How He Predicts Winning Ads Before Spending $1 (AI Workflow Breakdown)

0:00All right, I've got my friend Justin Brook here today. He is a media buying master. He's leveraging AI and a lot of cool ways to build different automated workflows and help people scale their businesses. Wanted to have him hop on and do a demo for us today and uh let us uh in on kind of what he's seeing in this automation space in the in the ads world. So, welcome Justin. Ah, James,

0:24great to be on man. Thank you so much. Awesome to have you. Cool. So, uh, yeah. What have you been, uh, what have you been cooking up lately? Well, I've been working on this idea. I first saw it on X., you know, just some guy was like, "Hey, you could predict winners before they start." And, and we'll talk about him here in a minute. You know, I like to give credit where credit is due. Um, but you never know what you're going to find on X. You know, there's a lot of crazy stuff on there. And I was like, man, this is crazy. So, I kind of

0:49followed the rabbit trail a little bit and watched some of his videos and I was like, "Oh, this guy's really cooking. He's got something here." And so, um, let me just break it down. I'm gonna just share my screen. It's gonna be easier to kind of show you than just tell you. So, perfect. Let me share, uh, let me just get my screen on the right one first. So, we want to go here. All right. Yeah. On X, you know, when you when you browse the feed, there's all these like workflows and people doing these giveaways. And I think one of the hard parts right now is like you got to

1:22separate the signal from the noise out there. you know what I mean? A lot of noise. So, uh yeah, that's why I was excited to to talk with you because I know you have a lot of experience in this space and uh your workflow is going to actually have like a real quality uh output. So, yeah, let's see what you got. Yeah. So, the idea here is to be able to predict winning ads before you spend any money. And if you could do that, it's like the way that I've been positioning it is like if you could predict the lottery numbers Mhm. Would you? I mean, who wouldn't, right? It's

1:55like, well, okay, so that might be illegal or I don't even know where we go there, right? But like what's completely legal is you could predict winning ads before they launch. And that's I don't know. That's pretty close to predicting the lottery numbers, you know? Yeah. Um, you know, if you can predict something that's going to scale, you know, what's it going to scale to? 10,000, 100,000, a million? You know, who knows, right? So, it's it starts with um there's three pieces of this and I'm going to walk through like I'm going to show you guys like the super super easy basic kindergarten version of it. Then we're

2:27going to walk up to the middle school version of it. Then I'm going to show you like the college level version of it, you know, um where people should be going. So, all right, you you start with a customer simulator, right? You need to get feedback. And this is the same thing people have been doing for hundreds of years in marketing. They would have focus groups. You do customer development calls. Um you you know you do surveys to your email list like you have to get feedback from a market and then from there you go and write new copy and then from there you would split

2:59test that copy. Well now with AI we can run that whole process you know ahead of time. And so one of the things, the big things, and I know your audience is going to remember this, but just for anybody that's like a super noob or anything like that, one of the things they have to understand about AI and these LLMs is they were designed to pretend to be somebody. They were designed to to figure out the next words in the sentence, you know, and so they're really good at pretending to be something. So you can have it pretend to be your customer. You can have it. I

3:33mean, that's one of the basic props we always do, right? You know, you are a world-class copywriter. You have 10 years experience. Like, that's what we're always doing. That's the So, this is a project in chatbt and I'm just using chatbt because I know most people have that, but you can build this in perplexity and abacus and any of the different programs. You're going to

3:52start with uh an instruction. Okay? So, you're going to have to put your customers into here. These are just my customer. This is leg legitimately my tool here that I'm using. So, pretend you are a struggling digital ad freelancer. That's my target audience. You are now part of a panel of prospects reviewing ads as a focus group. Your job is to give your opinion about how the ads make you feel and what they could do better. Right? So, we're just building uh a little guy that's, you know, sitting by our side every time I I'm using this for email subject lines. I started using this for my LinkedIn post

4:27and all all of a sudden my engagement's way up on my LinkedIn post because it's like I write something and then the guy next to me says, "Hey, that sucks." And it's like I don't have to find out after posting it that it sucked. I can find out before posting it and then I can tweak it. So, all right, let's dive a little bit deeper. You have two main problems and this is my customer problems. You have to know your customer problem. You have two main problems as a digital advertiser. need more. Oh, also I make sure you put but not limited to these two, right? Because this is the

4:57priority, but I also want it to be like, hey, remember I also have this problem. So they need more clients. They need more digital ad strategies, right? Yep. Then when I give you ads to look at, I want you to give me your raw personal feelings about that ad. Does it relate to you? Does it address your needs? What about it appeals to you? What about it turns you off? What would you wish the ad said? How could you can put all your questions in there. All right, last part here. Do not use generic information about copywriting or advertising. Stick to your persona and critique the ads

5:31based on your own desires, challenges, fears, frustrations, and goals because I don't the LLM will default to like best practices of advertising, right? I want it to stay in character. Yep. I like this a lot because you know what's funny is most people don't even use projects within Chat GPT. So like just having this as like this kind of available like project uh hub uh for for simulating customer responses is is is a

6:03timesaver alone in and of itself I feel. Yeah. Throw your email subject line in there before you send an email and get feedback. You know it's like I'm doing it for all kinds of things. you know, I was going to make a a new ad on Facebook. I dropped my ad in there and it was like, actually, I wish it said this. So, I changed that and and bam, it worked. Awesome. So, the the other part about projects is project files so that you can give it more information. So, I wanted to train it better. And so, I just used like a manis or whatever deep

6:31research to do to create a customer persona for my audience. I've got the ability. I've been doing this for 10 years in ad skills. So, I added a lot of survey data and and stuff like that. So, it now knows because of this PDF who my client is very, very well. So, between the project file and the prompt, it knows how to act. Yeah. So, let me pull up one of these and uh see some sometimes it's pretty brutal. So, I'm not going to give one of the ones where

7:00it just rips me a whole new one here. But you need to have thick skin when asking the AI for feedback. Absolutely. Um, all right. So, here's here's, you know, an example. I gave it um an ad and then it gave me so some impression. I immediately felt intimidated and a little turned off. Dude standing there jacked in a tight white tank top next to a two comma club award like he's trying to flex on me in more ways than one. It gives me alpha bro energy. I'm not saying he's wrong for showing proof, but

7:29right now I'm exhausted by juggling. He's just saying like, "Hey, it's cool that you're like, you know, next to your award and you're in a thing and you're at the gym and all that stuff, but like that's not addressing my actual needs." Right. Right. Um, and so what works, what appeals to me, the numbers. I like the dollar into $4.73. I like the 31% cold traffic conversion. I like the 63% drop in CAC. And we know our market like we know that those things would absolutely they love the specific numbers. Packaged everything into a blueprint. Fitness enthusiasm. So it said here's everything that's wrong with

8:05your ad. Here's the parts that I like. Here's what I wish you would say. If all it did was that that's valuable. Absolutely. But then there's more. Good. So what we do is we take that feedback and now we need to create new variations. And so I made a copywriter project, right? And so like I'm saying guys, this is the kindergarten version that anybody can build like really quickly. And so just a standard you're a copywriter prompt in here, your assignment. I will give you an ad we ran. I'll give you the feedback

8:38prospects gave us about that ad. Consider the feedback and the original ad. Then use your vast experience with copyrightiting to write a new to write three new. And if you want 10 new, 27 new, you know, whatever. Uh, just remember context windows. Um, three new optimized versions of the ad. Make sure when you write the ad copy, use my voice. I've got chatd already trained on my voice and custom instructions and all that. Yep. Um, never format your responses with dividers. I hate that

9:08chat does that with emojis and dividers. Like I just said, format it like an internal team email giving your brief insights plus the rewrit rewritten versions. This is so cool because it's literally like sending off feedback to a copywriter that's on your team and he just emails you back. I mean, it feels very real. So, uh, we drop the copy in in here, the feedback in here, and then the copywriter is going to take the feedback in the original ad and he's going to write three new variations. So, I mean, here's an example of it. Version one, version two, version three, you know, yada yada. All right. Yep. But

9:44then there's another. Now again, if all it did was this part, right, feedback and then write us new variations, that alone is a service that people could be setting up for companies because what that allows you to do is Cindy out in the front office can now write copy because Cindy's going to hack something together that's horrible. It's going to go to the customer feedback. Customers going to be like, "Change this, change that, change that." The copywriter is going to go write a good version of that. So now Cindy out front, all love to Cindy. Love you. you know, but now anybody in the company can be a

10:16copywriter. Yep. Yep. So, that's crazy. And like I don't know like just the power that these systems and these tools give, you know, one individual uh is is insane. Like they can write copy, they can analyze data, they can do all kinds of stuff. Like it's uh it's it's pretty incredible the leverage that you can

10:38gain from this type of thing. Yeah. Yeah. So last piece, we got to predict a winner. Okay. So what I do is we send this over to a I thought it was a cool name to name it prediction engine. Uh and so what I did is I gave an instructions. You're an expert data analyst. You have decades of experience using data to create predictions. When I submit ad creative to you, I want you to use your entire known knowledge of advertising and copyrightiting best

11:07practices to assign a score to each ad. And so that's important because chatbt already knows what good direct response is. It knows problem agitate solution and headlines and unique mechanism. It knows all that stuff. So using all of its known knowledge. All right. Best practices to assign a score to each ad. This score is called the success probability score and it judges each ad based on the original feedback from the panel of customers. All right. And the known best practices to create a probability. So, it's going to look at the feedback, the new variations, and all best practice, and it's going to start late, you know,

11:47assigning a score. Don't give me a long explanation. Just give me the success probability score for each of the ads. Score them zero through 100. And then I kind of tell it like, you know, a scoring chart right here. Yep. So then what it does, you know, I'll jump into one of these. And so it says, all right, version one, success probability score 91%. The reason why was this, that, and the

12:13other. Version two, 85, this one's 68%. And then I don't know why it didn't give me the variation on this one. Maybe the other one did, but essentially what it's supposed to do is it's supposed to tell you based on all known best practices, based on the feedback, this variation should be the one that's going to win. Yeah. And I mean to do that manually is impossible. you know, if you were just trying to do that prior to AI, like there's just no way to do that. Like to go out there and do all of that research and references and retain all of that knowledge. Like that's

12:47that's crazy, right? And that would that is quite literally like what an ad agency would do. You know, they're going to take your crappy ads that you've written. They're going to go write you better ads. They're going to go and split test them, find the best winners. And that's why people hire ad agencies for years. Well, this allows anybody to kind of install an ad agency into their business and they can predict which one's going to have the highest

13:13likelihood of winning going forward. Now, I've been doing ads for 20 years and this thing's at least as smart as me cuz every time and I've been like, I think that one's the best one, it also ends up choosing that one as the best one. Now then you have to run it and see how well it actually converts and then you'll tweak your prompts and stuff like that, you know, but for the most part, this thing's really good at coming up with new variations, telling you which one's the best one. Run with that. Any questions or anything about that before I go forward into like the middle school

13:45version of this? No, this this is great. I like how you're using uh chat GPT projects here. I think this is a very underutilized feature. Uh, I like the approachability. Like anyone can set this up. It does you don't have to know how to code or anything like that. You don't even know have to how to how to build a complicated workflow or anything. Like you can get value quickly uh following this process. I like how you know you have uh you're you're arming it with context, you know, uh your brand voice and and things like that. So you don't have to go back and prompt all those things from a fresh

14:20screen. it's already able to reference that stuff and uh and get to where you want to go quickly, which uh which is great. Right. Right. So, as you know, we've been training ad agencies for years and we're kind of telling them like, I'm sorry, but the stuff we've taught you for the last 10 years has changed. The market has changed. It's not our fault. It's not like it went out of date or anything. Just the market has moved. And so, we're trying to get

14:44people ahead of the ahead of the curve. And this is what we're showing them how to do is like the it's no longer really you're getting paid to run ads for people. You're getting paid to install these kind of things on the client's chat GPTs because now they have the power. You know, they don't want to go watch all the YouTubes and learn how to write all the different prompts and stuff, but they'll grant you $3,000 $5,000 to come build this for my business. That way Cindy out in the front lobby can now run all of our

15:12Facebook ads for us. Absolutely. Absolutely. like it it's about empowering and building these systems for for clients. That's like one of the biggest opportunities out there right now in the in the service and agency world. Like I I've seen it on uh on I I own a couple agencies. I I've seen that as well there. Like we've sort of pivoted from, you know, doing delivery to just installing and empowering these systems for clients. And that does a couple things. one, like if you're an agency owner or a freelancer or whatever, that removes some of that like performance pressure off of you in a sense, like you're empowering them to

15:48handle it and and that's cool. They're not messaging you like what's going on with this particular ad or whatever. It also elevates you, I think, like it it takes you from uh being seen as kind of like a task doer into a strategic consultant or a strategic imple, which means you can charge more money. you can attract better clients. Uh, and uh, you're valued more, which, you know, I started out as a freelancer back in the day. Uh, and that's probably when I found your newsletter, uh, a long time ago, probably 2013 or something. And I just remember like I hated how I was like talked to and and dealing with

16:24those clients and stuff. And now when you learn these systems, you're the expert in the room and it's a completely different ballgame. Yeah. and and people are willing to pay a lot more upfront because this isn't really a I mean you could attach retainers and do this monthly but it's kind of like an installation business you know and so it's easier to get a 5k 8k 10 15 um depends on what size company you're installing this for but like installing something like this is easier to get those bigger payments up front because it's it's a oneanddone it's not a monthly retainer. Yeah. and you know the

17:00the sort of like offering uh stack for you know uh digital entrepreneurs you know you have kind of like hey I'm going to sell you a course or a membership you're going to learn how to do it yourself then you have the done with you sort of like middle area and then the top is kind of like I'm going to do everything for you well now you can do like you can install it and then you can easily get to those other layers like once you install for clients then maybe you have like a group coaching program or something like that uh that's going to add value to a lot more people, uh,

17:29you're going to have way more leverage. You're going to be able to make more money. Your margins are going to be higher as well. Absolutely. All right. So, can we go on to the middle school version of this? Let's do it. Let's graduate. All right. So, now we're going to use I like Mind Studio. There's Make, there's N8N, there's, you know, Gum Loop. Like, there's a million tools, you know, today. Choose whichever one you are going to like using. I like using Mine Studio, so that's the one that I use. But you could do the exact same thing in any of the ones. Real quick, um, Mind Studio, one thing that I like,

18:02I'm not paid by them or anything like that, is they already have a bunch of bots made. And then when you click into any of these, you can click copy and it copies it into your account and you can start customizing that bot. That's how I learned in the beginning to start creating the beginnings of workflows and such, you know, like this one, turning content into a LinkedIn post and then you can add another, you know, then turn it into a Twitter and then turn it into an email so you can keep going. So, I have built that version that we just showed you in chatbt. I've built it and

18:36I call it advertiser Alex, you know, and so I can just run this thing at any time. Now, let me go into my builds and show you what Advertiser Alex looks like. All right. It's a cool cool platform. Like before we connected, I I didn't even really know of it, but the it seems

18:56like very direct. I like the simplicity. So, I'm going to I'm going to dig into this after after the call as well. Yeah, it's I I I know the founder would hate me to say this and I love your tool, man. I absolutely love your tool. It's kind of like the easy version of N&N, like the training wheels version, you know, it's very easy to get started. So, if you've been looking at some of the other tools and you're like, "This is way over my head." Try Mine Studio. Um,

19:19or maybe he loves that description. Okay, so let's talk about it. So, when you first jump into one of these automations, you're going to have like a start and then a plus sign, right? When you click the plus sign, it allows you to add modules or nodes, and there's tons of them. You can generate text, generate images, charts, text to speech, create a podcast, scrape URLs, um, you could chat, you can fetch YouTube comments, scrape Instagram mentions. I mean, there's a bajillion

19:49things. Throw it into a Google sheet. Like, whatever you want to do. Go find an email, summarize text, create a notion page. Like, there's a lot of things that you can do with these things. All right, but all I want to do is the first thing I want to do is I want to have a user input, right? And so all that is if I if I click into here, it's just a form field, you know. So I

20:09need one that's like a long text field. That way they can enter in a whole email or a whole Facebook ad or a whole blog post. And then I need one that's a an image upload. That way they can upload the Facebook ad or whatever. Okay. So we got to have a way to like dump the ad into the machine. Yep. Now, once it goes once the user uploads the ad, then what it's going to do is it jumps over to uh an LLM and you can choose. One thing I like about them is you can choose a bunch of different LLM. So, if you want

20:40to use Claude 4 or DeepSeek or Gemini or or Llama, whatever you want to use, they've got all of them in here. I like DeepSeek. Um I just like the way that it thinks a little bit better and math type stuff. writing, I'd probably use a Claude, you know, but um and also Deep Seek's pretty cheap. It's like 55 cents per million tokens. So, it's it's uh it's a much cheaper than Claude is at the moment. So, same exact thing that we built in chat GBT, but we just built it now in Mind Studio. We allow it to string it together. So, now I don't have

21:14to like copy paste, open a new window, drop it in there, copy paste that one, drop it in here. It's doing all the copy paste, drop it in there for me. So that's exactly what we built. So same exact prompt, but whenever you each of these steps allows you to assign I don't know if it'll allow me to open this. No, it won't let me allow to open that. But each thing then has a variable. And I don't I barely know what variables are, you know, but it's like when I when this does when it runs the prompt, the prompt the output of that prompt is then called

21:49this variable. Yep. which allows me to on the next step I just want to display the output of this step. So I just put the variable right. So it just takes it and then the variables allow you to carry things forward. Yep. Yep. So in this one Yeah. We have a variable called new ad. We have a variable called new image. We drop those into the prompt and then we have it reference our customer research which this is another cool thing about Mind Studio is it allows you

22:22to go and put in data sources in there. You can put up to 150 text documents in there or up to five million words. And I find when I've used this, it's better to have a lot of little files than it is one big file. these tools, they're able to read a big file. It's just harder for them, you know. So, in here, we put our customer avatar, you know, so it references that. So, if we go back, so you can see here it's referencing our avatar and then we're saying like here

22:55is the new ad or the new image. It does the feedback. It displays it. And now I don't have to make it display. I can have it just run the system and then give me a score at the end. Yep. I want to play along. I I find value. I learn in the feedback part. So when it's done there, I want it to show to me. I read the feedback. That's cool. Then I

23:17hit the next button on my interface. And now it's going to run the copywriter. The copywriter is going to do the three variations and it's going to send me an email. You know, the variable is called add variations. We display the content. So the same thing like it's a prompt then we would display the output then it's a prompt then we display the output then it's a prompt and then we have the finish you know so now we're displaying here's the three different variations and then finally it goes to the prediction and uh and one thing you might want to do is so so it's doing the prediction

23:53based on the customer sim fe feedback right we're pulling that variable along then we're pulling the ad variations along and then at the end here the output put behavior is displayed to user and that's where it goes to the endpoint. What do you want to anything you want to say about this before I go on? You know, I I like this because this platform seems so intuitive like uh you know the the output behavior. You're just letting it know how you want the output. Uh do you want it in a text format or or whatever? And uh that seems pretty pretty straightforward. I I really also like these model choices

24:29that you can implement into here. Deepseek or chatgpt or whatever like in in N8 like what I do and maybe this isn't the the the right way now but um I use open router uh which allows me to select which model I want to use for a specific task but that's another tool and another subscription and another node I have to go figure out and figure out how to configure. So, this uh this actually seems like uh it can accomplish a lot in fewer steps. Uh which, you

24:59know, is great. I'm all about that. One other tool that I want to show people is in the automation that you I think it's always by default called main flow. Um, and then right next to automation. So, as you're setting up your thing, by the way, you can add more modules up and down and left and right, and you can make whatever crazy thing you want to invent, you can you can do all of that, and then you just connect them. You can see here there's like a little dot on each side. So, you can kind of draw arrows to each thing. But

25:30the other thing is is a system prompt. And so, each step has a prompt, but then there's also a global prompt. And so that's where you would tweak it, you know, it's like, hey, never add emojis to it or always make sure to do this thing when you when you run these things, you know, and so you can as you run this thing, add to the system prompt so that it's it's working better as you go forward, you know. Interesting. Maybe you have a brand guide or whatever that that looks like. I I guess uh which is

26:01really really cool. I like that. Yeah. And then it lets you get super nerdy. I mean, there's functions and like if you if you're a coder and you know all the different features in here, like that's as far as I go. That's that's the top of my my pay grade. Um, but it will allow you to get deeper into this thing. You can run code, you can pass variables, you can do all kinds of crazy stuff. So, I think the the value kind of the the college level of this, right, is building these. You know, you I've built it. Now that I have it built, it's

26:34really easy because I can just go make a copy, tweak a couple of the prompts for another company and go install this for them. And now they've got this this tool. Now, one of the cool things about let's see where is uh let's see. So, here's all my published ones. And then it allows you so I could just quickly duplicate this. But Mine Studio allows you to embed the bot. So you can embed it in a members area if you build a cool tool and you want to say you can you can connect like a stripe PayPal and then you can just sell access to the bot. Y

27:08you know you can uh you know set up another customer's mind studio account and then you can just pass this thing on you know duplicate it into their account you know and so really easy to kind of turn this into like an agency where you're building this for other people. that could be a great lead magnet as well. Like uh I could envision like you know running traffic to an agent or whatever that's embedded on a landing page. Maybe it's a a snippet or a

27:36smaller version or something like that. They're able to run it, get value, and then you have a call to action afterwards and they're on your email list. I mean that like you know a lot of lead magnets I think have lost their perceived value over the past couple years. Like anyone can make an ebook, you know what I mean? anyone anyone can make an infographic or whatever that looks like, but this like I think people would really feel a lot of that like that initial value delivery and be really interested in whatever the next thing is. I I'm a thousand% with you. I think lead magnets have moved on from

28:11like ebooks and stuff. Prompts are one of the greatest lead magnets I'm using right now. And then you're right, like a pre-built agent that they can just go install is a hugely valuable uh lead magnet. And so one other thing I want to show you is that, you know, with this tool, sorry guys, I'm a little sick here. I'm trying to hold it. Hang in there. But um one of the things that this tool can do is you can see the history and you can turn that on and off. You know, some things are more private, chat bots and things like that, but it allows me to

28:43see every usage. That way I can diagnose. I can go, "Oh, okay. I see where it gave something wrong here. I'm going to change the prompt." And so, as you let your clients or your friends use this thing, you can come back in here and see all the things that they ran and then tweak it from there. Yep. Yep. Optimize. I mean, I I often view this as also like a continual process, you know, like once you build an agent or a workflow, there's always ways you can improve it and make it better and and and things like that. So, uh, this this history seems like a nice, uh, way to

29:16kind of get that information and and get that data. Yeah. Yeah. It takes a little while of time. That's why I didn't run it live for you guys. I mean, it is it's going to thing and it's doing a lot of thinking and sometimes it takes two minutes. Does have a very cool little animation there while you wait. But yeah, so this would be the outcome and you can design it to make it prettier or whatever if you want to, but I don't need it. I just need it to assign a score and then display the the winner for me. So, yeah. Yeah. Interesting. I think this is super valuable. Um, you

29:48know, if I was running ads right now, I would definitely want to know this ahead of time because you can waste a lot of time and money like guessing or running all kinds of creative variations and, you know, blow a bunch of money or or whatever. So, uh, you know, being able to kind of shortcut that process make it a lot quicker. Uh, and the fact that anyone can do it, anyone can set something like this up and uh, and get that data into their existing sort of way of working, like it's it's pretty amazing. Yeah. Yeah. It's it's phenomenal. I'm I'm geeking out with all

30:19the toys that are available now. Like I feel like I'm just playing these days. I'm not actually working. What I would tell somebody here is we had a guest on one of our workshops uh, named Raj, who is the guy that I learned this from. um at Raj A R A G R AJ J Ha on on Twitter. Check him out. He's amazing. Brilliant guy. And he made a really good point. Sorry, my microphone's falling. He made a really good point in saying that like if you just have one customer persona, then you could actually steer in the wrong direction. You're going to steer in the right direction of just one

30:57specific customer. But when we run our Facebook ads, we don't get to choose only that one person. Like if right, you know, for me, you know, targeting people who've read the book Millionaire Fast Lane or Rich Dad Poor Dad works really well, right? Well, when I target somebody who's read Millionaire Fast Lane, well, there's a 45year-old male, you know, and there's a 33y old female, you know, at the same time, right? and she's she's selling courses and he's selling services, you know, and so what he has done is he's built a whole company out of this where he instead of, you know, your panel is one customer,

31:35you know, it's legitimately 500 a thousand personas. So you get a little bit better market feedback back. And so like there are big boy college versions of this and he's he's going after big businesses and selling what we're talking about here at a low level. He's selling this at a high level to enterprise companies. Very interesting. Now like just to uh chat about Facebook ads for for a sec, you know, like you've been in it for a long time. I've been into Facebook ads for a long time. Like you know, remember the old days of Power Editor and and all that. Um, you know, back in the day it was all about like

32:14interest targeting, you know, and then we went through this thing where it was like, hey, just go broad and let the the algorithm do whatever. Where are we at today? Is interest targeting uh back. Is that valuable? I've seen some people talking about it. I haven't run ads for a bit. So, you know, where where do you

32:31stand on that right now? Yeah, pretty much the whole industry is moving towards this black box because I mean there's guys like us that are, you know, good enough to be dangerous, you know, when we're in there, but the average user doesn't know the difference between direct response and branding. Doesn't know what a problem agitate solution is, you know, and so they're trying, you know, it feels like they're taking away from us, you know, who are good. They're giving us, you know, they're taking away our levers and tools, but what they're trying to do is they're trying to make

33:01it easier for the their average user. And so their average user doesn't know how to split test, doesn't know how to write good copy. And so they're constantly taking away these features like interest targeting, keyword lookups, and you know, they're just making it more and more of a black box because they're like, "Look, we've hired billions of dollars of the world's smartest algorithm engineers and AIs and stuff. Just let us do it. just give us

33:25your copy and let us do it. Legitimately, I would say that I would say five years ago, I was like, "No way. Do not let their systems take over for you." Today, we're more in the realm of like, you know what, they're they're they're writer than they're wronger is the weird phrase we use. Um, they're right like 80 to 90% of the time, which means they're still going to be wrong, but they're going to be right more often than they're wrong. And so ultimately, you will optimize to be better. And so it is, you know, right now it's kind of moving more and more away from like

33:58specific niche targeting, which sucks for an OG like me. I want that control. But their AIs are getting really good and they're able to um yeah, just target broad and and and do well. Just you just want to make sure your copy does the targeting because when you go broad, the AI is looking at your copy and is going, "Oh, I know you want to go to these people." Yep. Yep. Uh, and a couple weeks back I I wrote this tweet and it was like, "What's the most uh underrated

34:28marketing channel out there right now?" And I think you had a reply that got like a ton of likes. Like it was like the top reply on there and it said Google Ads. People are I can't believe people are sleeping on Google ads. So you had like 150 comments and not a single one said Google ads. I was like

34:45how is this not even an option today? It's still like there's a lot of great advertising. Tik Tok ads, Reddit ads, Amazon now has ads. If you're selling books, Amazon ads is amazing. But it's like I start with Google ads, you know, right? Those people are searching for exactly what I'm selling. Like, might as well start there, right? Yeah. I sell a book about retargeting and I advertise on the keyword retargeting book. I don't have to be a brilliant copywriter to sell a retargeting book to somebody who

35:12just searched for a retargeting book. Yeah. They already have that intent. They're in the market and you just have to show up. Yeah. Yeah. which is great. So, um, man, this was awesome. Uh, I feel like I learned a bunch from from this. Hey, I'm definitely going to check out Mind Studio. I feel like I'm not using projects in chat GPT enough. Uh, that's awesome. And, uh, I like this predictive idea that that you showed. I think it's really powerful. I think there's a lot of use cases for it and had a lot of fun just kind of jamming on

35:44uh, the the nature of the ad business. So, where can people learn more from you? Like again, like just to preface, uh I think I found your content back in like 2013 or something like that and I was a struggling freelancer back then and you were sharing lessons like from, you know, running all these major campaigns for like, you know, direct response businesses and all that and got a lot of value from from you over the years and uh you know, where can people

36:12learn from you now? Yeah, so I'm on X. Facebook locked locked me out. I didn't do anything wrong. I just I couldn't log in and every time I'd have to like submit my ID and then they give me codes. I was like, I don't like you that much, you know? So, so I moved over to X and learned how to, you know, you guys talk different on X than Facebook. Thank god. Uh it's it's great. And so I moved over to X. So I'm im Justin Brook. So

36:36the letter I, letter M, Justin Brookke. And uh that's me over on X. And I'm I'm there most of the time, but I've also just gotten on LinkedIn. You just search Justin Brookke on LinkedIn and I'm sharing stuff over there. And then my company is adskills.com. Awesome. Awesome. And you you train people how to leverage all of these systems and workflows and leveraging AI kind of in their in their ad work, right? Yeah, we're showing people how to use AI in the profession of media

37:03buying. Okay, cool. Awesome. Awesome. Well, uh yeah, thanks again, uh Justin. I think people are going to get a lot of value from this and uh let's run it back sometime. I I really enjoyed it. Absolutely. I've got many more workflows, so definitely have me back. Let's do it, man. All right, Justin. Thanks a lot. We'll chat soon.