Go back

Episode 275: The Paid Ads Machine: Creative Systems That Scale in the Age of AI

54m 16s

Episode 275: The Paid Ads Machine: Creative Systems That Scale in the Age of AI

This podcast episode, featuring Ben Bennett and Melvin Cotton, presents the "paid ads machine" for navigating creative strategy in the era of Meta's Andromeda algorithm. Andromeda is intimidating because it understands both creative details (e.g., pacing, emotion, narrative structure) and user behaviors (e.g., subway browsing habits), creating a black box that makes advertisers feel powerless. To counter this, the speakers emphasize that while the user data side is inaccessible, the creative side is fully analyzable. AI, like a lab assistant, can be used to examine creative elements through a metaphorical microscope, but it requires human expertise to guide it. This guidance comes via prompts that teach AI specific frameworks, such as emotional or functional analysis, turning raw creative into actionable insights. The process is straightforward: feed creative into an AI system, apply a prompt (the lesson), and receive output. While automation can streamline this, the core value is in the prompt design and the principles behind it, not the technical setup. This approach allows marketers to understand their ads at a deeper level, reducing reliance on sheer creative volume and instead focusing on strategic, informed production. The session concludes by emphasizing that this system compounds over time, enabling continuous learning and improvement in ad performance.

Transcription

10002 Words, 53009 Characters

English
[Music] Hey, Swarm Marketers. Welcome back to the podcast. I'm Anthony, creative producer here at Swarm Marketer. Over the last few episodes, we've talked a lot about how paid traffic is changing, from marketing and retargeting to attribution to using AI for stronger ad copy. Today's episode brings those ideas together through the lens of creative. We're sharing a live session from our recent Mentor Table event in Nashville, featuring director of creative and Bennett and creative operations manager Melvin Cotton, the third. They walk through the paid ads machine they've built for building creative planning, and analysis in the age of Andromeda. You'll hear how they think about what metacies in your creative, what the algorithm knows about your audience, and how AI can help you understand your ads at a deeper level. If you're trying to create winning ads without just making more stuff, this episode will give you a much clearer way to think about the system behind great creative. But first, here's a quick message from Ezra Firestone, while he chooses Omnysend to power his email and SMS marketing. Hey, Ezra here with a quick note. You may not know this, but I just bought a new company. And you know what the very first thing I did was, was get them set up on Omnysend. I migrated them to Omnysend. That's how much I trust that platform to drive email and SMS performance. And you know, it's like Omnysend isn't as well known as some of the other brands out there, but I really believe in what they're doing. Not just the product, but the people behind it, their support team is next level. And the platform gives my brands everything they need to drive real growth. So if you haven't heard about Omnysend for email and SMS, I think you're really going to love them when it comes to comparing them to other things out there. Check them out. They're offering a special deal to smart marketer subscribers. So you can go to smartmarketer.com/omnysend. That's omni-n-i-s-e-n-d. Smartmarker.com/omnysend. And you can use the coupon code Smart to get 50% off your first three months. I don't think you're going to find a better platform that's going to help you get set up and integrated and optimized. So go check them out smartmarketer.com/omni-send. Use the coupon code Smart for 50% off your first three months. Thanks for listening. Okay, y'all. The most requested session. Is that put a lot of pressure on you guys? Yeah. [Laughter] All right, we already teed this up a bit. This is called the paid ads machine. You guys will learn about the content, but I want to introduce the people. Most of you know these guys, but if you don't, first I want to introduce Ben Bennett. What's up? Ben is our nicest team member. He is like our resident Golden Retriever. Melden is second nicest team member. Ben Bennett actually, we met because he was going through one of our trainings four or five years ago. And he went to college in the town that I live in Kentucky. Ben's going through the training and I can see like, oh my god, this guy has such a marketing mind. His storytelling, he just intuitively understood what we were doing. And he was working at a distillery. And I didn't want to poach him because I knew these people. We live in the same town, but I did tell Ben if you ever leave your job, let me know. So a few weeks later, somehow. [Laughter] It was really quick. Yeah, it like happens. Ben hit me up and I said, okay, I don't really know what you're going to do, but we'll figure it out. So we may have been a junior media buyer. Like, okay, he'll hop into ad accounts. And he'll help. He was awful. It was not going to work. This brain is not analytical. But what he was really, really good at was the creative side. That was four years ago. Ben now leads our entire creative team of 20 people. So it's like a 27 year old dude just making it happen. One of the leading industry experts when it comes to paid ad creative. So we're so lucky to have been, he's going to kick off the presentation. And he's going to be a great guy. Awesome. Thank you so much for the tee up. Is this close enough? Oh, I need to bend down. [INAUDIBLE] That would work? OK. That's a little closer. [BELL RINGING] Now I've done it. [INAUDIBLE] I was so sorry. Do you want the hand out? Yes. That's great. [INAUDIBLE] Back here, this top. Yeah, that's great. [INAUDIBLE] I need two microphones. So Ezra, this doesn't work. This one doesn't work if you like. Group it. OK. So anyway, paid ads machine. So thank you, Molly, for the introduction. It was wonderful. I'm so excited to get to speak to you all today. This is probably the easiest presentation I've ever done, because they were like, hey, Ben, you know that thing? You do eight, nine hours a day work. Go talk about it. So fantastic. I'm really excited to get to talk to you about all about what I've been doing. Before I continue again, for Melvin, man, it's like Ben leads a team of 20 people, not true. Ben leads a team of 19 people. And co leads a team of 90 people with Melvin. It could absolutely could not do without you, man. So super, super grateful for you, man. Absolutely. Rockstar. So to get into it, what are we going to tell you guys today? We're going to tell you about the whole entire paid ads machine that we employ and that we're building for all of our agency clients and sharing with you all here at Mentor Table. So I have marketing, you know, Lingot this up as the roadmap for navigating creative planning, production and analysis in the age of Andromeda. Boom, drop Andromeda in there. Super current within, all right. [ Inaudible Remark ] Prangum. I'm hitting that. Okay, sweet. All right, here we are. Okay, the full machine. So this is an overview of what we're going to talk about today. There's three parts to this if you want to break it up in the most simple parts. There's the diagnostic part, which is analyzing what's already in your ad account. I would start with production, but if you're in this room, odds are you probably produce some creative before in the past. So we're going to look at what you have, what's already been going on in your ad account, and how those dollars have already been spent. So that's what I'm going to spend the most time talking about today. Melvin's going to hop in and help us learn how to produce them because it's not enough to just see what's happened and what's going on in the past. We have to learn how to go forward and make them. So that's where the rubber meets the road. And then that process compounds. It builds. It learns on the learnings that we do. And we start the whole thing over again, right? You guys are familiar with this. Boom, down is down. Up is up now. So just make mental note, Melvin. You don't have to struggle with this if I give you the secrets. So diagnostics. So why is Andromeda so scary to advertisers in 2026? Don't raise your hand. I'm not going to make anybody out anybody in this room as to that you're scared about this. But odds are you probably are if you're in this room and you're tuned in locked into this presentation. I think that Andromeda scares us because we have the perception that Andromeda knows more about our ads than we do. We put it into Facebook and then there's this magical wizard that just understands everything about what's happening and it knows exactly who to show it to. And we can't possibly compete with that. So these are the two halves of basically as I understand what's actually going on and say he's going to come up here later. And maybe she just destroys this and like there's something that came out in the last two days that is that makes us irrelevant. But as it sits right now, this is how I see it. So there are really two halves to this thing, right? There's what Andromeda sees and what Andromeda knows. So the first half is what Andromeda is seeing. Now these are the things that it's looking at when you upload the creative, right? This isn't outside knowledge that it's pulling in. This is knowledge that it's seeing about the creatives that you made and that you upload. It's seeing things like for videos, what's the pacing like, what's the rhythm, what's the emotion behind it, what's the tone of the video, are using text, what's the imagery, that's the narrative structure, like guys it's getting this deep into what you're uploading, what are the demographic cues, what is the people in your ads look like if there are people, what is the behavioral context of the person experiencing it and that's what we're going to move on to and who knows what else, who knows, Sammy, again, I'm going to say Sammy's name a lot and she's not like a, she's not like don't treat her like some sort of meta inside her person that's like secretly a robot from Mark Zuckerberg but she's not, she's a wonderful human and you'll learn a lot from her. The other side of that, so what Andromeda sees is things that's seeing about our created. The other half is what Andromeda knows about the people that are using Facebook and so that's those things like what you do on the subway. I use that example to say Andromeda knows that if you're on the subway every weekday from 8 to 8 15, you don't wear headphones and so I'm going to show you a video that has big subtitles or that doesn't show a person talking to the camera or I'm going to show you image ads because guess what you end up shopping on the subway and because it's winter, I'm going to show you winter boots instead of these are the things that like this is where it does kind of get a little like trippy like Andromeda knows and sees all those things that you're doing right and this is why I think we see everything as this big scary black box. There's a wizard behind a curtain holding a box in which it's saying I'm going to decide which ads perform and I'm going to decide which ads get spend and which don't right this is how we view Andromeda. So we end up falling prey to the idea that if we can't understand as much about Andromeda as much about our ads as Andromeda does, why even try? And this is why I think that the narrative of creative volume is so prevalent right now because people are saying who knows? Who knows what Andromeda's doing or seeing in your ads? You just have to make a hundred and maybe ten of them get spend, maybe five of them get spend. And so I think that's why the narrative of creative volume has come around so strongly because we can't see this right half of the image. We don't have access to all of the information that Meta knows about the individual behaviors of our audience. And so as marketers, we're anticipating those things and we're thinking where they can experience this creative, but we don't have the incredible, you know, statistical, significant data that Meta has about every single person. But guess what? This left side of the image, what Andromeda sees in your creative, what it knows about what you're uploading is 100% fair game. So I hope that I can reframe for you over the course of this presentation that we can understand what is the real things that are actually happening. So how do we do that? How do we make that happen? We're going to level the playing field. And so this is a clip I found. This is like a screenshot from a YouTube thumbnail that everybody's, you know, hating on chat GBT right now. And I'm kind of in that boat until you hop on co work. I know as we mentioned co work earlier and that it's like the server's busy. Look, I don't know is anybody else like you're just hop on you're like, sick, I'm going to try this, this co work thing out, the server's busy. I mean, apparently this is not a room of people who are using co work a whole lot. Russ, I got to, I got to like shout out from Russ. Yeah. So at least Russ is running into this problem. But as of very recently, we have access to artificial intelligence that allows us to level the playing field with what Andromeda knows about our creatives. So this is what AI has changed for me. And this is what AI is going to change for you in terms of your creatives. So AI changes the amount of information we can access. That's the first thing. And then secondly, obviously it changes the amount of data that we can digest. So first, digesting data. This is the less interesting part, right? This is from office space. Has it, do we have any office? Russ, look at Russ. It's just me and Russ talking here. Well, me and Russ will just go in a different room and then hang out. So office space is basically they're just chewing up data. They have their like coachy little desk job and they're like, we want more out of life. So they invent this scam and then they end up getting in trouble. And they're like, well, we want our desk job back. This is what I picture AI doing is digesting the data like these two people from office space to hate their job. We understand that part of the process, right? It basically transcends our ability to look at the data and to understand what's happening with a massive amount of data. The second thing that it helps us do I think is much more interesting. So AI gives us a new way to see our creatives. It gives us a microscope through which we can evaluate our creatives. And so this is what that microscope looks like again. A microscope is an amazing tool. But I don't know if you all had a similar experience when you sit in chemistry class and you're like, all right, like now we're going to look under the microscope. You're like sick. That thing looks small. I guess I don't know like what's happening here. We have no idea what's going on under the microscope. The magic is in the scientist ability to look at that and say, that's a mitochondria. Can you be a point to the mitochondria? Russ, no, I'm just kidding. So the magic is in understanding what we're actually looking at. And that's what I'm going to show you how to do today. I'm going to show you the trade knowledge of what can I use AI to understand about my creatives. So our role versus AI's. So you're going to see these two characters a lot. I've got beaker here to my left, which is AI. So beaker is super eager, over caffeinated, ready to dive in and to really, you know, he's going to chunk up all that data because he's on like caffeine and whatever, whatever else, right? But he has zero in the field experience. He's not a creative strategist. He's not a scientist. He's needs somebody else to go and guide him, right? That's where honeydew comes in. They're on the on the right honeydew Bunsen. So any Muppets fans Russ. I'm just going to keep going to you rest the whole time. Anyway, we're honeydew Bunsen here on the right who has been doing this since since Beaker was born, right? We know the ropes and we've been doing it. And now it's like, hey, we have this incredibly new powerful assistant. But we need to teach it everything we know. And this is the procedure of how that works within our creative data lab. So we take our creative, which we already have in this case because all you guys are in this room with successful businesses, you have your creative. We're going to take that. We're going to chew it up. We're going to put it under a microscope. Then our lab assistant is going to say, okay, what am I seeing here? What am I, what have I learned from you in the past about what I'm looking at? And then they're going to write it down. All they do is just write it down in their little lab log and move forward. That's the process we're going to go through today. We need to guide it, right? This would be chaos if we didn't know how to guide our senior, our lab assistant. So how do we guide it? Prompts become lessons. So when we set up our system, when we're going through, how do we set up this creative lab that I'm talking about? The point is the prompts actually have to include the context for the framework for which I'm going to talk about. And that's going to make sense when I go over some examples, but it's not just, hey, here's how you're going to look at this creative or here's what to do. It's by the way, you're not a marketer. You don't have creative strategy education. You don't know what emotional frameworks we're going to apply. You don't know what functionals, you don't even know what that means, right? In your face, AI, finally, somebody had to say it, but this is where we come in. Thank you for saying that, dude. Beave and Russ, I'm coming back, I'm coming back and forth between y'all. Okay. So the process is basically creative in. Then we give it the prompt. That's the lesson for the AI to understand what's happening in our creative. And again, when I say creative, this is not ad manager data. This is the creative itself. This is the image, this is the video that you've produced. So the prompt's going to chew it up. And then we simply have output, right? So you get the system going. It looks great and you're like, wow, this is phenomenal. Let's add more. So then you get to this point where you're like, okay, I'm going to take my creative and I'm going to have it analyzed this way, this way, this way, this way, you're like, whoa, look at all this output we have. And then you end up with something like this. So this is a literal screenshot from the AI automation flow that I have set up very rudimentarily, if that's even a word. This is a screenshot from the AI flow that I have set up. And basically what we're looking at is you have the creative that goes in on the top. Then you have this massive middle section that takes that creative and applies an analytical framework to it. It chews it up in a way in which I taught the AI to do so. It does a little formatting. And then it has an output. So creative in prompts output. So I'm going to stop there. I know that might have been intimidating for some people. They look at that and they're like, oh, I can never do that. First of all, yes, you could. I does most of it anyway. Do not get overwhelmed by the automation or the technical setup here. The value in what I'm presenting is not in the automation or the setup itself. It's in the principles and the way and the lessons that are contained in those prompts. So I'm going to show you how I've automated this, but this process is just as effective as a manual process. So if you can copy and paste, which again, I'm very confident that all of you in this room can do, this will benefit you. So this is what we're looking at. It starts at the top with a Gemini prom. This is not the font's got a little change so this is not like a prom for AIs. This is a prompt for Gemini. So thank you. That was off the cuff. Yeah, oh dude. Yeah, Claude and whoever else go into this side tangent. I love the name Claude. My wife is having a baby in July and like I'm thinking about like it will Claude be like raise your hand if you think Claude is too like popular to name a baby dang it Molly put her hand up so it's over. It's over. Clyde. Okay, please remind me after because I'll forget Clyde. Clyde? No, I don't. I think we're different generations here as they're yeah I'm too young. Clyde Owen anybody? Am I saying that right? Clyde Owen? Okay, okay so yeah I mean if he takes after me he'll be long. You know, am I right? No, definitely not. So anyway the Gemini prom. So step zero is we're going to take our creative from a photo or a video and turn it into something that AI can understand right? This is like you come in to the lab with your specimen. Now we're going to put it onto a little glass sheet and we're going to drop a glass of or a little drop of water on it so we can put it under the microscope right? We have to prepare the creative to be in a readable format for the rest of the process. So this is a massive big long Gemini prompt just to show you that this is the type of prompts that we're working with and this is why this system took so long to set up is because building these prompts to rely to run reliably like time after time takes a lot of a lot of iteration and a lot of patience. So that's why I put this slide up here but basically what it's doing is unfolding the DNA of the creative. So I'm again not a science guy but when you look at DNA you have your G's and you have your A's and your C's and your T's and I only remember that because of the movie Gatica and our science future was like oh that's all the DNA things anyway. It's not relatable do you remember that movie Gatica? Yeah shout out okay sweet so this is basically unfolding the creative DNA to say hey this is a format that everybody else all the other analyses can understand so super simple we basically at the top have that Gemini prompt that choose it up and then feeds it to our analyzer so super simple and these are the different just to put this into perspective for you all I'm going to show you a couple of those analyzers like I'm going to share all of them with you in a goody bag that Melvin and I have put together that is full of like all you'll ever need for this but I'm going to share all these prompts these so don't furiously write them down it'll be in a it'll be in a giveaway but there are two types of variables when I'm looking at creative and again I'm talking about images and videos there are closed variables and you can basically think this uh this has like this is a drop down box if you did a Google form you're like choose one of these I'm not leaving you an open text field to write all that you want these are closed variables there's a limited amount of options so these are the first things that when I was like okay well how can we use AI to analyze our creatives these are kind of the first things that came to my mind so for example create a style I mean I get this question all the time I'm like what style should we be doing for our industry like what what is the you know the the fit and polish that we should put on our creatives per our industry right this is a prompt that will analyze what Gemini spits out and tell you what style what style the creative should be so I put a sample here just to show you again how in depth these prompts go and a show an example of if you want the system to be good you have to do a really good job teaching the AI so you see influencer here and we're going to describe that in as much detail as we can per prompt so what does an output look like from the scenario super simple to drop down you just put style graphic boom awesome next one emotional motivator if you guys heard me talk in blue ribbon in like three years ago I spent a lot of time talking about this right because I think it's super important but I basically worked this into our creative analysis flow to say I want to know the emotional motivator every single creative that we make and I want our AI flow to be able to tell us that so again here's a sample what does the output look like competence it's just one it's just a drop down thing it's super simple okay now we know we have a graphic and now we know that the emotional motivator that we're targeting is competence wow the world opening up to us right awareness level this is another one that you guys are probably thinking about when you're just going through and you're making your creative report what what warmness what level of the funnel are we targeting with our creative right unaware problem aware solution where deal output super simple just the drop down box so we know that we have a graphic that targets the competence motivator for problem aware people right we're getting a full picture do you see what we what we're doing here we're building a machine readable picture that rivals that of andromeda we're knowing what's happening in our creatives and we're putting it on paper we're putting it on a spreadsheet but I show you these massively complicated prompts in this system just to get three closed variables and I show you this for the reason of saying that you can do this by yourself without without AI right these are all evaluations you can make of your creative independent of having AI make this decision so that's still super important and your creative tracking sheet should include all those things this is where it gets a little more juicy right these are the open variables so think of this as you're making the google form you have your drop down and then you have your long text thing where you can let somebody go nuts right describe and however much detail you want what is actually happening so these are the long text the narrative analytic data that we're making just for the purpose of AI we're not putting this into a spreadsheet for us to read we're putting this into a spreadsheet for AI to be able to digest and understand what's happening so let me give you some examples this is where it gets really juicy because as a creative nerd as a marketing nerd I can take all the super fun books in the fun analytic frameworks that I've learned over the years and without any manual effort start to evaluate our creatives from that so first one Donald Miller fans in the room absolutely I got to be a lot of people in the other raise their hand it's phenomenal phenomenal marketing education so I was sitting there building this flow like what if I could have Donald Miller's framework applied to every single one of our creatives what if I could pick his brain about what's happening in every single one of our creatives and have AI be able to analyze that so I took basically a prompt that says here's what he thinks here's his framework can you apply that to every single one of our creatives that goes to this flow and this is what came this is what comes out again there's a lot of stuff happening here this is not for us this is not for our eyes but now in that spreadsheet and you can see spoiler alert there's some other there's some other categories up there as well that this spreadsheet is growing into a knowledge base for AI to be able to be a creative expert about what we've already produced it already exists and so now this thing is getting a lot smarter so diving into more of them functional versus emotional think back to TMTP the example you have a hundred it was like literally like a hundred megabyte iPod back in the day so like you know that tells us how far we come so you have a hundred megabyte iPod that's the functional benefit the emotional benefit is this is nostalgia in your pocket it's music everywhere you go it's life in your pocket right there are both there are strengths and weaknesses to both of those arguments some people respond to functional arguments some people respond to emotional arguments what if I could take everything that we produced and have a score that tells me this ad is super emotional or this ad is super functional and not only that but again paired with data I'm getting a little ahead of myself but this is what the output looks like to where I can say one being super emotional zero being super functional this ad is right down the middle it's a point five it's right down the middle I wonder how that impacts performance right and so on you can go so nuts with this philosophy and I'm sure that's what I'll be doing for for a huge part of my career right I'm so glad that Ezra took the time to talk about you know AI and working it into the philosophy of what you're doing because this has allowed me to basically have you know start to to build out the framework for my team being able to know or to at least see all of these different frameworks that would take years of training and years of explanation and so on and so forth so what do we end with we have the process that we started with which is the analytical process and then it ends in a Google sheet it's so simple like it's so easy and again you can copy and paste all these things so if you've never built an automation in your life this is still this is still available to you so we end with this Google sheet that looks a little something like this now this is a screenshot of a real test case that I ran on an ad account of a creative spreadsheet that has so much a wealth of data of information of what's actually happening in the creatives I mean you can see just going down the headings of those columns there all the things that we can see have data for and test against our ad data what is the goal you know it's great to know all of the stuff about our creative and it's fun to play with it. AI and it's fun to play with automations. What is the goal here? We're closing the knowledge gap between meta, what meta knows about our creative, and what we do. So we're basically reverse engineering, the process that Andromeda is going to look at your creatives with, all of the different frameworks that it's using to look at your creatives. We're gonna just take our strong human marketing brains and put those down into AI and use it in a spreadsheet so that we can have all of this knowledge for our benefit and for our use when we look at our own data. So we end with this machine readable format that matches or rivals that of meta. And we can, you know, we end here, right? We're at the data out section. We have this massive spreadsheet that looks so, like satisfying, it looks like a kaleidoscope when you look at it. So much data going on, what do we do with it? What do we do with all this data? The next step is plug it into AI, right? AI is going to, this is the digest part, the office space part. This is where our AI is going to start pairing the data that we've learned from our analysis flow. So you take your creative data and then the other part that we're introducing is our add account data. So how did it actually go? What got spent? What was the row as? What was the CPC? What was the CPMs? All those fun metrics, we're going to start connecting them with the data that we made, with the data that we used AI to make ourselves. And so we're taking our choice. Cloud code work is the best one. And Cloud code work is the best one I've found. If you can get it to server to work and not be busy for you, even though we pay for it, what is that about? That is the-- you just walked in at a certain point. I was like, you just told me to use this, even though we've been using it for weeks. But I appreciate the advice. But we're taking our AI of choice. It's good advice. It's good advice. That wasn't a slight. So I'll be hearing about that one later. That's going to be the-- I'll be in our next, whatever meeting. So we're taking our AI of choice. And we're creating now a creative report that will blow your mind. And so I have some examples. First, let me talk about how to do it. Super simple. Just go to Windsor AI, go to coupler.io, not sponsored, no free cloud. I had no choice over-- no preference over either of these. But it's super simple. You can figure out yourself. And I saw you were trying to take a picture there. And I got you. Hold on. I got you. Did you get it? Did you get it the first time? OK, hold on. OK, never mind. That's all good. Wow. Thank you. Just the random interruptions and sidebar conversations. Is that what you mean? OK, thank you. That's what he means. I'll cut that one short. So what are we looking for in this report? So now this report is being created. What do we actually get out of it? There is-- guys, like this has changed so much of how I view. Like we're onboarding a creative client right now. And yesterday, I, again, caught it's great. The usage thing, man. It's like, I don't even use Opus ever. I use the lower one so that I don't go over it. I still go over it. I feel fix it. It'll be all good. But I was just sitting there with all of this data done for this new client that we're onboarding and asking, called these questions. What are we? What are they? Over-testing. What are they under-testing? You're going to start to see. It's going to be like, hey, look at your awareness scale. All of your ads are solution-aware. All of your ads are problem-aware. You need to branch out. And I did a presentation like this two years ago at SmartMarketer Live. And I said things like this, like look and see where your gaps are. Now we can have AI do it ourselves. Like, nothing has changed except for the mechanism by which we can get it done at scale. So things you can start asking it are the questions that, for us, people who have to deal with Facebook data, you dream about it night. What if you could ask it? If we do an image ad that is targeted towards beginners, and then we also mix it with the achievement, emotional motivator, what happens? Have we done that before? What's the row ads look like? What's the cost per click? And better than that, AI is better at looking at data than we are. Like, we can just ask it, what are you seeing? Tell me what to look for. Like, you're going to look at this report and see all these connections that I never even would have thought of. Tell me what's going on here, right? Put, you know, let's get some money out of that call subscription finally, even though I'm not paying for it. Once again, thank you, Ezra. Getting more juicy. These are the deeper questions that you can get into. Like, imagine if you could just sit and ask a club, what happens when I show the product in the first five seconds of my video? Not just how much spend does it get? Not just how much row ads? How long do people stay compared to the other videos? How do they drop off if I show it? Is the conversion rate better? Is the AOV better? These are all questions that you can ask just as if you were sitting there, basically, like with an unwilling, like hostage participant that knows everything about your ads, and they're just tied up and you're like, tell me why my videos are bad. And then they have to do it, right? That's what you get out of this kind of scenario. So these are just all examples of what you can see in a report like this. So things like beginner persona, if you're looking here on the winners, the beginner persona, guess what? That persona is outperforming the one that we've tried most of our ads with, which is the struggling persona, right? Like, you think about how this changes when you go to have your creative conversations. You're saying, okay, it used to be, okay, I'm gonna sit at the table, what problems are people struggling with? How do we get emotional pain point footage that covers the problems that our avatars have, right? That's a great discussion, and we're still gonna have that. But this one is gonna tell us, it's not so much about that. The ads that we see performing are about beginners. So a person who's not necessarily struggling with something, but who's excited about it? Who just doesn't know? They're like, hey, I'd love to try a new sort of natural hair product. I've never gotten into the natural space before, and not it's, oh, all the other natural ones don't work. Like, that's who we think is watching. This gets to tell us who is watching and who is actually converting based on all that data. And so this is kind of the example that I just shared. But this was a big insight for me. I looked at this and said, this is counterintuitive to what I would have thought. I would have thought to put forward really good pain point footage and to put forward really good problem scenarios for people is the most effective thing. This is showing us that people haven't started the journey that we assume that they're super far on. So let's make some ads geared towards beginners. Let's move up to funnel. By the way, it's a cool finding because, you know, talking about getting new customers, you move up to scale. That's an amazing way to do so, the awareness scale. So other things you can see are not just single variable analyses. So there's simple things like, okay, you have the list of personas, which personas perform the best. That's simple, we all understand that. But where AI comes in is it can tell us, not only like, here's a persona that works super well, but from this list of 20 different variables that you have, when you connect these two, that's when you've really got something, that's something that like, oh, boom, that's a formula that we can take in and then multiply with our iterations. So these are just some examples like, oh, by the way, that stuggler thing that we were talking about, that really works for video, it doesn't work so much for images. And again, you don't leave that to the AI to say, okay, now we never ever make stuggler videos again. You go back, you say, what videos are the stuggler videos? What's going on in them? What's similar in the other videos that are working that we can take and use? And I say, go back and do that. Go back and ask, "Claude, man, it's just like, "it's the cycle that you go through and the insights "that you'll find are amazing." So moving forward, I could talk about that all day because it is just kind of like, stalling for the click here. That might be nothing, did I turn it off? Oh, anyway, oh, I stepped on the punch line there. So, if you're, I know. So anyway, the joke has been ruined. This is you all sitting there, a lot of you, I'm sure. Some of you are like me, you're like the golden retriever, eternal optimist, let's go do it right now. I'm gonna piece out and go figure out this report. Some of you are sitting there and you're saying, I could never do this. This could never be something that works for us. I'm not tech savvy enough. I don't know the creative frameworks. Punch line, beat says, boom. Melvin, you're gonna struggle, man. It's gonna be brutal for you if this doesn't work. Stop it. Stop that philosophy right now because we have good news for you. The good news is that first, I'm in the process of building it out. This is what I'm doing. This is what takes up most of my mental energy. This is what takes up most of my time. This is the work that I'm doing. And when it becomes a process that's shareable, you all will be the first to have it. Do you people in this room specifically? I'm speaking to clients, I'm speaking to mentor table. You all will have access to this. Secondly, your sheet doesn't have to be that complicated. You don't have to start with all the like, mind blowing nerdy stuff that I was talking about. You can start with something as simple as style and motivator or style and product, style and skew. You can do whatever is important to you or whatever you've seen from your experience has a strong correlation with performance in your data. So this is completely customizable. I'm here to tell you more of a framework for how to think than what to actually do yourself. So you can do this yourself. And lastly, the real IP, the real value that I'm sharing with you here is not been as so cool he made this. automation that took way too long and probably breaks every time that you want to change anything. The cool thing is that you have this list of prompts and this creative strategy knowledge that we've put together for you and that I'm sure you've been learning over your time advertising to take and fit into an AI flow that works for anybody. So to use the Avengers thing, I've never even seen the movies. You have the Infinity Stones and now go snap your fingers or whatever the thing no sky does, right? So go figure that out and use take this stuff and use it yourself. So now we're going to move on to we've opened the black box, our minds have been blown, we have so many ideas about what to produce when it comes to our creative production, where do we get from here? Melvin is going to come up and let us know how do we actually act on all these ideas? How do we take this from? Ideal land, which is where I have you know prime real estate and ideal land, that's where I live, I pay a really big mortgage for that. But, um, to interact. Yes, can you hear me? I got you. The clicker? I'll take your witness. All right, thank you so thank you all so much for giving us your time, giving me your time to share a little bit of what I know here about the creative process. Thanks so much, Ezra Mali, John and Pepp who's not here as well and of course, Fearless Leader Ben. So let's oh wow, we're at the very end. Well that was quick. Let's start building. Awesome, so here's our goal. Since we get to pull them. I got you. I'm going to work over my head. Okay, perfect. All right, to establish a system that gives us good creative and also strong feedback. Now, how do we do this? We've already addressed the problem. We're going to take everything that binge is shed and we're going to build on top of it because the creative landscape can somehow, oh my goodness, wow, I'm having issues here. Okay, the creative let's say one more time. So, one more time? So, we'll work it that way. So the creative landscape can sometimes feel like this. You know, it's a big chaotic random creatives, you know, inconsistent messaging. All of these things, you know, go ahead and click that next one, but if we were to take a step back, really, the creative aspect, the creative landscape is really this, the different styles, the different angles, the different hooks, all of the things that make up our creative. And when we look at it from this perspective, go ahead and hit that next one. It looks more like this. And this feels a bit chaotic, but it's creative diversification. And this is exactly what we're after. So now I'm going to talk with you all about how to build a system that allows us to be able to leverage all of these options and be able to input them into our creative. So, go ahead and hit that next one. And we should do that by first starting with our creative menu. Now, what you should think of this kind of like a buffet. I'm from Florida, while I live in Florida, I should say. And I like the crews. And anytime I go to the buffet bar, I come there with my plate and I try to get everything, everything that looks good. You know, I want a little bit of this, a little bit of that, sample that because that looks weird, but I'm gonna try. Yeah. And so that's kind of what we want to think about with our creative menu. We want to take this opportunity to define the types of creatives that we want to produce, we want to try and then flush things out here so that our team has a bit of a roadmap of how to move forward. Now, we really want to take our time here because we're actually building our, what we would call our engine for our entire creative machine. Go ahead and hit that next one. So this is what that will look like. So as you can see here, we've defined or we identified one type of create. This is going to be an Instagram story model. We're going to label it in our first column of what this is. Then we're going to provide some of the example, type of ideal, social proof that we want our creative to, or our team to be able to kind of know the idea and have the direction of where they want to go. Take the time to then, take the time to then identify certain key things that we've identified that are successful for this AS style. So things like buzzwords, you know, creative phrases, things like that, that we've identified to provide a bit of guidance for our team. We then want to also share when's the best time to use creative style like this. We want to take our messaging and marry it with the proper style to then present to our audience. Next, we then, well, thank you. Next, we then want to think about formatting as well. As we kind of share with you briefly, we have the entire list that wealth of creative options that we can leverage. So we want to be able to kind of think of what styles can this translate to in order to present to our audience. And then lastly, of course, distinct our variations, the different styles that we've already tried. So this is what we really want to spend our time developing, flushing out the details because ultimately this will be our engine for our entire pay-to-ass machine going. So now we talked about the creative, the creative on menu. Now let's talk a bit about content creators, you know, understanding that this is one aspect out of our entire list of creative options. But if we're going to leverage that, we should have some type of system in place of how to do so. So there are a couple of options that we can go through with our creative, our content creator system. The goal here is that we are to identify the best one that fits us. There's no wrong answer here. All have their own unique benefits. We just want to identify the one that works best for us for our system and our brand. So here we have, we have TikTok Shop, Creative Platforms, Direct Hires, and also Brand Embarrassed. I'm not going to spend too much time on TikTok Shop because Dilly-Hon, it has an amazing presentation. She's going to share it later. So in a nutshell, TikTok Shop product and exchange, free product and exchange for content. Sender creator, a product, they create the content, and then they post on their page, earn a commission. I like this idea. I like this particular option mainly because it allows, it incentivizes the creative to make good content. The option for the better it performs, the more they have, the opportunity to make and their commission. There are a few challenges that I've listed here that we want to be aware of. But I will say this and you know, Ben, correct me if I'm wrong. What we've seen is that these type of videos tend to have to be most authentic and to be most native for the platform. So it's a really good option. It's really good options that you're looking for. UGC creator platform. So think of this as your one stop shop for all things creators. You can hire, find hire, pay creators, and then of course receive content. There are a few challenges there where you want more so on the technical side of things. But if you're looking to scale and we're talking about scaling content and doing it quickly, this could be a really good option for you. Next we have the direct hires. This is the option that we tend to use the most here at Smart Marketer because it gives us the most creative control. It is a bit time consuming which can be a challenge but for the level of creative control that you have, we get to dictate the terms. We get to dictate what transpires. But for the level of time that it does take to actually do that, it can be beneficial or we'll find, we find that most beneficial I should say on our team and it can be beneficial for you as well. So something to think about and then brand ambassadors. This is what we have a dedicated group of creators who create content for us on a monthly basis. And this is where we, it allows the good thing here. It allows them to actually learn your brand while also providing you with a consistent flow of content on a monthly basis. The challenge here, the biggest challenge that I found is that when we talk about creative diversity, it can limit your brand's exposure to different diverse creative styles. But again, as I said before, it allows us to have a consistent flow of content and all of these options. We just need to identify which one best suits you for where you are. Go ahead and hit the next one. Now, thank you. I know you got, I'm a little bit nervous, but I'm working my way through it. So now that we talk about the creative process, right, you know, these are a couple of things we want to keep in mind when we're dealing with creators when we're working directly with them. First things first, creative tears. We want to keep it in mind because this typically impacts how they expect to be paid. We always recommend to stay within the nano space, which is one to one to 10k followers, you know, just because it gives you the option of more control. So just something to think about there. When we talk about compensation, what do we actually pay them? That is a big question. The better question would be what do we actually want from them? That allows us the kind to dictate where our pay expectation should be or where their pay expectations should be. I always suggest including the product value, your product value, and your compensation offer. For example, here we have your paying them $300, you have a $200 product, that's a $500 package. Explain that to them. We've had the opportunity to work with a lot of career creators in all of these tears by simply just taking the time to explain, hey, we're giving you $400 worth of product value. We're paying you $200. You must $600 offer. That's pretty fair. We're also providing you with the concept. We're also doing all the editing. This is a pretty fair offer. Most of the time, the creators usually understand and you get a chance to work with them and they actually continue wanting to work with you after that project. So just take the time, explain that to them. Again, a lot of them don't know what they don't know and when we shape their perspective, it also makes for a great work of relationship. And lastly, contract one more. That We go lastly, contracts. So here, I always suggest that when you develop in your contracts, have a conversation with your legal representative, your lawyer, just to ensure you're including everything that you need within your contracts for your brand. But these are a few things that we've listed in ours just to properly set expectations and outline the goal of the creative project that we're working over with our creative partners. Thank you, by the way, Paul. Now, we have our content, whether it's outsourced or created in-house. This is all of what we're managing. There's so many moving pieces when it comes to all of the creators and tracking our content and what we're having with talking dead lives, we're talking what works, what's not working, what's in revisions, what needs to go like, all of these things we have to keep in mind when we are-- we're in the mix, we're in the weeds of our creative process. So having a strong tracking system allows us to have a central location to be able to know where things are and the status that each piece of creative is actually in and in the given time, which brings us back to our goal. To a step, you go ahead. It brings us back to our overall goal, to establish a strong system that provides us good creative and strong feedback. We use that feedback to inform our next creative decision. And that keeps our engine going and our entire creative system moving our title, paid-ass machine moving. You will hit that last one. So now, Ben and I have shared all of this information. What do we do? Well, if I were starting right now, the first thing I want to do is identify where I am on the creative spectrum. Am I experimenting? Am I testing different ideas, trying to figure out what works? Am I doubling down? I'm scaling, doubling down on the winners that I already have, or my systemizing, documenting workflows, role maps that allow us to streamline the process. Wherever you are on the creative spectrum is completely fine. The goal here is to identify where we are so that we can implement a structure that best suits you wherever you are. SquarePuff. And lastly, let's start building. Here, we provided a few downloadable templates that you can use to help jumpstart your process, where you are on your creative journey. Our goal here today is to provide you with some tools that you can use to help you build your system and continue doing the amazing things that you're already doing. Good. Yeah, that's huge, man. Give it up for Melvin, big help. That's really huge. And this stuff is so important because you and I know, it's like we can have the best system for analytics and diagnostics like I was talking about. And then the rubber meets the road and the content doesn't come through. If this stuff breaks down, then you don't have-- Absolutely. --what you just spent invested all this time and money and people man hours into. And if you get-- these are just like-- this is why it's so valuable. It's like, Melvin has so-- not just with us. Melvin-- this is his first time at SmartMarker. Like this isn't his first rodeo in terms of presenting and being an expert. Melvin is a production expert. And so if we didn't have him on the team, it's like we'd be going down this rabbit hole where I'm like, frokin' through the flowers and bouncing around. And then like, you know, we don't end up having any content and getting sued because we said that we could only use it for like three days and we ended up paying the guy like $1,000, right? Like that's the type of person this mindset and the information that he's sharing is like, if you don't have this, you won't have the scale and the growth that you want out of the whole thing. So it's just incredibly valuable information. So thanks for-- Thank you. Thanks for 10. [MUSIC PLAYING] Thank you so much for listening to the SmartMarker podcast. If you enjoyed what we shared today, please help us by subscribing to the show. Not only will you be notified when we have new content, it also helps get the show out to many more people. We appreciate you and thank you so much for listening. [MUSIC PLAYING]

Podcast Summary

Key Points:

  1. The podcast episode features a live session from the Mentor Table event, focusing on building a "paid ads machine" for creative planning, production, and analysis in the age of Meta's Andromeda algorithm.
  2. Andromeda is perceived as scary because it knows more about ads than advertisers, with two halves: what it sees in the creative (e.g., pacing, emotion, imagery) and what it knows about users (e.g., behavior, context).
  3. AI can level the playing field by providing a microscope to analyze creative elements, but it requires human guidance through prompts that act as lessons, teaching AI frameworks like emotional and functional analysis.
  4. The process involves three steps
  5. The value lies in the principles and lessons within the prompts, not the technical automation, making it accessible to anyone.

Summary:

This podcast episode, featuring Ben Bennett and Melvin Cotton, presents the "paid ads machine" for navigating creative strategy in the era of Meta's Andromeda algorithm. , subway browsing habits), creating a black box that makes advertisers feel powerless. To counter this, the speakers emphasize that while the user data side is inaccessible, the creative side is fully analyzable.

AI, like a lab assistant, can be used to examine creative elements through a metaphorical microscope, but it requires human expertise to guide it. This guidance comes via prompts that teach AI specific frameworks, such as emotional or functional analysis, turning raw creative into actionable insights. The process is straightforward: feed creative into an AI system, apply a prompt (the lesson), and receive output.

While automation can streamline this, the core value is in the prompt design and the principles behind it, not the technical setup. This approach allows marketers to understand their ads at a deeper level, reducing reliance on sheer creative volume and instead focusing on strategic, informed production. The session concludes by emphasizing that this system compounds over time, enabling continuous learning and improvement in ad performance.

FAQs

The paid ads machine is a system for creative planning, production, and analysis in advertising. It has three parts: diagnostics (analyzing existing ad account creative), production (creating new ads based on insights), and a compounding process that repeats learnings.

Andromeda is scary because it knows more about ads than advertisers do, analyzing creative elements like pacing and emotion, plus user behaviors. This creates a perception of a 'black box' that controls ad performance.

AI allows advertisers to analyze their own creatives at a deep level, similar to what Andromeda sees. It provides a 'microscope' to evaluate ads, digesting data and revealing insights about creative elements that were previously opaque.

Prompts act as lessons that guide AI, teaching it frameworks like emotional and functional analysis. They include context so AI can properly evaluate creative, making the system effective even without automation.

The first step is preparing the creative for AI analysis, like putting a specimen under a microscope. This involves using a prompt (e.g., for Gemini) to convert photos or videos into a readable format for the AI system.

No, automation is not required. The value lies in the principles and lessons within the prompts, not the technical setup. A manual copy-and-paste process can be equally effective.

Chat with AI

Loading...

Pro features

Go deeper with this episode

Unlock creator-grade tools that turn any transcript into show notes and subtitle files.