the last, like, few weeks. People have exchanged a total of 40,000 messages with Runeth.

Just to give you a scale of, like, how far this thing has gotten. And I was like, okay, there's no way I'm going to read 40,000 messages. But, like, what if I could understand what the hell is going on here? Like, what are people asking?

What do they want? And, like, what are some themes and trends? And I'm like, I could probably get AI to go and read every single one of those conversations and come back to me. But obviously, you don't go and just, like, dump 40,000 messages into, like, ChatGPT or whatever.

Like, blow up the context from this. So, like, Vibe Coded some stuff with Cloud Code. We came up with a plan together of how it was going to go and, like, do it one by one. By the way, I did this, like, several weeks ago.

And I just had one agent do it. So, it took, like, a few hours. And I felt like an idiot now for doing that last time. Because now what you could do is, like, create a swarm of agents.

Right? So, I just told Claude, can you spin up, like, 40 agents? And then you just divide the work. And then they just, like, attacked it.

And they were done in, like, 20 minutes. That was, I can't remember what it cost. But it wasn't cheap because I was using an expensive model. Like, I really did want to understand what's going on there.

But, like, that is an example of the scale of work that is possible to do right now. But it's a different paradigm than, I think, ChatGPT has, like, screwed things up for us a little bit. Where we are used to sitting there and talking with AI back and forth as, like, a chat partner. Which I actually think, like, that is a great place to start.

It creates a ton of value. But Sequoia, if we can link to the, I'll send you guys the link and we could share it in the show notes. But Sequoia released this article where they basically claimed that we have achieved AGI for, like, practically speaking, we have achieved it. And the framing that they were describing was around this idea of long horizon agents.

So, like, how do you set up a work environment for an agent to go and do work for, like, an hour or two or three? And I think that is the paradigm shift that when we start thinking about the world that way, it starts to be like, okay, this is really powerful. Like, the kind of work outputs that I can get done from AI starts to be really interesting. As opposed to, like, going back and forth and talking about, you know, what's the history of this or that, you know?

All right, friends. This episode is brought to you by Northbeam, the marketing attribution solution we love over here at the pod. And, you know, Zach, good news, Northbeam is launching Northbeam Incrementality. Have you guys used this?

We have. We've gotten a little bit of access to it. And so far, so good. Max is stoked.

He's my CMO. He's digging it. Yeah, I mean, out of the gate, we're super happy. Yeah, I mean, Northbeam Incrementality gives you easy, automated, self-service incrementality tests right there in the platform.

And it gives you sort of, like, actionable data and allows you to operationalize a lot of these. I think that's a big difference between Northbeam's incrementality tool and what we have other places in the market. It can, you know, run, basically, it runs end-to-end with lift testing with your MTA and your MMM. So you can scale what works and cut what doesn't, which, you know, I know for you guys, your companies is huge, right?

Yeah, I mean, the fact that we use Northbeam as our North Star for all of the dollars that we spend, to have this functionality built in right into the MTA is just, like, not only a time saver, but it just makes the workflow much easier. So, big fans. Yeah, and it's something that you trust, which I think is huge. So they're giving right now really good deals, pretty cool 50% off offer to listeners of the Scalability School podcast.

You just go to northbeam.com.i forward slash incrementality and you get to be the first on the list. So tell them what we sent you, 50% off all tests for a year for those who join now.

So check it out. Lots of awesome features that you have and look forward to hearing how it goes for you. I mean, give us the tea, Rezo. What were some of the things that, you know, we're being, what were, like, thematic things that people were talking about it with?

Yeah, what was cool is, like, it gave me some insights about things that, what people were trying to do that, that it couldn't work. And so people are trying to do a lot of image generation, video generation that we don't support yet, but we will soon. But probably the biggest one is around concepting. So I think there, there is this, when we think about the creative strategist job, we've had this view for a long time that downstream of concepting is basically producing briefs that other people are going to execute.

But the act of concepting is basically to try to find the, like, higher level themes that are worth pursuing, that are worth creating five or six or 10 briefs around. And so it really is around, like, what themes are working, what themes are not working, and how do I explore ideas at a concept level so that I can basically go and create, like, a few versions of that concept. And so that was, like, by far, by far the most, the most common use case. But it's super interesting, because we're starting to see the agent, like, part of our view on this was that on a long enough time horizon, the agent is going to disrupt our core business, because you might spend a lot of time inside motion, poking around, looking at reports, trying to gather insights.

But now you can just, like, tell Runneth to do it, and it'll go and, like, gather stuff. So, like, it did a lot of, one of the things that we've done in preparation for this moment is that when you connect your ad account to motion, when I say, like, underhyped products, like, this is kind of what I mean, we're not loud enough about this, but our view is that, like, we have such a big community, and once people start to experience the value, like, it'll just spread organically.

It's not worth us, like, shouting from the rooftop. So we're spending a lot of time trying to get our customers to adopt this stuff. But basically, when you connect your ad account to motion, what we immediately start doing is, like, watching every single one of your videos, and we're summarizing them. So you know how people are, like, have figured out that you can upload your video to Gemini, and you say, hey, Gemini, like, watch this video.

And, like, as soon as you connect your ad account to motion, like, we are off to the races doing that for you, processing every video, summarizing every video, most importantly, like, categorizing every video based on AI tags that we know really matter, not, like, was there a dog in this image, in this ad or not, but, like, what were the hook tactics used here? What messaging angles are found here? What visual formats? Like, the key elements that we know creative strategists use to test.

And so all of this rich data is, like, ready and cleaned and just, like, perfectly LLM friendly. And so when you type into motions agent some kind of question, like, we've already prepared all of that so it can fly, right? Like, I think when a lot of people are taking this use case and trying to vibe code their own agent, which I think is an excellent idea, by the way, everybody should do it. Because then when they come to what we have, they'll appreciate it more.

I actually think there's, like, an appreciation gap that you really get. Like, when we talk to users that have vibe coded their own systems, like, oh, my God, this is crazy. But when we talk to people that haven't done that yet, like, oh, this sounds fun. When all of that is clean and ready to go, then the agent can really, really fly.

So that was another really big use case is people asking questions like, hey, what's working? What's not working? And, like, giving it more and more challenging questions. And, like, it delivers and does a good job.

And so, like, what I was trying to figure out was that where are the places it gets stuck? Or where are the places where, like, based on the conversation patterns, it seems to be going well or it seems to not be going well? And the parts where it didn't go well, what can we learn from that? And, like, there's just so many interesting use cases, I think, around processing large, a very large body of data that we couldn't possibly do right now.

But it would give us so much more clean information to make better quality decisions. So a lot of interesting stuff that I think this year, people are going to be pretty blown away how different it was to last year, where last year is, like, everyone was kind of made to feel dumb, all these really, like, splashy videos and, like, expensive courses and whatnot. But I think this is the year where, like, people are going to experience the, like, wow, stuff is real. It works.

I can't believe it. So let me ask you this. So basically you have this pattern of, you know, right now people are really looking for iterations and, or rather, sorry, personas and new angles to go after and everything. That's, like, a huge part of what, you know, you're trying to do as a meta advertiser.

There's this shift and it sort of, like, started when you were, like, launch ads and then you were using Motion to look at creative analytics. You're like, oh, that makes sense. And now it's, like, the next phase of it, which is, like, yes, you're looking at creative analytics, but it's also then unlocking the analytics now is not just numbers, but it's also the concepts. And so that's unlocking these new ideas and new concepts and new groups of people you can get in front of.

We know with creative diversity that style is important as well. So is the next phase of this, you think, like, we'll be able to go and not only unlock new angles and, you know, that type of idea, but look at what's trending on certain social platforms and be able to weave that in. And then it's pulling in that along with what people are responding to on your ads and it's saying, hey, by the way, you should run this ad. And I think that's what Meta's trying to do, but they're, like, not going to do that very well themselves.

But is that where we're going, in your opinion, on this, you know, in terms of, like, what the next phase of it is? Well, yeah, and to add, like, one more thing before that, it's like you said we have this sequoia, like, they're saying you should think about how to do two to three hours worth. It's like, okay, I pull up Replit and I'm saying, do this, and I think I have this super long plan. I'm going to hop in the sauna and I'll be back in 20 minutes.

And I look and it's like, it ran for five. I'm like, shit, okay, I don't even know how to get there. So I'm curious, like, as a part of that, it's like, is the way that we do work right now, in D2C land specifically, it's like, is there even things that AI can do in two to three? Like, is there even enough to be done for to think in that way right now?

Or are they talking about, you know, like bigger, bigger projects and we're just, we're just small fish. So I don't know if you could somehow wrap that in. Yeah, yeah, yeah. So on the style piece, I think they're kind of related.

Is that, so one of our AI tags is called visual format. That's the tag that we've kind of thought about as this, like this kind of spans across brands because like a podcast ad is a podcast ad or like, you know, before and after us versus them. Like these are the basically the formats that people are using. So we do categorize those for all of your ads.

And we also have a very, like our inspo product is one that we've been working on quite a bit for the last six months. And it's now very, very good, partly because we're tagging all of inspo as well. So like not just, we're not just tagging all of your own creatives. We're tagging the entire ad library.

We've taken a bit more of a curated approach where like, there are lots of ads that are not interesting on the inspo library. So we've tried to like rank the brands that we think do a good job, but we're also watching all of those videos and categorizing them based on visual formats. So we have a pretty good pulse on what brands are shipping, what visual formats are new or trending. And most importantly, like what are the visual formats that you have not tried yet, right?

Because if we can categorize all the visual formats that other people are doing, then we can just immediately give you the delta of what your competitors are doing that you're not. It seems we'll see if things change over time, but it seems like right now, the formula is basically something like figure out what messaging angles work, the kinds of messaging angles that if they work, you could probably spend millions of dollars against that messaging angle. And then try to find as many visually diverse ways of expressing that messaging angle until either the messaging angle runs dry or you've run out of like visual ways to express that.

And then just repeat and repeat and repeat where you basically go from like big messaging anchor one to the next and really floor it in terms of the amount of like visual diversity of ways that you can express that. And so like in some sense, there's like a simplicity in that formula that like even that, you know, it's so interesting that as an industry, we are one foot into like the depths of AI and then another foot into like, wait, what's the right way to do marketing? Like, you know, a lot of brands are trying to get their bearings about this, like these just like 101 marketing principles that I actually think the more algorithm-based the targeting becomes, the more the principles of marketing are just like kind of what has always worked.

As part of that, I think because of, you know, the question you asked Brad about the like, is there even like multiple hours of work? I think it's helpful to think about that as like a mental model, but we have a lot of like five, 10, 15 minute chunks of work that we can get through before we get to like deploying AI that not just runs for two, three hours, like it should just be running 24 seven. You know, like there actually is no reason why we don't get to a point soon where AI is just kind of working around the clock.

But I think the idea of thinking about multi-hour work is a good breaking of our chat GPT paradigm where we, because like I had a big conversation with a lot of our engineering team and product team as we were designing the new version of the agent that has completely taken off. There was a trade-off between how fast it could be, right? People are like, you know, people are going to get annoyed if it has, if you have to wait like a minute, they're so used to chat GPT, like we're going to get killed if they type in a question like, please analyze my last 300 days worth of ads and we don't give it an answer in like 10 seconds, people are going to get pissed.

And I was like, no, people, people probably wouldn't mind waiting one, two, three, four minutes if it was clear that like hours of work just got done, that would have taken you hours, like in that like two, three minute query. And so I think that actually is like, I really pushed our team to think about like, it's actually the complete opposite. It's not how fast can we get our agent to respond to you? It's how long can we get it to work and be productive?

So like the question of latency is like literally the opposite. Like I don't want it to respond in 30 seconds. I want it to go and do 10 hours worth of work for you and for users to be happy with that 10 hours of work. So it's more of like, I think anybody in knowledge work, if there's one paradigm shift that they make is that if you can get an agent to go and do work, five minutes of work for you, great.

Then try to do 10, then try to do 15. Keep trying to push that until you have an agent doing work for you for a longer period of time. I think it's useful to do all this vibe coding stuff, but I will say not just us, but like every buddy who's been building software for the last year now has like models that really work. And so I think like software products are just going to get so magical this year.

And so like for people who don't feel like going through the learning curve or figuring out cloud code in your terminal, like basically it's all coming to you in real time, really fast. And I think people are going to be able to skip the learning curve part of it, but I think it's useful to experiment and play and try things out. But yeah, software products are going to get magically good this year. Can we dangle some use cases?

This doesn't need to become a motion pitch. I have some, like I have a specific example I can share to like kick this off of like, if you logged in, if you're connected to motion right now, as an example, and you could, you could, before motion had the, whatever version I have access to of the agent in the chat window, like before that was, that was in there, you know, it's like you could, you could use, to your point, you could use Gemini to analyze things one by one. And then you could try to drop it into cloud or chat and you could try to like stitch all these things together.

So it's like, it's not like you need motion to do this. I'm just saying it makes it a little bit easier to do it. Um, maybe, maybe more than a little bit, but I think giving people some maybe specific examples, like where they can find extra value today. So I'll share one.

I'm basically telling our entire team, it's like, we should do this for every single client today because it's super easy to do. And I already wrote the prompt for it. My, my example. So like we, we have clients that, um, I think there's a lot of teams that probably feel like, hey, we're, we're kind of limited on assets.

And maybe we don't want to tiptoe into video gen AI creators. We think that's a leak, whatever, whatever your opinion is about creating it. So like we have to do with what we have. I think people get stuck thinking if I'm going to iterate, like I can't, I've got a video.

I've only have, I only have a two minute video. I can't do that much with this. And I think like that's terrible thinking because you can do two minutes is insane amount of footage for like what you can do with that. So I can share this prompt.

Maybe we could save it somewhere for people to copy and paste and drop it into motion. But it's like, take your top spending video of the last 30 days and just go ask it to iterate. But you can be more specific about how to iterate. So it's like, take this video using only the footage that's in here, only the voiceover that's in here, anything that's in the video, and then come up with a couple of different things.

One, how do I, how do I change the visual hook in a bunch of ways to make it packaged up like it's a new video? Two, how do I rewrite the narrative of this? Like, is there something later on in the video that is more compelling as a hook that you can pull up to the front, but it changes kind of the narrative and the framing of this? Three, ditch the voiceover entirely.

Just make it an organic looking video. It goes from two minutes to being 15 second clips and it just pops up the different features on the screen. And there's like a million, like there's so many other ways that you can do this. And I've done this several times.

At some point last year, I put together a case study of us doing this very manually. And we took an ad that's like, the first time we launched it, it spent 16 grand. And by the end of the 60 iterations, literal 60 iterations, we made this ad, it spent a million dollars over the course of nine months. And it's like, I'm not saying tiny iterations to death, like don't just change the text overlay.

But if like you run this, Motion will give you those examples. This is the Andrew Fox will guarantee on it on my behalf. It's like if you don't find something that like gives you additional ad spend or becomes a new winner, I would be shocked. Because you can get, if you have two minutes of footage, you could probably make 50 unique looking ads with that.

Throw on an AI voiceover that again tells a different story on top of that and you're extending it even further. So what are, I guess like, I'm curious how you feel about that. And then two is like, what other specific examples are you seeing people run? Yeah, that's really cool that I've seen a similar example where basically like repurposing content and footage is a really interesting idea.

I saw another use case where someone was doing that, but on a larger scale, like it doesn't have to be, it doesn't have to be one video, right? Like you can tell it, I saw one that was like, go do that, but for the last 100 top spenders that we've had. And then try to like mix and match and take only the stuff that you've seen from these like 100 scalers because we know we have the footage and so on. And how do you kind of Frankenstein together a new set of ads that are from like all of these different past footages?

And so like the thing that's really interesting is you can take any example and just multiply the scale on it and say like, okay, if you could do it for one, you could do it for a hundred. And the thing that is super interesting, I mentioned like underhyped idea, like the agent that you have and everyone is using and the one that is like basically gone vertical, we've been referring to it as like the legacy agent at this point. Just like that's how fast things are moving in AI.

And like there is a new paradigm that I think you guys saw this like open claw, clawed bod, molt bod, whatever we're calling it these days like that. The reason that thing took off, there's a really interesting principle in that that I think you will see applied to basically the idea of AI agents from now till the rest of time. Is that AI age and it's the reason why clawed code is really powerful on your desktop. It's the reason why clawed co-work is really powerful because it has access to your desktop.

Basically, it turns out that if you give these powerful models access to a computer, they can do unbelievable things. Like give them access to a computer and give them some time and they will do unbelievable things. The open claw is an example in the most extreme sense because like they could do crazy things. They could like, they can get you hacked.

Like it's a very, like be careful with that one. It's fascinating, but be careful. But the new version of the agent that we're releasing is basically has its own computer in the cloud, its own virtual machine. Like for every single company that's a Motion customer, we are giving their agent its own dedicated virtual machine dedicated to that brand and only that brand just so that it's like safe and sandboxed.

And so that means that any command that you type into the agent, it has like an entire computer at its disposal to go and do that for you. And the thing that when you watch these agents, how they work when you give them like a file system in the computer, the thing that starts getting crazy is that they start writing files for themselves. So like write himself a to-do list, come up with a plan, write like reference documents. And it basically is the solution to the context window problem, right?

You know that if you just throw a ton at AI, it's going to get confused. So then you're like, well, I want you to process a lot of information, but if I just throw all that information at you, like you're going to get confused and hallucinate, but when it has its own file system and you watch it, it'll like leave itself notes and give itself like reference points and jump back and forth between files. And it's kind of the solution to the context window problem. And so like when, hopefully we'll, I don't know why the team didn't give you early access.

So we'll make sure you have early access to try it out, but it's just like a different scale of example. So when you think about what can this thing do, like you can basically say, if you think about, for example, like a weekly cycle, right? Like how would a team work? How many assets do you need to produce this week?

And almost like working backwards from a weekly body of work, that starts to get really interesting because, okay, how am I getting to decide like the concepts that I want to pursue? Well, probably I need to do a crap ton of research in my own data, probably in the inspo library, which by the way, the agent has access to the entire inspo library and it's social coming soon. So it's getting to the point where if you can come up with an SOP that somebody on your team was going to go and execute that this week as part of the process to produce like, I don't know, 20, 30, 40 ads that week, you would be surprised how many of those steps is possible to just like get AI to do it.

So it might be, you know, go and look all of our top performers and all of our bottom performers and try to find the themes and trends from each, right? Like that's, that's a very common use case we've seen people do for years. Go look at all of our ads, see the top performers, see the bottom performers, watch them and try to come up with themes and trends. Like, okay, that's gone.

You can, you can be, and you know, when I say that's gone, it's like there is some back and forth to get the agent to do that work to your level of satisfaction. So I think that basically starts to become what work feels like for us. How do I give the agent a long task and go back and forth on like the structure of that task until I'm happy with the output? I'm not saying that like every time you do that, like one shot, it's going to be excellent, but you can iterate on that and then let it go work for 10 minutes and then see what that does and iterate on again.

And some of those, like what you're finding is you have a prompt that works and now you can kind of reuse it or iterate. It's, I think, I think we're going to be experimenting with different approaches to like research, let's say. Like, you know how right now we probably do a lot of work that acts as like the inputs to our decisions, but we have limited time and we're probably just kind of winging it a little bit. Like what happens if you could do 10 times, 100 times the amount of research work before you make any decision, right?

Like that, that's the kind of example that I think AI could create the most amount of leverage is, and it's so funny, we really believed this last year, that's why we never got into asset generation because we're like, that's the least valuable part of the process. The most valuable part in the process is like back office creative. It's like all of the things that go into the inputs of producing creative and that's the stuff that's very time consuming anyway. So like, how can you scale the amount of, it's kind of like the example I gave of with the help of AI, you can go and like understand 50,000 messages, you know, and like use that as an input to make better product decisions like that is helpful as opposed to like, hey AI, go build me three more features.

It's like, okay, it could probably do something, but it's like, no, hey AI, like go and help me sift through all this craziness to help me come up with my next decisions that I'm going to create, which by the way, I'll probably get your help on too, but you know, I think what it also starts to expose to is, and this is why I was saying like as an industry, we're like moving forward on multiple fronts. Probably the most important front is still just like creative strategy, best practices.

Like often the reason why people don't know how best to deploy an AI agent is because they still need to like master what good creative strategy means and the more you get good at that, the more obvious it becomes, which by the way, plug for Motions first ever virtual bootcamp that we're launching in, I think we're already starting to promote it and doors open in March. This is going to be a really cool program, mostly centered on just like fundamentals, best practices, just trying to become an excellent creative strategist.

The more people spend time becoming an elite creative strategist, the more obvious it becomes like how AI can help you because a lot of it is about like gathering inputs to make really high quality decisions and AI is very good at helping you gather inputs. So I think, I think it's going to be a fun year as people try to master two fronts, being an elite creative strategist and getting really good at AI. Yeah, I actually created a little content about the bootcamp and some of my content around it.

I'm singing for the ad, just FYI, in case you want to ask your team about that. I don't know if you saw that, but it's going pretty viral right now. I would like to transition us to a new segment of the show that I've never even told Brad about called After Hours, which is what we would talk about if we had hit stop because it's always really interesting what happens when you aren't recording and what we get into. And I want to ask a question.

So welcome to the After Hours segment of this podcast. Right now, it sounds like, you know, your opinion is people need to be focusing on longer agentic shifts in terms of like they need to be, you know, giving more hours, more tokens, and that's like the metric of getting into it deeper. You see across a lot of creative teams, like what are, what are the things that people need to be doing right now that there's not, like that are, they're missing on or they're stuck on or that you see all the time that just drives you crazy that you're like, look, like you got to get better.

Like this is something that's happening. So one of the things that we've been talking about internally quite a bit is just how important hooks are. And I think that's not a new idea. I think like everybody who's doing anything in direct response, creative strategy, there is this obsession around hooks.

But one of our views recently has been that hooks are so much deeper than this like tacky pattern interrupt that you might do in the first three seconds. It's almost like the hook is like the DNA for the whole thing. And if the team gets good at hooks, they've got it all figured out. Like everything basically becomes downstream of being really good at hooks.

And so I think it's so interesting seeing like a wide range of teams that are elite at creative strategy. And I actually think like that's been a minority of people. We live in this like bubble of exceptional D2C marketers, which has been like, I just feel so fortunate that we've been amongst this like group of elite advertisers. But like you stick your head out and you talk to like the everyday brand like, wait, what's a hook?

Like, what do you mean? Like, what's going on? Like, and so I think there's a lot of people who are just kind of learning direct response principles for the first time. And I think if that piece of the equation is not in a good place, then everything else is going to suffer and struggle.

So like as people come up with like concepts and ideas, but they're not kind of thinking about them in direct response hooks, then like, then things are really going to suffer. And like, this is probably a bigger conversation, but one of the things that I've started to notice, even with the elite direct response teams who have completely mastered hooks, I find it mind boggling that there's still this brand team that hasn't understood what's going on here yet and is disconnected by it.

And I have a theory on this that I feel like the brand team is in for a bit of an awakening because you go back like seven, eight, ten years, the people that were doing direct response, they were like destroying the brand, right? They were doing like crazy things, like tacky things, like things that were just like really easy. All right. Well, you know, like, but partly it was because the lever was audience targeting.

It's like, listen, I just need to get in front of the right person. As soon as I do that, like we're printing money. And so it was just like things that were not necessarily rooted in, let's say, creative taste, right? Like it was, it was not very creative.

But now you actually look at direct response is a very, is a very creative sport at this point. And the people that are really good at creative, direct response are actually producing incredible content that in many ways is indistinguishable from what's going on in organic. And the reason I make that point is to say that it used to be that the brand team was basically like they understood the customer. They didn't, they understood the narrative that connected the brand with the customer.

Whereas it seems to me like now the direct response creative strategists, they are the brand, like they are the ones that are actually building these like connections with customers through ads, through social. And the brand team is actually the one that is more disconnected at this point, not to create a feud amongst teams, but like I've seen it quite a bit at this point. Like if you have a team that's like starting to think about TV or starting to think about 100% by the way, totally agree.

Billboards, it's like you obviously should take the best performing hooks from paid social and obviously that should be the basis for whatever billboard campaign you want to do, right? But it's like, not really. That's like, there still is disconnect where people are, and so I think our community of creative strategists are going straight to the top basically that the more power they get, the more influence they get, the more revenue that the companies make and the lines between what is direct response and what is just like good marketing at this point is like those lines are gone and so my hot take is our community needs to go straight to the top and basically get more power and influence at companies because they know what's working, you know, and like the, that's the kind of messaging.

Yeah, and it's being tested like millions of times a day, you know what I mean? I actually, I agree with you and I literally had this conversation like the other day with somebody about that exact topic about I can't believe that the number of brands I see that are big brand, big D2C brands that they know all the data on the creative and what drives like revenue for them and then they still put out, my example is connected TV, this like connected TV thing that looks nothing like it and takes none of the elements and I'm like, I don't understand, like this thing's over here is making you a jillion dollars and none of that shit's over here and this other thing looks super polished and they, you know, work with some connected TV company and whatever and it's like, doesn't make sense.

You brought up a new segment which was After Hours and then you went into beefs and I just love how that flowed together and it's great. Let's get into beefs. For sure, that's a beef. One last point on that, I think Dara Denny started talking about this and I think she's, we're like completely on the same page with her is that the creative strategist job is not necessarily to just like make more ads, it's to grow the business and I think in tech, there is this idea that the product manager is basically like the mini CEO, right?

And the CEO of their product area or whatever and I think the creative strategist is basically the mini CMO and the skill that I think they need to learn and like organizations need to learn is like last few years was around bridging the gap between the creative and the data and now it's like an altitude gap where it's like how do you bridge the gap between what's going on on paid social and social in general to basically the business and the more rapid the feedback loop between higher altitude business decisions and what is going on in direct response, I think that's the flywheel that will enable companies to grow really fast and so I think to all the creative strategists who've been like learning and leveling up the next year or two I think is leveling up to be like you should butt your head into what's going on at like the leadership level because the amount of value that you can create to get them on board with the ideas around direct response like really could move the entire company so like get the leadership, get the CEO, get the board involved in ads and like thinking about that because it is the most powerful signal for go-to-market for any consumer company is like taking the things that are working here and applying the entire go-to-market strategy around it and not being like oh there's this ads channel like no, no, no this is like the whole thing basically like this is the feedback loop that you drive.

Yeah, not to like divulge into a whole separate topic but like a lot of brand is viewed through this like layer of consistency across messaging and how things show up but social feeds are everybody has a personalized progressively more personalized feed to them and so like if you want a level of consistency consistency doesn't necessarily mean that all of a sudden you're saying dramatically different things and your values are changing it's just like you're making them show up to that specific person in the way that they care about it and I feel like that's a pretty big adaptation like that's what performance mark because we don't care like okay, I'm gonna be a little dramatic for a second it's like we don't care we just want to make money so it's like we were on that extreme and now like that's starting to be true and on the brand team you know, they're thinking well, we need it to be consistent and it's like there's a way to be consistent without being dilutive but also speaking to the person on the other end because like if there's trillions of pieces of content out on the internet and people are spam scrolling their feed every single day it's like what's getting them to stop and pay attention is the thing that is more unique to them which is why Meta is spending billions of dollars on GPUs or whatever the hell that I don't understand so anyways it made me think of that if the consistency originates from what's going on in direct response and we use that as like consistency across everywhere else I think that achieves brand goals of consistency while making revenue for the business and actually like because direct response is such a powerful beast at this point you can't you can't fight it anymore because then you're going to be inconsistent so like if you if you don't fight direct response and let it consume the brand and make sure that you're doing it in a way that's like authentic to the brand like I think those are the brands that will do well both in terms like near-term revenue and long-term brand is you got to work the hooks right into your brand identity you know it can't just be like our brand identity is X and then we're going to go try all these like unrelated hooks on paid social that's not going to work like the hooks are the brand basically the hooks need to be consistent to the brand and it's not going to be the brand that's going to set the hooks I think it's like kind of the other way around so there is a merger that needs to happen that I think I think is going to be really interesting and our community is very well set up so I think this is the year for our community to be like a lot more ambitious with the amount of value that they create in companies and not think about ads as a channel but like the work that we're doing in paid as basically like the microcosm of the entire business is go-to-market strategy yeah yeah there's many other things that I'd love to dig into so we'll have to save that for a follow-up or a separate time especially as it pertains to what hopefully I get some early access to but this is my after hours questions like in the short term it's like this is awesome dopamine is perpetually popping off because you're just unlocking new things day after day but like do you ever get like sad just like shit what's going to be left for us like are you there yet do you have that do you fall into that at all or are you kind of like yeah we got time I don't go there for myself partly because I think that if you are a person who is like really high agency and really curious and like loves tinkering and playing around with stuff like we are in the moment where like that person is going to have a blast and like personally we've grown motion we're like 75 people now and like we have lots of different teams and so on like I'm still like deep in the weeds of like playing around with AI and doing cool stuff like just if uh if we didn't have motion and I wasn't running this company I would be having a blast just individually playing around with all this stuff um and I I think there's a real superpower to the the people who are doing that it is concerning I will say it is concerning if you are not leaning in that direction because um increasingly that is going to be the way to create value which like I really do think about um we are allocators of AI tokens I kind of love that job right so like from my standpoint I'm like that is exciting and so I'm fired up but I think uh where I get concerned and nervous if if you know there are some people who are like I don't know what that means I don't want that job like I have a different job I've gotten used to a different job and I I think maybe it's not like this month next month maybe not this year maybe not next year but I think we're getting to the point where that's the only job left is being an allocator of AI tokens and if you're not doing that job I don't I don't think there is a job that is not that interesting so let's get allocating giving us not just food for thought but a meal a meal for thought an entire buffet for thought uh with this episode uh Reza very much appreciate your time and um thank you for joining us on the scalability school podcast thanks guys always a pleasure this episode is brought to you by the Foxwell founders membership that Andrew and his wife Gracie run it has been absolutely pivotal for not just the homestead team but the easy street brands team we've had I don't even know how many members are currently in there that are part of our ecosystem but when it comes to anything from learning ads to understanding what's going on to building an agency to knowing retention it's been absolutely useful for our team when they get stuck or they need help to just go there and resource all the other experts so definitely would recommend it for anybody that's looking to you know take it a step deeper try to get a little bit more knowledge on growth marketing and all the world DTC is yeah I think one of the most incredible things about it is you can just like open up the Slack group every single day you can pin your favorite channels for the topics you care most about and like every day there's going to be somebody who just like because they want to contribute something valuable to the group you can go learn something every single day and it's going to be extremely useful there's some ballers in there that you just get like the benefit of learning from that like for the cost like you couldn't pay them that for their time but through the membership like you get access to some incredible people and tons of resources yeah I mean I think the biggest resource to me too is like the events that you know Foxville Founders does they've been able to do some even in Wisconsin even in the boring state of Wisconsin which is pretty awesome getting people together in person and able to have really just like honest conversations of what's going on what's working for them now you know where they're at in their business and knowing that there's going to be like Brad said some real killers in the space in this membership that can help and are willing to take the time and help so that's been a huge part of why a lot of our team have really enjoyed it as well and the applications are now open if you're looking to join so Foxville Founders yeah foxvillefounders.