Should you act on AI marketing advice without checking it?

Transcript

Charles Travers: Hey guys, welcome back to another week of e-commerce marketing chats with myself Charles and Jimmy. Jimmy, how’s it going mate?

Jimmy Dalgleish: Hey man. Yeah, it’s going good. So yeah, a little bit late to this call. So I apologize, but, yeah, no, it’s good to, it’s good to chat again. And, ⁓ actually I saw something actually today, Charles. So I’m not even sure if you would have seen it yet, but I just jumping straight into it, ⁓ Google search console. They’ve got now a section where you can ask AI questions about the data in your Google search console.

Charles Travers: Yeah. Yeah, I’ve actually experimented with that a bit. It was in beta at one point. I don’t know whether they, you know, extended it out now to full roll out. I’ve used it a little bit. I think the first time I used it, didn’t actually work. So I probably didn’t use it, decided probably wasn’t going to use it again. But I will delve back in. I didn’t realize that that’s Because often with Google Search Console, and same with Google, they release a feature and then sometimes they get rid of it. So they’re kind of testing as they go along.

Jimmy Dalgleish: Yeah. So yeah, I am. So I heard about it today and one of the, one of the interesting things that people are using it for. if anyone wants to go in, when you actually log in, it’s actually a bit hard to find. next to the filter butter button at the top, there’s like this tiny little AI symbol and you can click on that and then ask questions, but how people were using it that I saw that were testing it. They were using their report to say, Hey, give me the whole list of like ⁓ informational queries, like who, what, where, when type queries, and then a whole list of transactional type queries around buying intent. And yeah, it was just, it was interesting. And then it just pulls up the data straight for you, saves a lot of sifting through and the AI is pushing it out. So I haven’t had a full test of it yet, but just for the audience listening, I like to keep them. in the loop with what’s happening right now. And it’s literally just been released fully from what I can see one or two days ago. So yeah, definitely worth experimenting. think if you haven’t had a look in Google search console, it’s, it’s one of the, ⁓ let’s say underrated heroes, I guess of knowing what people are searching for, what organic content you should be doing, how your blog posts are doing. Yeah. It’s, I reckon it’s still really underrated. I use it a lot to get keyword ideas for Google ads. And yeah, if people are looking for particular blog content that they should write, sometimes it’s a good place to start in there and see how people are actually searching for you and how you could service some of those queries, I guess, a little bit better.

Charles Travers: Yeah, absolutely. What I’ve been doing recently is using Claude. So I’ll actually extract the, the last 28 days of data. I’ll upload it to Claude and get it to kind of analyze that Google search console data ⁓ or I’ll filter down. So for example, with some of my larger e-commerce clients that have got a lot of collection and product pages, I’ll go granular. So you can actually filter by say collections or product pages, and then I’ll extract and export that list, upload it to Claude and get it to start analyzing and see where the opportunities are. Literally just earlier today, I gave it two different spreadsheets. So I gave it all collection pages over the last 28 days compared to the previous 28 days. But actually, more importantly with e-commerce stores, I prefer to look at year over year. So I did the last 28 days compared to the to the previous year, year over year and uploaded that as well. So I then got a full report of all the potential opportunities. And, you know, it just, it just saves so much time doing it manually, really. That’s the biggest thing. And, you know, Claude, or any of these platforms, they may not get it 100 % right all the time. but that’s where I still think there’s value in having an SEO behind just to double check those things and make sure that the suggestions are legitimate and are worth actioning.

Jimmy Dalgleish: Yeah, we were talking, I was talking about this today. Like the AI is great and people that know what they’re doing, make it sound easy. But then when people that don’t know what to do or what to ask or the questions to ask, for example, if you’re creating an app for your Shopify store, you can, ⁓ basically do it at the touch of a button. But the thing is people building the better apps are the ones that still know coding. And still know what to ask and what it should look like and how to present. So, I think, I think the gap is like shortened. Like it’s not going to take you a long time to understand what to ask it properly. But if you’re just starting out trying to replicate what someone that’s been doing it for months has done, then may take a little bit longer. But what I have seen today too, just how powerful the Shopify AI is becoming.

Charles Travers: Yeah

Jimmy Dalgleish: ⁓ sidekick is amazing for, ⁓ I saw it today. So someone editing a part on their website and they just went to someone else’s website, found a little feature that they liked on their website, took a screenshot, ⁓ went into the theme editor in Shopify pasted the screenshot in and said, ⁓ build me this on my store. it, took that, it took that little widget and it built out.

Charles Travers: Wow.

Jimmy Dalgleish: everything in the theme editor too. So you can edit all the parts directly in the theme editor. You don’t have to go into the code. and it didn’t just copy it directly. It kind of tailored it to the theme of the Shopify store, which I thought was nuts. what I would say is it takes a little bit of tweaking because I saw this done live and it looked great on desktop, but the, the mobile optimised you needed just a tiny little bit of work. if

Charles Travers: Yeah, that’s incredible.

Jimmy Dalgleish: I will say to anyone, if you are doing any Shopify AI stuff on your theme or changing anything like that, make sure you’re checking the mobile view too, because you want to make sure whatever you’re building on your desktop view is more importantly optimized for mobile. You might even want to start mobile view first. to, um, that’s where, honestly, that’s where the majority of your traffic’s coming from. I’d say most e-commerce stores now 70 plus percent in. mobile traffic compared to desktop. yeah, like if you, if you’ve been holding back on like fixing your website or changing it, like just go into Shopify, you don’t have to edit your live theme, go into Shopify, duplicate your current theme that you’re using, and then just play around with it with the AI and see what sections you can change, what you can add. I think, It’s going to go a really long way to improving some of these really average looking websites, conversion rates. And you can basically go to a really high converting website that you know is doing well, copy different features on there, implement it for your own site and tailor it to your theme and design. And then, yeah, you, you’re good to go. So I haven’t tested it. So I’ve seen it work. haven’t tested it on a live site yet, but from what I What I saw, there was no add to cart function or anything on these particular things that we’re building out. So, and it worked in the live editor. So I, couldn’t see why it wouldn’t work on your live site too, but yeah, I think AI is helping e-commerce stores a lot. I think I watched Alex Hormozi like interview a website, a design company and the guy was saying, you know, we’re, losing. We’ll we’re losing all our customer base. You know, they’re all going on AI and doing it. And Alex, he kind of reframed it a bit. He’s like, all I’m hearing is your job just got a hundred times easier. Like, and he said, the people that are coming to you for a, for a website design, weren’t these vibe coders anyway. like the, the vibe coders weren’t coming there. It’s like these tech literate people that.

Charles Travers: Yeah, that’s true.

Jimmy Dalgleish: Wouldn’t have a clue about AI that just want a new website. And the guy’s issue was overhead. Like he had too many staff and low margins. Alex and Mozi’s like cut half, cut off your head count and implement AI yourself. You’re worried, you’re a web design agency. You’re worried about other people using AI and you’re not even using it. It’s, ⁓

Charles Travers: No, it’s a good point. mean, and it’s like, for example, in the in the SEO industry. Like I said, the the small companies, they’re not coming to me, they would have gone to Fiverr in the past. Now they’re just trying to experiment with these, you know, AI agent hacks. And yeah, I saw one the other day, which was like, you know, I, there’s a lot of There’s a lot of posts out there which you do really need to verify first. Like I saw one which was saying, I fired my SEO agency that was charging me $10,000 a month. Firstly, you know, is that actually true? Probably not. And two, I fired them to, and now I have a whole team of AI agents running my SEO campaign. The biggest thing that I don’t mention is the results they get. It’s all well you fire your SEO agency, but.

Jimmy Dalgleish: Yeah.

Charles Travers: Is it actually leading to growth? So you need to, you know, there’s a lot out there and I’m worried that a lot of information is going to lead a lot of e-commerce stores down the garden path. And they’re going to think, okay, we need to fire our, you know, our agency that is doing a good job at the moment as it is. And yes, there’s, you know, it’s always good to kind of review the staff, where you’re at, can you save costs potentially? but equally it depends on what stage you’re at. Like if you’re talking at like a seven to nine figure e-commerce store and they’re gonna sack their Google ads guy or their SEO agency and then try and bring in a bunch of AI agents to run their brand. Yeah. I would suggest that’s probably not the.

Jimmy Dalgleish: Yeah, don’t get me wrong. Don’t get me wrong. think it will, I think it will definitely get there. Um, but I’m still not sure. I still, I still think the agencies that build like the agents of stuff will still be getting client. don’t think, I don’t think you’re right. Large companies like that will be risking it internally. They’ll be going again to someone who’s tried and tested all these methods, at least, at least for the, for the foreseeable future. Anyway, the rate that the technology is changing, who knows? that looks like in, in a year, year or two, but at the moment we’re at the stage where, cause I am coaching a lot of students in e-commerce. I’m having them come to me and show me the advice that, know, chat, JBT or their, their AI agents gave them. And a lot of the time you need to be very, very careful of, the data that you’re, ⁓ that you’re feeding. You’re, feeding AI. So let’s, let’s use this as an example. I had someone coming to me saying, they’re, ⁓ the, the AI agents telling me like I should scale Google ads. It’s going fantastically. And I’m looking at the account and the return on ad spends less than one. I’m like, what if I was, if you, if you scale this, you’re, you’re just going to be like bleeding money. So Then I had a look into what they were asking AI. So what they did with AI is they said, ⁓ I’ve done, ⁓ I’ve done this much in revenue on Shopify and this is how much I’ve spent on Google ads. So the AI has taken that as all the revenue that’s in the backend of Shopify has been generated by Google ads, which was not true. ⁓ so it’s like, yeah, if you’re only, if you’re only spending a thousand dollars and you’re getting $20,000 in Shopify, that’s a 20 return on ad spend. should scale Google ads through the moon. But Google ad that one, they had such a small budget, it wasn’t profitable, but the AI didn’t have that context. So you need to be very, very careful about double checking the data that even AI is referencing. in that case, AI just… sees, okay, this much revenue, this is what you said, how much revenue, this is how much you spent on Google ads. And they’ve basically said, you know, all the revenue in your Shopify store has been as the result of this small amount of spend that you’ve got on Google. So you should just scale it through the moon. Um, and I’ve had a, I’ve had a few different instances where students have shown me the recommendation for what they should do in Google ads. And the context has just been completely off like that’s pulled data from the wrong place or, ⁓ it’s gone over all data rather than just your Google ads data. So yeah, people need to be really careful about kind of the advice that it’s taking from AI because the advice that they’re getting is only as good as the context and the data that AI has that they’ve provided. And they, sometimes they don’t even think about that. yeah, be, be careful in how you implement it because yeah, I’ve seen a couple of recommendations over the last month or two that they had to just blindly implemented that they would have lost hundreds of thousands of dollars in some, sorry, in some cases. So yeah, don’t just follow the advice that you’ve been given blindly. Still make sure it’s subjective, double check it. Um, have a look into, uh, have a look into if you, if you think it’s too good to be true, sometimes it probably is. Um,

Charles Travers: Yeah, It’s a really good point. It’s a really good point, Jimmy. I think, you know, there’s nuance to these situations, you know, especially when it comes to the actual strategy side of things, which, you know, an experienced specialist will pick up that AI isn’t quite at that level yet to do so. ⁓ But I think more the specialists train the AIs, the more we’re going to get, you know,

Jimmy Dalgleish: Yeah. I don’t even think it’s the, I don’t even think it’s the AI. I think it’s the user. just like the AI would probably give better recommendations if they have full context, but the user that’s asking the questions is not giving it all the data or the full context. So the results and the feedback that they’re getting back is, is poor. ⁓ so I think. Yeah, exactly. Yeah. A hundred percent. So I think if they, ⁓ I’ve.

Charles Travers: way we modernized knowledge. Yeah, that’s true. Yeah, it’s only as good as the inputs that you give it.

Jimmy Dalgleish: I honestly find the AI advice pretty, good if you’ve given it like really good context and the right data to, to go off. But the issue is at the moment, a lot of the times that data isn’t good enough to get the, get the recommendations that you need from AI. So, ⁓ yeah, it’s, just worth keeping that in mind.

Charles Travers: Yeah, for sure. A couple of things I’ve been working on this week, Jimmy, I’ve been kind of seeing how I can further execute AI into my services and my workflow. And specifically, Claude skills is something that I’ve been kind of experimenting with this week, which is basically building out a workflow or instructions that you can effectively install into your Claude account and you don’t have to, whereas usually when you’re creating or you’re asking Claude to do something, you’re having to, if you start a new chat, you’re having to tell it what it needs to do every time. Or you need to, to input a new prompt every time with Claude skills, you create this kind of workflow or set of instructions. You install it into your account and then all you have to do is just ask it. a question or ask it to do something and it’s already got that skill preset into your system. So I was been building out an e-commerce SEO audit at the moment and simply just ask it, know, Claude, my client’s domain is this, can you compare it against their top three to five competitors, run me a full SEO audit of the site and I don’t have to. go into any more detail than that because I’ve already created that skill. So yeah, really, really valuable. And one thing I would suggest is if you’re unsure how to get started with Claude skills and you want to kind of systemize your workflow, create a new project in Claude if you’ve got the paid account, open up a new chat and say Claude, I want to create a new skill around X, Y and Z. Can you give me the framework on how to do that. And it will just build you out the skill for you. And then you can start talking to it and customize it as you go along. And then once Claude has delivered that skill, you simply just have to click copy to skills and it will just automatically add it to your repository. So there’s no real, you you don’t need any real technical skills now. It used to be that even you just have to upload a file. Now it’s just, yeah, it’s pretty much automatic.

Jimmy Dalgleish: I think too, think, I think the skills parts, it’s becoming valuable too. Cause I see some of the best in the world creating skills and then you can base and some of them are just giving them away to other people. you can take a skill that someone at the elite level has built and implemented into your own systems and then tailor it to like, I find that with everything, even if the best person in the world’s built it, whenever I run something from someone else, it’s like, ⁓

Charles Travers: Yeah.

Jimmy Dalgleish: This is great, but this is how I do it. And you just want to change a few things in it to your own nuances and to your own workflow. So you can do that now. So, but yeah, like Charles said, you can start that way and build your own, or you can go out and find someone else’s skill and then tailor the run it, see what it does, and then tailor it and add your own nuance to it.

Charles Travers: Yeah, I think there’s a whole GitHub isn’t that get hard depository of like skills

Jimmy Dalgleish: Yeah, it’s almost like it’s becoming like its own little marketplace. I think it might become its own marketplace at some point, like people building out skills. We were talking about this the other week where ⁓ SaaS providers will basically almost provide just a template for the SaaS. And then you can go out and build your own SaaS product on top of an already working product. So yeah, I think you’re going to see a lot more of that. think SaaS businesses will be in trouble. Especially if they’re relying on ⁓ like kind of high monthly costs and ⁓ like buy your seat type SaaS models like that, that’s gonna be gone sooner than people realize. And I think the level of SaaS that are coming out into open source, it’s not gonna be too far off. you being able to build an enterprise level SAS and tailor it to your own needs. So you can think about kind of the softwares that you’re using for e-commerce. Now it’s going to get to a point where you’ve been asking for all these features that have never come and you’re still waiting and like parts of the parts of the software annoy you. You’re basically almost going to be able to take a template of, that software and then implement all the, changes that you’ve been asking for and just have your own version of that software. So, ⁓ I’m looking forward to it. I think it’s, ⁓ I think a lot of SAS companies are going to be in trouble, but I think a lot of people’s workflows are going to be enhanced massively. So it’s just, I just know all these people that have had so many ideas that never have never been able to build it and now building them now. So you’re going to get a lot more stuff coming to market because there’s not that technical co-founder constraint anymore. Sure. You might need them just to like polish off the last five, 10%, but compared to where we’re, I think that five to 10 % will be gone too soon. You would just be able to ask the chat GPT basically everything or the AI, whatever I’m using Claude code too. So you’d be able to just use Claude code. So, Hey, fix this. And you’ll be able to do it all yourself. But yeah, if you’re. If you’re having a frustration or there’s something at the moment that there’s not a solution for, you can just build out almost your own little software as a service. And then I’m not saying it’s going to go mega. And if it does go really big, then expect to have copycats of your, this is what people don’t realize. They think, ⁓ I’ve got this great idea. I’m going to build it. Soon as it becomes big in any capacity, you’re going to have a hundred people copying exactly what you just did. And they’re going to be able to do it with ease as well. ⁓

Charles Travers: Yeah, I think it’s more of an add-on SaaS, it? More than the actual, the main business. I think it will become a, know, potentially an additional revenue stream for some businesses opposed to the core service or the sorry, the core product. Yeah, I must admit one that I’m very close to wanting to cut is SEMrush. SEMrush is in crypto.

Jimmy Dalgleish: ⁓ You’re not the only one that said that to me this week. A few people have said that to me.

Charles Travers: Like, SEMrush, like for those listening, a great, for many, many years, it’s been one of the best all-in-one SEO tools out there. The problem that really frustrates me is they just, at every level, they want to increase your subscription. So they’ll add these new features and then you go to click on them and it’ll say, ⁓ yeah, you can have this, but it’ll be $67 extra a month. And before you know it, you’re spending a small fortune. So yeah, I’m contemplating seeing if I can replace that. But really the only thing that’s holding me there is I’ve got all my clients, you know, keyword ranking data on there. They do have a pretty good competitor analysis, but equally, you know, there’s other tools out there now that do pretty much the same thing. And if you can plug it into

Jimmy Dalgleish: Yeah.

Charles Travers: some of these larger LLM tools. You’ve got all the data there anyway.

Jimmy Dalgleish: Yeah, it’s true. So, um, I was thinking like this software that I wanted the other day, I only wanted like a small portion of it. So it was only going to use like a very small portion of it, but there’s like all these other features that I will just never ever use. And the price of the software just didn’t justify the small section that I wanted to use it for. And I think you’re going to see a lot of SAS companies just fill that needs like, don’t pay 150 bucks a month for this. If you just want this small portion.

Charles Travers: So.

Jimmy Dalgleish: just come here and spend 10 bucks a month. And I’ll be like, yes, that’s what I, that’s the only part of it that I want. I am, I’m just going to go there and get that. think you’re right. They just keep adding all these features that no one’s asking for, on a lot of different sassas. And then they almost do that just to justify increasing their price. So it’s like, well, I prefer to just keep paying the same price and using what I had, cause I’m not going to use any of these new features that you’ve just implemented.

Charles Travers: Yeah, exactly. And it’s very easy. Software is, you know, it’s almost like software creep, isn’t it? You can, yeah, before you know it, you’re spending a small fortune just on software alone. So I think, yeah, certainly over the next few years, that’s going to be one cost area that’s going to be hopefully significantly lower for many e-commerce stores where they can start implementing more custom software for their own store. and they won’t have to rely on as much third party SaaS products. So yeah, exciting times.

Jimmy Dalgleish: Yeah, I mean, I think the main ones I’m using at the moment on agency side is obviously monthly reporting software. ⁓ I’m using Swido at the moment, but I can see myself replacing that in like just when I get time to build out something for clients. I think that’s the thing at the moment. Well, once people have time just to have a look at what they’re spending on SAS and be like, ⁓ and those who don’t, SAS is just means software as a service sometimes. Charles and I use acronyms and I forget that not everyone is aware of all these acronyms, but yeah. So SAS just means software as a service. And I would say that yeah, people, especially, especially lean businesses, they’re to be reviewing all their softwares that they’re using and be like, okay, which ones can we rebuild internally? Sure. It cost us a little bit of time upfront, but then once we tailor it to our own solution, it’s more often than not going to be better than the solution that they’re using anyway, because it’s specifically tailored to what they actually want it for. And I think you’re going to see a lot more of that. And it’s already happening. ⁓ I’ve seen it everywhere, the ecommerce coaching community, they’ve already built their own internal AI platform, all the students are in there, they’re using the AI, they’re asking, this is where I’ve been seeing a lot of the AI feedback and where it’s been giving some wrong information based on the data that they’ve been given, but, ⁓ it’s like, it’s version one. So there’s going to be, there’s going to be a few, a few little linkups in there, but yeah. And, ⁓ the speed of implementation now, a student will ask, Hey, I need this in the software. And you know, that used to be submit a ticket. It might get visited a month down the track. Now it’s literally like, okay, I’ll implement that this afternoon. It’ll be live tonight. ⁓ that’s speed of it now.

Charles Travers: Yeah, yeah, no, for sure. Sure. Yeah, I think actually the only thing holding some e-commerce and other companies back is actually getting approval on certain things. That’s the thing that I sometimes find frustrating from my end as service provider is I can, okay, great, I’ve used AI, I’ve created this report that we need to do X, Y, and Z, but I need to get approval in order to get these things live. which can take me matter of minutes to get done, but I’m waiting on an answer that might take me two, three weeks. And I just think in that space of time, the more leaner e-commerce stores that don’t have that line of approval are gonna get things done a lot more quicker. especially for those larger stores that have got bigger teams, you certainly need to be aware of that because you’ve got these small players that are kind of solo. that if you’re not careful, are going to be your competitors in the future because they’re going at a much faster rate because they can get things done a lot quicker because they don’t have that red tape.

Jimmy Dalgleish: Yeah, I agree. It’s been a bit, you even look at it now, you look at some of these older industries and you think to yourself, I could rebuild this over the weekend and probably get more leads than what they’re getting right now. Like that’s where it’s coming to. ⁓

Charles Travers: Yeah, for sure, for sure. Yeah, I think like you said, it’s allowing your team and the service providers to be able to execute ⁓ on the task because the tools are out there to get things done a lot faster now. It’s just a matter of doing them.

Jimmy Dalgleish: I think, ⁓ I think if people are keen to jump into using AI a bit more likely we discussed earlier in, this episode, I think they should go into their Shopify store and start to add some additional features that they’d want on their website homepage. Start with editing the homepage first. It’s going to be the easiest win. Go find, ⁓ go find a competitor or a high converting website and start to. Take a screenshot of the little segment that you want to use. Plug it into your theme editor, use the AI to build it. Take about three minutes and then you’ll have it on your website. And I think once, once you see that it’s going to click, you’re like, ⁓ gosh, I can rebuild my whole homepage probably over the weekend. And before like you’d have to go to dual developer, go through a whole design process. Honestly, sometimes it’s like four to five weeks of just. back and forward before you see anything, you can rebuild your website homepage now over the weekend. and not only that, you can model websites that have really high conversion rates or that you know is selling, ⁓ like hotcakes. So there’s really nothing holding you back to at least if you’re looking to improve one thing this weekend, don’t have a play with that and try to improve your conversion rate. ⁓ this weekend. So yeah, because I know we talked about AI apps and probably some people in ecommerce like, oh, I use any of that for my business yet. Or they might be paying for software as software at the moment, but they’re like, I’m just in the weeds from ecommerce business. So if you want to get a kind of a small win from your ecommerce business this weekend, Go test that in the theme editor and make sure you duplicate your live theme. Please do not edit your live theme. You may run into issues and it may stop the website working potentially temporarily. So make sure you’re doing it on a copy version of your theme when you are doing it. yeah, I think, um, we’ll kind of leave you guys with that action tasks this week. Go and, uh, go and experiment with the homepage and, and, and see what you can do.

Charles Travers: Yeah, good point, good point. Sounds good to me. So, do you to see us off?

Jimmy Dalgleish: Yeah. Well, yeah, we haven’t posted in the Facebook group for a little while. Charles was just messaging me about that the other day, but you can reach us there. So if you want to ask questions or chat, go check out Shopify marketing secrets on Facebook. Otherwise everywhere else where e-commerce marketing chats on all the podcast platforms and on YouTube. until then, we’ll catch you guys next week.