Passing The Torch #09: The Future of Marketing and AI, with Mike Taylor

Marcos Bravo: Hello, everyone, and welcome to the Passing the Torch podcast by Growth Mentor. My name is Marcos and as you can see, I am not in the studio. I’m also taking a little time off to visit the family back in Chile. So yeah, what you can see in the background is something nicer than my regular place. Okay, that being said, Today I’m bringing you a very golden nugget is one of the best interviews I’ve done with one of our mentors, talked to Mike Taylor. Mike is one incredible person but someone who has a really special marketing approach. That’s all I’m gonna say. Because you need to watch this episode. Don’t forget to subscribe to our podcast, follow our social media channels, and enjoy the following conversation. 

Marcos Bravo: Hello, Mr. Mike, how are you doing today sir? 

Michael Taylor: I’m good. I’m good. How are you? 

Marcos Bravo: Good, good. I’m it’s a pleasure to be with Michael Taylor here. Especially because 49 Man, like when we were talking about like, when can we minds that you need to bring Mike in? Because he’s the OG growth mentor. It’s like literally the essence of what GrowthMentor is. So I don’t want to put the standards. I know right? Now, but I’m seriously I mean, Mike, you’ve done tons of things, you literally have half of the years of experience that I do, but 10 times the amount of experiments and experiences that they’re crazy. And I’m gonna let you talk about those. But I need to start with my regular sort of linking all of these stories to growth mentor, I want to ask, like, how happened that you ended up in growth mentor? How do you feel that you have something to give something to say? Just tell us about it. Like how everything ended up in Growth Mentor? 

Michael Taylor: So growth mentor caught me at a really good point, I think I was looking for stuff to do for the first time in my life because I was leaving my agency. I founded the agency and grew it to 50 people, actually, a lot of the people on Growth Mentor X ladder, you’re gonna have like a growth mentor, like a lot of Mafia, and grow up just kind of nice to see. But, but yeah, I went from, you know, running an agency with 50 people to then just like, you know, learning how to code. I didn’t have anyone, bothering me anymore, like asking questions, and, asking for advice, or asking for my input on different campaigns. And I really missed it. Obviously, I didn’t miss it in the first week. But like, after a few weeks, I was like, Oh, I kind of like miss being needed. So really, from my own ego, that joined growth mentor, I think Foti kind of reached out to me. But I think someone else on the platform, maybe Matt Eisner, I think it’s like, it’s really legit, you should talk to Foti. And you know, you’d have a good time that so I joined and, really enjoyed it. I did. I think I’ve done over 200 sessions, something like that. 

Marcos Bravo: And oh, my god, man. Like I literally, I mean, for years, I think I have 24. I’m not even raised about the big stage.

Michael Taylor: Do you get to count these podcast interviews?

Marcos Bravo: Right, like I just reviewed now.

Michael Taylor: Yeah, so yeah, I mean, it started off I was doing, you know, I’ve kind of gotten ramped up in some periods of the past few years, and then ramped down in other periods where they’re super busy. But I’ve consistently had a profile and changed that profile considerably. Now, it’s all AI stuff. But I’ve done a bunch of these sessions and met some really interesting people. I think I’ve written a blog post for Greg Venter on like, why I do it. But we can get into that, but I think, really, one of the big things is that it just gives me a lot of energy. Like there’s a really big difference between taking like a client call with shops or energy sometimes, and you know, taking a call from a super smart person who shows up, prepared, they have interesting insights, they actually kind of teach me a lot more than I do.

Marcos Bravo: That’s what we talked about before with some other guests is that is never a one-way conversation is a whole learning experience that you get, because sometimes you end up with someone with tons of experience or someone new or whatever, but there’s something there’s an exchange, it’s not just like, I give you and you get nothing, just completely both ways.

Michael Taylor: I think that’s a mistake, right? Like, I get out of touch, right? Because I’m not actively running campaigns anymore. So it’s really actually a big benefit for me to hear from people like what they’re struggling with, what issues they have, what’s working now what’s not working? Because I need to stay plugged in otherwise, I’m probably going to work on the wrong things or, kind of have misguided notions about how the world works today.

Marcos Bravo: So you been evolving on the sort of, I don’t wanna say services, but the way you help write you set a program on the marketing side pure, and now you’re fully going AI?

Michael Taylor: Yeah, how it’s pretty recent past few months? Yeah.

Marcos Bravo: So how did you change it? I’m sure it’s not because AI came up. I mean, did you have experiences and doing things with AI? From before? I know, like you guys in the ladder are crazy about the experiment and things. So how are these evolvement? Are these evolution works into the stuff that you offer?

Michael Taylor: Yeah, good question. I’m honestly still figuring it out as I go along. But the transition was when I left the agency, I really wanted to get technical, again, like, kind of get deep into the technical. So like, I’m a marketer, right? But I’ve always been really interested in growth marketing, specifically. And I just kind of see that or define that as, like the marriage between engineers, like software developers, and know the kind of the product side as well as the marketing side become integrating those two things together. I’m always really fascinated by people who can code and build these big systems that automate parts of the work. So when I left the agency it was my opportunity to do it. So I did like a data science boot camp, data quest did that like 20 hours a week for the first three months, and then did some consulting on the side and, and there were three bets that I made, that were at the intersection of things that I knew that people needed, based on my experience that ladder, and things that other people didn’t want to do that I like to do. I think you can make money there. Right? And, and enjoy doing it for the longer term. So I mean, I kind of made like three bets. One was just analytics and tracking. You know, nobody likes setting up tracking Google Tag Manager, or Google Analytics. I’m one of these weird people who actually kind of like it like I did over 200 times that ladder. So I got really deep actually set it up. I set up Google Analytics for Google once. For Google Play.

Marcos Bravo: There’s a burger for McDonald’s, like, don’t worry. 

Michael Taylor: I also ran the Facebook ads for Facebook, but that’s it. But yeah, so I was deep into that. And I knew how to code and I thought this was gonna get me more technical. So that was the first thing where I was making money and, and helping people and kind of know, I still own the shares in my agency, I haven’t sold them. So they’re still going, they’re still going concern still growing. So I started to work for a living, like having the salt. So I was, you know, that was what brought in income initially. But I found that those services were really tough. Like, nobody expects tracking to be as hot as it actually is. So I think it should be easy to track the performance of campaigns, it’s actually really, really difficult. So, I found over time that no matter how good I did it, in terms of like, implementing the tracking, it was never like, the cleanse was like, never really impressed. You know, they’re never like, really happy with it. I felt like that was like a, you know, just gonna get worse with iOS 14, and all of these kinds of changes to privacy legislation. So I started to go in the opposite direction and go into marketing mix modeling. That was the second bit. And marketing mix modeling is a statistical technique where you can work out your marketing attribution, but without having to track individual users. So it suddenly blew up, like, actually, Facebook released a marketing mix modeling library, like an open-source one. And that really validated the whole industry. And it’s been going for people who have been doing this for 50-odd years. But that really blew up and actually became then my, my primary source of income. And we can talk a little bit about that as well. But then the third bet was AI. And I didn’t make any money on AI for like, two years, right? So I was part of the GPT, three beta, and in 2020, I was still messing around with AI. I thought it was amazing. But it wasn’t that popular at that time. And I was still messing around with AI in 2022, and mid-journey and Dali came out and staple diffusion. I was technical then I knew how to code. So I got really deep into stable diffusion, all that stuff. And then it’s only really recently that I started to make my first dollar like giving people training but yeah, like I now have an online course on AI. Prompt engineering. It’s on Udemy and then I’m writing a book for O’Reilly Media on prompts engineering. It’s the second book. We can talk about that as another one of the better made I wrote their self-published book as well, but this one’s like, I call it like the real book. So it was just self-published on Gumroad. And it was a super valuable project. But yeah, this one’s like a real publisher. And you know, it’s going to, like, hopefully, be pretty big in terms of that distribution network. So yeah, anyway, that’s a long rambling story, but like that, you know, eventually, I got to the point where I was like, I know, I can make good money in athletics and track, but it wasn’t particularly satisfying. I could make really good money in marketing mix modeling. But I kind of reached a limit of like, if I wanted to get better at it, I’d have to get a PhD in statistics. I didn’t really want to get back to school. AI is at one kind of intersection where I really love doing it, it’s moving really fast, kind of feels like being around at the beginning of the internet, you know, it’s like a real platform shift.

Marcos Bravo: I made a post on LinkedIn. It was, this is before I saw your, your bio that says from engineering, but it was making fun of people becoming product engineering because a lot of people just telling completely. Oh, for sure, yeah. Right. And everyone’s like, you’re using it wrong. It’s just become a joke. Eventually, he’s like, Yeah, you’re using AI wrong, because so it’s hard for people to really find the value when a lot of people say, Yeah, I’ll sell you 100 prompts or whatever. And then you read them, it was like, I could have saved that says it’s nothing. So what’s really behind a real prompt engineering?

Michael Taylor: I experienced this, it didn’t feel like my first rodeo when I saw this come up, right? And it felt like the early days of growth hacking, right? Because, you know, we were the largest growth hacking agency, I think, at least to my knowledge, we’re 50 people, right? So you know, and then growth became a sexy word. And then everyone was in growth, right? Like, even if you just run Facebook ads, you know, you weren’t, you know, I used to say, if you can’t push to production if you can push code, then then you’re not a growth team. Right. And, and you know, that obviously, that feels a little bit snobby. But equally, like, it’s very, very hard, I think, to actually make a company grow to the level of expectation that growth teams have set like the growth teams at Facebook, you know, Airbnb, and all these other companies, it’s very hard to achieve that if you can’t make changes to the product, right? So I think it’s, it’s kind of a mistake to flood that term. And I see the same thing happening with prompt engineering. Yeah, like, the funny thing you see on Twitter is that people are there the post always starts with like, 99% of people using chat GPT, wrong 2000 prompts and I’m like, I don’t need 2000 prompts, I need to know how to make my own prompts.

Marcos Bravo: How do you structure that’s it? I don’t need you to give me 2000 prmots.

Michael Taylor: Exactly. So what I’ve tried to focus on is more Defining Principles. The way I see prompt engineering is it’s about talking to an intelligence and other intelligence, that’s artificial instead of biological, right? So in the same way that you learn management science, in order to become a better manager, and kind of understand how to manage a team. I think that in the future, prompt engineering, whether we call it that, or something different, is going to be a discipline for working with artificial intelligence. So you know, it’s like management science, but for AI, and I think you no matter how good the models get, you’re still gonna need to learn how to communicate with them. Because people are pretty smart. Relative to these models. We are, you know, we are AGI right, and there’s still like a whole host of issues that we have to solve like growth mentor exists, partially because, you know, people don’t know the best way to approach tasks, they don’t know, the best way to approach situations. So I think being able to give that guidance to ChatGpt, or whatever AI model you’re using is going to be a thing, you know, now and far into the future.

Marcos Bravo: Now, this question is Mike, it might sound like a joke, but do you really think people are using AI wrong? 

Michael Taylor: I mean, it’s 99% of people are using? Exactly. I mean, they’re actually right about that.

Marcos Bravo: Of course, there’s that space of, for example, I don’t know, like when I started using chat, DBT was like, Alright, give me 10 ideas for Facebook posts, whatever, right? And then you get stuff that sounds really bad. But if you ask the right way, you can get way more than ideas. You can get a whole structure, a whole distribution plan, the whole strategy behind what you’re asking, but it’s really, is it? So the question is, is it really about the way you ask the questions, or there’s something else behind it?

Michael Taylor: Yeah, for sure. It makes a really big difference. And I think the big difference between what the Twitter gurus are doing with that 2000 prompts, and like what I’m doing is that I test stuff, right? So it’s very similar to growth in that respect. I can AB test like I would run a prompt one time and see, oh, the results seem good. I’ll run it like five different ways. And I’ll run each variant, you know, 30 times, and then I would look at this Cisco significant sense of how high have they, you know, like, what parts of the prompt are really making a difference and like, what, you know, what, what is actually getting a better result. And often what you find is this, these big along look like magical spells, you know, like, incantations, like, you’re some sort of wizard, you know, using the AI, let’s malevolent God, you know, like, I think a lot of these prompts are crap, to be honest, like they’re not, you know when I’ve tested them, they don’t work as well as, like some of the simplest stuff. So I think that said, there are other techniques, which are super important that people forget to do so. Like, one of my favorite reframing things here is you, because you talked about how you approach it. First thing, I think the initial approach everyone has with AI is they go, yo, give me five ads, right, like, give me you know, like, 10, SEO tactics, or whatever it is. But you’re not going to get great results, because it’s just trying to guess what you want. And if you don’t give it much guidance, it’s gonna give me back the average of the internet, right? Because it was trained on the whole internet, and it can actually give you pretty much anything I can give you, you know, 10 SEO tactics that, and Hemingway would have given you something, it’s like it can give you like, you know, Shakespeare in prose, like, you know, like, pretty much anything it can give you back, you just have to kind of know how to unlock it and kind of get asking the right way. So the key I find is in kind of you providing the secret source. So like, when I’m writing blog content, I won’t say write me a blog on value-based pricing. Instead, I’ll ask it to interview me on value-based pricing, and then write a blog from that, right? So like, what people really care about is your unique experiences, like what you know about your business, and what you know about marketing or whatever discipline you’re in, and ChatGPT can then expand that information, I can take that, you know, that nugget of information that you have that unique experience that you have, and then expand it to a full post or, give you 10 different ways to say in terms of social media posts, or whatever it is. But I think you have to provide that secret sauce.

Marcos Bravo: When you talk about the title of real or another, whatever we want to call it in the future is prompt engineering. That means AI is not truly going to take over every marketing job. Because I mean, I’ve been worried about that, like I already doing the planet B and started making ice creams and but yeah, but there’s obviously there’s a space for collaboration between AI and a human side, and the example that you just gave is a very good one that you’re asking AI to extract your own experience into creating something. So how do you think this, this symbiosis or whatever you want to call it, how is it going to work between humans and AI, especially on the marketing side?

Michael Taylor: Yeah, the analogy I like to use is, I don’t know if you’ve ever been in, like, say, St. Pancras Train Station in London, or like any actually, they have this in a lot of train stations, but they have a piano, and people can sit down and play at the piano, which is pretty amazing. And whenever someone is playing, they know how to play, then everyone will stop and watch them. But like, if there’s nobody playing the piano, the playing piano actually can play itself. And it will just play a tune in the background. But nobody stops to watch that. Right? Like nobody cares about that. So, we’ve already made the technology to completely automate piano playing, but nobody cares. Because there’s no unique human experience, right, like people want to see, almost like the imperfections. I kind of stretch that analogy when I’m using it, but by like, I feel like there’s a similar role for humans in a post-AI world as well. Right? Like, yes, you know, AI can automate a lot of the stuff, but that’s just going to kind of free us up to, you know, fill the gaps in between those unique human experiences. At the end of the day, like, I was just listening to a podcast, the quiet podcast, talking to Daniel Act from Spotify, and, he kind of explained how music has kind of changed in this way, you know, used to be in Beethoven’s time or Mozart’s time, they would have to, like be able to imagine how old are different instruments with sound, because they can actually hear someone playing an instrument. And you know, they couldn’t do trial and error, like had to imagine so you have to be an amazing conductor in order to create music. Whereas like today, a DJ can play a few sounds and use the software.

Marcos Bravo: You have flipping loops or whatever.

Michael Taylor: So like most musicians today, like don’t know how to play most of the instruments using their tracks. You know, like, you know, like Yeah, how many how many DJs know how to play the, all these different instruments? Like how do they know how to play the drums now, they’ve known to press the buttons in the right order. So I feel like, you know, that doesn’t make the music any worse. Like, I mean, some people might I’d argue.

Marcos Bravo: Yeah, that’s debatable. 

Michael Taylor: Yeah, but like, you know, it’s still really interesting, like, still really unique human experience when they do find that right combination. So I think that that’s like not going away. Even if AI automates pretty much everything. It’s going to still give us a lot of space to express ourselves. And I think it just kind of raises the bar. Like, nobody makes bad music, you know, it’s pretty hard to make bad music now. Let’s say it was it was always guaranteed that you would make bad music.

Marcos Bravo: That’s a great analogy, though. They weren’t about the piano. You just started I mean, I’m not sure if you started fully with the ladder. That was your first marketing experience.

Michael Taylor: I worked on a couple of startups before that. A couple of growth roles. So it’s in the travel industry, a company called Travelzoo like daily deals kind of company works in Shopstyle, which is like a big fashion affiliate, and a few others. So I kind of learned peak growth, learn SEO, you know, landing page optimization, that sort of thing before I started ladder.

Marcos Bravo: okay, because you’ve been sort of doing something very, right, which is sort of smell the next not trained, but the neck. Right, yeah, you’re nailing that. So what are you doing right now? What’s yours? Because obviously, you’re doing AI? But are you planning to create a startup around it? Or is Jason consulting? How’s it working at the moment?

Michael Taylor: Yeah, good question. I’m still trying to figure that out. Right. But generally, I try and, when I’m spotting these trends, like things, things to focus on, it’s usually, you know, some situation where most people don’t realize or don’t want to do the thing. But when you do the thing, it gets better results than anyone expects, right? So, I think growth kind of fits into that. Because actually, most of the traditional marketing people I talked to, at the time, when growth hacking came out, they were laughing at it, they’re saying, it’s stupid, actually, digital advertising was only like, 5% of media at the time. So like, even the whole bucket of digital to them was a rounding error. And like the concept that you would put a developer in the marketing team, it didn’t make any sense to them. Right? So if you worked at Procter and Gamble for like, 20 years, and you know, you’re like, you know, you got like a reputation industry, you’re not going to like to change your LinkedIn to growth hacker, right? So there was a real barrier to like, people wouldn’t adopt it. But like the people who did adopt it, bringing insane results, like, you know, the Facebook growth team, got the company to a billion users pretty quickly through a few different growth hacks. And there were like, kind of instances where I saw it just like working really well, for me as well, when I tried it. So, you know, AI feels the same way. And specifically within AI, what I’m focused on right now is around the testing side. So you know, we don’t have all the frameworks. There’s no Optimizely there’s no, there’s no VWO of Google Optimize for prompt engineering right now. So you have to do it pretty manually. So one of the things like email I was working on today was just kind of like a, just an interface for doing that. And it’s an open-source library. So I’m working on that. But then also kind of using that, again, you know, most people are using Chat GPT wrong. They’re not testing your prompts. And therefore, like, I feel like right now I have a superpower. Because I am testing my prompts. I know, the power of testing, I’m just applying it to this new domain. And you know, rather than kind of blindly messing around with trial and error, I’m starting to see what actually works and what doesn’t. So as well as kind of making the testing side easier. I’m also exploring three different product ideas and I’m finding a prompt that works better than average. You know, and like, other people probably wouldn’t try that prompt. 

Marcos Bravo: I mean, I really find it cool that you’re finding all this like missing spots that happening, especially with anything coming up, you know, how do you see it in the future? Right. I mean, we talked about how the relationship is probably going to work if you do it together between AI and humans. But what about next, like, how do you think that’s going to move forward into marketing like is just going to become a lot of automation happening or, or there’s something’s going to top what AI is doing right now? I mean, again, I like your vision of what’s coming. So what do you think is going to happen sooner than later with all of this working on?

Michael Taylor: Yeah, good question. I think if it hasn’t happened already in your organization. The real no-brainer is to use AI for all prototyping of creative ideas. So, you know, I think there’s probably some resistance to using it for the final production version, right, mid-Gen is still really good. And they’ve solved a lot of problems, but you can kind of still tell it’s AI. And the same thing with the chat GPT. Chat GPTfour is really good. But in a lot of instances, you can still tell it’s AI unless you do a lot of front engineering to make it sound more like you. Now, I would say, you know, that doesn’t mean you can’t use a prototype, right? Like prototyping, it is a very low barrier to entry. There’s no risk to using AI in prototyping. And there’s a huge benefit, which is, rather than going back and forth with the copywriter, and back and forth to the designer, now you can like actually make the thing that you want pretty quickly, even as a nontechnical person, and you’re even someone who doesn’t know how to use Photoshop, you can kind of generate an image. And even as someone who doesn’t know how to write, or copy, you’re not an expert in that you can still generate some decent ideas and kind of pick the ones that resonate with you. And then it’s just like a smaller barrier to entry for the designer and copywriter to kind of access your thoughts. Because otherwise, they’re just trying to second guess like, when I say, you know when I say Batman, you know, like, you have a different image in your head of Batman. So I say draw Batman like he might like the comic, so you might draw more like the comics, I might like the Christopher Nolan film, so I’ll draw it like that. You know, and I think AI in the prototyping phase can really help, you know, bring those two visions together. And kind of you can show someone something, and they can criticize it much better than they can imagine it in the first place. So I think that’s an immediate no-brainer, I think it really cuts out a lot of communication issues between design and creative teams and the management.

Marcos Bravo: Now as a human being, do you think is going to have a big impact? Also, like in the day-to-day of people? It’s because obviously, yeah, not just AI is there for business, right? Eventually, it’s like, it’s supposed to do something for humanity supposed to be one of our greatest advances, like, how do you think is going to affect the rest of like, all of us? 

Michael Taylor: Yeah, I think it’s going to have a similar impact to Google. Right? So if you imagine a world before Google, you know, I remember, pre-Google, maybe dating me a little bit. But you know, before search engines, you just have to ask people. We look something up in an encyclopedia, which is really painful, you go to the library and check out books, and hopefully, you find it in the book, right? And, then you just couldn’t get anywhere near as much done. And, you know, unemployment is at a record low, despite the fact that everyone has Google now. Right? Like, you know, that librarians saw more librarians in the world, and they were able to internet, right? I think that the immeasurable impact on people’s quality of life, it’s just as you know, this is going to happen again, with ChatGPT and AI, right? And it’s already affecting me right now. So, you know, like, actually, you know, I was, like, just lifting some weights. And that was like, curious, like, what is a good amount to lift? Right? Like, what should I be aiming for? So I just asked, you know, like, what is and you know, it actually gives me a really good answer is, like, you know, beginners should be waiting, like this percentage of the body, body weight, and like, you know, intermediate is this, and like, the world record is like, two times body weight, you know, things like that. So, it just gives me a really good, interesting kind of context around this task. I’ve been doing a bunch of construction work in my house, and I know nothing about it. But I’ve been asking, like, which type of paint do I need to get, you know, for, you know, if I’m going to paint my room, so like, it’s already like, a really big part of my life. And, you know, like, obviously, I code, right? So like, I’m using this a ton for building stuff, as well. I think it just really takes away the fear of trying new things. I think that’s the biggest impact because I would never get involved in construction stuff. Like I would just pay a deal based on. Yeah, because because, like, you know, I just didn’t know anything about it, right? But whereas like, now I can at least ask and kind of get a sense of whether it’s hard or easy, and then kind of outsource the parts that are harder or need a bit more expertise, right? And, the same thing with coding, right? Like I am building stuff as I would never thought to build my own open source, like a Python package before. But I honestly just said, like, here’s the script that I’m using, how would I turn this into a Python package on pi? Dog and it knew, right, like, and it gave me like, here, like you need to add like the init file here. You know, so like, I think it just makes you fearless. Because, like, what’s the real downside of like, she’s asking and seeing if it’s actually easier than you realize?

Marcos Bravo: But I definitely make a difference because because you code so tight. ChatGPTtell you things that you understand. And I think that’s, again, that’s why people are using a wrong because they ask for something they don’t understand. Like, I remember there was a founder of a startup that was helping as I look, I don’t need a marketing prisoner anymore, because I just asked ChatGPT. What are you understanding what he’s telling you? Like, when he talks about SEO, or Google Analytics, or whatever you really get is like, no, no, I just let him do whatever. Like, that’s, that’s not going to help. It’s not gonna help. 

Michael Taylor: Yeah, just, you know, they’re like, I actually think you can get pretty far doing that, you know, like, I have friends who don’t know how to code and they have like, whole applications, like automating stuff. And they’ve literally like, I don’t know, man, I just keep asking ChatGPT, and it keeps working. And I think I’m all for that, you know, like, I think it takes away the fear because it’s pretty easy to hire a developer to make something work a little bit better, right? Like, it’s very, very hard to explain to a developer, like what you want in the first place, or like, you know, the vast majority of all code written is just thrown away, because, you know, the product didn’t work out, right? Or it didn’t, it couldn’t, like didn’t actually get good results, right? Or because they didn’t even finish the product in the first place. So I think you kind of need to treat these tasks as more exploration, right? Like, you know, maybe that guy that you’re talking about, yeah, he’s gonna make some mistakes, because you didn’t have an experienced marketer, but like, was it realistic to have an experienced marketer in the first place? And, you know, maybe didn’t have the budget, right? But maybe you can get to the point where he now has the budget, through using chatGPT, he’s kind of, you know, you tried all the obvious stuff. And then you can go to a more experienced marketer and say, you know, this is working, this is working, that’s not working. I’ve tried all these different things, the ChatGPT recommended. Now, what should I try, and then that experience, marketer just has a lot more to work with, and, you know, that clients now afford them. So I think it’s really bridging that gap. It’s not between, it’s not like replacing the experienced marketer, but it’s like getting more people up to that level where they can actually become a client.

Marcos Bravo: Yeah, so let me just, I’m gonna stop closing the conversation because we only have a certain time. But, there are two things I want to cover. So I want you to tell me a little bit more about the books that you’ve been writing. What’s your approach? Because especially like, marketing mimetics. You went really? Into the biology and psychology of marketing? Yeah, I haven’t read it.

Michael Taylor: That was a real, real passion project.

Marcos Bravo: Looking at it, I’m definitely gonna read it now. Because it’s like, that’s, that’s what it really like, is like the human nature behind all the businesses. So tell me a little bit about that book. And Tim, a little bit about the book that you’re writing, and then we can move into that. 

Michael Taylor: So like, funnily enough, like, I wrote a book probably at the wrong time in history. Just because, you know, if I’d started it a year later, I probably wouldn’t use chassis meat eater. Right. If you don’t, I mean, and, you know, like, I really started to blow up, you know, as I was writing, like, in the last few months, and I had to kind of change a few things about even being early on AI like that, I started to realize that wow, like, this actually changes a few different things about what I’m writing about. But it was really beneficial because the topic that I wrote about was mimetics. So memes, you’ve heard of memes, funny internet videos. But the word meme actually predates the internet. And what it really means is any piece of information that gets passed on from person to person, so anything that gets imitated, so like, you have, like this fluorescent light in the background and says on air, right? Like that is a meme. Because, you know, other people have had that, or similar to that, right, fluorescent lights is one meme on air is like another meme. And like, pretty much everything around us is like something from somewhere else. And that’s what gives it cultural significance. And, and, you know, whether, like, consciously or subconsciously, you are, you know, kind of saying something, by having that in your background, right through, you know, so it really expresses something about your culture about what you value about how other people should perceive you. And, and mimetics is really just studying that. And the really cool thing about AI is that, you know, prior to AI, it was really expensive to these types of studies. Because, you know, if you wanted to see like, maybe you could do a mimetic analysis of podcasters. And you could say, Okay, how many of them have on air in the background? Right, like, maybe it’s like 10%, right, and other people who have on air in the background before are better or worse than average, right? So you can kind of do these correlations and start to see like, Okay, what’s important to have as a podcaster. What’s not important to have, you know, what type of mic the podcast is using, like, I’m using blue Yeti and you’re using that Riverside.

Marcos Bravo: It is the Behringer.

Michael Taylor: yeah, there we go. Okay, so swag, isn’t me but now with AI, it’s super easy to do this type of analysis because you can get the AI to tag this stuff, right like you can and use some of the components of stable diffusion, which is like open source to be able to tag different things and images, you know, you can tag different things in customer reviews, you know, in a blog post, so you can kind of get a sense of what are the types of memes that I should be including in my marketing campaign, and what ones shouldn’t be leaving out, because they’re not performing. So as I was reading the book, I was like, it was a real slog, like trying to convince people to motivate people to do this type of analysis, and that this type of analysis is valuable. Some of the cost of doing it dropped by 1,000%. So so so like, I actually kind of like, I think, made a lot more people interested in the topic. And I hope, like, it’s certainly a self-published book, right? So I don’t, you know, the need to make a lot of money from it is more like a passion project, something that I learned. That was I think, when I look back on my agency experience the key to all of our best successes, that ladder was when we did this type of mimetic analysis. So that’s why I wanted to capture in a book, but yeah, I mean, it kind of, it’s like, a little bit ahead of the curve, but maybe a little bit too far ahead of the curve, because I still pretty niche. Not many people have heard about when I was growing up, hopefully, it’s growing, you know.

Marcos Bravo: And the second book that we’ll be reading now.

Michael Taylor: Literally, I just finished this crazy editing process, like I took the whole year to write my first book, and the self-published, you know, so it didn’t like, you know, it wasn’t New York Times bestseller the first week. But then, I was doing some training for O’Reilly Media at the time. And then I’ve been out to a conference, they hosted a pretty cool conference out in San Francisco. It’s like an unconference. They call it that fruit for cough is the name. And it’s like, invite only and they just, they they figured out the agenda. Like they’re at the actual event, right? So it’s a conference, but like, everyone who’s invited could be speaking, oh, cool. And they have, like, Nobel laureates, they have like, crazy, like, I have all these authors, like pretty amazing, like, researches, I have a lot of meta employees because of meta campus. And, you know, I gave a talk and a few other kinds of met some interesting people. And I don’t know if that helps my case. But then, you know, they reached out a few months later, and they were like, do you want to make a book on prompted during all this training that you’ve been giving? For us? It’s online training. So I was like, Yes, of course, like, this is a real publisher doing it, you know get into that textbook game. Hopefully, it’ll go well, but, yeah, like, I actually just finished writing the first draft, and writing with my co-author, James, who also works on a couple of other projects, but, but it’s on promise engineering kind of teaches the principles of prompt engineering, have approached it like this management science. That I think it is, also, it’s not like, you know, 99% of people are using it wrong. And it’s not a big Yeah, exactly. I mean, maybe I’m missing out on money, probably by not making that the title. But, you know, I don’t want to be associated with that. 

Marcos Bravo: No, no, that’s fair enough. Mike, so just to close the loop, we started with talking about growth mentors, we’ve been through some of your stories and have liked how your experience has been shaping. I like to ask, and I know, it’s not very original, but a little advice for the people who are thinking of calling for a mentor call with you, or with someone else on there that states that they need help. What will be your your to go tip that you always drove out there, for example, for me is to test everything. That’s just, I’m not gonna tell you. Yeah, that’s gonna work. Just test, right? But what will be a tip that you think is super valuable in your experience to people either starting or with experience, but they’re having the sort of moments of doubts or problems?

Michael Taylor: Yeah, good question. I like to think from first principles. So, you know, one of the one of the really big things that helped to ladder was, you know, we would just break it down and say, like, okay, like, where is the real problem here? So yeah, the overall problem is that you’re not growing fast enough, right? Like, nobody’s growing as fast as they want to be. So, there are all sorts of components that go into that, right? Like, it’s, you know, do the audience know who you are? And if they know who you are, like, what do they think about you? And if they like, they think favorable things about you, then then, like, what’s stopping them from coming to your website or checking out your product? And if they’ve checked out your product? Is there any friction that’s stopping them from continuing? Or if they’ve, you know, used it in the past? What’s what got them to stop using it? I quite buy they’re not using it now. So, you know, every big problem is just a series of smaller problems. And as you decompose that into the individual parts, you start to spot areas where you can make a difference. And rather than being overwhelmed with this big nasty problem, breaking it down into individual parts, like you actually can make some progress. and feel good about the progress that you’re making. And this is something I got from learning how to code because, you know, in software engineering, it’s like a key part of it, right? Like, you just kind of have to end all engineering, right? Like you take a big problem, you turn it into lots of small problems, and then you start going to those offerings, right? And, you know, it’s true in AI as well, right? So, you know, the first thing I’ll tell you, if you’re struggling with your prompt, the first thing I’ll tell you is, like, try and break it into multiple tasks, maybe you’re trying to do too many things all at one task, right? And, and if you break it into multiple prompts, and then you can chain those prompts together, and overall, like maybe you can achieve something much greater than, you know, than the individual parts. So like, I think that’s like a universal advice is like, you know, break it down into what are the things that need to happen? And like, what are these sub-problems, and then just reason from first principles and think about, like what, you know, like, it’s either this or it’s this, right?  can you break that, like, you’re either this is true, or like, if that’s not true, then this other thing has to be true? And if you can get it down to those fundamentals, then you should have the problem become less overwhelming. And you might not immediately know a solution, but it feels more approachable. And you can make start making more progress. So you know, who you need to talk to, you know what you need to try?

Marcos Bravo: Awesome, Mike, where’s your course in Udemy, right?

Michael Taylor: Yeah. on Udemy. So, I guess you could provide a link in the show notes. But the first one, when you go for prompt engineering work with the top one, now, we’ve had over 2000 students take the course and format it. So it’s 6000 of them were like, this month, which is kind of nuts. It has doubled. So yeah, I mean, it’d be, I wouldn’t have to, you know, I won’t have to do anything else if it keeps doubling, but check it out. And I think I will say it’s more technical. So, you know, there was, like, you know, about 30, or 40% of the material is, is, you know, you don’t need to know how to code but, but we’re focused more on the typical use case.

Marcos Bravo: I’m gonna add all of the all of the details, all of the links, so you guys can go and check. Not only that, you can also book a call with Mike and find out more about how AI can help you and your business. So Mike, man, this is time we have for today, we definitely gotta be in touch. I’ll see you around the world of Growth Mentor. 

Marcos Bravo: So again, another pleasure to have people like Mike in the show. And usually, if you’ve been watching the show, you know, I do a swoop, when I talk about the things that I’m taking from this conversation. But this time, I want to do something different, I want to invite you to go and watch this all over again. Because the conversation about AI and how AI is changing everything that we do, not just as professionals, but also as humans, is something that you need to bear in mind. And I think the way Mike would send it in there is very not unique, but very, very clear for you when you try to understand if AI is going to replace you, or what’s going to be the value proposition that you can have as a company using AI is not the automation anymore, is you, you’re gonna make the difference is the human relationship with AI. That’s going to make a huge difference in the way we grow together with AI. So that being said, my soup for today, which doesn’t have the soup, is to go back and listen to this because it’s really important for all of us professionals. Doesn’t matter where it’s marketing or whatever you want to do. So watch it again. subscribe to the podcast. My name is Marcus and I’ll see you next time. Cheers.