Transcription
So Claude looked at this clothing brand's Instagram account and cut their cost per purchase by over 30% in 18 days. Not a consultant, not a $5,000 agency. Audit Claude in about four minutes.
And here's the part that stings. The brand was boosting their posts. Please don't ever do that. That was their entire paid Instagram strategy. Boost the real to got the views. Hope it sells product, watch the cost per purchase climb and wonder why I've spent more than ten years in this industry working with over 300 clothing brands in the last 12 months alone.
In the gap between the brands winning on Instagram right now and the ones bleeding money has never been wider, you would think it would be closer because of all of this AI, but honestly, it's because of all this slop. But if this sounds familiar, boosting post inconsistent returns, no real system scaling up and then your cost per purchased absolutely skyrockets. I'm going to show you exactly how to fix it using Claude live today. So stay through the campaign structure section. That's the part that actually determines whether you scale or bleed. And it's actually much simpler than people think, but they overlook it.
All right. So let me show you the end result before we actually build it. So watch the left side. That's Claude, okay. And watch the right side. That's the ads manager. Here's what happens when they are connected on the left. Claude's analyzing the brand's existing organic content, their actual reels, and identifying exactly which one to run as a paid ad and why on the right campaign built from that analysis. Right. This brand I mentioned cut that cost per purchase dramatically, not because we found a magic audience. Really, you shouldn't be using audiences inside of Meta. A little hint later because we gave the algorithm the better inputs. That's what this video is about. Let me show you how to build it.
So most clothing brands running Instagram ads are doing it backwards. They create organic content. Something pops off, they boost it. They wait nothing of the way they hoped. The brands scaling right now are doing something different. They're running ads behind content. Claude, help them build in using Claude to structure the campaign before they ever touch ads manager.
Now here's the three part system. Part one Claude analyzes your best organic content and tells you exactly what to run as an ad and what the hook angle should be. Part two Claude writes the ad copy, the hook stack, and the caption built specifically for Instagram's format. So that hook is really could be a visual hook. It could be the very first, you know, ten or sorry, three to 10s of the video. Part three Claude builds the full campaign brief before you open ads manager. Right. It knows what objective you're going after, which is sales. It knows who to use as the targeting for the content. Right. What type of content are we going to go after? Because the targeting does the content. You do not need to pick some warm audience. Clarity knows that if you've trained it correctly and what to watch really in the first 72 hours.
Now, one brand we work with had a very strong organic content, right? The reels were getting tons of views, but their paid ads weren't converting at all. We ran all three Claude layers and the cost per purchase dropped. Let's start with the foundation part one. Before you run a single Instagram ad, look at what you already have. You've been posting. Some of it is working. The question is, do you actually know why? Not that the reels got views. That's not the answer. Why did it get the views? What specifically in those first three seconds made someone stop? And would that same hook make someone buy? That's the difference between organic reach and paid conversion. Most brands never bridge that gap. They boost the reel with the most views and hope the algorithm figures it out. Claude figures it out specifically and fast.
Okay, so watch the top part of this prompt. This is where I'm going to put in the prompt. I'm using a dictation software that's really great. You can technically do voice inside of Claude. It's good. There's a voice mode. However, I like to use this tool called Whisper Flow. I'm not partnered with them, but if they do watch this, I would love to partner with them. They are great. Here's prompt one. So copy this. I'm going to use Whisper Flow. And actually, before you run, this property is exactly what to give Claude. Pull up the last 90 days of reels. Pick your top three by views for each one. Give Claude three things the view count, the first line you set on camera, and how the comments felt. Did people tag friends, ask questions, or drop fire emojis? Right? That's it. You have to have a little bit of insight here, and you really got to send that data over to Claude so it can build off of it.
If you are not familiar with Claude, I would 100% tell you to get the pro version. It's like $20 USD per month. And then make sure that when you're creating something, you add it to a project. The project allows everything. I think of this as like a folder for like you have a filing cabinet. You put everything inside of that folder that you're talking about with your brand. So for us, we'd want to create a folder specifically around Instagram ads for clothing brands and put everything inside of their inside of your folder. Okay, for this, I'm not doing that right now because this is just for the YouTube video, but for you, create a folder. I mean, if you don't know how to or add a project, right? So click add a project and you can add it into it that way.
Here is the prompt. So you are a direct response marketing strategist for a clothing brand running Instagram ads. Here is what we sell, our price point, and how we describe the brand in three words. Here are our top three organic reels from the last 90 days. For each one, the view count, the first line spoken on camera, and what the comment engagement looked like. So with this prompt one, I'm going to create a. This is more of like a creative analysis based on these three organic pieces. Tell me which one has the strongest paid ad potential and why? What is the specific hook mechanism that made it perform? Rewrite that hook for a three-second paid ad opener and tell me what audience signal this content is already attracting, so we know who the algorithm is already finding. This is just an example. For a YouTube video, you can go out and find me any clothing brand out there. Maybe you can check out Satisfy running, or you can check out Bandit, or really any clothing brand out there and analyze those posts and give me back numbers and data. Again, this is only for a YouTube video, so it doesn't matter a ton as far as the accuracy, because I'm just going off of this as an example to give you guys context as well.
I gave it a fake brand, or actually it's a real brand here. So Bandit running. And then it broke it down here. The three reels, the trailing last 90 days. The run club energy, 22-second clip of a Friday evening run club in Brooklyn. Packed streets, matching kits. Reel two. The founder POV, 38-second talking head from a founder Tim West. Reel three, kit drop revised reveal, 15-second product drop clip. All right, then it's out here. One strongest paid ad potential. Real two wins, but not for the obvious reason. For 112,000 views on a reel one. Looks like the easy pick, but view count alone doesn't sell. You tell you add potential. Look at real two, 180,000, 198,000 views with 1200 comments. That's a comment to view ratio of about 0.6%, roughly three X the platform's average. The level of commitment density signals algorithm is serving this to people who have a strong opinion about the subject matter. Again, that type of information and insight from Claude, very powerful, very and powerful. And more importantly, the hook. We never pay for an ad. Here's why. Creates a pattern interrupt for an exact audience you want to pay to reach. The hook mechanism. The mechanism is credibility and version. We never pay for an ad. The psychological trigger, cognitive dissonance. All right. So the viewer here is we never paid for ads while watching a paid ad. That tension demands resolution and they keep watching. That's great. The secondary mechanism is authority positioning the founder on camera, speaking plainly. No cuts, no music. That format signals confidence. And if you're watching this right now, I mean just yapping type of content where you're just literally talking about a subject for a long period of time is working really well right now because so many people are trying to get away from this AI generated only type of content.
Before the audience signal who the algorithm is already finding. So basically the comment patterns across all three reels, the algorithm has self-sorted the content into three distinct audience buckets. Bucket one, the identity runner. Bucket two, the founder, the marketer. Bucket three, the performance forward runner. All right, so that's the full framework in action. Aaron, thank you. Clean enough to walk through on camera and teachable enough that any clothing brand founder watching it can immediately apply it to their own three reels. Okay, thank you, Mr. Or Mrs. I don't know, but it's talking through which pieces to run. Why specifically the hook rewritten for paid and the audience signal already embedded in the content. This is your creative brief, and it took you less time than it would to take a scroll on your own analytics, which I encourage you to do as well. You're watching a paid ad for a brand that built 50,000 loyal customers without running a single one. Read that back. That's a scroll stopper. Not a template, not a not a swipe file. That's a line built from content that already proved it could hold attention. Now it's engineered to convert, and your competitors are not building their hooks this way yet. That's the analysis.
Now, the actual ad in part two. Claude writes the full ad, the hook, the body, and the caption. And the caption formula is different from anything you've seen. Most caption templates sound like an ad, and the one Claude builds from this analysis sounds like a post, and that's exactly why it converts. I'll show you the exact prompt. Part two. You have the analysis. Now. Claude writes the ad. So Instagram copy has its own rules. The hook is not a headline, it's the first spoken word or visual in your reel caption. Doesn't carry the sell. It reinforces it for people who are already stopped. When your copy comes from the creative analysis in part one, none of it is a guess. Every word is a direct response to something your audience already proved they respond to.
Okay, so right now I'm grabbing the creative analysis. It's over here working. While that's happening, I want to give you the second prompt, the Instagram ad script. Here's the prompt to grab. So using the creative analysis, write me three versions of a 30-second Instagram Reels ad for this clothing brand. Each version needs a spoken hook for the first three seconds that stops the scroll, body that connects the product to what the audience is trying to become, and a call to action that creates urgency without being pushy. Also, write the captions for each and keep it under 125 characters, making it feel like a native post, not an ad. And make sure the first five words would stop someone mid-scroll. Write each hook as if you're interrupting someone scrolling on a on a Tuesday morning.
All right, so we've got three ad versions, three different hook angles. And look at the captions, okay. The caption doesn't say shop now or link in the bio. It reads like something a real person would post. That's what makes it, like, feel native. That's what makes it convert. All right, so version one. The brand that never ran ads. The hook here. We built this brand without running a single paid ad until now, which we talked about. Body. So we grew because runners wore our kit, tagged their crew, and showed up again next Friday. That's it. No influencer deals, no discount codes, burning holes in our margins. Just people who wanted to look like the person they were becoming and found gear that matched the version of them. Call to action. New season drop is live. Let's see. Version two. You already know who you are. You don't run to get in shape. You run because this is who you are. That's good. If somebody who runs in a community with a lot of people who rock this brand, I can say that's 100% true. They're not running three days a week and 25 miles a week, which is completely fine if you want to get in shape. Honestly, it is. This is the people like me are a little bit crazy. It's like, why are you running 60 to 100 miles a week? There's, you know, why would you do that? The kit question. All right. It's good. This is called to warm. So it says most running gear is designed for someone who isn't you. It's designed for the mass market. The person who needs motivation, the person who might quit. You're not that person. You're already running 40 miles a week and wondering why your tights feel like a compromise. Bandit started in Brooklyn in the Brooklyn run club scene, because serious runners needed gear that took them as seriously as they take the sport. Cool. Okay, so version one, the brand that never ran ads targets conviction buyers, best for cold traffic. We build a brand without running a single paid ad. Version one, two and three. This is good. So this is your production brief. You hand this to whoever, shoot your content and tell them to deliver it. If you're handing it to yourself, great. Real quick. If the last two prompts just changed how you think about your Instagram ads, please hit subscribe. Every week I break down a new Claude workflow or a real brand result just like this one. You don't want to miss what's coming up next.
So if you're still watching, you're the kind of brand owner who's actually going to do this. So let me show you the piece that most people skip. Most running gear is designed for someone who isn't you. Actually, before we move on, read that hook one more time. Most running gear is designed for someone who isn't you. That's. I mean, that's great. So that line right there is what stops the scroll on a Tuesday or Monday morning. Right. For someone who's never heard of your brand, Claude wrote that in under two minutes from three reels you already shot.
Now the part that determines whether that hook actually reaches them, because the creative is only two-thirds of the system. The third piece is what actually determines whether you scale or stall, and it happens before you even open up ads manager. Hey, look, if you're this far into the video, I know that you probably have a clothing brand. You're probably trying to scale it. If you want somebody to take a look under your hood to see what's working and what's not, so you can get over any type of obstacle. We have no cost strategy session. You can check out the link in the description down below.
All right. So part three. Most people open ads manager and start building budget here. Ads at their pace, in the copy, pick in the audience. Please don't pick an audience and hope someone right and hope it finds someone who will buy. We open Claude first. Before I used to open up Google Docs, but right now I open up Claude first because it knows everything that I'm trying to go out. I'm giving Claude everything we've built so far, the creative analysis, the ad scripts, and the offer. All three go in together. Now, here's the prompt. Number three. Copy this. So based on this creative analysis, these ad scripts, and the offer, build me a full Instagram ad campaign brief. Include recommended campaign objective, which again should always go for sales. Which audience, again going broad to start with and why? How many creatives test in the first ad set at $100 a day budget? And more importantly, what to look for in the first 72 hours? And what metrics tell us this is working versus not working. If you have purchases, that's what we're looking at. If you do not have purchases, that is not what we're looking at. We are looking for purchases. If you don't have purchases, we're looking for this other data, which it will give us.
All right. So this is what we got here. This is the offer. The free same-day shipping plus limited edition branded kit bag valued at $38 bucks. The purchase conversion. This is not awareness, right. We're going for sales. We're not doing traffic. Our budget is $100. We're doing three ad sets, basically $30 per ad set. It's going to be set up as a CBO. So campaign budget optimization, which means Meta chooses what's going on. Our ad set one is we never ran ads. Okay. This is we're going broad again. So this is all a prompt. And I would give you kind of the updated version of this right here. I'm not going to do it look like and I'm not going to exclude purchases here. This is just Claude being Claude. My AI is not perfect, okay? But I'm not doing any of that. I'm not excluding and I'm not doing it. Look, like I said to this is who you are. This is kind of that warm identity for the runner. So Instagram engaged is website visitors, video viewers. I'm not doing any of that, right? The content will do the targeting for us. Okay. So that audience targeting piece, I'm not really worried about that. The creatives in that ad set though is right. So identity reel, the run club B roll, the 6 a.m. runners caption, the UGC style, that's all good. This audience targeting. Look, you just need age, gender, location for targeting, and that's it. We only need one campaign. That's part of it is right here. 72-hour check-in. This is what it says. Don't make decisions at 24 hours. Please do not. The pixel needs time to exit the learning phase. And even if it doesn't, it just needs more time and data. So you're looking at gray through play, link click-through rate, cost per CPM, okay? You're looking at these things and this is your first campaign, okay? This is something to go off of. But really, I'm looking for purchases. This is what's working versus not working. Gives you kind of a framework. And then day seven decision framework. I think you should launch new ads every single week. Doesn't have to be a ton. So many people I think lie and think that because of Andromeda, you've got to have all these new pieces of creative. You do need different angles and you don't need things to look the same. You don't need 100 pieces of creative every week. That's slop. Don't do that. This is good. Came through the text version of the breakdown of what the ad set is without the visuals, so you can have that as an option as well.
All right. So here's what the structure looks like when Claude has informed every layer. So we have got one campaign, okay. So I'm gonna create a campaign, go down to sales. I'm not going to worry about the audience because we do not care about the audiences, age, gender, location. It's going to be a new sales campaign. I am literally not going to work at this catalog. Look at this catalog right now, because I'm not trying to do a catalog ad. I'm choosing campaign budget optimization. That means, again, it's distributed through the ad sets. I said my budget was $100 per day. The audience segment and reporting. Again, I'm not doing anything else here, okay? I'm not segmenting at all, not worrying about categories. Okay. So for clothing, you don't got to worry about housing or anything. Now the ad set, this is where a concept happens. Okay. So a concept could be what we talked about over here. Whoops. All right. So the gear built for you, that's a concept right. That's an ad set. And so all the gear built for you and who that who that is going after and that types, those types of creative are going inside of this ad set. Right. So the data set, which is formally the pixel, we're going to choose that and the conversion event is going to be purchase. The attribution model is going to be standard seven-day click, one-day view. It's completely fine. We're letting it start date and end date. There is no end date. We're going to let it run. We'll turn it off when it doesn't work. Now you can go into these placements and controls. I always tell people to review that yourself. Make sure that everything is formatted correctly, because there are different formats for things that look slightly different. They talk about saved audiences here. Don't worry about that. Here's what the full instructions are. Again, I'm doing age into location. If I know I'm going after only men, then I'm going after only men. If I don't go after 65-year-olds, then I'm not going to. Okay, choose the age range, but keep it broad. And also, if you don't know who your audience is yet, one, probably shouldn't run ads yet. But even if you do, keep it as wide as possible because Meta is pretty good at what they do. You can also hit this right and further limits. If you run into an issue where age and location isn't an option, you can scroll down to a little button. It says further limit the reach, and you could go and it will give you another option for age, gender, location. I have just already turned this on for this account.
Now, the ad. You're going to go through here, and you're going to make sure that the Facebook page and the Instagram page and the Instagram profile are connected on here. Remember, your job through the creative is to go through our creative brief and to create that content. And your job is to not pick a bunch of targeting options. Your job is to show the algorithm who your customer already is through the creative that Claude helped build and the content that they've already responded to. Now we ran all three Claude layers, creative analysis, ad scripts, campaign brief. Then we rebuilt the account around Claude that produced this. Okay, make sure the website URL that you put in either a product page or collection page. Don't send them to your homepage. And then make sure that you choose the website and you have a green light here ready to go. You are parameters. Okay. So if you're spending more than $100 a day, you should use what's called a third-party attribution tool. This is a fancy term for basically saying pass the data back to the right system. Most of the time, ads manager data inside of here is completely inaccurate and wrong, because Shopify and Meta and all of those other tech giants are not sharing data with a tool like Triple Whale. You get that through Line, you get the connection between them, and it gets a much smarter model so that you know if an ad is actually working or not. If you're not spending $100 a day and you're only spending money on Meta, don't worry about Triple Whale. You can sign up for free, actually. But if you are over $100 a day, 100%, I absolutely recommend for you to set that up. So there's a link down in the description down below. We are partners with them. You don't have to use Triple Whale, you can use somebody else. But there is a free version that you can use as well. And the whole point is that you have accurate data for your business. It tells you attribution of sales through email and SMS and ads and other multiple platforms. If you have those, you're going to put that information here. That is really it. If you have offline events or app events, if you don't know what those are, does not matter. Right? Means you don't have an app, or it doesn't mean you may not have a brick and mortar. If so, you would want to turn those on.
So the brand is using Claude to analyze their content, right? Their creative and structure their campaigns. They're going to own their category. So I put together a Claude prompt pack built specifically for Instagram ads and clothing brands. And most brands spend weeks figuring out which content to run, rewriting copy that doesn't convert, and guessing at the campaign structure. These three prompts compress all of that into an afternoon. The Creative Analysis Prompt, Ad Script Prompt, and the Campaign Brief Prompt, plus extra writing for captions and reading your first 72 hours of data. If you want this, just comment the word Claude down below. I'll send it directly over to you for free.
Look, if you have a clothing brand and you're trying to grow, YouTube knows you need to go watch one of these videos right here. Next. Y'all. See you next time.