Our guest on the sellerboard show was Anthony Lee – AI automations & psychological marketing specialist. Anthony has also been a consultant, brand owner and author on the topic of scaling a business leveraging the Amazon platform.
We talked about:
- what is AI? Chat GPT, Open AI tools;
- list optimizer bot: Title, Bullets, Description;
- AI and images, A+ content;
- challenges and opportunities for 2023;
…and much more!
What the full interview on our Youtube channel!
Speaker1: [00:00:07] Hello, everybody. Welcome. This is your host, Fernando from sellerboard. I have a very special guest for you that’s going to talk about AI on Amazon. It’s going to be very cool. It’s very interesting and you’re going to love him. But before we start the interview, you have to go to sellerboard.com and test out our software. If you have an Amazon store, you’re going to need a software just like sellerboard. Go there, click on the demo, play with the software. You’re going to thank me. Ladies and gentlemen, I represent you now, Anthony Lee.
Hello, everybody. Welcome. I have a very special guest. I have the pleasure to work with him a long time ago, and I’ve been chasing him down on LinkedIn to get here on the podcast. He finally said, Yes. Ladies and gentlemen, I present to you Anthony Lee. He’s the content strategist for Analytic. That I got it right? Yes. Perfect. Yeah. Anthony, thank you so much for being here.
Speaker2: [00:01:11] It’s a pleasure. Thank you so much for having me. Sorry that it took so long for us to finally connect. Life’s just been crazy.
Speaker1: [00:01:20] No worries. I’m glad you have your own point. And I’m sure you’re going to give us a lot of good content like you always did. Yeah. So go ahead.
Speaker2: [00:01:28] I will try. I appreciate the vote of confidence.
Speaker1: [00:01:32] All right. So first question, as usual. Tell us your story.
Speaker2: [00:01:38] Okeydokey. So. 2014, I was a copywriter who commuted an hour one way to work every day, and then on the weekends I worked at a restaurant. So I was working many, many hours and I wasn’t making a whole lot of money at all. And then all of a sudden, I just randomly ran into a video about Amazon FBA and Alibaba and just the light switch came on and I jumped headfirst. I didn’t have any money, so I bootstrapped my entire business on credit cards. Literally, like had to. Here’s a hack for anybody that’s broke. I couldn’t pay with a credit card for my inventory from China, so I had to set up a credit card with my wife’s PayPal account and have her pay me so that I can then take the cash out and give it to to my Chinese supplier. And then when I was up and running, I did really well, sold like $25,000 worth of product my first three weeks. That went super well for a while. And then from there, because of my fast success and the fact that I was kind of figuring some stuff out, my mentor hooked me up with some consulting clients at the exact same time. Another friend of mine wanted to start a service which later became a SAS known as Zen Blast. And I was the first employee. I rose up and then some time passed, I got rid of or I moved beyond having so many clients and then left Six Leaf, which was the rebrand is on blast and went to helium ten for a little while and then helped Howard tie found Signal lyrics. And then when that was over, I was looking for a new adventure and that’s when we worked together. And then after that, you know, now I’ve gone back and I’m helping Howard and signal it again and really just kind of tinkering with stuff all the way. 90% of it is ways that can help e commerce, Amazon sellers that still my heart and my love. But yeah, that’s that’s my story in a nutshell.
Speaker1: [00:04:00] Nice. And where are you working now?
Speaker2: [00:04:05] No. Well, aside from doing. A lot of the same stuff I’ve been doing for the past couple of years now, which is marketing and content. I’ve been really falling down the rabbit hole of trying to build. I solutions, right? So we have access to these artificial intelligence tools now, and I have been loving it. So I’m in there constantly figuring out ways to kind of utilize it and in ways that can help people create content, create a. Amazon. Content, for example, listing stuff, ideas, rewrites, etc.. So that’s that’s like been my world probably for like the last six weeks now.
Speaker1: [00:04:54] Nice. Is that that is our main topic of our discussion today, right?
Speaker2: [00:05:00] Yeah. Yeah. That’s, that’s definitely something that I enjoy talking about thoroughly and.
Speaker1: [00:05:07] I.
Speaker2: [00:05:07] Wouldn’t mind.
Speaker1: [00:05:08] Let’s do it. So where do you want to start?
Speaker2: [00:05:13] So I think the best place to start is what exactly is it that we’re talking about, which is we say A.I., but like, you know, everybody kind of knew what I was before in the context of, Oh, you mean like that movie about the little artificial intelligence kid, Right? Is that what you’re talking about? But over the past couple of months. It’s really it’s really been. Very different because this company called Open Eye has come out with public access to a chat bot called Chat GPT three. So it turns out actually that was already available. And how I know this is because I completely forgot. A year ago I signed up. But I and I used it. That was back when the model was called Babbage, I think. And it wasn’t very good. It was, it was normal chat bot stuff like I build chat bots for years and this was just like, Oh yeah, I could, I could build this. You don’t have to have an artificial intelligence for this. And I forgot. And then when chat GPT three came out, I went to sign up and it was like you already have an account. I was like, do I? And I went in and they were like, Yeah, sorry, but your trial is already expired. It’s like, Oh yeah, because I’ve had an account for like a year now, but people weren’t using it. It didn’t blow up because there wasn’t anything unique about the older versions of the chat bot. They were very similar to anything that you could mechanically or manually kind of code into like a mini chat, right? But then GPT three came out and now it’s taken the world by storm because this thing is way smarter.
Speaker2: [00:06:53] It’s just way, way, way smarter. And everybody is kind of blown away by it because it’s not just like some silly chat bot anymore where you’re asking it weird questions and an answer. It’s like it’s, it’s giving you real advice. It’s, it’s giving you context on things that you can that you need help with. Like if you don’t know how to code. I did this the other day. I had a widget and I’m trying to center it into a WordPress page and the widget like it just it’s not cooperating. So I went to chat GPT three and I said, Hey, I need to center this in my WordPress page. Can you like help me figure out what to do with the code? And it added code to it and said, Here, try this. And it worked. It’s fricking crazy. So anyway, so that’s what, that’s what we’re talking about is, is this because there’s a lot of tools now, there’s a lot of tools and early adopters for probably the past five or six months have been using things like Copy, AI and Jasper. But now we have this. Less expensive for now, free for most people. When you sign up, you get three months and seemingly more powerful option that that’s that’s really kind of pushing this whole narrative of utilizing AI to its fullest capacity to the forefront. So. Yeah. Sorry I could rant about that forever.
Speaker1: [00:08:28] Do you think it only became popular because Elon Musk tweeted about it? Or do you think it was already popular before he tweeted?
Speaker2: [00:08:37] No, I don’t think that his tweet so interestingly. There’s a bunch of science behind, like how ideas and things go viral. Most of the time, the biggest figures in the center, they actually are. They don’t. They don’t catch on. So, okay, here’s a here’s this is this doesn’t have to do with AI, but it’s an interesting brain fact. So it turns out that we humans are influenced by our network, but it’s the percentage of our network, not the number of our network. Right. So if I have ten people in my network and a 40%, 40% of them doing something would be enough to sway me, then I need four members of my of that network for that to work. When you have a huge network like Elon Musk does, he’s not even going to pay attention until like a bunch of them are talking about it. By the time a bunch of them are talking about it, it’s already popular. So a bunch of people were already talking about Openai and Chad GPT three because it was blowing minds and then you get celebrity endorsement and now it’s even bigger. But yeah, no, it was on its way there regardless because the technology is mind blowing.
Speaker1: [00:09:46] It’s brilliant, right?
Speaker2: [00:09:48] It really.
Speaker1: [00:09:48] Is. What is the craziest thing that you have or tried besides the widget?
Speaker2: [00:09:54] Uh, I’m. I’m doing. Everything with it. Now, I don’t know if any of it would be considered crazy, but it’s cool, so. I built a bot that automatically at the push of a button, it automatically generates content ideas. Right, because it’s it’s the eye is the best for for research because it has access to so much data, so much information. So it’s really good as a brainstorming tool, right? So I ask it for ideas, and I built this in a no code application I love to use make, which is formally integrate that and essentially write out prompts. And that’s something else that I do a lot too, is I just test different prompts like, does it work better if I use a lot or a little information, How many examples do I need to give it, etc.. And a prompt is the is what you tell that I for it to give you information back what they call a information. So I built that. So basically how it works is I have this app where I create a button and all the button is, is a web hook, right? It just sends a signal. So the button lives on my phone, I connect the button to open. I sorry, I connect the button. Yeah, I connect the button to a database that has five of my favorite topics to talk about and then that is connected to a randomizer so it can pick one of the five. Because I learned the hard way that open. I doesn’t do really good with random like if you give it a list and you say pick one at random, it’ll probably 90% of the time we’ll pick one number one. So I had to build a randomizer into it.
Speaker1: [00:11:59] And then.
Speaker2: [00:12:00] Connected that to the AI. And it’s got this prompt where it says, Hey, give me five ideas for content I can create on this topic. And then every time I cook the button and it sends the message to a Discord server, because these are I used all free tools to do this and click the button. Boom, my ideas pop up in my my discord and they’re pretty robust too. I like it a lot. So that was one thing. And then just today I built an auto Twitter bot. So I have this Twitter account that’s it’s really into crypto and I connected some feeds from crypto news, their blogs, and then I take the summary of the of the blog and I put it into the AI and I tell the AI, like give me some commentary, use the same tone as these examples. And I gave an example of like three of my, of my original tweets and now. Twice a day. It’s going to tweet for me using my tone of voice, a commentary on the news from those from those blogs.
Speaker1: [00:13:13] That’s crazy. Where did you get these ideas wrong? I just.
Speaker2: [00:13:25] I’m a tinkerer, man. I’m a tinker. I just tinker. And I’m like, What else could I build and what can I do? Like, could I automate that somehow? So I think the biggest and coolest thing that I’ve done, which is relevant to this conversation is I built an A listing optimizer bot.
Speaker1: [00:13:43] Right? Oh, yeah. Yeah. Now we’re talking. Let me write this down. Listing Optimizer bots. Yes. So that’s what I wanted to go to. Money.
Speaker2: [00:13:56] Yes. So. Basically I use the no code applications and then I built out I use an interface and it’s already had two iterations and I’m working on getting better and better or better. But essentially through the interface, it just asks you what your product is in the simplest terms, what your brand name is. And then what your primary keywords are. So there’s a lot of assumptions in this first model. I hope in the future to have a model that automatically can pull keywords from competitors, but that requires some coding and I don’t code. This is the stuff I’m doing with no code, but you put in your keywords and your product and. The AI creates a title using the structure of a handful of examples that I’ve given it from across categories. It doesn’t use the the content of the title. It uses the structure just so it knows. Like this is about how long an Amazon title should be. This is this is the structure, right? Like we use dashes or bars to separate ideas and and then after that it takes the title, it knows the title, and then the next prompt is, okay, this is the product using the same keyword and this title provide five bullet points, make them a minimum of I think I put like between 140 and 250 characters and and make it benefit driven copy, and then it pushes out copy based on benefits because it had access to the World Wide Web up until 2021.
Speaker2: [00:15:37] So it’s going to know about products like nothing that’s brand spanking new. But anything that has existed in the past, it’s going to know how to describe it. Benefit driven. And then the final prompt is and here’s the title, here’s the bullets. Create a description, mention the brand name, make it personal and personable, and then it writes, you know, up to 2000 character, product description. And there you go. So what this gives you is a workable template. It doesn’t mean that you have to copy and paste it word for word, but there’s a lot of people in our industry that don’t feel comfortable writing copy. They don’t think that that’s where their biggest skills lie. So I create a tool like this, for example, to help those people, right?
Speaker1: [00:16:26] You don’t know any confidence. If you go to apparel, you have like hundreds and thousands of hundreds of SKUs. So that’s the time saver for that.
Speaker2: [00:16:34] Yeah, well, exactly. That’s that’s another thing too, is, you know, exactly timesaver, creativity, help. And then you just go in and you make minor edits rather than having to write everything from scratch. You just go in if you need to make any edits. And the cool part is, is if you don’t like what you got, you run it again and it’ll give you something completely different because the AI doesn’t ever give you the same. It depends on how you set. There’s a little thing called temperature in open AI and it tells the bot how to be creative, how how creative to be. And if it’s set to zero, then it’s just going to give you the same thing over and over again. But if you set it to one which is the max, it’ll give you a different answer every single time. So you just use it until you.
Speaker1: [00:17:15] Get Are you, are you are you using the page version?
Speaker2: [00:17:21] Okay. So I have to because my because my trial ran out. That’s on my personal stuff. I actually built this particular bot using the signal Linux account. But we’re real close to being run through our through our because we’ve been we’ve been testing it. There’s a lot of people that are going through it. So we’re real close to running through our free trial anyway. And then after that it’ll be. It’ll be. It’s really inexpensive, though, Like really inexpensive. Like I think producing one listing is a roughly. Like a thousand, maybe 1500 tokens, and you get in and that’s like $0.06, right? Because you get like 1000 tokens for $0.04. So it’s like $0.06.
Speaker1: [00:18:08] Cheap and then 1000 and then tokens and then it turns out to be $0.06. Oh, my God. Yeah, Yeah, I can afford that.
Speaker2: [00:18:18] I really think that everybody, everybody can. So it’s it’s really cool and it’s accessible to everybody. And I, I want to encourage more people to play around with it because these things are, you know it’s they’re they’re fantastic so chat GPT three the chat not that open I I use open I because that’s the only one that has an API and you have to use an API to automate stuff. But if it’s just you chat. Gpt three is free. It’s free, totally free. Right now they’re not. They’re going to monetize it, but for now it’s free. And so you can actually just go in and start create an account. They do not require credit card or anything, just go to beta, dot, dot, dot com and create an account and you can go to chat attribute three and open up the chat window and then. One, two, one. You can work with it, right? So you go in there and you say, Hey, here’s an example of like a really good competitor title to it. It’s iterative because it’ll it’ll talk to you the whole time. So you go in there and say, Hey, this is an example of a really good title. I sell the same product. Can you help me? These are the keywords I want to focus on, and then it’ll take that and it’ll go, okay, well here, how about this? And then you say, I really like that. Can you provide me some bullets and then spit back the bullets? Most of the time they’re too short. So then you go back and say, Hey, can you make those a little bit longer? Can you can you include some benefits in all of the bullets and then it comes back with something longer and then you can do the same thing with the description. It’s all iterative, right? The eye remembers a little over 8000 words of your conversation right now. And that’s like almost a novel, a small novel. Right. So you can have a lot of conversation with this thing before it starts forgetting what you guys are talking about. And then when it does forget.
Speaker1: [00:20:09] I thought everything I had to include include everything, all the details. You know, you would catch Cappy catching up with the conversation.
Speaker2: [00:20:17] Yeah. Now it’s GPT three. You have to include everything if you use the open eye playground. But with chat. No, it’s all there. It’s in the history. Up until a couple of days ago, actually, it was keeping a log of all of your history so you could even go back and continue a conversation you had with it days ago. But it got overloaded and then that wiped out all the history. Now, it started over again, but it’s it logs the history and every instance has a memory of over a little over 8000 words. When chat fork comes out, it would be way bigger. But right now. A little over 8000. Still, that’s enough to get an Amazon listing. That’s enough to get an Amazon post. That’s enough to get literally anything that you could possibly script for an Amazon video. Like anything you can make on Amazon, you have enough in that instance to have a back and forth mastermind communication with AI, which is, in my opinion, the smartest friend you’ll ever have and then come up with something really good.
Speaker1: [00:21:22] Yeah, I like that idea. You talked about posts. I was thinking about social media posts. You can give. Like, ask me. Give me an idea to put something on social today. That’s so cool. I have to choose is more important. Question How do we use this I for listing on Amazon.
Speaker2: [00:21:43] How do you use the the three.
Speaker1: [00:21:47] Yeah. How do how do we access. Do you have is there a way we can access the vault that you built. Oh yeah. Just go through it.
Speaker2: [00:21:55] If you want to try the bot that I built. 100%. Absolutely. So site analytics dot I is the website and then it’s forward slash. Ap wl dashpot. So I call it app low because that’s Amazon product listing Optimizer right? Apolo But.
Speaker1: [00:22:14] Okay.
Speaker2: [00:22:15] If you go there, you’ll see the little chat window, the widget window that I had to use chat GPT three to help me figure out how to send her. And then that is the widget that takes your information. Three keywords, your product. And then and then you get once you put in your email address, you get emailed the, the listing and you can do that however many times you want, right? For however many products you want. Right now it’s for signal.
Speaker1: [00:22:43] Eric K Right.
Speaker2: [00:22:46] Dot I signal. Dot, dot. I do. You have. Yeah. Let me put it in the chat for you.
Speaker1: [00:22:53] Oh, got it. It’s okay. Think analytics dot I and, and this. Yeah. The first one is the I, then a Y, then I, I’m going to put that on the link below only or until.
Speaker2: [00:23:07] I just put it on the I just put it in the chat.
Speaker1: [00:23:09] So anyway that’s going to be a lot easier.
Speaker2: [00:23:12] Right now it’s free. We’re. We’re going to take it down eventually. But right now it’s free. And your audience. Absolutely. You guys can use it. And then I also encourage, you know, because here’s the thing with the tools like this, and then if you use any of the currently available commercial tools like Jasper or whatever, a lot of them, they access the same the same technology, like it might be open. I it might be. There’s a couple other ones that are open source and really all they do is they just hardcore hard code. They’re prompts, right? So you go in and you pick stuff from the dropdown that’s just hard coding there prompts. If you want to practice your own prompts though, go to chat GPT three and start talking to it. You’ll figure out here’s a cheat code, ask it what prompts you should use. So this is the conversation I would have. Go to chat GT three and say Hey look, I sell. An olive oil sprayer. Right. I am looking to optimize this. With persuasive copy on my Amazon listing. What prompts should I use? To get openai and chat GPT three to provide the highest quality output and then it’ll tell you use these prompts and then you go and use those prompts say, okay, this is because it’s going to ask you for information. That’s basically what you’re doing when you say What prompts should I use is telling you, okay, well you need to provide me with this, this and this, and then ask me to do this.
Speaker2: [00:24:47] And then you follow those instructions and then do it one at a time. Like do your title. Don’t do it all at once. Do a title Bullets description. If you’re looking to optimize or if you’re talking about social posts about it, you can say, okay, this is my target demographic. Follow the instructions that it gives you for the prompts and then ask it like, you know, can you give me some ideas for Facebook posts, for Twitter posts, for YouTube? So if you’re doing videos, for example, asking for ideas on what the video can be about and then turn around and ask it for like a like a script or storyboard on what what you should do in that video. Once you choose what you want it to be about, then ask it what you can do in the first 3 seconds as a hook to get people’s attention. It will give you all these ideas and then you just go in. Now you’re like creativity to the max. You can go in with a piece of paper and say, okay, this is how I want it played out. I assist. It’s the future. We’re going to have AI assist us with everything this is.
Speaker1: [00:25:49] I have. I have so many questions in my head I can even. First question, what is the percentage of humanity that you think that understands how this is going to change everything? Like one two years?
Speaker2: [00:26:04] I think there’s this is actually a really good question. Thank you. I think I think there’s a lot of skepticism right now. And the reason is because a lot of people are resistant to big tech changes for a number of reasons. Right. The media always paints big tech as the big bad guy, right? I mean, look, social media changed the world, but we’re still sitting there talking about how Zuckerberg is a terrible because he spies on all of us and Google, you know what I mean? Like, these things change the world. We use them every day, and yet we’re always talking bad about them in the media because that’s just that’s that’s that’s how we are, right? So we already have this distrust for big tech. And then also we romanticize stories like John Henry. I don’t know if you’re familiar with John Henry. This the story of the railroad road worker. Right. So they were they were laying railroad by hand with a big sledgehammer, knocking in the nails, and then the steam engine was built. And John Henry said, steam engine can’t do what I can do. And then they set up a contest and to see who could lay this this line of rail faster and just just by a hair.
Speaker2: [00:27:18] John Henry won. And that’s the story. We’re told this as kids write. But the part that it doesn’t tell you because in the story shortly thereafter he died, but the part that his heart stops because he worked too hard. The part that they don’t tell you, though, is that is that the steam engine continued to lay like 3000 more miles of track, which means the steam engine actually won. But we have this romantic idea that we want humans to be super like super powered. That’s why we love Marvel movies, right? We want to be super powered. And the idea that a machine can come in and do it better, it just doesn’t sit right with us. But what we don’t like. So a lot of people are skeptical. A lot of people don’t understand. That technology is there to hold, you know, to make you better. There was a dude driving the damn steam engine, right? Like it wasn’t it can’t do it by itself. Right. And this is the same thing that I will not be able to do any of this stuff by itself. It needs you. So it’s just. It’s like a cybernetic arm. It’s just.
Speaker1: [00:28:21] Not.
Speaker2: [00:28:22] But a lot of. Yeah, right. Oh, gosh. Right. A lot of people are skeptical. So I think that the majority of society right now, just like everything. They won’t they won’t embrace it until it’s a tidal wave, until it’s too late, until it’s like, oh, well, you know, like, think about it. Think about how many people refuse to get to have a Facebook account. For the longest time, every one of them have Facebook accounts now. Every one of them. Not only Facebook, but now they have Facebook, LinkedIn, Twitter, Right. Like. But there was a time not that long ago where everybody was like, no, I don’t do that. Yes, you do. You just aren’t going to do it until it’s too late. And this will be the exact same.
Speaker1: [00:29:04] All right. I have I have three I have three partners and, like, three men in my head here. Number one, can we do a listing optimizer based on on specific competitors? Like make it better than those competitor?
Speaker2: [00:29:19] So. Yes, sometimes. All right, here’s a thing. The AI is programmed to recognize when like like, I’ll give you an example. I tried to give it somebody else’s terms of service the other day and say, Hey, rewrite this for us. And the first try, it was like, No, I can’t. I can’t do that. You need to write your own like, or else your you might not be compliant with the law. Then I reworded it and tried it again and it was like, Oh yeah, sure. So it really depends. Like sometimes you just have to coax it sometimes and figure out how to communicate with it the better. Rather than say, Hey, rewrite this competitor listing the better way would be. Here’s an example listing This is the basis. This is what I like. All right, now I’m writing my listing. Can you write me a similar one? It’ll do that because it understands that we’re not copying, we’re not rewrite and it’ll end up being a rewritten anyway. That’s just how it works. But but yeah, you can train it, train it right with your examples and you should actually because that helps it not like completely screw up and give you something that you can’t use.
Speaker1: [00:30:29] Wow. Second question. Is there any chance we can create the image already? No, not really.
Speaker2: [00:30:37] So this is something that we’re that we’re working on at CES Analytics. But it’s going to take some time. So right now, most of the AI driven image stuff is like. It’s either you feed it text or you give it an example image, and it’s making art. It’s making art. That’s what it does. It makes art. Dolly is really good at abstract stuff. Mid Journey makes like epic looking, realistic art for like storyboards and for novels, but it’s not specifically designed for product stuff and there’s a lot of people that want it to be, and I think they’re working on it, but they’re working on it, in my opinion, in the wrong way. They’re working on it. Like, can I tell the AI, Hey, I have a black mouse and it’s going to show me and it’s going to make that image. It’s like, But then it won’t be branded, right?
Speaker1: [00:31:31] No, no. But how about how about you get, like pictures of your product? It can be like a 3D render or super cool picture. And you say, okay, you build me a listing with those pictures.
Speaker2: [00:31:43] That’s that’s the plan, right? So the plan that we’re working on is we want to be able to upload an image. So we want to be we want to make it to where it’s like you upload it with your phone, right? You don’t even a professional picture, like literally just take your sample and click it and then go, here’s the image and then the AI. Ideally, what we’re hoping to program it to do, it will will make it pretty well. They’ll realize what it is, right? They’ll go, Oh, okay, I know what this is supposed to be. Let’s let’s, let’s polish it. And then you tell it. These are the scenarios. I want it like I want it in use by a brunette woman in a kitchen. I want it just sitting there, like in a in a sunlit window. And essentially it’ll be like a fake photo shoot. Right? You give it the you give it the prompts, the scenarios that you want the product to exist in, and it’ll provide it to you the same way that it would that a photographer would had it taken those pictures. We’re probably a little ways off from that, I would guess. At least a few months, if not a year, because and I know we’re not the only ones working on it, I promise you, there’s probably 20 other people that are doing it. Something will come out. But as of today, that’s that’s not the world that we live in yet.
Speaker1: [00:32:54] But it will get here. Have you thought about doing? Caption and image Altec Altec true to the A-plus content. Because when you upload an image on A-plus content, you can put like a caption you can put like tag. I you thought about doing that too.
Speaker2: [00:33:17] So connecting. Yeah. I mean, obviously anything, anything text related will be stuff that we’ll be able to to work with. Slowly but surely we’re getting there. A big part of optimization is obviously the A plus two. So that will be in the works in the future for sure. Once, once we can build a more robust UI. And then the ideal scenario is that eventually we’ll be able to like make it so easy that you don’t you don’t have to receive it in an email. It can just be like, All right, well, your your account is connected. Click this button and then it’ll, it’ll publish everything. But that’s all this stuff takes development. And you know how I don’t know if you’ve ever worked with developers before. A developers are amazing. They’re so smart and I love them, but they’re never, ever, ever, ever on time.
Speaker1: [00:34:08] Oh, yeah. So I don’t know exactly right.
Speaker2: [00:34:12] It’s. Yeah, it should be done next, next week. And then we had a snag. Maybe not.
Speaker1: [00:34:18] Yeah. Okay. What about, you know, how soon somebody will come up with a good solution for PPC management? Because all the AI software that we have for PPC management, they need a lot of supervision, a lot of an added contextual analysis. And you can just not you can just like let them run. You know, a lot of times PPC professionals, they be AI software on BBC. Do you think that’s going to change very soon? A little bit. Down the street.
Speaker2: [00:35:00] So I’m biased because Cigna Linux offers an AI tool. Okay, But but here’s the thing. Here’s my thoughts on it from an outsider’s perspective. You’re never going to be able to trust the quality of something that doesn’t have human oversight and doesn’t matter how good AI gets. That will always be the case because AI is a human collaborative tool. It is not a tool that’s made to replace humans. Yes, it can replace humans and some very basic automations, but otherwise. And so for that reason, even the tools that exist now for PPC, like, like there’s always going to have to be oversight. And the less oversight you’re asking for, the more you’re risking that the quality, the output won’t be as quality. So they are constantly, even before open AI, constantly trying to improve. I guess some of the some of the outliers where like bids, for example, for keywords that don’t make sense happen, but for the most part. Eight, the 8020 rule. Right. 80% of what these tools focus on is pretty good. They pick good keywords. And then you got to also make sure you don’t ask it to do too much. Like you don’t want something that’s just set and forget, Here’s my product, make a bunch of ads. What you do want is something that’s all right, We’re going to set these ads up. And then the whole goal of it is to scale out bidding and then move the keywords into like from broad to phrase to exact, so that we can we can identify the keywords that are actually making money, right? So you don’t ask it to do too much because it’s smart, but it’s it’s still like a child. It’s like a it’s like a genius five year old, right? It’s still not going to get things right.
Speaker2: [00:36:54] So you don’t ask to do too much and then and then monitor it enough. And I mean, it can be a great tool because it catches you like I get what you’re saying when you’re like, Yeah, it can be outperformed, but what do you mean by outperform? Like, I don’t think thing metrics like ACOSS and stuff like that are really good in comparison because it’s always situational, the keyword always different and then, you know, the bid type, the, the everything right is always situational and different. Like you could also be a human that gets outperformed by another human this month and the next month it can be flip flop. So it’s no different I think, with the AI tools. So I think if you manage your expectations and you go in there and use it as as the tool that it’s supposed to, what it does help you do is it just helps you automate a bunch of the really small, the micro actions that are, that are going to take you time like no more bulk uploading and and no more more like adjusting bids for a billion different things. No more having to dig through. If you have too many campaigns to make sure you don’t have keyword overlap. So these are the things that AI is good at, right? Because then you just hey, I want you to I want you to get rid of all instances of overlap. And the AI says, okay, I can do that as long as you manage your expectations and use it for what it’s supposed to, then yeah, our experience with AI and PPC is going to continue to get better.
Speaker1: [00:38:19] Okay. I have a I have a stupid question you don’t have to answer. If you don’t want, you can just skip. Can we use have you ever thought about maybe we’re far away from it, Maybe we’re close. But I was thinking we go to an Amazon products and there’s like, a bunch of. Branded products are all the same coming from all the same factory. Maybe I can try to look at the list and say, Hey, I have a unique idea for this product, something that can be completely different and better than everybody that’s just putting their brand on the product.
Speaker2: [00:38:59] 100%. We will see it. We will see it. And I know we will see it because if nobody else builds it first, My my good friend and CEO of Analytics, Howard has been literally wanting to build that since before we knew what Chad was. I’m not kidding. A couple of years ago. He’s like, A.I. is a future and one day I want. An eye that can tell me what products are going to be hot. So, yeah, entirely. I think we’re going to see stuff like that in the future, being able to ideate differentiation because it’s important, right? Like it’s it’s we get so caught up. First of all, humans, here’s a thing. We humans are limited in our capacity, right? We’re wonderful, amazing creatures. We have huge brains, but we’re limited in our ability to juggle a million things. Right? So when you’re learning to sell on Amazon, you’ve got to figure out how to do the accounting, how to do the sourcing, what you have so much going on to sit down and also simultaneously be able to ideate a million different ways to differentiate. Can be difficult.
Speaker2: [00:40:07] And then like you might pick one or two ways to be different, but at the exact same time, there’s 20 other people who thought of the same thing. And so when you end up launching, you end up launching with 20 other people a different version of the same product. And it’s like it’s just it’s, it’s a tough game. So yeah, I assist will be able to help in situations like that because it’ll be able to provide you with statistical analysis that you don’t have access to and say, Hey, look, we’ve looked at the size of all of these products and this has the highest demand, but these are things that the reviews say that indicate that it could use improvement in these areas. This is a color variation that has performed really well, but only on one other brand like these are the kinds of stats and data that you’ll be able to pull. And I fully believe that you’ll see that one day, and that’s going to help in this process tremendously. We’re only going to get better. It’s going to be great.
Speaker1: [00:41:03] Mm hmm. Perfect. If you ever get close to something like this, please let me know.
Speaker2: [00:41:09] Yeah. No, for sure. I will publicize everything that we’re doing with. With I constantly and myself. Right. Like, just look at my LinkedIn feed. I’m going to share later. My Twitter bots.
Speaker1: [00:41:23] Yes. Cool. Nice. Anthony, what keeps you. What’s. What’s keeping you up at night and what makes you excited to get out of bed in 2023?
Speaker2: [00:41:36] It’s definitely been trying to build more I automations. I just like I literally have a list of my phone of, Oh, I can make a bot I want to make I want to make a copywriting bot. Right? So you know how like copywriters have like all these different formats, right? A To pass, like these are formats the way that you structure. I want to build a bot that has those built in. I bought the has this built in. So then you just give it like your scenario like this is what I’m trying to, to sell or this is what I’m trying to optimize and it’ll just spit it out and be like, well, here are the formats. Which one do you want? That’s what keeps me up. Like literally, I paced the kitchen thinking about this stuff and then the next day I’m like, How can I get work done? And also like, go tinker and try to build this bot?
Speaker1: [00:42:21] Mm hmm. Mm hmm. I, I can imagine already you were like, you were studying you really into behavior of purchasing. Right? So I can I can only imagine the things that are going to be doing. 2023 is going to be very amazing.
Speaker2: [00:42:38] I appreciate that vote of confidence. I hope so, because these are definitely passions of mine. I look since 2017, I’ve been building chat bots in an attempt to replicate human psychology so that the consumer to bot conversation leads the consumer down a comfortable path where where they still connect with your brand without you having to be present. So I only makes that easier. But literally this is something that I’ve been doing for many, many years now because it’s fascinating to try and automate some of those psychological triggers. So yeah, I’m excited about it too.
Speaker1: [00:43:16] Nice. Do you have any final thoughts for us?
Speaker2: [00:43:22] Yeah. I just want to encourage everybody to have an open mind about the advancements that we’re we’re seeing in this technology. Remember that every tool that comes out is an opportunity for you to leverage it as a business owner. That doesn’t mean you have to use all of them. Gosh, don’t get rid of the shiny objects. But don’t be afraid to leverage something that’s going to make your life as a business owner easier, whether that be A.I., whether that be chat bots or other kinds of automations, or if what works for you is still using a piece of paper and a pencil. It’s actually been proven. Retention is higher whenever you write something down with a pencil. Whatever form of technology is going to help you, I definitely recommend find that and leverage it and make it your strength. Double down on your strength and higher out your weaknesses. That’s that’s that’s my advice.
Speaker1: [00:44:17] Good. I like that. We like that. And how do people connect with you, Anthony?
Speaker2: [00:44:22] Well, you can actually email me directly at Anthony Ask Analytics Dot I or I’m also psych marketer pretty much all over the internet on the social media. Is Twitter LinkedIn still have the TikTok YouTube. I’m actually posting stuff about I now on those so I’m not hard to find.
Speaker1: [00:44:44] All right Anthony, thank you so much for your time. I really like the conversation and I really hope to have you again in the podcast.
Speaker2: [00:44:53] Absolutely. Thank you so much for inviting me. It was it was an honor.
Speaker1: [00:44:57] Thank you, guys. I’ll see you guys in the next interview. Thank you, Anthony.
Speaker2: [00:45:01] All right. Take care.
Speaker1: [00:45:07] And that was our interview with the Supper, Anthony Lee. I hope you guys liked it. Please. Now, do me a favor. Go straight to the sellerboard.com. Test out the software. Click on demo. Play with all the features that the software has. If you have an Amazon store, I’m telling you, you’re going to love the software. Let’s go to sellerboard.com right now and test it out. I’ll see you guys in the next video. Thank you so much. Bye bye.
