EP #317] [ENG] - How to leverage UGC to scale your brand on Amazon - John Gargiulo

EP #317] [ENG] - How to leverage UGC to scale your brand on Amazon - John Gargiulo Unknown Speaker: Welcome to The Ecommerce Lab By Ecomcy. This is the place for everything related to Amazon private label and e-commerce. Learn exactly what you need to start or scale your business. Get insights from the top industry experts who will discuss the latest trends and best practices in the world of Amazon. From choosing products and sourcing from a supplier to setting up your Amazon account and marketing your business, you will hear it here. Let's get started. Here is your host, Vincenzo Toscano. Speaker 2: Hello, guys. Welcome to another episode of The Ecommerce Lab By Ecomcy, the place of air related to Amazon FBA, private label and e-commerce. My name is Vincenzo Toscano. Founder and CEO of Ecomcy. And today we bring another special guest. His name is John Gargiulo and he's the founder and CEO of Airpods, which is an amazing solution when it comes to leveraging the technology AI to figure out how to create the perfect video for your ads. Right. And this is something that actually I'm very interested to speak about with John today because AI has been, you know, changing the game when it comes to a content generation. And there's so many new things out there that you need to be aware of, because if you're not, most likely your competition is doing it. And then it's going to be a matter of time before, you know, they're eating into essentially your market. Right. So, John, it is definitely a pleasure to have you on the show today. Thank you so much for coming. How are you doing? Speaker 1: Yeah, I'm doing awesome Vincenzo, and you actually are a rare person that pronounces my last name perfectly correctly the first time. Speaker 2: I wonder why. Yeah, I mean, I think we're pretty much neighbors. You are from Sorrento. My family is actually pretty much from Positano, Majores, so it's like, we're pretty much, you know, very close, you know? Speaker 1: Positano is one of the most beautiful little dots in the world. I'm really impressed that your family's from there. Speaker 2: That's incredible. I know, I know. I don't know why I'm in London, to be honest. Anyway, John, thank you so much for coming to the show. I guess, you know, let's start with a quick intro by yourself, what you do, and then of course we can dive into the topic. Yeah. Awesome. Speaker 1: I'm excited to talk about AI. Yeah, so I'm John. I'm the founder of airpost.ai. We make video ads for paid social. I feel that in not that many years from now, all of this sort of influencer platforms and creators and All of the people that manually make videos over weeks and months and cost advertisers thousands of dollars will be sort of melted into AI. We'll be doing those things. It's coming, whether we're, you know, however we might feel about it. And we want to do it well and ethically. You know, with creative at our core. So that's what we're working on. My history is I was at Airbnb for a few years in marketing starting in 2016. I come from the ad world originally. So I was a creative I went to learn copywriting at Miami Ad School, worked in the agency world in New York, doing TV commercials and things like that in the past life. And then here we are in the era of AI, changing everything, trying to keep that authenticity and creativity that's so important for video ads as things progress week over week. Speaker 2: I love it. Yeah, I think, you know, especially with the background, I'm sure you had at Airbnb, like you experienced, you know, what is, I read the few elements and tricks that you have to embed into, you know, a commercial, a video to make people engage with the app and the service. And I'm sure that's some of the secrets of how you implement the AirPost, you know, like really understanding what is the ecology behind what a piece of content has to have to drive a click, which we know is all about that conversion. Like I see so many videos out there of images or banners that people think that that's, you know, Effectively, it's just a matter of putting something out there to drive a click, but it's beyond that. It really has a science. You really need to understand your avatar, and on top of that, you have to make it look beautiful, which is something that with AI now is easier than ever. I guess going back to the foundation of, you know, AI, I think some of the conversation I've been having lately is that what, you know, potentially before could take you weeks, could take you months, now within a few seconds, you know, you can make a few videos that you can repurpose across, you know, different platforms to running ads and everything. So let's start with that. I guess you now, you know, Airpods and everything that you experienced, like how do you feel AI has changed the game when it comes to creating ads and videos for that, you know? Speaker 1: I think it's starting to change the game. You mentioned avatars. So first of all, everything you're saying is true. You know, what I'm hearing you say is The expertise is important. So prior, after Airbnb, before AirPost, I founded an agency called ReadySet that grew very big, very quickly, still doing well. We just had our best quarter in Q4 of last year. And it works with clients like DoorDash. Kim's and hers, we've worked with Walmart directly before, to make video ads at scale manually. So we have a studio in LA, we have all these creators we work with, we have all these amazing creative strategists, and we crank out hyper-efficient, we call it hyper-production, shoots every week for these companies. And that's where we incubated AirPost, and that's how we learned to do this. I find that most people sort of tweeting some demo where, you know, out there saying they make AI ads. You know, if you scratch the surface with your fingernail, the thing falls apart. If you actually use it or talk to people who've used it, it's not very good typically, because it's really, really hard. So back to your question. Yes, AI is changing things. I think it's happening slowly right now, despite what you see on Twitter. And then it's going to happen really fast. So right now, if you want to go make a video ad with AI, there's a few things that work pretty well. Like you can Go work with an avatar you mentioned, right, from one of the sort of avatar talking head as a service companies and build something. Mostly you'll do most of that work yourself, whether it's in CapCut, whether it's just, you know, in Premiere, you'll have a green screen, you'll decide what shots go in the background. It's going to take you quite a while, but at least you can get an actor that looks like you want to look, sounds like you want them to sound without having to go, you know, send a product to somewhere and do the PayPal and have a bunch of meetings and briefings. It's like you can really control them. That's pretty much there. It's still the lip syncing is a little uncanny, but within months that will get just really, really good. Beyond that, in terms of, you know, we mostly orchestrate using AI. So we take All of your footage, our footage, so we are unique in that we have the largest vertical video library in the world of native UGC type shots. We've built and built and built that over years at ReadySet, and it's become, it's part of the IP that's in AirPost and it's become really useful because You know, bad footage in, bad ads out. Like if the customer doesn't have enough good footage to fit every type of script, it's really actually hard to make a good ad. So by having hundreds of thousands of clips of, you know, people having trouble sleeping, sleeping well, looking in the fridge, you know, on a run, pushups, whatever, basically like stock, but there's really no stock out there publicly available that's like native UGC, you know, we invested millions of dollars shooting all of that stuff for real clients. So I think the combination of those things is what's working best now. And even then, like we're on a journey to get better. We have a Notion document called 99 Problems. That's all the big and small things between the automated sort of orchestrated ads our AI engine can make at AirPost and like the ads you see on Foreplay or Facebook ad library from top brands, like what's missing. And we're just cracking away at that every week. Speaker 2: Yeah, I think something that you mentioned there that is super, super important to bring forward is the shooting and the footage that is very important to come up with all this content. And the reason why I mentioned that is that before AI was as powerful as it is right now, You could have the argument and sometimes the excuse to say, you know, I'm not able to really run successful ads, so I cannot successfully scale my brand because I don't have the resources to do a shooting, a high model, so have appropriate, you know, kind of a studio. But now that excuse is out of the window because as you're mentioning, live within You know, I mean, it's using solutions like Airpods, you have amazing footage that you can essentially repurpose. And it's just a matter of, you know, how you can tailor and connect that to your avatar, right? Speaker 1: That's exactly right. And, you know, again, most of our ads don't have avatars. Like, I think it's having the avatar in there makes it difficult. Product footage is still really important. So, you know, smaller companies often have some product footage or product photos. I would give people the advice, like, pick up your phone. iPhones have beautiful cameras now. Go outside where, you know, God's lighting is always going to be better than whatever $10,000 worth of lights you try to buy inside to pretend to be the sun. Put the product outside in a situation that makes sense, if you can, if it works, and shoot with your phone. Shoot lots of different hero shots, we call them, of your product. That's really what's needed for these kinds of engines. And then, like you said, yes, you could use an avatar. You could use just different shots that you have orchestrated together. Solutions like ours, you know, use computer vision to understand all the footage you have, all the footage that we have. We can talk a little bit, if you want, about AI footage. Like, you know, you could take a photo and animate it with AI. You can make a photo, which I can explain, and bring it to life with AI. All of these are like major ingredients to the soups that we're making. Speaker 2: And I think something that I would like to hear or take is sometimes I keep hearing, you know, the feedback from people say, yeah, but sometimes people can tell that this footage has AI embedded or it's not 100% like real footage. Like from your experience, and I'm sure you're doing this now with hundreds of clients and brands, like, do you really see that the transition into content being generated with AI creating some kind of phenomenon where people now want a more like organic content or is creating some kind of maybe We can call it like uncertainty when it comes to shopping because now people don't know what is true or what is not true. If you see what I mean, like everything now can be fake with AI. So have you seen any shifting behavior now with this new trend? It's still early. Speaker 1: You know, a couple of things to note. One, most AI avatars, AI lip syncing, AI made footage from scratch, generative AI is still a little weird. A year ago, it was triple weird. And two years ago, it was completely weird. And next year, it won't be very weird at all. And the year after that, we won't even be talking about this. So that is affecting this answer to your question. Because it's still weird to just have AI footage, AI talking heads, we are focused not on like, to be clear, Virginia Post, you're not seeing any AI footage, you're not seeing any gen AI weird stuff that was made with AI. We're taking all of the footage the customer has. This is all real footage, real actors, really shot with a real phone. And we are coming up with the idea. What do we want this script to be about? Oh, it's a voiceover about gift-giving. And the voiceover narrator, her brother, always gets her terrible gifts for Christmas. And this year he got her something that really surprised her because she said a major problem she didn't even know that he knew about. And you lean in and you want to hear the story. And then it's the editing. And it's the shot selection. And it's the emojis and the captions. It's orchestrated with AI, but there's nothing in there other than the voiceover. That's AI, right? And the voiceover technology is already not weird. Like, it's like 94% of the way there. It's weird sometimes, but really rare. So that's what we're doing now. And our philosophy is to wait until those other things you're alluding to, like the avatars and the generative AI footage, get better, get almost perfect, get almost indistinguishable before we start throwing a lot out there. Because throwing a lot of that stuff out there looks really cool on Twitter and LinkedIn and little video snippets, but in market for real advertisers like we work with, it's not good enough for them. And I don't blame them. So we're trying to just ride the technology waves as they develop. Speaker 2: Awesome. So now. Something that, you know, I get this question a lot and this now goes, going back to avatar and all of that. The question that people come to me and say is, Vincenzo, like, okay, like now I have these amazing tools that can create, like, very crisp content, things that are really engaging, like, you know, they're really now making my brand look like a, you know, like a very premium, successful brand. But I'm trying to figure out what are the elements in terms of, The transition, the song, the font, whatever the context behind the video that is achieving virality across the market. So is there something that you give us an advice or as a guideline for people when they're creating all this AI content? How you come up with the piece of content that AI can come up back to you that could potentially reach virality? Like how you identify those elements, if you see what I mean. Speaker 1: I definitely see what you mean. The simple answer is those elements don't change just because AI is here. You still need a hook. You still need to see this product in the first frames. What is this thing for? Do I care? Is there some compelling and interesting hook like headline there? I'm in the video ad. Am I being educated? Am I learning something? Having an ad saying, you know, you're going to love this. It's a game changer. I'm so happy about it. That's not useful. Having something for a skin cream that says, you know, this is the only one that, you know, we use. So-and-so oil for moisture, and only 2% of products use this, but it was just discovered in a paper last year, and this one uses it, and it causes the pores. I'm leaning in. I want to know, what is this? I'm learning something here. Or they say that in a sale, if you can identify with a problem, you've made 90% of the sale, no matter your search. So if you're talking about, oh, you have this problem with, again, some skin condition, dry skin, and you feel embarrassed and you wonder why you're in a conversation, is my skin flaky? If you could relate to the person, well, that means having a good script. So none of those stuff changes, Vincenzo. You still need to have authenticity, which is really hard with AI, especially AI-generated footage. You still need a hook. You still need a call to action. You still need to educate and bring the person along down the funnel. And you don't want to lose them. And I think if there's one word in all the paragraph that's the most important, it is authenticity, which is ironic because we're talking about artificial intelligence. So there's artifice there. And that's why I think that we need to close that gap. And that's what we're trying to do between what looks and feels real and And right now, manual is winning there. Like, sending your product to a real person who has this real problem, who understands the product, really used it for a week, is talking about it, is definitely miles ahead of AI right now. I call it the last five miles of AI, but we're trying to close that gap. Kind of like self-driving cars, right? It's like, they drive, you see lots of videos on Twitter, but they're trying to close that last mile for that really weird situation to make it all the way there. That's where we are with AI, even further back. Speaker 2: Love it. Now, I guess in terms of split testing, because at the end of the day, especially with videos and content for ads, it's all about split testing, trying to figure out what is effectively the piece of content that can consistently bring you the right results when it comes to growing ads. So with the clients that you're working on and helping you split test all the different pieces of content using AI, Is there any level of frequency or piece of content that you recommend people to generate on a constant basis? What is your advice on that to come up with that kind of balance? Speaker 1: Meta has advised that you need to be testing at least 50 different concepts, five zero different concepts a month, even if you're only spending 25k a month on meta. Almost nobody is doing that. Almost everybody is still running the same seven static ads that sort of have been working best since last November. You know, smaller advertisers, it's really hard. We work with smaller advertisers and bigger ones, particularly bigger ones when it comes to anybody, when it comes to iterations and variations and split testing. Yes, you want to have multiple things that you're testing at once, especially the opener, the headline, you know, to do four different things with aimed at different persona with a different value proposition. Find sort of what works a lot faster than if you only do one. That's something we're doing at scale for large companies. So you know, Walmart has hundreds of thousands, millions of SKUs, making a video ad for each of them, making a video, we're not working with Walmart yet, but you know, making an ad you could imagine for any random coffee mug that has A voiceover, talking about it, why it's valuable to them, call to action. Like right now they probably have some GIF or something with some pink background or a blue background that slowly moves and it says a line. A much more dynamic ad, split tested with different orders of the shots, different personas that it's speaking to. That's 10x more efficient in terms of ad spend than the old way that we are in today. Speaker 2: Awesome. Now, how seasonality plays an effect here when it comes to generating UGC content, right? Because something I was talking with a person a few weeks back is like apparently the proper way of doing it is ideally you should be adapting this, you know, if, for example, your product could be good for moms and then you have to adapt it for Halloween and then you have to adapt it for summer. Things like that. So I guess that comes also a little bit aligned with what you just mentioned, which is you need to proactively be revising your content. So when it comes to using and leveraging AI for that, do you see also like a different approach of attacking this methodology? Like you can be more fast and all of that. Like what is your take here? Speaker 1: Yes, absolutely. Like even big companies with big creative teams, we're working with one right now, and big agencies, They told me it takes about three to four weeks from idea or like, oh, we should do something for Mother's Day or summer or and having the actual finished. If we can compress that time from 30 days to three, or to three hours, or to three minutes, that abundance will have a big effect. The way we're doing it at AirPost is a little different than everyone else. Everyone else is doing, and there's pros and cons to both, pros and cons to both of these. Everyone else is doing what I call one by one. So, you know, here's a platform, here's how to build an ad, and you spend 20 minutes, two hours typically building one ad. We're doing many. We call it an instant gallery of hundreds of video ads. Like you said, if Valentine's Day is coming up, there'll be some that are about Valentine's Day without you even asking. It can understand your ads, understand that that's something that you do. You can put in that you have a sale for Valentine's Day, 20% off. You can tune all of this in our platform and it can do that. So I agree. The speed, cutting the speed down or making things a lot faster will be very important. Speaker 2: Cool. Now, in terms of what do you see the future of AI going with this is like, of course, now, I mean, we're getting as close as you mentioned that, you know, the avatar as an integration with, you know, your original content combined with a stock content is getting more and more real. Like, is there some exciting technology or trends that you see that are happening? I mean, you are on a daily basis sourcing this. I'm sure you see things coming. In coming to the action faster than traditional people, is there any trends or predictions you would like to share with the community that you think, OK, I think people should be aware of what is coming when it comes along? Speaker 1: Yeah, I mean, the biggest thing I tell people is we'll talk about what's coming and how fast things are coming. And it's hard to predict also, like, you know, month to month. There's so much hype and discussion of AI because it is going to change everything. It's the internet in 1994. I mean, it's coming. It's going to change everything. It's going to change everything faster than the internet did after the early 90s. And even more so than the internet at that time, there's so much awareness, there's so much discussion of AI that one might think as a marketer, for example, or a founder or an ecommerce company, oh man, I am so behind. I don't know. I'm using ChatGPT. I hear there's all this stuff I'm supposed to be doing. I read about this guy who uses AI for everything. Oh, how is he doing it? Don't fret. It's not you. Most of these things are not fully baked. Most of these demos that you do from different companies are not going to work out. Just like trying to, I don't know, have an ecommerce company in the internet in 1994, right? It's just not baked yet. But it's going to be amazing. And so your question was, what's coming? I think, again, voiceover is probably the furthest along. You would think that text would be really far along. And it is far along. But it is amazing to me that that's another good example. It's been going the longest. It's been going years now. AI text, make me a blog post about this, educating someone on that, talking about this or that feature. And anyone who's tried to write those posts and SEO stuff at scale, it's really hard to make it seem authentic. It's not there yet. And so I think that should be a message to everybody like, oh, that still has five miles to go and that's years old. It's been good. Video, images, again, it's like it's tantalizing, Vincenzo. With 4.0 now, you can take an ad and make an image ad and you're like, oh, that looks pretty good. But when it's actually your butt on the line and you're spending a million dollars or a thousand dollars of your own money and you've got to make these ads work and you don't want that CMO saying WTF, this business is off-brand. Hey, where's the fine print the lawyers say? It's got garbled or the AI. It's not there yet. So what's coming? You know, I think it's true what Zuckerberg said recently in their earnings call, like someday you will just give them Business goals, and if you really trust them, which I think people will, as much as they say the opposite now, they can go make ads for you, do the creative, do the media buying, do the reporting. That's a lot of trust, by the way, but I do think people will do it if they can see the sales. They're just owning the outcome. I think other platforms will try that as well. That's years away, years away from now. Even Google, as someone pointed out to me, I'm no SCM expert. For years now, they said, stop trying to pick keywords. It's not 2015 anymore. We'll do it all for you. And any self-respecting SEM leader or agency doesn't do it that way. They don't trust Google to do it. It doesn't do it perfectly well. So all these things are developing. I can't tell you which one is going to land first. Video still has a ways to go. We're not even generating audio with an actor yet. I mean, we're starting to, but it's really bad. But it's fun to watch things happen. Speaker 2: Yeah, for sure, for sure. So I guess, you know, John, it's been a pleasure to have you. And I know, you know, that's everything you specialize on at Airpods. So tell us more about, you know, how people can work with you guys, how you guys can add value in that department. Speaker 1: Yeah, thanks for asking. I mean, we're a really young company. If you are an ecommerce company that has a physical product you can hold in your hand, And you need more video ads and you're exhausted from working with agencies and churning through influencer platforms and all this stuff and spending a lot of money. I mean, for hundreds of dollars, not tens of thousands of dollars, you can get a huge gallery of ads with real people really shot with real cameras talking about your product, seeming to talk about your product. You might see them, you might hear them. It's someone else, but you can get a lot of ads that are good and getting better every week. Yeah, just go to airpost.ai and onboard. You can check it out for free and see what it looks like to actually get the ads. It's hundreds of dollars a month. If you email me, john at airpost.ai, I can get you half off the first month, but we're very much in a feedbacky phase right now because again, Well, I'll put it this way. I don't think any marketer is going to use AI tools, whether orchestration, avatars, whatever, to replace the ads they're making today manually, unless the quality is 95% of what they're used to, at least. We're not at 95% yet, and we're better than anybody right now. So, you know, we need your help. We'd love your feedback. And yeah, just come to airpost.ai and let's jam. Speaker 2: Thank you John. It's been a pleasure. Thank you so much for the value. For everybody looking to find more info, find it down in the notes below. Other than that, see you in the next one. It's been a pleasure. Thank you so much for your time. Speaker 1: Thanks Vincenzo. Speaker 2: Thank you. Bye bye. Unknown Speaker: Thanks for listening to The Ecommerce Lab By Ecomcy. Be sure to subscribe so you don't miss an episode. While you are at it, we would appreciate it if you could leave an honest rating and review on Apple Podcasts, Spotify, or wherever you listen. That will make it easier for others to find out about the show and benefit from it. Want more? Visit our website at www.ecomc.com where you can get your first consultation for free. Or find us on Instagram, Facebook, and LinkedIn at ecomc.