Expert Marketers Give Masterclass in Creative Research & Execution

Expert Marketers Give Masterclass in Creative Research & Execution Speaker 2: Welcome back to another episode of Chew on This powered by DigiCom. As you can see, we have a ton of guests today. We have the team from DigiCom here today to talk through creative research, execution, how you guys should be thinking about making creatives, testing them, iterating on them, scaling them. I think this conversation is very, very important, especially post this Andromeda update on meta, TikTok shop becoming this massive phenomenon. For those of you who haven't met the guys at DigiCom, we'd love quick intros just around the table and then we'll get into it. Speaker 3: Sure. So my name is Hemant. I'm the CEO and founder. Speaker 4: I'm Danielle, the COO and co-founder. Speaker 5: My name is Mikey, and I am head of partnerships and co-founder. Speaker 1: And I am John, director of growth. Speaker 2: Love it. Well, thank you guys for being here today. I'm very, very excited. We haven't done a podcast with this many people on it. So I'm really interested to see the dynamic and how much we can actually get into. So first things first, right, let's let's kind of dive into the fundamentals of Creative, right? When we talk about paid channels like your Metas, your TikTok, YouTube, the content that typically does well in these areas is that typical direct response type of creative, right? There's a time and place for more maybe branded, you know, when it comes to maybe brand awareness channels like a TV or, you know, Linear Connected, et cetera, et cetera. But I think for the best part of the last decade, I think the best creatives have been very direct response, right? 0.5 seconds to just get somebody to stop from scrolling, right? So let's dive into direct response. What is direct response and why is it important for paid channels? Speaker 1: Yeah, so I'll take that one. Easy. You're trying to elicit a response in that moment, right, Ash? You're trying to stimulate in that moment. And it's a very psychological and behavioral thing to stop somebody from what they're doing, to draw attention to what you're doing. And you need to continue to stimulate those people with the first five seconds, obviously, but then through the remainder of the ad. And you're continuing to stimulate them through the journey. And that's what the right response is, in all of its essence, is How do I get them to elicit an action on the stimuli that I'm providing them? And those stimuli, it's so many different versions of it, right? There's audio stimuli. There's the way that they're feeling, emotional stimuli. So you got to build off of that. And I think that's what's really important about direct response. And we get into this rut of like, okay, we need to have this very kind of like high level branded Design or creative, but it's like, is it actually doing what we want it to do, which is driving somebody to make a purchase, that behavior that we want to draw from them. So yeah, that's how I define direct response. It's very straightforward. It's what marketing's built on for centuries now. Speaker 2: In terms of now, the first thing that you said is you got to kind of evoke this emotion from somebody, right? And you have to stop somebody from scrolling in the first, you know, we've said five seconds. Maybe it's three seconds. Maybe it's even one second now, right? What can you guys do in that first, you know, that one second that will actually stop somebody from scrolling? It's by Hitting them where maybe their subconscious is thinking about something but they haven't really thought about it or said it out loud and that's probably the only way you can get somebody from scrolling. Now to actually understand that and get to that point, there's a lot of research that goes into doing this. You can't just set up a video and say something generic and hope somebody stops scrolling. A lot of this comes down to research. Now from your guys' perspective, What does that research process typically look like and how can you guys actually get to the root problem that you guys can actually solve with whatever the brands that are selling? Speaker 3: So first is we are always focused on selling a product but really you're selling a problem and that's very important. Understanding the root of the problem. Understanding the persona of that person with that problem. Creating that chart, that graph, that persona is super, super important so you can understand who you're selling to, right? There's so many times we will work with brands and they have an idea, this is the target customer, only to find out that it is someone totally different, right? And it's happened time and time again, whether it's gender, age, You know, just from like a demographic understanding and with your data that you're starting to receive back, like, yes, you need to first create a persona and that persona is living and breathing. That persona, who you think it is, that's your hypothesis. And once you start getting data back, that persona might totally change and that's okay. Make the shift. Figure out who that actual persona is from the user or the customer that is purchasing your product and then build off of that. And I think a lot of times, you know, brands often fail to identify the customer. They do, like, brands will do a good job of saying, hey, this is who we think the target audience is. But as that brand is growing and evolving and shifting, that audience might be different. Based on what the data is saying and then when they have new product releases, that could shift even more and you might have two or three or four personas based on the product type that you are marketing. And so I think the core is understanding the persona and then shifting from, hey, we're not just selling a product. We are solving a problem for someone and then focusing your messaging and creative to solve that problem for that respective persona. Speaker 2: Makes sense. So with certain brands and maybe even products, and at least for us, especially being in the supplement space, certain products can even have multiple benefits. And so at least for us when we try to Advertise that we do offer multiple benefits. I feel like when you try to advertise multiple things at one time, you're not really talking to anybody specific, right? So to your point, it's like your one product may have multiple personas, right? And obviously at the beginning when you're starting to figure things out, you as a founder or operator or marketer is just like, well, I think my product will solve this problem for this person. I want to get you so far, right? Then for you guys to be able to scale, it's actually finding the unique personas that you can really nail the message down and figure that out, right? So question for you guys is actually figuring out who that persona is. What is your process to actually get that insight and then allowing that to create the creatives from there? Speaker 4: I think there's a lot of different ways that you can approach it. I mean, easiest place to start is your customer feedback, your existing customers, really like combing through the reviews. And I mean, there's a lot of generic reviews. This product is great. This product is life changing. But like finding the specific nuances to tell you, okay, this is This solves like a very unique problem. It's not just about, you know, clearing up your skin or losing weight. There's something like very specific that you can then start to build a persona around and kind of like, you know, use that to really craft your messaging, your engagement on social, even like your ads, right? Like ads that Perform really well and have like longevity and a ton of spend behind them tend to also have a lot of engagement behind them. So what are people saying in the comments? People who are skeptical about this product and will it solve the specific problem that they're having or is it going to give them the need or aspirational whatever it is that they're looking for. Reddit threads, you know, it doesn't have to be a thread about your specific product, but maybe you start doing some research around the problems that your product are solving, look in the communities, find the really like specific use cases and nuances that are moving the needle for consumers and then start testing around those personas. And that's really how you start to build these very like specific Targets that are not just based on like your usual demographics, like what are their behaviors? What moves them? What is the driver behind their purchasing decisions? So yeah, there's a lot of different ways that you can start to find these details. Speaker 1: And to Danielle's point, I like to think of the comments as like a microcosm of like the general populace. If 10 to 15 or 20% of the people in the comments are saying something in particular, probably 20% of your audience is thinking something in that same manner. So it's a way of kind of like breaking it down and you're creating those mini personas and those mini problems that each person has. And Reddit is a great place because it's anonymous and people will tell you exactly what they're feeling without any repercussions because it is completely anonymous. So it's also a great place for research in that kind of like state of mind too. Speaker 3: Last week John was speaking about this one specific example, right, of using Reddit for a product where the price point was extremely high. So John sat down, pulled all this data and then started developing ads that Literally called out. Speaker 1: Addressed everything that was in the Reddit thread directly. Speaker 3: We know our product's expensive. Speaker 1: The cost is $599. Why is it so expensive? Like yada yada. And then I started making ads that literally called out, it's this expensive because we do X, Y, and Z. And it's like addressing the problem straightforward. It's like, okay, I know that 30% of the users, 40% of the users, the price point is like the immediate problem. Okay, how am I going to address that? I'm going to give them every reason to justify the price point in an ad. Straightforward. It's not cheap because yada, yada, yada, yada, all of it. It mixes, it grinds, it cooks, it weighs, it chops, it makes you 40 pounds of food every serving that it builds out. So yeah, I definitely think it's like, all right, What are these people thinking that are commenting in the reviews and stuff? How are they feeling? And then how do I take that and address it immediately and create a persona out of that? Speaker 2: One thing that you just said was how is the consumer feeling, right? They don't care about what features the product has, right? That's how they're feeling. Speaker 1: That user journey starts before they even see the ad. Speaker 2: Exactly. Speaker 1: Like how are they feeling in that moment, in that day? How is their self-esteem that day? Like how are they feeling if they're excited, they're more likely to purchase something spontaneously, right? If they're having a bad day, maybe they're looking for something that's like chocolate and that's gonna sort their day, right? So it's definitely different. Like people are going into this looking at your ads with different emotions and you got to justify like, hey, I'm showing this ad in this state of mind and that's why I have the solution or the after feelings, right? Like the after emotion, after you buy my product, what are you going to feel and how am I changing the way that you're feeling right now? And that's going to elicit that direct response is because now I've changed your view on how you're going to feel by the end of the day. Yeah. Speaker 2: Right. No, it makes perfect sense because I mean, let's just say you have, you know, again, we'll bring it back to supplements. It's like, okay, well, my product has this ingredient, this ingredient, that ingredient. You should buy it because it's a great formula, right? It's like nobody knows what the ingredients do, right? But I think what matters the most is, okay, well, what What is the outcome of what those ingredients can do, right? What is it going to do for me? Why should I care, right? So I think emotion and invoking emotion is very important when it comes to, you know, especially in video ads and especially, you know, doing it on static ads. What are some of the emotional drivers that you guys like to hit on when you guys actually create these types of ads? Speaker 3: So just also going back to like, what Danielle was saying, we actually all approach creative in a very different way. So, you know, Danielle has this structure, like the structure that she kind of goes through to tie the persona and the messaging and I'll let her jump into that and then John is thinking about Maslow's hierarchy and we should, you know, we should definitely like touch on those points, right? For me, it's always like, What does the data tell me and how can I chop up the data in every which way to say this is what's working like, you know, and I think it's just, it's, we just all approach it, you know, in different ways. And then Mikey uses just everything we do, which is really cool. Speaker 2: Yeah. I love that. In terms of now, let's call it more, let's get into like maybe the video side of things, right? Like the actual audience that's watching. How, like, niche are you guys getting, right? It's like, are you going after 35 plus moms who are stay-at-home moms or are you going after, you know, that 35 to 40 mom who's got, you know, one kid, another kid on the way, juggling this, that, like, very specific, right? Like, how in-depth are you guys going when you're trying to talk to a specific audience? Speaker 1: So, I think that depends on the product and Hemant brought it up with Maslow's Hierarchy of Needs, right? are just physiological that people need and that's the bottom layer, right? And then there's safety, that people are looking for safety, that's the next need. You could just basically use demographics for that and not get into the subcultures or the ICPs because that's a physiological need and that's a safety need and that's basic, right? Like, I need that, you need that, we need water. And it works. Beyond that, then you start getting into the subcultures that you're talking about, where it's like stay-at-home moms with kids that, for whatever reason, they want to make bath time more fun. That is a more direct thing that comes after those basic needs are made, and that's where the audience types are starting to filter out themselves, basically, because each one of the next layers of Maslow's hierarchy of needs, which is going to be social belonging and esteem and self-actualization, You're now going into less of a gender breakdown and it needs to be more of like who is this persona that wants to belong to this category of people, right? So like an apparel brand like Gucci, right? Like it's going to have a much different persona. It's not going to be just every female that's 30 plus. That's basically what it alludes to with Maslow's Hierarchy of Needs. It's just like, where does my product fall into this? And then the further up the pyramid it falls, the more directed my ICP needs to be, the more thought out that persona needs to be. So that's kind of the way I think about it for sure. Speaker 5: Yeah, I think also there's not only one persona per brand. I think that's a misconception on the brand side that we obviously like to Test into on the performance side, you know, we have examples of brands who are in the CPG food space. They want to be targeting specifically, you know, moms with two kids who want to be able to eat more health conscious. And then me from someone who is, you know, not a mom of two kids and enjoys, you know, working out and health and wellness. I see it from the perspective of, you know, someone who's into fitness and wants to do, wants to be able to eat this product and not sacrifice. Their diet. So kind of been able to find for that particular brand like two different styles of creatives, two different styles of personas that both work. So I don't think you should be so focused only on, there's multiple personas per brands and you should really be testing into what those personas are because again, you don't know what they are until you test into them and looking at the comments, the Reddit threads and all this, you'll be able to find those as well. Speaker 3: Fair. Speaker 2: So I want to get into the tactical side of this, right? Which is, okay, we've done the research. We understand what these potential consumers actually want, right? And we know what they care about, how they potentially speak, their tone of voice, etc., etc. How are we actually Using that information and creating content that shows them A, what they want and talking to them in the way that they like being talked to. What does that look like and how can brands make sure they nail that? Speaker 4: Yeah, I think talking to people in the way they want to be talked to is really important. And again, that can be done in the research stage. You're seeing, you know, you're taking note of the language that people are using in their feedback, in their comments, or like on TikTok videos, you know, do a little bit of research on TikTok, look up certain Topics, problems, products, see how people are talking about the products. What's like the slang that they're using? And really, you know, find ways to incorporate that into your creatives. And that kind of also brings us into like the importance of hooks, right? Like you said, the attention span is getting shorter and shorter. You only have a few seconds to grab people's attention. So much of your focus needs to be on the hook of your video. Or your static, the headline, whatever it is. But a lot of your focus needs to be on that hook because if you're losing people's attention within that first one to three seconds, then What you're saying in the rest of your ad doesn't really matter because that's it. So really doing, we like to do a lot of our testing around the hooks because we do find that that moves the needle the most and there's different ways to do that, right? You can do it through a visual hook, what the first few frames include, whether it's like person talking head, maybe it's like a product shot, an unboxing, whatever that might look like. Then you have your audio hook. What is the person saying? Text hook. If you don't have a video that's like, you know, maybe you don't have a voiceover, how can you manipulate the video to say different things and appeal to different users or to different pain points with the text hook? And really use that as like kind of like your first line of testing and then from there you can really like start to build and once you kind of like get the data and you you're really able to like break it down look at your hook rate, your click-through rate, your conversion rate, understand like where the drop-off is or where maybe you know you can improve the metrics a little bit then you can start to like refine that and and really understand how to like isolate the right variables and and use that to put together the The content that will drive the best results. Speaker 2: I love it. Speaker 1: And I don't think there's like a more evident place than like the different platforms on how they speak differently. Speaker 2: Right. Speaker 1: Like people on Twitter and Meta and TikTok and Reddit all have their own like voice and just even marketing to people on different platforms you have to have a different voice. So before you even get into like the ICPs and the subcultures of like who you're actually targeting, it's like, what platform am I serving these ads on? Yeah. And how am I going to vocalize what I want to? You know, like TLDR on Reddit is big, AITA, right? Like all of that kind of like the defined cultural Reddit language is working on Reddit. It might not work on Meta. It might not work on TikTok. So yeah, that's kind of step one too, as well as like, where am I actually serving these ads? Speaker 5: A good example of that is with an apparel brand of ours. Our YouTube DemandGen campaign ads are completely different than what performs within Meta. Meta, they want to see someone wearing the product, maybe them talking to the camera. We tried those ads on DemandGen, did not find success there, but actually laying out the product and showing the graphic In the video was more effective and we found that through like the what I would consider like demand gen hook rate. Essentially first 25% of the video. So it's just interesting to see that those hooks do vary not only per persona but also per like platform. Speaker 2: I want to shift gears into how you guys are leveraging AI here, right? Because we're talking about a lot of different data sources, right? You're talking about your reviews, your comments on your ads, Reddit threads, post-purchase surveys. There's a lot of data to go through. And sure, like, you know, you can kind of comb through it and find different things, but, you know, we have ChatGPT now. So how are you guys actually leveraging, you know, AI? I'm actually very curious because, like, even at Avi, I've, like, tried and messed around with, you know, creating a A GPT just for like obvious, like certain products. So like I'll upload, you know, the post-purchase surveys, I'll upload our reviews, our surveys, comments, everything else, right? And then I'll ask it to kind of give me an analysis of who is our target persona and get it to a point where like, all right, well, let's come up with some ad copy, right? My problem is that the ad copy I get is bullshit, right? Speaker 1: Are you uploading the data too? Speaker 2: Yeah, yeah, yeah, uploading everything. Speaker 1: Like click-through rates, everything. Speaker 2: Oh no, no, not even, like this is just like to conceptualize creative. Speaker 1: No, but you should do that. You should upload your data. Speaker 2: Okay. Speaker 1: Right, like you have all these data points. ChatGPT can comb through. We recently made a ChatGPT for one of our clients and I uploaded all of the different data points. I gave it the best copy in terms of click-through rate on Twitter, on each platform. I gave it the best copy, the worst copy too, on the other spectrum. And then it'll be a little bit more prime to give you the results that you want. So it's not just about giving it the brand assets. You also have all this data that you can feed it and let it make its decisions based off of that data. So yeah, I think that would be the next step for you would be to... Speaker 2: I mean, walk me through this as if I haven't done this before, right? Because I've done it. I haven't really gotten the best results out of it. But I know there's brands out there that are like, all right, How do I really leverage AI? What is a good starting point for brands and how should they be thinking about this? Speaker 1: Building GPT is a good idea because then you're going to stop the bleeding between everything else that you're doing in your ChatGPT or wherever you're using Gemini or Claude or whatever it is. The next step would be to upload all the brand guidelines, brand assets, your best creatives, also feed it the data, give it the charts of the data. Speaker 2: So how are you actually preparing this to submit it? Are you creating like a PDF? It's like, hey, here's our static ad that had X amount of data. Speaker 1: Excel's. Speaker 2: Excel's, okay. And then you're just uploading that? Speaker 1: Yeah, just uploading it. Yeah, I think it recognizes it immediately. And what you're doing from there is you got to obviously the input. To get the certain output, you need to put in a certain input. It needs to be very detailed. I recently started making very hyper-realistic images and it's because I've started to understand The camera film, the film that's needed for a certain shot, the camera that's needed for a certain shot, and without that prompt of the camera and the film and the aspect of the shot, it's not going to deliver what I'm looking for. So you need to be very nuanced with what you're providing, ChatGPT, and detailed, and giving it a role. Let's say, for example, I did this with one of my clients' copy. I had this copy that was all about social belonging, brotherhood, and all of that. And I wanted to understand what about it was working and I asked it to analyze the copy, the two pieces of copy, as if it was an Ivy League marketing professor and give me each one of the different points on why it's working. And then now I see that there's five or six variables that I can, within the copy, adjust, like in and out framing and like being part of a group versus another group and all of that. You have to give it a role, give it a task, and then understand that you're not done. It's going to take 10, 20, 30 prompts to get to the final product, which you've Edit, again anyways, because you're the editor. So it's just a tool to get you to where you want and I think prompting it correctly and appropriately is a big part of it and also building separate GPTs where you have all this data. You doing it by per product, that's exactly what I'm doing too. It's like, okay, each one of these products is going to have its own information and it should be split out into its own GPT. So yeah, that's kind of how I think about it. Speaker 3: And I just want to talk a little bit about the prompting process. So we've been building our own AI engines for internal use, not so much external, where we're uploading essential links with data g-sheets, slides, with information. I'm on version 49 now of our prompt, like little version 49, like going back and forth over and over and over and over and over again until it's like this data is important. So like something that John and I just worked on with this tool is adding into the prompt causation and correlation structure, right? So like, we lifted spend for like, spend increased on a campaign, conversions went up. CVR is through the roof, right, compared to all the other campaigns or creatives, right. And for us to get the information that we need about creative or campaign or information has been slowly adding in and figuring out that we need to give the AIs a framework to essentially think within right because these AIs can tap into so many data points right so like for us it's a What's actually interesting is the prompt that works better for us in terms of the information that we need to pull is creating the role of a business analyst versus a growth marketer, right? So business analyst is giving us a lot more information with correlation and causation, but growth marketer is just five word bullet points that we're getting back. And so working on Like John is saying, that chain and prompting for AI, right, is unbelievably important. Speaker 2: I love that. So, first step is to obviously upload everything that you possibly have, right, including your existing assets, the data that you already have, especially on both winners and non-performers. The second step is actually getting the prompt into a position where it's giving you exactly what you need it to, right? Now, the third thing here is actually asking it for what you need, right? What are you guys, you know, leveraging AI for? Like, are you guys, like, ad copy, the actual concepts, hooks, like, what are you guys using it for? Speaker 4: I mean, I think it's like a little bit of everything. Like one, I think, primarily as a research tool, because it's just makes research so much easier and faster. I think like hooks and ad copy, like you're saying, a lot of what you get from ChatGPT is like, it's like this weird, ambiguous, like language that's like, okay, what does this mean to anybody? So it's good directionally, maybe sometimes you'll get like, a really great headline that you're like, okay, this I can actually test as is, but I think it's, it's a good starting point to like, give you the idea of you know, what you can delve a little bit into. But again, then it's just like, it's applying like your research, what you know, as a human and how people talk. And I think it's also important to remember that like, You don't need really flowery or elaborate language in your performance marketing ads. This isn't Mad Men. We need to be very direct and to the point. How are people talking about things? John was saying earlier about making bath time more fun for kids. This is an actual ad that we tested for a client. It's a soap product for kids. But there's like a novelty aspect to it and so the whole concept of this ad was making bath time more fun. And it's amazing like what the smallest changes in language will do for an ad. There was one ad that absolutely crushed performance, spent tens of thousands of dollars. Every other ad flopped. But the messaging was the same across the board. It was just the way that we formulated the hook. Something along the lines of you know like Mom hack, make bath time more fun with one product versus POV bath time is now your kid's favorite part of the day or something like that. Makes all the difference. You're saying the same thing but you're just saying in a slightly different way and that's also why like the hook testing is so important because sometimes it's not that your messaging isn't you know the right avenue it's just maybe the way that you're saying it isn't the best way or it's not really like resonating and hitting with the target so just trying to formulate it in like a few different ways to see okay like what's what's gonna hit what's gonna grab people's attention. Speaker 2: Makes sense. The one thing that you mentioned was you know being able to Understand why you're testing something I think is very important. For a lot of people out there, I feel like people just throw shit at the wall and see what sticks, right? And I feel like that's probably like okay in the beginning, but then once you find something that works, how can you learn from that? Learn from what worked and learn from what didn't work, right? So walk us through how brands should actually be thinking about, you know, a proper creative testing strategy versus just Testing whatever and anything. Speaker 3: Sure. So I think first is with a lot of the brands that like especially the audience we're speaking to on the pod, right? You're already running marketing. You're on Meta. You're on Google. You have a plethora of like data sets that you can pull. So pull that data set. This part is probably the most annoying part of what we do where it's an ad, it's the link to the URL, and then it's defining each attribute. So video static, the type of language we're using, if it's a video, the length of the video, maybe the camera angle, the product that is in the video. Tag every single attribute of that one ad and then do it for everything. Then combine the actual data, whether it's triple wheel data or third-party platform data or platform data to those attributes. Look at the insights, find the patterns from that and that's going to give you a good base layer to start. After that, you can think about, okay, this type of messaging is really well. My hypothesis is that I want to run like a multivariate test or just an ABC test on this hook. But like Danielle said, we're going to reframe how we provide this messaging. And then test them against each other, right? And keep it clean, run it in its own ad set. And you'll eventually systematically figure out a way to say, yes, this works, no, this does not work, this bombed, this crushed. And like, that's very much my way of thinking and like approaching this first base layer of making sure that there's data integrity, there's clarity, and then To create your testing roadmap, it's okay if this test wins, here's what I want to do next. If this test doesn't work, here's what I want to do next. Oftentimes, I think with a lot of marketers, it's, hey, like you said, let me throw a bunch of shit at the wall. One thing will stick. Okay, what do I do next? And your ad fatigues and you have nothing next because a lot of times you're not really sitting down and thinking about the end user. You're not looking at the data. It's just, hey, I have this concept. Let me throw it up. And that creates a lot of barriers and then you get into a cycle where things are working really well or things bomb and it's like then you're trying to play catch up and catch up and catch up and it's really difficult. That's difficult for any brand. And so it's systematizing and organizing. Organization's a big part of what we do. Talking about this, it's like, oh, this is such an easy, simple thing. 90% plus of the brands we speak to, doesn't matter if they're a $20 million brand, $100 million brand, One million dollar brand. They are not organized. They don't know what hooks work. They think about a specific messaging and persona of a person, which is not what the data is saying. They're trying to push products in specific ways. It just comes down to something so simple as, hey, my video assets are in this folder and my static assets are in this folder. Within that, it's here are this influencer or this messaging type. It's just, hey, here's an asset folder with everything in it and they can't point to, hey, this is actually what's working most of the time. Cleaning that up, creating a structure, looking at your data, and then taking that and creating a systematic way to say, if this test fails, cool, I learned something. What did I learn? Writing that down. If this test wins, here's what I learned. Here's what I'm going to do next with it. Speaker 2: Love it. I mean, there's definitely a lot of brands that are not doing this, right? And they're probably like, and I fall to this too. I have an ad set and I'm like, V1, V2, V3, and that's it, right? So I have nothing to go off of. But I love this approach where you're tagging every single attribute. Then you're figuring out, okay, what is the commonalities here? And then, all right, now I actually have a hypothesis of, okay, well, this hook worked, this tone of voice worked, this length worked. All right, now let me go and test something similar to that. And then on the other side, it's like, well, this didn't work, this didn't work. Well, what can I test? Next to figure that out, right? So I love this approach. Speaker 1: It's like pulling on a thread, right? Speaker 2: Exactly. Speaker 1: You just keep pulling and closing doors behind you on something that didn't work, right? You can revisit it at some point. I'm not saying close the door forever. Speaker 2: Yeah. Speaker 1: But like keep pulling on the thread that is working moving forward because you know the color orange works, right? And the color black doesn't. So you put black in the rearview mirror. You keep pulling on orange and through different testing you're going to find other elements while you're doing it. But at least Let's say the black ads were getting a 1.1 and now they're getting your orange ads are at a 1.9. You've removed that 1.1 from the data set. Like you won't get a 1.1 because you're not doing black anymore. And I know it's not as easy as just like, okay, we're switching the colors, but each one of those attributes will do that where you're just lifting the base slowly just, all right, I'm eliminating things with Bad performance. I'm not doing it again, and I'm continuing to do what's working. So that's growth, right? That's obvious and evident growth. Speaker 2: I mean, from your guys' perspective, I mean, by being able to do this, have you seen a correlation of an increase of hit rate for when you guys are testing new ads? Speaker 1: Yeah, especially as time passes. Speaker 4: As you're learning what works for a specific brand, it can take some time. But yeah, once you learn it, I would say the success rate definitely increases. I'm not going to say it's 50%, but it improves. Speaker 2: If you can go from a 5% to a 10%, it's better off, right? Speaker 4: Yeah, it's a win. Speaker 2: I want to get into the actual tactical part of testing creatives and what are the kind of the metrics that you guys are looking at, right? So let's look at meta, right? We have a batch of creatives. How are you guys testing them? What's like the strategy there? And then what are you guys looking at metric-wise to see if something's performing, not performing, or going to perform? Walk us through that. Speaker 4: Yeah, so usually there's like a set of ads, right? That's also a very important part when you're, you know, trying to figure out what works and what doesn't. I see a lot of brands who are just like, oh, I'm gonna create like one or two ads and we're just gonna run this, which, what are you learning from that? Speaker 3: Right. Speaker 4: So there needs to be a set of ads. We like to do many, many variations and For a video ad, let's say we're breaking it down by the hook rate, the average watch time, the click-through rate, the conversion rate, and really trying to pinpoint what is working and what isn't. You have a really high hook rate, great. How do you apply this concept or this message to another ad? You have really high click-through rate, but your conversion rate is awful. Maybe your ad is misleading. You know, there's a mismatch between what you're saying in the ad and what you're promoting on your website. How do you align those two a little bit more? So yeah, really just like breaking it down and trying to find, you know, what the different metrics are telling you and comparing it against each other. You know, this hook rate was better than that hook rate. Okay, that's how we start to eliminate, you know, these messages or maybe these like visuals that aren't working as well. I started with a product visual versus like a person, you know, what's the hook rate between the two? What's the click-through rate between the two? So really breaking it down by all these parts and then Using that to create like your next hypothesis and how do you plan on improve whatever the metric was and making sure that there is a testing component for the next step as well. It's never ending. You're always testing some component or variable. So yeah, that's kind of how we usually break it down. Speaker 5: And I think just going off of what Danielle was saying with The different metrics, the numbers don't lie at all, right? I think something that we notice all the time is we have a different idea or we know what works compared to what the brand thinks. They know what works. An amazing example is a UGC versus an extremely polished video. Started to work with the brand back in September. All of the video assets were super polished. Heyman took it upon himself to kind of dive in and create a UGC video because we didn't have any for the brand. That video performed exceptionally well for months. And we, when we filmed it originally, you know, it wasn't even, it was, it was horizontal. We had a cropped it vertically. Speaker 3: As a marketer, I should have just done it on my phone, but I didn't have a stand. Speaker 5: Yeah. It's like, we know how, how we'd like our brands to position the videos, but we're not, you know, videographers. Like their sauce, I don't want to throw you under the bus here, but there's sauce on the side of the stove. There's the wooden spoon still in the shot. Unknown Speaker: And these ads crushed. Speaker 5: And the brand, you know, our contact there was like, not going to lie guys, this brand, this video is pretty lame. Like this is not a great video, but you guys are the experts. Let's try it. That video crushed. And then we created about 17 different iterations, whether it's Hemant's face in the beginning of the video, it's a finished product in the beginning, or it's still in the pan, or he's twirling, like, The food around the fork, it was exceptional how well this worked. But from the brand's perspective, it was, this isn't a polished video. This isn't a good look for us. But its performance is so different. Speaker 3: And that's like the, that's always like the catch 22 for us, right? Where it's like, oh, this needs to be like prim and perfect. And like the reality is, guys, we have to test it. Like it doesn't matter if you're a super on brand, you know, creative director or founder, it's, Hey, you're trying to build a business and learn and figure out what is working and what's gonna drive you forward. Just because it's not perfect doesn't mean you shouldn't give it like a little bit of a test. Put a couple hundred bucks behind it. You might find something unbelievable and then after that you build an entire program around it and that could net Such positive results. And you know, we always have the conversation with especially new partners like, oh, this is not to our standard. And this happens almost every kickoff, right? And it's like, well, can you just give us like $1,000 to test this? Like, just give us that test. And Oftentimes it happens, sometimes we get the pushback and then we go back with the data and say, hey, this is the direction that we might want to consider moving in. For that brand and a multitude of other brands, it's been, oh man, I should move away from creating content in the way we have and let's focus on creating a lot more UGC content or founder-led content. Test it. Just test it. What do you have to lose if it's a couple hundred bucks? If anything, you have so much to gain, right? And it's opportunity cost for you in the long run. Speaker 2: This episode has been super, super tactical all the way from consumer research all the way down to leveraging AI to actually get that done. Understanding why you should understand What the test that you're actually running is going to do for the brand and understanding how to iterate and actually create a creative strategy that works for the brand. So thank you guys for joining us today. Quick, would love a 10 second takeaway for the audience that's listening and watching. Something to implement in their business starting today. What would that one thing be? Speaker 3: Take the opportunity to test because if you don't, you lose out on opportunity costs that you might have never known existed. Speaker 2: Love that. Speaker 4: I would say get to know your actual customers on a very deeply personal level to understand what moves them and what they're looking for. That's really, I think, the best way to develop winning ads. Speaker 5: Yeah. I would say get organized. It's impossible to look at the data when you're not. And the data doesn't lie and trust it and continue to work off of the information you have. Speaker 1: To Mikey's point, it's not about your feelings. It's about making the best ad that you can and it doesn't have to be the most polished ad. It has to be the best ad. For that moment, for what you're doing, for that subculture, for that ICP, for that demographic that you're targeting, it doesn't have to be the best ad to you. Speaker 2: Love that. Well, thank you to the DigiCom team. You guys are absolutely incredible. For those of you who are potentially looking for somebody to help with growth, I mean, these guys handle growth, performance. And they're super knowledgeable on creative itself, which is the biggest needle mover when it comes to performance in your ad account. So if you are looking for somebody to help out with creative strategy, increasing spend and efficiency in your ad accounts from Google to Meta as well, definitely reach out to them. We'll leave the link in the bio and thank you guys for coming. If you want more from us, follow us on Twitter, follow us on Instagram, follow us on TikTok, and check out the website, JuwanThis.io.