ADS TO AI — 10X YOUR VALUE AS A MARKETER | Mike Rhodes | #6

00:00:00: So welcome to a new episode of the Masters of Search Talks.

00:00:03: I'm super happy to introduce today's guest, Mike Rhodes, who has become one of the most influential voices around PPC and AI.

00:00:13: After building and selling for businesses, Mike has spent a lot of time, probably he will tell us today how much time, building products like the famous PMAX Insights Script, I think like in PPC.

00:00:26: everybody knows this or at least ninety percent of the people and teaching people how to do the same.

00:00:33: So Mike, can you give us a quick intro about your journey and especially your move from agency founder or a business builder to AI builder?

00:00:43: Yeah, yeah.

00:00:44: Yeah, I fell into Google ads by accident.

00:00:48: I went to a conference in two thousand and four and Perry Marshall was in Australia for the first time speaking.

00:00:56: Ripping into yahoo or overture as it was back then and talking about this new thing that he was starting to look at called Google ads.

00:01:03: and Yeah, well, yeah a quick twenty years later.

00:01:08: Here we are.

00:01:08: I ended up doing all of the training for Perry's group and writing the book on Google ads with Perry which I believe is the world's best-selling book on Google ads, but that was all Perry not me.

00:01:19: And then yeah built an agent fell into an agency by mistake too.

00:01:22: I took me sort of a year of helping people here and there and swapping advice for a case of beer or a round of golf.

00:01:30: And all these people kept saying, yeah, but mate, I don't care how it works.

00:01:34: I just need you to do it for me because apparently my MO is just.

00:01:38: I like to explain things and I like to teach people how it works.

00:01:43: So I fell into agency land and all of a sudden turned around and my kids were still both in primary school a couple of years ago.

00:01:52: And I just thought, you know what, like this I've always been very careful of my time.

00:01:56: I was never working seventy hour weeks.

00:02:00: Maybe right at the beginning before kids.

00:02:03: But I just thought, you know, this is a time I do not want to miss.

00:02:06: I know I'm never going to get this back.

00:02:08: It's time for someone else to take the baton and take this forward.

00:02:12: The plan was to basically do not too much.

00:02:16: And that lasted about three weeks.

00:02:18: And then I rediscovered how to code.

00:02:22: And it was right around the time that GPT-IV was coming out and growing.

00:02:27: And I just, yeah, the PMAX script, as you mentioned, just went gangbusters, which gave me a lot of leeway then to just explore.

00:02:38: And since then, I've been building tools.

00:02:40: I've always built tools for the PPC community.

00:02:43: going back to two thousand and eight, I think was the first one that I built.

00:02:48: I've always, as you said, I had to teach other people how to do what we did.

00:02:52: back in the agency world.

00:02:54: Again, starting in two thousand seven, two thousand eight, I've had various sort of membership groups and, you know, spoken at conferences and so on.

00:03:02: And I absolutely love it.

00:03:04: I love building stuff.

00:03:06: I'm not particularly good at managing people and stuff and things.

00:03:11: So like the running of a medium sized business, definitely not my forte.

00:03:17: I was.

00:03:18: I was definitely much happier once I had a sort of a proper general manager come in and she managed all of the day today and allowed me to travel more and speak on stages and so on.

00:03:30: But yeah, I just love building stuff and so now it's just such a wonderful time to be a builder because you can go from idea to impact in the course of a day really and then get an MVP out there and get feedback on it and see if it resonates.

00:03:48: And that's what I've really been doing most of the past two years.

00:03:51: Before we dive into more of the PPC, the AI, the building stuff, I can imagine that there are a lot of people in the audience also either from maybe a freelance background or maybe also agency owners, which I'd just like to take a very short moment.

00:04:08: to recap this journey when you were selling the business or when you were transitioning out of the business.

00:04:14: Can you just give us a little bit of an overview of things you were doing in the run-up of you going out of the business?

00:04:23: Were there certain key milestones you set for yourself that you have to achieve so the business is sellable in a way?

00:04:31: I'd like to hear a little bit of this process.

00:04:38: Yes, I suppose going back, gosh, more than twenty years, I was an EMIF consultant.

00:04:44: I paid Michael Gerber a very large sum of money a long time ago to become an EMIF consultant.

00:04:49: I'm not sure if you've even heard of or read the book, the EMIF.

00:04:53: It was the small business Bible, you know, twenty five years ago.

00:04:57: It was the thing and it was the thing that really helped my first business be a success.

00:05:02: And I didn't know there were these consultants.

00:05:04: I went off to the States.

00:05:05: I paid him a huge amount of money and trained and basically it's all about work on the business, not in the business.

00:05:11: That's essentially it, right?

00:05:13: So I've always been a fan of systems and processes that run the business and the people run the processes and the business is just a system of systems.

00:05:23: So the business was pretty systemized.

00:05:26: I mean, it's a professional service business, so it's always going to be a bit messy and especially in the industry that we're in.

00:05:32: There's a lot changing all the time.

00:05:34: It's not we used to joke.

00:05:35: You know, it's not like insurance where you could probably do the same thing that you were doing ten years ago.

00:05:39: You know, the half life of information in the PPC world is two years.

00:05:44: So I don't know that there were milestones as much.

00:05:47: I mean, getting a general manager, like I said, which was probably five years to selling, but it wasn't like I was like, oh, I'm going to sell in five years.

00:05:54: So I'll get a general manager now.

00:05:57: If you'd asked me then I would have said I'm never selling.

00:06:01: The way I pictured the business, imagine a big sheet of paper in front of you, like a three sheet of paper.

00:06:06: I've always said that it's easier to work out what you don't want to do than it is to work out what you do.

00:06:12: So you picture all the stuff that you do as this huge sheet of paper in front of you.

00:06:16: It's much easier to go, oh, I don't want to do that half of the page anymore.

00:06:21: Let's spend the next year getting rid of that half of the page.

00:06:24: Okay, now I'm down to a smaller piece of paper.

00:06:26: Now next year I want to get rid of that half of the paper and then this bit and then this bit and then this bit and you end up with just this little.

00:06:33: and that's where I was.

00:06:34: I was just doing the little bit that I felt I was quite good at and that I really liked doing but I don't know.

00:06:42: it was always a what I really struggled with.

00:06:44: I'm sure many people will relate to this but I always struggled when when a team member quit or when a client left I was supposed to be kind of removed from that you know.

00:06:57: A couple of layers away from that.

00:07:00: That's why you put in a proper general manager, you know, but I could just never quite remove my My brain and the feeling of that.

00:07:08: it always felt like a kick in the guts And I thought gosh if I if I'm still feeling like this now I'm always gonna care and it's always gonna hurt.

00:07:20: Yeah, I just like maybe maybe I should try something different after after seventeen years and yeah, I had a chat completely by chance, really.

00:07:30: I was actually considering maybe expanding and buying a creative agency.

00:07:34: Like there were a lot of different sort of options floating around.

00:07:38: And then I just, yeah, back to two kids in primary school.

00:07:41: Actually, I don't think I want to do that.

00:07:44: I really want to hang out with them and spend time with them.

00:07:46: We worked out back then that my daughter had a hundred and thirty four Monday mornings left in primary school.

00:07:54: And we've been counting those down.

00:07:57: We don't actually have a chart on the wall, but we both know roughly where we are on that list and we've got about fifty five to go at this point.

00:08:03: So, you know, it's been recounting that down.

00:08:06: I was just very, very aware of that number.

00:08:09: So, yeah, in terms of milestones, I mean, not really.

00:08:11: It was just had an amazing team.

00:08:13: That was the worst part, you know, being away from your team and missing those conversations and missing the problem solving part of it.

00:08:21: That was definitely the hardest part of dealing with it all.

00:08:24: But No, it's yeah dive into a new thing and don't look back.

00:08:29: Keep moving forward

00:08:31: Awesome.

00:08:32: I once saw this video of yours on LinkedIn where you were telling an analogy or a metaphor about playing table tennis.

00:08:41: You probably remember this.

00:08:43: maybe and after that question we will definitely move into the AI and the PMAX and the PPC stuff.

00:08:48: But maybe can you quickly share the story again with the audience because I think it's so.

00:08:53: it's so eye-opening.

00:08:54: Yeah, I like this one.

00:08:55: I wish I knew where this came from so I could give proper attribution.

00:08:58: It's definitely not mine, but you always feel like you have a lot coming at you, right?

00:09:05: And basically, I get your picture of that as you're playing table tennis and there's just constant.

00:09:13: There's a ball coming at you all the time and it's all you can do to just keep trying to pick up the bat and just keep smashing the ball back.

00:09:21: But no matter how much you do, it just keeps on coming at you.

00:09:25: And the idea of systemizing a business really and putting some automation in place and letting go a little bit and having other people do some stuff, it's all related.

00:09:36: But you notice one day that down by your feet, there's a pile of bricks.

00:09:41: Oh, I've noticed those before.

00:09:42: That's handy.

00:09:44: The trick is that you have to bend down to pick up a brick to put it on the table to start building this wall at your end of the table.

00:09:54: The downside is when you bend down to pick up that brick, inevitably, a couple of balls go flying past you.

00:09:59: Shit, I missed one.

00:10:01: And that's you feeling like you literally dropped the ball on something.

00:10:05: You've missed something in the business.

00:10:07: Something has gone a little bit wrong.

00:10:09: No worries.

00:10:09: We'll repair that.

00:10:10: We'll fix that.

00:10:12: Now I've got to bend down and pick up another brick.

00:10:14: And if you keep on doing that, you keep picking up these bricks, you eventually build a wall such that the balls coming over the net bounce off the wall.

00:10:23: and go back and to begin with not very many bounce back.

00:10:27: but over time if you keep building the wall yes every time you pick up a brick you miss a couple but eventually you've got this whole wall there and all of the balls bounce back.

00:10:35: and now freedom.

00:10:38: you know freedom to do what it is that you want to do.

00:10:41: freedom to do the thing that you set out to do in the first place.

00:10:45: you know Gerber used to tell the story about the carpenter.

00:10:48: he loves carpentry.

00:10:49: he works with wood he loves the smell of the sawdust But now he's running a carpentry business and he hasn't touched a piece of wood in six months.

00:10:57: But he's wearing twelve different hats and he's hiring other carpenters and he's dealing with invoices and dealing with clients and But he's not doing carpentry.

00:11:04: He just wishes he could get back to the carpentry.

00:11:07: That was why he got into this in the first place and it's it's a similar sort of thing, right?

00:11:11: But you've just got to you've got to build the wall you've got to let other people Do the stuff that they're great at and you're probably not and be okay with missing a couple of balls going flying past.

00:11:22: yeah i'm no idea where it comes from but i love the picture that builds in your head.

00:11:27: maybe at some point in time we will find the person that originally came up with the story but yeah i found it so inspiring.

00:11:33: so now let's move over to a little bit more ppc and a. i related questions.

00:11:38: so your pmax script.

00:11:40: which we obviously have already introduced has become somewhat legendary.

00:11:45: I like the term you used.

00:11:47: It went gangbusters.

00:11:48: I have to use this too.

00:11:51: And like for me, I feel like all of the time, especially in startup world, people are talking about, yeah, the idea is one thing, but execution matters a lot.

00:12:01: But now we have this great new tools and obviously execution is still very important, but we have stuff where basically our imagination can become reality in a much shorter period of time.

00:12:13: So I just want to get a feeling of how did you come up with the idea for this product or for the script?

00:12:20: Oh, gosh, that's probably going back about three years.

00:12:24: It would have been a conversation with Stacey and Calvin and Brent and the wonderful team that I had back at WebSavvy back then.

00:12:32: And just really solving that problem.

00:12:35: So we worked out that you could work out how much of your total Pmax spend how much you were spending on shopping.

00:12:42: Because you could go into the product groups, but you couldn't just.

00:12:47: Export that and some it all up right because each line was repeated to maybe three times.

00:12:52: you couldn't even just some it all up and drop it in half.

00:12:56: If it was a very simple campaign set up you could just about eyeball it and go.

00:13:01: it's probably around about twenty three grand last month but.

00:13:05: It was a bit painful to do and a more complicated setup with more product groups.

00:13:09: It was just hard.

00:13:12: But we kind of lived with that and we eyeballed it and we moved on.

00:13:15: And then we worked out one day that, oh, actually, if you add the columns for the number of views and the CPV, you could multiply those together and you could work out how much money you're actually spending on YouTube.

00:13:29: But gosh, this is really testing my memory.

00:13:33: that worked for accounts that had a YouTube campaign running or had ever ran a YouTube campaign.

00:13:40: But if that campaign had never turned on any YouTube campaigns, those columns weren't available.

00:13:46: But they were with scripts.

00:13:48: And so we've sort of found this magic hack we thought of like, huh, well, let's save doing all of the maths for the product groups.

00:13:55: Let's let the script do that.

00:13:57: Let's let the script work out how much we're spending on video, because it doesn't matter if they've got a YouTube campaign or not, we can work that out.

00:14:05: And then whatever's left is search and display.

00:14:07: That was version one.

00:14:09: And then I ended up chatting to a wonderful bloke in Copenhagen.

00:14:13: I actually met him in speaking at Bologna, Gian Paolo's event, the wonderful Adworld Experience, one of the best events in Europe.

00:14:25: Antibias was there.

00:14:27: We had chatted maybe two and a half years ago and Tobias showed me how they had figured out how to Get a rough estimation of that display spent.

00:14:36: so now we could split apart display and Search.

00:14:40: I said, oh, can I build this into easy?

00:14:42: Of course you can like go for it's not my thing.

00:14:44: I just but he discovered that not me and so built that in gave away a free version of this two and a half ish years ago and still felt like it was a bit of special source for us as an agency to do the big version of this that we had and that bucketed products very similar to the labelizer script and product hero now do, but we had sort of our own janky version of that back then.

00:15:14: And when I sold the agency, I kept the IP for the script.

00:15:19: And so, you know, four weeks later, I was speaking in Italy at Giampaolo's event.

00:15:24: I thought, yeah, bugger it.

00:15:26: Why not?

00:15:26: Let's see if more people want this and are willing to pay a couple hundred bucks for it.

00:15:31: That's all.

00:15:32: And it went gangbusters.

00:15:34: So I quickly built a little site and started selling it through that.

00:15:40: Then built an MCC version because a lot of people like, yeah, this is great, but I have to do this for every single account.

00:15:46: And every time you keep changing the lemon script, which is all the time because I kept having new ideas and everybody kept telling me what else they wanted.

00:15:53: and hey could we build this in?

00:15:54: and could you add this and like yeah here's a new version.

00:15:57: oh great now i've got to update fifty accounts again.

00:16:00: um so i created the mcc version and sold that and um yeah it just went went ballistic.

00:16:05: it was great um but yeah it's it's the it's the community around it but i'll have like all these people it was back to those conversations that i missed from the web.

00:16:13: seven days now all of a sudden like i'm having conversations about Oh, how are you using this?

00:16:18: Oh, that's really cool.

00:16:19: That's really smart.

00:16:19: And what would you like me to add to this genius idea?

00:16:23: Hang on.

00:16:23: Let me just let me go and add that.

00:16:25: So that's how it sort of built and built and built and brought up to version eighty eight now.

00:16:31: So keeps on going.

00:16:33: It's a great story.

00:16:34: Now, if someone is listening and feels like, hey, I also want to get started with building automations, building scripts, or maybe even building agents, although.

00:16:45: Like there are different specialties to these different stages of automation, but what would you recommend someone that says, okay, Mike, I'm down.

00:16:56: Next weekend is all like exploration time.

00:17:00: How should you get started with building that stuff?

00:17:05: Start with a problem.

00:17:07: Start by thinking about what are all the tasks that I do.

00:17:12: Rather than trying to solve this for someone else, think about your own tasks that you do.

00:17:18: of those, which ones are sort of data related?

00:17:21: And is there anything in there that you're doing on a regular basis?

00:17:24: Let's say more than three or four times a month.

00:17:26: So maybe you're pulling the same data for your boss.

00:17:29: Maybe you're creating some sort of client report and they want to see that every week or every month.

00:17:34: But you've got a few clients and it's kind of the same report.

00:17:36: But you're doing this stuff repeatedly and you find yourself doing the same sort of things over and over again.

00:17:43: That's the sort of stuff that a script is perfect for, assuming that it's Google Ads data.

00:17:48: If it's some other type of data, you can need some other type of automation.

00:17:54: To answer that question, a year and a half ago, I would have said something completely different to what I'll say now.

00:18:00: Because what I'll say now is just turn on your favorite AI, chat, GPT, code, whatever, and just start having a conversation with it.

00:18:10: Here's this thing that I do.

00:18:12: And I get this data and it looks like this.

00:18:14: And I find myself doing this and this and this and all the time.

00:18:18: I've heard of this thing called Google ad scripts.

00:18:20: I've never created one.

00:18:21: I've used other people's never created one.

00:18:23: How could I do this?

00:18:25: What script could we create?

00:18:27: So rather than just diving into chat, you can say, I need a script that's going to pull this data and do this.

00:18:32: Maybe you don't know what you don't know yet.

00:18:35: So have a conversation with an AI first.

00:18:37: Be prompted by the AI.

00:18:39: Say, can you please ask me some questions so that you can figure out, dear AI, what it is that we need to build?

00:18:46: Great.

00:18:47: Now, can you build that for me, please?

00:18:49: Because guess what?

00:18:49: It's seen millions and millions of Google Ads scripts, and it's really quite good at getting it right.

00:18:56: Sometimes it'll get it wrong, but the wonderful thing about code, it's a bit like maths, right?

00:19:00: There is a right answer.

00:19:02: So you take the script, as you well know, and you go chuck it into Google Ads, and it either runs or it doesn't run.

00:19:08: And if it doesn't run, there's usually some sort of messy error thing down the bottom.

00:19:11: You don't even have to understand what that means.

00:19:13: You just copy the whole thing, chuck it back into your favorite AI and say, your script got this error.

00:19:19: Fix it for me, please.

00:19:21: And it does.

00:19:22: And you'd usually only have to go around that loop.

00:19:24: these days with the powerful AI that we have now, maybe two or three times.

00:19:29: And all of a sudden, you've got a working script that solves that problem.

00:19:33: Awesome.

00:19:34: Use the time you've saved by automating that thing.

00:19:38: If it was ten minutes a week, it doesn't matter.

00:19:41: Use that time you've saved to go build something else.

00:19:43: I have a little calculator on my website now that shows, like, if we only spent twenty minutes a week building some sort of automation, and even if it takes three weeks, it takes three lots of twenty minutes to build one automation that saves me twenty minutes.

00:19:59: So I'm going to spend an hour building the thing that's going to save me twenty minutes.

00:20:04: It just shows this exponential chart.

00:20:05: if you keep doing that week after week after week after week and then you plug in What's my time actually worth?

00:20:13: And you very very quickly find out that you can save or make thousands of dollars a year depending on how you value your time or Dozens and dozens of hours per year if you prefer to think in terms of time saved just by continually going all what next?

00:20:31: i'm just and once you get the bug for this it becomes quite addictive right?

00:20:35: because you just start seeing problems everywhere and i just love solving problems.

00:20:40: i love the the intellectual part of that.

00:20:42: i'm like how am i going to do this?

00:20:44: and then you just keep pushing the edge of that.

00:20:47: i'm like okay what are some more problems?

00:20:50: yeah not that oh that's.

00:20:52: that seems like a hard problem to crack.

00:20:54: how am i going to solve?

00:20:56: oh yeah okay let's just go attack that problem and it's just addictive.

00:21:00: like I said before going from idea to to impact so quickly is just so much fun.

00:21:06: and yeah you don't need to learn to code.

00:21:08: it helps to know a little bit about code but I think it's much much easier to learn that as you go and you don't have to.

00:21:17: there are plenty of people using scripts that have no idea about code.

00:21:21: that's totally fine.

00:21:23: but if you do want to sort of understand a bit about code it is infinitely better to start by trying to create something, even if it's like a little thirty line script that Claude writes for you and you use, and then you start asking Claude, like, what does this part of the script do?

00:21:38: And why did you do it this way?

00:21:40: And give me three other ways to do this part.

00:21:44: And you learn by doing rather than learning the way.

00:21:46: I learned the code a long, long time ago, which is like learning if statements and for loops and variables and all this.

00:21:54: It was so abstract.

00:21:57: Back then, learning from a textbook for six months before you could actually write any kind of workable code.

00:22:02: Now it's like, give me workable code and I'll follow along with you and learn as we go.

00:22:07: It's a much, much more interesting way to do it.

00:22:10: But of course, you don't have to learn to code if you don't want to.

00:22:13: But you just have to get the bug for solving problems.

00:22:16: Yeah, I also remember myself ten years ago Googling all the time and hanging out on stack overflow and always searching for the part of the code that would help me solve this Nitty gritty problem in the WordPress plugin or WordPress theme.

00:22:32: It was such a different time, but

00:22:35: it was so painful back then.

00:22:36: right because I remember when I was a kid type copying out computer programs from a magazine to create some computer game.

00:22:45: It's like eighteen pages of code.

00:22:47: Of course, yeah, one semicolon out of place.

00:22:49: Nothing's going to run.

00:22:51: I'm like, mom, you're a touch typist.

00:22:53: Could you just could you actually do like two, three pages and then go, I'm done.

00:22:58: Back to me.

00:23:00: But it was so painful to debug those things back then.

00:23:04: And you got these weird error messages and you didn't know what they meant.

00:23:07: And this is before you could get a stack overflow and actually look it up.

00:23:10: But Now, it's just so simple.

00:23:12: It's like, here's the error.

00:23:14: You know what that means.

00:23:15: And even if it doesn't get it right first time, do that enough loops, like I say, two, three, maybe four times, it's going to solve the problem and write working code.

00:23:25: And that's the wonderful thing.

00:23:26: That's why AI is so good at code, because there is this kind of binary it works or it doesn't.

00:23:32: We thought that AI would take years and years to write creative things like poems and songs.

00:23:39: Turns out to be quite good at that too because it's just language.

00:23:42: but running code and just does it work.

00:23:44: or doesn't it work?

00:23:45: great Give it the problem.

00:23:47: It just goes around around that loop until it solves the problem.

00:23:49: It's great.

00:23:51: now give us a little bit often Behind the scenes look into the text that you're using because I mean people I think see chat GPT Now releasing GPT five.

00:24:03: they have they.

00:24:04: Some people, I feel like it's still quite unknown, have used Claude, then there's Gemini, then there's whatever, all sorts of stuff.

00:24:12: So what are you

00:24:16: using?

00:24:16: What are you using?

00:24:17: Or what would you recommend?

00:24:18: Maybe even if you're using something that you feel like you wouldn't recommend for somebody that's just starting out.

00:24:23: So where would you start?

00:24:24: And maybe

00:24:25: someone was just starting out exactly.

00:24:27: I would not recommend the text that I'm using today, but that has evolved to this point.

00:24:32: But also, I see very much Google ad scripts as just the first step on that bridge of going from ads to AI, right?

00:24:43: That's the gateway drug to go, oh, this is possible.

00:24:47: And the bit that I think might be useful for your audience, the bit that kind of got me over that hump at the beginning was this realization that code can run in parallel.

00:24:58: So you take the time to.

00:25:01: You found that ten minute task and you sit there thinking, is it really worth?

00:25:06: Mike said it's going to take me an hour?

00:25:07: So it's probably going to take me two.

00:25:09: Is it really worth me spending two hours of messing around with something I do not understand to save ten minutes?

00:25:16: Let me just try it.

00:25:17: OK, it's going to take twelve weeks to pay back that ten minutes, right?

00:25:21: Because of lots of ten minutes.

00:25:24: But it was that realization that I've built that little ten minute thing and that little ten minute thing and that little ten minute thing and I don't have to sit there and wait for them all to run one after the other.

00:25:36: I can come in at nine o'clock on a Monday morning, hit the button, and by ten past nine, fifty of these things are all complete.

00:25:43: And I've just done a day's worth of work, and it's only ten past nine.

00:25:46: I haven't even had my second coffee yet.

00:25:48: Oh, my goodness.

00:25:50: This is quite fun.

00:25:51: And then, again, you re-invest that time, right?

00:25:53: So at the beginning, yeah, pick one.

00:25:57: I think it's very useful.

00:25:59: to be aware of a few tools.

00:26:02: It's kind of that T-shaped marketer type idea really, isn't it?

00:26:05: Of like being aware of lots of different tools, but probably going fairly deep into one because you get a bit of an intuition, you get a bit of a feeling for a particular model, which is why last week there was all this uproar about GPT-IV going away and GPT-V replacing it because people felt that they kind of, they understood GPT-IV that the way it Spoke to them.

00:26:29: It's sort of it had a personality if you like.

00:26:31: I know we tend to anthropomorphize things but It does feel like these things have a bit of a personality.

00:26:36: you get used to it You get used to the way of working with it And then when that disappears and a new thing comes along people just aren't very good with change usually right.

00:26:45: so Pick one.

00:26:47: Yeah, I agree the vast majority of people it's yeah most stats are made up on the spot But it's something like ninety five ninety ninety five percent of people using chat gbt.

00:26:57: at this point Claude is a very, very, very distant second or third.

00:27:02: And then, yeah, Google have their Gemini.

00:27:04: Gemini is really good at code.

00:27:06: And he would think, oh, it's Google ad scripts and Google Sheets.

00:27:09: So maybe I should use Gemini, but their interface, like the website where you go to use this thing is just horrendous.

00:27:18: I hate the Gemini website.

00:27:20: So I personally would use Claude.

00:27:23: I love the way Claude works, but also the Anthropic team are just.

00:27:28: They are shipping stuff like all the time.

00:27:30: So I now use a tool called Claude code.

00:27:33: I probably went from just writing code.

00:27:38: See, when you start, and you wouldn't do this today, but when you start, you start doing it all in the Google ads interface.

00:27:43: And it's a horrible place to try and edit code or see what's going on.

00:27:47: Definitely don't do that.

00:27:48: So today, if you were starting out, you'd probably start in chat GBT because you're probably listening to this and going, well, that's what I use.

00:27:54: So start there.

00:27:56: But if you're open to change, then have a go with claw.ai as well.

00:28:02: And you do that for a bit.

00:28:03: And then when you get to the point of, okay, I want a place to save all of these scripts that I'm downloading from LinkedIn and I'm making little changes too.

00:28:13: I just want to build up my own little library of these.

00:28:16: I don't just want to save them in an account.

00:28:19: What happens if I lose access to that account, right?

00:28:21: You might be saving those.

00:28:22: Don't save them as Google Docs.

00:28:25: That's when you sort of graduate to using a tool like cursor.

00:28:28: So that's sort of the next step on from just doing everything in a browser, like in a chat GBT.

00:28:34: A tool like cursor is fantastic because you can save all of your files in different folders down the left hand side.

00:28:40: Your code is in the middle of the screen.

00:28:42: And then you've got this AI agent over on the right.

00:28:45: So instead of copying and pasting code between all these different screens, you've got everything you need in front of you.

00:28:52: And then eventually you'll get bored with cursors limitations.

00:28:57: And that's when you might use tool like Claude code, which I use inside of cursor, because I like the visual aspect of cursor and being able to see all of my files and see the code changing.

00:29:07: But Claude code is just the most powerful coding tool.

00:29:11: And it's kind of almost programmable.

00:29:13: It's hard to explain and we definitely don't need to go there.

00:29:15: But you can basically take huge big things that you do and automate them down to a word.

00:29:22: And so typing a series of a few words, you can do an amazing amount of work in a very, very short time.

00:29:29: It's just incredibly powerful.

00:29:30: It's what the Anthropic team use to build Cloud.

00:29:33: So, you know, they're quite good at it.

00:29:36: They've made it basically freely available.

00:29:38: And it's just an incredibly powerful tool.

00:29:40: But I wouldn't start there.

00:29:41: I would start in cloud.ai or chat.gbt.

00:29:44: Or if you really, really love Google, you might try and start in the Gemini app.

00:29:49: But it's just, it's a bit ugly, and it's just a bit.

00:29:52: Again, not very nice to use.

00:29:55: Probably you're right the session we have today is not enough to unpack all this and go as deep as you could.

00:30:02: or maybe we should also for the audience, but you already dropped.

00:30:07: You already dropped an important keyword which was from ads to AI Because you recently launched this membership program.

00:30:15: I think it's quite new or at least fairly new.

00:30:19: Yeah,

00:30:19: yeah, so actually For me, let's maybe talk a little bit about that because I feel like you have a vast amount of knowledge and you were already telling the audience that you like to teach people stuff, etc.

00:30:33: So I can imagine this being a really, really valuable program.

00:30:37: But what's the vision here?

00:30:39: What have people, what's to expect?

00:30:42: So I picture this journey from ads to AI.

00:30:47: I think that, I think a lot of businesses, a lot of agencies are going to be just fine.

00:30:53: But a lot of businesses are going to feel the squeeze of commoditization over the next two, three years.

00:30:58: I think that for many, just offering the services that they offer today, maybe ad management services, I don't think that will be enough in isolation.

00:31:07: I think there'll be pricing pressure on those services as businesses sort of start to say, well, isn't the whole thing automated?

00:31:15: You know, Pmax and all these things in meta, like, what do you actually do?

00:31:20: They'll be saying.

00:31:21: They still, they were already saying that.

00:31:24: So I see this sort of path to offering new services to the same clients.

00:31:35: I mean, fifteen, twenty years ago when I started, small businesses were crying out for somebody.

00:31:40: If you could spell AdWords back then, you got a client.

00:31:43: And it's kind of the same now.

00:31:45: All these small businesses, they kind of feel like, yeah, we know MetaAds, we can do that ourselves.

00:31:50: Thanks.

00:31:51: But this AI thing, They're confused, they're overwhelmed, they're short on time like never before.

00:31:57: They need help desperately.

00:31:59: So, agencies that are on the cutting edge here and you don't really need to be too far ahead of everybody else to be on that cutting edge right now have this huge advantage to solve those problems for the same clients they want to delight and surprise today.

00:32:16: but they're just going to have to transition their services a little bit.

00:32:20: They may still offer those ad services, but there's probably some adjacent services there teaching those small businesses how to get the most out of AI.

00:32:28: The question I hit all the time is, yeah, but where do I start?

00:32:32: And so I think for those agencies, they need to start on themselves.

00:32:35: You know, what Americans would call dogfooding.

00:32:37: You've got to be doing this to yourself and with your team and upskilling and lifting up your whole team.

00:32:45: completely changing probably how your agency approaches things.

00:32:51: And it's not just about efficiency.

00:32:52: It's not just about cramming more work into the same number of people, but it's definitely not about, oh, how can I replace all of these warm bodies with some cold hard code and save more money?

00:33:03: Mr.

00:33:03: Burns, it's not about that at all.

00:33:05: I don't want to work with those agencies.

00:33:07: But the agencies that are like, how do we take the amazing skills that we have in front of us?

00:33:13: and use these incredible tools like AI and automation to augment all of this to make these guys so much better at what they're capable of doing for our clients.

00:33:25: Do that to yourself first and then go provide those additional services, maybe extra products to those small businesses because I think those businesses, they're not stupid, they're going to be crying out for that.

00:33:37: And if you're going, no, we just do Google Ads Management because we're really, really good at it.

00:33:42: And they're going, yeah, thanks.

00:33:43: I'm going to jump over to this guy who can do all of this stuff and teach us how to use AI and completely revamp the way we do content and change how we do all of these things, build these little tools for us internally.

00:33:55: We used to spend twenty grand a year on this SAS tool.

00:33:59: And this agency over here just built us a little version of that in a weekend, exactly to our specs.

00:34:06: It's the same with the agency, right?

00:34:08: then have to, if you know how to do this stuff, you don't have to do kind of things to like, eighty, ninety percent of the way you want to do them because you're trying to fit your system into a tool, build your own tool.

00:34:20: And now you've got it doing it exactly the way that you want.

00:34:24: And you have the skills to keep changing that and evolving that.

00:34:27: And as your team keep coming up with great new ideas of how you can use it, that's a great idea.

00:34:31: Let's build that in.

00:34:32: I'll just spend half the weekend on that.

00:34:34: And now we've got an even better tool that everybody can use.

00:34:37: And so that for me is that journey.

00:34:39: I picture it as a bridge from ads to AI.

00:34:42: I think the scripts are just the first step on there.

00:34:45: Like I said, the gateway drug, that's the bit that makes you go, oh, wow, this is fun.

00:34:50: This is possible.

00:34:51: I save all of this time and I get to solve problems and build stuff.

00:34:55: It's figuring out who on your team might be you.

00:35:00: It might be somebody three levels down.

00:35:02: It might be somebody three levels up, but figure out who the people are.

00:35:06: in your agency, in your business, or if you're a freelancer, it's probably going to be you.

00:35:12: Figure out who those people are and invest in them.

00:35:15: I have this line I use on stage of five percent or a hundred percent.

00:35:20: So think about the amount of income, top line, not bottom line, top line number for last year and take five percent of that.

00:35:28: Or if you're a freelancer, think about your income for the last year, take five percent of that.

00:35:33: That is the amount I believe you need to be investing right now in your AI literacy, in your education, because if you don't, you're risking the hundred percent of that number.

00:35:44: You're risking being obsolete and being gone in two, three years.

00:35:49: So I think that's a pretty easy thing to weigh up, five percent or a hundred percent.

00:35:53: But I still see so, so many people just sort of head in the sand of like, no, it's all a fad.

00:35:57: I don't need to learn anything about this.

00:36:00: It'll all blow over.

00:36:01: I'm going to keep going the way I'm going because it's the way we've always done it.

00:36:04: I think those businesses are going to be in trouble.

00:36:06: I'm not interested in working with those, but the ones that are really interested in like, we know we don't know, but we really want to learn, that's micro.

00:36:14: I love working with those people.

00:36:17: And what can people expect from the program?

00:36:19: Like, how does it work?

00:36:21: Because I mean, I can imagine a lot of people thinking like, hey, this sounds really great, Mike.

00:36:26: Where can I join?

00:36:28: How much does it cost?

00:36:29: What does it include?

00:36:30: It costs about two bucks a day.

00:36:32: That will go up at some point.

00:36:33: So I love pricing that goes up into years.

00:36:35: So I had the first hundred people got it for five hundred euro a year.

00:36:40: Currently it's six fifty and it's about to go up to eight hundred for a year.

00:36:44: Or you can pay, I think it's about three hundred bucks a month.

00:36:46: So it's clearly aimed at people that want to be with me for a year because I want to massively over deliver on that year.

00:36:53: So I have weekly videos where I talk about tools.

00:36:57: and show off how I'm using various tools and how small businesses are using tools.

00:37:01: We have a monthly office hours Q&A webinar.

00:37:05: I just had one couple of nights ago.

00:37:06: It was fantastic where people can come in, ask questions, maybe show off a little thing, but just get unstuck.

00:37:13: And then I'm creating a new course every quarter as well.

00:37:16: So I drop my courses in there.

00:37:19: They get courses before anybody else.

00:37:21: They get the courses for free.

00:37:23: And those courses, yeah, maybe I go on and sell elsewhere.

00:37:26: redone recently my entire Scripts and Sheets Mastery course for twenty twenty five completely redid that.

00:37:32: Last week I just dropped a Claude code course showing how to use Claude code inside of cursor, which has proven to be quite popular in the past week and a half that it's been out.

00:37:43: I've got so, so many plans.

00:37:45: I'm in the process of creating a basically an AI literacy course.

00:37:50: not just for them, but for them to be able to use with their clients so that they can take it, white label it, pass that on to their client and say, here's a course that will teach you and your team in small business world how to use AI.

00:38:03: And they just get that for free.

00:38:05: They get the transcripts, they get the slides, they can tweak it and change it.

00:38:08: They can use my videos or they can rerecord their own.

00:38:11: I just, yeah, I just, I love over delivering on value.

00:38:15: I hate feeling ripped off.

00:38:16: And so I'm obsessed with over delivering on value.

00:38:20: Yeah, build the agent stuff.

00:38:21: So I've done a few.

00:38:22: build the agent courses or that stuff will be in adds to AI in the nearest future by the end of the year.

00:38:29: That'll be the next course that I do in Q four.

00:38:32: And then loads and loads of plans for more stuff.

00:38:35: Yeah, couple of bucks a day, basically.

00:38:39: I think it sounds really good.

00:38:40: So I think we'll just drop the link to the landing page in the description of the video below so everybody can can check it out.

00:38:47: Yeah, adds to AI.com.

00:38:49: Pretty easy.

00:38:50: I found the shortest domain I could.

00:38:52: It doesn't really matter if you type in from ads to AI, TO, or the number two, I've got them all.

00:38:57: I like buying domains.

00:38:59: Really, really great.

00:39:02: You already mentioned built the agent and I also saw you doing this in-person workshop in, I think it was in Amsterdam, right?

00:39:09: Yeah.

00:39:10: Yeah.

00:39:11: So now we were talking about scripts, we were talking about Writing code, we're talking about cloud code, automations, agents, all that stuff.

00:39:21: Can you give us a little bit of your perspective and your mindset around the differentiation, for example, between automation, general automation, maybe also automated workflows and then agents?

00:39:36: Right.

00:39:37: Yeah, I think, so firstly, the definition of agents has changed a lot over the past one month.

00:39:44: There was a quote I had once, technology.

00:39:46: Technology is the stuff that doesn't work yet.

00:39:48: You know, it's the stuff that's just out of reach.

00:39:51: And so the definition of an agent, twelve months ago, today we look at that and go, yeah, of course it can do that.

00:39:58: But an agent should be the thing that does this.

00:40:00: It's the thing that's just over the hill, just out of reach.

00:40:03: I think that, well, when I first got into AI, I'll put it this way, when I first sort of went deep into AI about seven, eight years ago, I came away thinking, you know what, most businesses don't need AI at all.

00:40:17: They really need automation.

00:40:19: I wrote this whole business plan out for like an automation agency and then realized I didn't want to do that because automation kind of bores me.

00:40:27: And there were lots of those.

00:40:29: There's a great guy here in Australia, a good mate who's built the automation agency.

00:40:33: That's literally the name of his business.

00:40:35: And it's really good.

00:40:36: And it does really good stuff.

00:40:37: So we just would send people his way.

00:40:41: Where am I going with that?

00:40:42: Automation, I think, is the place to start because most of the time that will solve the problem.

00:40:50: You don't need the latest shiny object.

00:40:53: You don't need to be like sending a whole bunch of stuff off to a large language model to get a response in text back.

00:41:01: Actually, you kind of need something that works every single time in exactly the same way and just looking for those opportunities, looking for things we can go automate is a really good place to start.

00:41:13: Really easy way to do that, by the way, if someone's listening.

00:41:17: Next time, first of all, think about the tasks that you do and think about those that maybe you do fairly often and don't require huge amounts of human judgment.

00:41:29: So you can sort of plot all your tasks out on that scale of really frequent to infrequent and loads of judgment to not very much judgment.

00:41:36: Find that quadrant on that little chart.

00:41:39: Stuff I'm doing three, four times a month at least.

00:41:41: not massive amounts of judgment.

00:41:43: Think about all the tasks in that quadrant and then record yourself, turn on Loom and record yourself doing that task and just talk your way through it.

00:41:53: Maybe it takes you twenty percent longer than it usually would be, just talk out loud as you're going through.

00:41:57: When you get to the end, Loom will not only create the transcription of that for you, there's a button next to that transcription that turns that into an SOP or standard operating procedure.

00:42:09: Then you take that SOP, Over to chat GPT five and you turn on thinking mode up at the top and you say here's a task that I do all the time.

00:42:18: Can you see any opportunity?

00:42:20: To help automate any of this or are there any steps that we could eliminate or are there any steps?

00:42:24: You think I'm missing but basically how can you help me?

00:42:28: Automate this task and it'll give you a bunch of ideas.

00:42:31: some of them will be made up and wrong Yes, but most of the time a lot of it will be right and if it's Google ad space stuff you might even Nudge it in the right direction say like are there any ways to use Google ads scripts to help with any of this stuff?

00:42:44: or Google apps scripts, which is a very close cousin?

00:42:48: But if you're working with spreadsheets a lot, you definitely want to look into apps scripts.

00:42:51: They're incredible.

00:42:52: They're powerful.

00:42:54: So you start there.

00:42:56: Hey, you mentioned AI workflow.

00:42:57: I think most things in AI world right now are workflows rather than truly autonomous agents.

00:43:05: There's lots of arguments in the AI world about what actually is an agent and how autonomous is it and how agendic is it.

00:43:11: I really don't think that matters too much.

00:43:13: What matters is does it solve problems for you, for your business?

00:43:17: And we've all seen those flowcharts on LinkedIn and drop the thing down below in the comments to get the thing that gives you all that rubbish.

00:43:27: Those flowcharts, something like a make or a zappier or Today's flavor of the month is N-Eight-N, which I know you know very well.

00:43:36: Maybe one or two of the boxes on that big flowchart are large language models and AI doing its thing.

00:43:43: So you get to that step and you take all of that information to that point and you send it off to Claude and it comes back with some text.

00:43:49: You go, great, thanks for that.

00:43:51: And then you use the output from Claude as the input for the next step.

00:43:55: So you just, oh, okay.

00:43:57: So that's the bit where a human maybe made some sort of decision about something, and that was the point where we had to go, is it more this or more that?

00:44:05: Is this email really urgent or can it wait till tomorrow?

00:44:09: If it's really urgent, then do this step next.

00:44:12: Otherwise, let's go off on this other trail.

00:44:16: Or is this customer review really positive, really negative?

00:44:19: Do we need to respond to this urgently or not?

00:44:22: Do we need to thank them?

00:44:23: Do we need to send them a gift?

00:44:25: So I think thinking about it in those terms, that's exactly the right phrase that you used of an AI workflow.

00:44:32: So it's just an automation workflow.

00:44:33: But some of those steps might be AI.

00:44:37: But I think they get conflated.

00:44:39: And I think that what a lot of people call AI is really just automation.

00:44:44: And that's totally fine.

00:44:46: And more people should be using a lot, lot more automation.

00:44:51: I've heard some horrendous stats.

00:44:53: recently about just the average number of SaaS tools that a small business uses.

00:45:00: Apparently, it's on fifty on average, and they're copying data from a sheet into zero to create some invoice, and then they're copying something from Gmail into Asana, and they're copying from here, and it's just all copying and pasting stuff.

00:45:15: If you find yourself copying and pasting, or if you find yourself saying things like, To do that, we go over to this spreadsheet and we look over here to look for those little phrases being used around the office or by your workmates or by your clients.

00:45:32: That's a really good automation opportunity because if you're going over to there to get some information and then going over to there to get some information, then maybe you stay being the human in the loop there, synthesizing all of this together.

00:45:45: Maybe we don't try and use an AI for that yet.

00:45:48: It's worth trying.

00:45:50: But please, please, it's twenty twenty five.

00:45:54: Please don't be like going and fetching all of this data bit by bit by hand every time you do that task.

00:46:00: Have everything brought to you in the format that you want so that it's always there at the ready whenever you need it.

00:46:09: And then, sure, keep making the decision yourself for now, but also in parallel, be testing that, you know, be testing.

00:46:18: Can we use an AI to make this decision that I make at this point before we bounce over to the next step in the automation?

00:46:25: Let's run it in parallel.

00:46:25: I'll make the decision ten times.

00:46:27: Let's get the AI to do it ten times and let's compare results.

00:46:30: And when we get to the point that the AI is kind of like nine out of ten, good.

00:46:34: Maybe that's something else that we can get off our plate and we reinvest that time into working on the next bit and on it goes.

00:46:40: I

00:46:42: remember so well like eight years ago I was working as a student at a big internet company in Germany and I felt like okay I want to.

00:46:49: so besides all freelance work I was already doing I wanted to get an idea about how actual business and internet business works and remember this colleague and I liked him a lot.

00:47:01: but I always felt like when I was looking at his screen he was literally copying.

00:47:06: impressions and click data from one Excel sheet to the other Excel sheet, which was then connected to a PowerPoint slide deck, which was then manually presented to a client and sent out.

00:47:18: This is exactly even at that point.

00:47:22: AI for me was not a thing back then.

00:47:25: So eight years ago, I was like, Hey, this can't be the case.

00:47:30: And then when I, for example, stumbled upon Zapier and felt like, Oh, cool, you can exactly what you were telling just right now, you can connect these tools and send data over there and there.

00:47:43: And I feel like a lot of people just have to train this muscle more, right?

00:47:46: This sensitivity for

00:47:51: it's been.

00:47:52: Yeah, for me, it was kind of a boring thing.

00:47:54: I could work out how to do at this point.

00:47:56: But I think for a lot of people up until now, it's been very hard to do that.

00:48:00: You've obviously got the brain for it, right?

00:48:02: You see it and you go, oh, I can see what the pieces of that are.

00:48:04: I can see how they fit together.

00:48:06: And tools like that, like NA-TEN, there's a bit of a lift there.

00:48:10: And a lot of people don't want to spend the ten, twenty hours to learn that up to a decent level.

00:48:15: It's the same with Photoshop, right?

00:48:17: You've got to tip in the time to learn how to use the tool.

00:48:20: The wonderful thing is now, So many of those tools just run on code in the background, right?

00:48:26: So Claude can create, or Chatcha BT if you want to use it, or Gemini, bless you, can create that JSON file or that code that builds all of those little boxes on the screen that tells you how to automate the thing.

00:48:39: Maybe you need to go in and tweak it a little bit.

00:48:41: You still need to know how to use the tool, but it's never been easier and it's going to keep getting easier and easier.

00:48:49: So I think it's about like finding the way to solve the problem that you enjoy.

00:48:54: I just love solving the problem.

00:48:56: I like the intellectual challenge of that.

00:48:58: I like seeing people's face light up and when they get the bug and it's like,

00:49:02: can we

00:49:04: really we can do that?

00:49:05: Oh my God, that's so cool.

00:49:07: And that's that's why I did build the agent.

00:49:09: because it was just like sheets were getting a little bit.

00:49:12: And straining, you know, like the UI of Sheets, having to do it cell by cell, I would come up with a new version of the PMAX script.

00:49:19: And that was the easy bit.

00:49:21: And then I'd spend like three days fiddling around in the sheet, getting all the formulas in the perfect place.

00:49:26: That's why I kind of moved to doing stuff inside of an app.

00:49:30: And I needed to build the app first before I ran a program called build the agent, before I tried to teach that.

00:49:36: And it's just, oh, it's just so much fun, like having a slider.

00:49:42: I've

00:49:43: got a minute.

00:49:43: I can give you an example for your listeners.

00:49:46: If you go to AT-twenty-agent.com, you do not have to sign up for an account.

00:49:49: Accounts are free, but you don't have to sign up.

00:49:51: I've put loads of dummy data in there so that you can play with the tool just to experience what it's like.

00:49:57: So go to AT-twenty-agent.com, and then you can use the sample data.

00:50:01: You go up to the top and you find the keyword page.

00:50:03: I think it's under Analyze Account.

00:50:04: There's a few pages on there now, and there's a Keywords page, and it shows you a tree.

00:50:10: of keywords.

00:50:11: It's just a really intuitive way of thinking about campaigns that branch out into ad groups that branch out into keywords.

00:50:20: And as you move the slider for cost or impressions, you move the slider up and it shows you only keywords that have spent more than ten dollars last month.

00:50:28: Move the slider up.

00:50:29: Let's just show only the keywords that have spent more than fifty, more than a hundred, more than five hundred.

00:50:33: And so it's a really intuitive way to say, what are the most important keywords in my account?

00:50:38: Now, yes, you can go into Google Ads and you can set a filter there, but it just feels a bit more cold and impersonal.

00:50:44: I love the UI and constantly trying different ways of displaying data and turning it into useful information.

00:50:52: And then, yeah, you can try hacking AI over the top of that and turn useful information into insight, into actionable insight.

00:51:00: We're kind of right on the border.

00:51:03: to finally answer your earlier question about agents.

00:51:05: We're kind of right on the border between insights and action.

00:51:09: I think when we're really in agent mode, maybe next year is actually the year of agents, we'll be trusting the agents to then take those actions and do much more of the stuff for us.

00:51:21: But right now we're at the point where the AI can help us surface those insights and sort of solve the blank page problem and give us an eighty percent good enough, hence the name, eighty twenty agent, give us an eighty percent good enough insight.

00:51:35: But then it's kind of still on us at that point to go.

00:51:38: actually let me let me bring my human brain to bear on that and Synthesize it with some other data and think about that and then I'll go take the action.

00:51:47: for now I'm not at the point where I'm ready to trust The agent to do everything from where to go, but I'm sure yeah, I'm sure that that's gonna come.

00:51:59: I think the keyword tree example is so cool.

00:52:01: I also obviously tried the AT-Twenty agent.

00:52:04: So I can just recommend everybody signing up for an account and connecting there or one of their own Google Ads accounts or a client Google Ads account.

00:52:12: I did it and I found it really useful.

00:52:15: I think this is something that I just want to do a plus one for the data visualization part because I think a lot of people still feel like they have to just Tussles through these spreadsheets etc.

00:52:27: and uncover stuff whereas.

00:52:31: Well done data visualization can help you see stuff that is just impossible to see from just like campaign campaign campaign.

00:52:40: some cost some conversions really hard especially if you take in a dimension of time like developments over time.

00:52:47: so if you think about like okay hey my campaigns went down the last month why did it happen.

00:52:54: Probably there's something that happened either to the keywords, to the ads, maybe something like in a particular week and like trying to track those changes down and then revert it or come up with optimizations.

00:53:07: I think this is where people confuse their job.

00:53:10: They think their job is to hunt down this stuff, whereas the job is to actually uncover this and then come up with ideas to solve it.

00:53:19: Or what's your take on that?

00:53:20: Do you find this a reasonable

00:53:21: approach?

00:53:22: I love that.

00:53:23: I thought of it in those terms.

00:53:26: It's adjacent to one of those things I was saying before about.

00:53:29: you shouldn't have to go get the data from there.

00:53:32: I want the data brought to me.

00:53:34: I want those insights presented to me in such a way that I'm going to get it straight away and then we get to action and then we get to what clients really want, which is how is the business going to change?

00:53:46: What's going to happen to my profit?

00:53:48: when you do all that stuff.

00:53:49: They don't want to report with a whole bunch of insights.

00:53:51: They definitely don't want a dashboard with a whole bunch of data.

00:53:53: What they really want isn't just the list of actions, but what they really want is like, what's going to happen to my business?

00:54:00: How much more profit is this business going to do?

00:54:02: How many more people are we going to be able to serve?

00:54:04: How many more people am I going to be able to hire when I grow this business?

00:54:08: That's what they care about.

00:54:09: And yet, you know, Avinash, who was Google's digital evangelist, I think was his title for many, many years at Google.

00:54:16: He's left Google now.

00:54:17: wonderful guy.

00:54:18: If you ever get a chance to see him speak, please do.

00:54:20: He's a wonderful, wonderful speaker.

00:54:22: He talks about the data vomit.

00:54:25: All these agencies are just vomiting data over their clients.

00:54:28: Clients don't want that.

00:54:29: What they want is expected business outcome.

00:54:31: What's going to happen when we do all of these things?

00:54:35: But yes, it's very, very similar.

00:54:37: Like, have it brought to me make the getting of the insight way easier and way faster so that you can, as you said, then go do something about it, make the change, run the experiment.

00:54:50: I think most agencies would be greatly served trying to outdo the number of experiments that they ran last month.

00:54:59: That should be the leaderboard of like, no, we run a hundred and thirty six experiments.

00:55:04: That's the thing I think that clients want to hear too.

00:55:06: Like we have a a system for running experiments, a system for always trying new creative, trying new things.

00:55:13: Sam Tomlinson talks about this a lot in his wonderful Sunday newsletter, like the velocity that you can test new creative at is going to be one of the key things that matters, not?

00:55:26: do you have a really good way of looking at engrams?

00:55:30: That helps, but it's about speed, really, and we're all under pressure to be more efficient.

00:55:38: That's what I love helping people with.

00:55:41: We will get to the point where like, okay, so then everybody's using the same tools and everybody's thirty percent more efficient than they were last year.

00:55:48: Yeah, true.

00:55:49: That becomes a new baseline and then you just have to keep on going to the next bit.

00:55:54: I mean, this is the bit I don't like saying, but it's true.

00:55:56: This is the slowest it's ever going to be.

00:56:00: It only gets faster from here.

00:56:01: Sorry, this is the slowest the world's going to be for the rest of your life.

00:56:05: But it's also, this is the worst AI we're going to have.

00:56:08: And yes, I'm completely red-gilled on the AI, but I honestly believe if it never got better from here, we would spend the next five years putting in place the AI that we have available today, because businesses tend to move slowly.

00:56:21: It's incredibly powerful.

00:56:23: It's stupid too, and it does weird stuff, and it hallucinates.

00:56:26: I get all of that.

00:56:27: It's not all singing or dancing, brilliant at everything.

00:56:31: But your job, if you're still listening to me drone on an hour later, your job listening to this is to figure out what it can help with and what it's not good for.

00:56:40: And you do that by playing with it, by experimenting, by testing it out.

00:56:43: Go spend ten, twenty hours with it.

00:56:46: Every single thing that you do throughout the week, every task that you do, you're like, hmm, how could AI help me with this?

00:56:54: I wonder.

00:56:54: And if you're really unsure, turn on voice mode and chat GPT, go for a walk and have a chat with it.

00:57:00: Ask it, explain what you do, and ask it where it thinks there might be ways to automate, ways that AI can help.

00:57:08: I love chatting to the... So I use the chat GBT mode on the phone for a conversation.

00:57:15: I use Claude mostly for code, but I love that chat GBT app.

00:57:21: It's quite amazing to listen to.

00:57:23: My kids love talking to it as well.

00:57:24: It's quite bizarre.

00:57:26: What do you know about Mike's daughter, Frankie?

00:57:31: I did it too.

00:57:32: I had a walk in the park and just was talking like half an hour to chat about some, I don't know, random problem or challenge in the business.

00:57:44: And I felt like it's just also help structuring your thoughts.

00:57:48: And I think there's just this key insight.

00:57:51: It goes back to an article that a German, I don't know if it was a philosophist or like doesn't matter, but Heinrich von Kleist is his name and he wrote this article about the, I don't know the English term, but basically the idea is that while talking, you have to structure your thoughts in a way, if you talk to another person, you have to structure your thoughts in a way that the other person can understand what you're saying.

00:58:18: And if you do this, and it also applies to public speakers, for example, you have this, you have this this wild idea about where you want to go with your thought.

00:58:28: But like on the go, you have this energy of just your thoughts, structuring like the unstructured thoughts in your head, structuring around this, this really cohesive message.

00:58:39: And this, this thing, I just, I can't stress this enough.

00:58:43: If you have a complex problem in your head and you feel like, Hey, this stresses me out.

00:58:47: It's like so confusing.

00:58:48: Talk to someone, talk to a person that at least generally gets the context.

00:58:54: What is this?

00:58:55: Okay, you have to tell me what this is in a second.

00:58:57: or talk to AI.

00:58:58: It goes, it also helps.

00:58:59: There

00:59:00: is a concept in computer science called rubber ducking.

00:59:04: The idea being that you have this object on your desk and many people have an actual rubber duck on my desk.

00:59:09: I really should have a rubber duck sitting here for this example, but I have this little three D printed metal thing.

00:59:14: I've had it for years and years.

00:59:15: It's one of those kind of shapes that you can't make, but it's been three D printed.

00:59:19: And I just have it here and I sort of sit with it.

00:59:21: It's not that I talked to this, but.

00:59:23: Yeah, you're dead right.

00:59:24: talking out loud helps.

00:59:26: the idea with the rubber duck is you explain your problem to the rubber duck and the act of talking out loud and explaining it.

00:59:33: So it's like fifty eighty percent of most of the problem.

00:59:37: Is a rubber duck that talks back.

00:59:40: And it can ask you questions and yes, it can be a bit sycophantic sometimes and it can blow smoke up your ass as we would say in Australia, but you can also prompt it and say I want you to play devil's advocate.

00:59:50: I want you to push back on me.

00:59:51: I want you to think critically about this.

00:59:53: Don't just say no to everything, but think critically about this.

00:59:56: And maybe poke me a little bit where my argument is weak or you can use it to role play.

01:00:02: If you're going for a job interview, you can use it or if you're interviewing someone, you can say like, what are they likely to say?

01:00:08: And what are my questions?

01:00:10: What were a bunch of good questions for this particular candidate?

01:00:13: Not like, give me five questions that I'll ask everybody.

01:00:17: Here's the CV.

01:00:18: is their work history.

01:00:19: Here's everything you know about our business and the role and the PD and the job description.

01:00:24: And what should I be asking this specific person as wonderful for stuff like that?

01:00:31: It's so personalized and tailored.

01:00:34: And yeah, having a chat to it to get clear around your thoughts, I think it's one of the most underrated uses of AI.

01:00:41: I'm just in the process.

01:00:42: I'm trying to build this new app just this week.

01:00:45: just been playing with a bit.

01:00:46: So I've no idea when this comes out and if it will be ready on time.

01:00:50: But I'm going to keep going with it even though it's doing my head in as a way to show people the power of just adding that little bit more context of telling the AI a little bit more about yourself, about your business or about the products that you sell or the services that you offer and just how much better the response can be if you take the time up front, that upfront investment to educate the AI.

01:01:16: It's the old, what's the saying?

01:01:17: I don't remember who it was.

01:01:20: A Lincoln or Benjamin Franklin or somebody, if I had six hours to drop down the tree, I spend the first four hours sharpening my axe.

01:01:27: It's that upfront investment.

01:01:29: I get that nobody has the time to do that.

01:01:31: But oh my God, if you do, the payoff is massive.

01:01:36: Yeah, we also use this concept at our business that in also for our clients that we call a super prompt and it's exactly What you were just emphasizing.

01:01:45: so we have this document where we just say Hey, this is the business.

01:01:49: This is the target group.

01:01:50: These are key demographics.

01:01:51: These are the products, these are the services.

01:01:54: These are the core objections that people have about the service.

01:01:57: This is how this differentiates.

01:01:59: And this is like it, depending on the client, it ranges from a thousand five hundred words to something like three, three K words.

01:02:07: So there might also be like text examples in there.

01:02:11: Yes.

01:02:11: And the quality of the output that you get out, especially like.

01:02:15: Yeah, around strategic thinking, it's already great.

01:02:18: But especially if you write like content, for example, if you work with landing pages and all that stuff, like it's just, it's a night and day difference.

01:02:27: Yeah, you mentioned the strategy, right?

01:02:30: AI has basically read the entire internet.

01:02:32: So it's seen every slide deck that McKinsey and Bain and BCG and every consulting firm has ever put out online.

01:02:38: It's read all of that.

01:02:39: As you say, it is a genius management consultant, but If you don't tell it about you, about your business, about the stuff that you do, you're just going to get bland average answers because it's just predicting the next token.

01:02:53: That's all it is.

01:02:53: It's just a big bag of words and it's just pulling words out the scrabble bag and it's just predicting the next word.

01:02:59: But the more context you give it, the more information about all of that stuff around you, then that helps sort of ground it.

01:03:07: It helps move it away from that.

01:03:09: That average has this really strong gravitational pull, that average.

01:03:13: push it away from that center and push it out to the edges by explaining all of that information that you're talking about and drag it away from the boring average center.

01:03:24: That's where the gold is.

01:03:25: That's where the really interesting answers are, where it can be the best management consultant in the world.

01:03:30: But on your specific business, on that particular problem at that moment in time for that person, it's incredible.

01:03:37: But again, it's an upfront investment.

01:03:39: It's a bit of a heavy lift.

01:03:40: So I'm building this tool to sort of say, like, here's some dummy data again.

01:03:45: Play with it.

01:03:46: See what's possible with and without this context.

01:03:48: This is the answer you get without.

01:03:49: This is the answer you get with.

01:03:51: Try it with different models.

01:03:52: See which one works best for you.

01:03:54: Now, bit by bit, add your information to that because nobody wants to sit down and write a three thousand word document in one go.

01:04:03: But you can add that bit by bit.

01:04:05: Oh, okay.

01:04:05: Now I can see the benefit of doing that.

01:04:08: I'm willing to do the next little bit.

01:04:10: And yeah, I use a tool, WhisperFlow.

01:04:12: That's part of my tech stack that you were asking about before.

01:04:14: I love WhisperFlow.

01:04:16: Roads.ai, if you want my affiliate link.

01:04:19: It's a wonderful tool to just hold a button down and you just talk.

01:04:24: And you can get so much more nuance out that way.

01:04:27: Yeah, it's four times faster than typing, even if you're a good typer.

01:04:30: I'm pretty fast, but.

01:04:32: i can just ramble a bit.

01:04:34: as you've noticed today i tend to go on tangents and i tend to ramble a bit but just just doing that for three four five minutes gives the ai so much more to play with than just.

01:04:45: i'm got this problem and i'm typing it out and i'm going to stop as soon as i can because i'm typing and i'm not very good at typing and go.

01:04:53: if you just ramble for three minutes and just talk to your computer gives it so much more to play with and the answer will be so much better.

01:05:00: i'm so glad we got to that bit that you remind me about that because it's just it is night and day.

01:05:05: I totally agree.

01:05:06: And people need to experience that.

01:05:08: They need to see that and go, Oh my God, really?

01:05:11: I did not know.

01:05:12: I thought it just wrote Dr.

01:05:13: Sue's poems.

01:05:13: I didn't know it could do that.

01:05:15: Holy shit.

01:05:16: Okay, I mean, what else can it do?

01:05:18: And that's the best question you want to get to like, what else can it do?

01:05:21: Not where do I start?

01:05:23: Just start by playing.

01:05:25: Yeah.

01:05:25: couldn't agree more with that.

01:05:27: So now to be respectful of your time and wrapping our obviously very insightful, very great conversation up a little bit, you already mentioned speed and velocity.

01:05:39: And a big struggle I hear from either friends or colleagues and also other business owners is The the speed of change in the AI world.

01:05:50: so there's all the time a new model coming a new feature release.

01:05:54: then there are people obviously on LinkedIn and you have so you have the The confidence to say yeah, yeah, this is an N&A workflow and it Yeah, it's rubbish.

01:06:05: I go past that, but I feel a lot of people have a fear of missing out.

01:06:09: They feel like maybe I should get that.

01:06:12: So how do you handle this or what's like your mindset around all this news and information that's just coming at you?

01:06:22: It's a really good question.

01:06:26: Answer it this way.

01:06:27: Um, so I love books.

01:06:30: I've got a thousand books over there.

01:06:32: I've read eighty percent of it.

01:06:35: Um, We tend to read stuff or even watch a YouTube video Just in case it's just in case learning.

01:06:43: We'll read this book.

01:06:44: It's come out.

01:06:45: We've heard good things about it.

01:06:46: Just in case maybe one day six months from now two years from now I've got a problem where I can use some of that information and apply it.

01:06:53: Yay feel better about that problem And I think we're moving to that world of more just in time knowledge.

01:07:01: so That's certainly what I do with YouTube now.

01:07:04: I'm not just watching video other than all the sports highlights and stuff that I watch there.

01:07:08: It's mostly like I've got this particular problem right now.

01:07:11: I'm trying to figure out this new thing and I can watch something that solves that problem right now.

01:07:17: And so my use of books has changed dramatically over the past two years.

01:07:22: And so to turn that around to answer the really good question, I would say like, focus on the problems.

01:07:28: Don't don't get caught.

01:07:30: in the headlights and looking at all of the shiny objects and thinking that you need every single one and you need to type checklist into LinkedIn fifteen times a day to get all of those things sent to you maybe.

01:07:42: Instead just go like what is the problem I need to solve in the business?

01:07:46: right now?

01:07:47: I picture businesses like a pipe and there's always a thin point in the pipe.

01:07:52: somewhere might be over here and it's fat pipe over here and it goes thin again.

01:07:56: it's fat over here.

01:07:57: there is no point making a fat part of the pipe.

01:08:00: even fatter.

01:08:01: there's no point taking this part of the business and making this part better if there's a really.

01:08:06: In part of the part is the theory of constraints.

01:08:08: it's the bottleneck in the business.

01:08:10: you have to identify the bottleneck in the business and work on that.

01:08:14: there is no point working on other parts of the business.

01:08:17: you could have three times the leads coming in but if you have one person calling people back it ain't gonna matter.

01:08:24: you need to work on the bottleneck first.

01:08:26: so then okay what specifically can I do right now.

01:08:31: To work on the bottleneck?

01:08:32: and yeah, okay I suppose I'm at the point now where I am Consuming a lot of AI stuff because I'm fascinated by it and I want to stay on the cutting edge and I have a community of people now to serve and I feel that sort of part of my job is to And I've always loved this.

01:08:48: I've always loved running four or five hills out into the future and then coming back and not just Reporting on it never been really interested in that but like showing what's possible and building stuff to say like this and making it really useful and really practical and then running off again into the future.

01:09:04: I kind of live my life six months in the future, four or five hills that way.

01:09:08: And then every now and again, pop back into the real world and go like, oh, look at what I can do.

01:09:12: It's amazing.

01:09:13: I just love that.

01:09:14: So I like consuming a lot of stuff, but I get it.

01:09:17: You're wearing twelve hats.

01:09:18: You're running around like a chicken with its head cut off.

01:09:22: It's like you're super, super busy running a business.

01:09:24: Where the hell do you start?

01:09:26: Pick the bottleneck and Exactly what you were saying before.

01:09:30: Talk it through.

01:09:32: Have a chat.

01:09:32: It may feel ridiculous the first time you do this, because why am I talking to a computerized voice?

01:09:38: that's just a big bag of words, apparently.

01:09:40: It doesn't actually think.

01:09:41: It doesn't really know.

01:09:43: And yet it's actually really, really good at that.

01:09:46: So talk the problem through and like, give me always.

01:09:50: I have this thing of one for one.

01:09:52: Here's my one problem.

01:09:53: Give me four.

01:09:55: possible ways.

01:09:56: make them as different as possible as diverse as possible to be four different ways that we might be able to solve this problem.

01:10:02: you're not saying like tell me how to do it and i'll follow what the a i says.

01:10:05: Thank you for different ways that i might be able to explore to maybe try and solve this problem and then give me your one recommendation.

01:10:12: so one four one.

01:10:15: Pick a couple of bottlenecks and try that and then go okay shit let's do it.

01:10:20: Let's try and solve this problem.

01:10:22: Let's ask AI for help.

01:10:24: There's probably some automation in there to free up some time.

01:10:27: And also, if you listen to this, it's not just you.

01:10:30: Lean on your team.

01:10:31: I think it's absolutely an all hands on deck moment for most businesses.

01:10:37: It's got to come from the top.

01:10:38: It's got to be like, this is important.

01:10:40: And yes, we are going to fund this.

01:10:42: We are going to give you the resources to do this again, back to five percent or a hundred percent.

01:10:47: You need to invest in teams right now.

01:10:50: But also those people on the front lines of people dealing with customers all day.

01:10:54: those are the people that are going to have the best ideas how to implement this.

01:10:58: They're all using these tools anyway boss sorry to tell you they're just not telling you about it because they're worried that you're going to tell them off or you're gonna fire them.

01:11:06: All they're going to end up creating the system that replaces them.

01:11:10: the word about that the word that the breaking some a i policy and maybe they shouldn't have pasted all of that client data in and it's against some policy that your lawyer wrote for you and.

01:11:20: So they're worried, but they are using it in their jobs already.

01:11:23: They're just not telling you about it.

01:11:25: I think you're going to get the whole business together and say, right, we are taking this seriously.

01:11:29: We are investing in to start with just basic literacy.

01:11:34: We just don't need to just be on the same page and then let's invest in, right?

01:11:37: Who's going to be that sort of like the advanced team?

01:11:41: Who's going to run one or two hills into the future with me as a business?

01:11:44: But it has to come from the top.

01:11:46: There has to be a vision of where we're running to.

01:11:49: We're not just rushing off in different directions over different hills.

01:11:52: it's like here's the mountain we're going to climb.

01:11:54: this is what the business is going to be like two or three years from now if we all go.

01:11:59: do this now.

01:12:00: I need a few like scouts to come with me one or two hills into the future.

01:12:06: who's going to come with me from base camp up to camp one and help build the trail for everybody else that's going to follow along behind?

01:12:13: like this is the view from the mountain top.

01:12:16: But now let's focus on just getting from here to here.

01:12:18: Let's just get to camp one.

01:12:20: And then let's go build that path and then bring everybody else up to camp one.

01:12:24: Okay.

01:12:24: This is the new version of the business.

01:12:26: Okay.

01:12:27: Take a moment.

01:12:27: Let's have a big party.

01:12:29: Right.

01:12:29: Scouts, you come with me.

01:12:31: Now we're going to build the path up to camp two and we're going to keep on and it just keeps on going and going and going.

01:12:37: But yeah, I get it.

01:12:38: Most, particularly small business owners who are doing twelve different jobs.

01:12:44: It is utterly overwhelming.

01:12:45: They don't have a spare minute, plus they're trying to be super parents, plus they're trying to be friends, plus they're trying to do everything else.

01:12:52: It is really hard.

01:12:53: So just don't get distracted by the shiny object.

01:12:57: Focus.

01:12:57: And focus stands for follow one course until successful.

01:13:00: That's what focus stands for.

01:13:02: Follow one course until successful.

01:13:04: Figure out that thing, that bottleneck.

01:13:06: Fix that bit, get that part of the pipe wider, and then find the next bottleneck.

01:13:12: and don't get distracted and just keep on going.

01:13:14: The game is just staying in the ring.

01:13:17: You will get knocked down.

01:13:18: The whole game is getting back up again and staying in the bloody ring.

01:13:21: My mentor told me that once.

01:13:23: He said there are many, many days where you just want to quit and walk outside the ring and just go, I'm done.

01:13:27: No, get back up and you stay in the bloody ring.

01:13:32: Mike, I can't tell you this has been such an enlightening and insightful session.

01:13:36: So actually I want... So everybody that has listened until now, I can actually recommend you just listen to the whole thing again, because there has been so much to unpack.

01:13:47: So actually this metaphor or this image of going like two, three hills into the future, I will... If you ask my team, we will hear this a lot now.

01:14:00: I've got a whole document.

01:14:01: I've written around that.

01:14:02: Happy to send it to you.

01:14:03: Link in the show notes, whatever you want to do.

01:14:05: Yes, that's how I picture it.

01:14:07: I'll send it to you.

01:14:08: Let's

01:14:08: do it.

01:14:09: So, Mike, thanks so much for this insightful session.

01:14:14: It has been a blast.

01:14:17: Everybody that wants to follow you, see more of the stuff you're doing.

01:14:23: Try some of the tools that you're doing.

01:14:25: What's best to follow you, where to find your stuff?

01:14:28: I'll

01:14:30: make it easy for you.

01:14:31: Nine characters, eighty twenty agent.

01:14:34: So eight zero two zero agent dot com.

01:14:37: Um, all the courses are there.

01:14:38: The link to the ads to AI community is there.

01:14:41: That's the best place to find me.

01:14:43: I don't really do social media.

01:14:44: I'm on LinkedIn a bit, but I don't do anything else.

01:14:47: But yeah, eighty twenty agent dot com.

01:14:49: Everything's there.

01:14:51: Awesome.

01:14:52: Mike, thanks so much.

01:14:53: And, uh, yeah, everybody else.

01:14:56: Go explore, try to use some of the concepts and some of the ideas and some of the thought starters that Mike has given us today and just go try it out today or tomorrow.

01:15:08: Thanks very much, man.

01:15:08: Appreciate it.

01:15:10: Thanks for having me.

01:15:10: It's been great fun.

01:15:11: Thanks so much.