Jade Nangah: Hello, all you marketers out there. Welcome back to another episode of The Algorithm Isn't Enough. I am your host and founder of Voice Scale, Jade Nanga, and today I'm honored to be joined with Philippa Games. Philippa, tell the marketers out there a little bit about yourself. Yeah, so ⁓ I've been around in the digital space for a while. But right at the moment I'm working with senior business leaders, CEOs, et cetera, around the real role that the website plays in the overall business strategy and some of the information that business leaders can get from their website ⁓ that can feed into that business strategy and whether they're relying on the right evidence ⁓ to make decisions. I love that. And I I was doing some lurking on you in your background, twenty-five years of experience of talking about website and the evidence that they gives for business decisions. What's some of the biggest things or what's maybe the biggest thing that leaders are still getting wrong about their own website? Well, I would say actually two things. one is it it's it's it's really bad and I don't want to insult anybody, but in some ways it's it's the real basic stuff, right? you can't tell if your website is successful if you don't know what it w you want it to do. And so exactly what are you aiming at are not only the big goal which is usually either to sell something or to get leads or, you know, the the revenue piece, ⁓ but also all the other i in inter interim goals. but you know, but when I work with a client, I mean the first thing we talk about is what's the goal, what are you trying to do here, what are all the different things you're trying to do with your site? And and still to some extent people have difficulty in articulating that, as I say, beyond the sort of really obvious. ⁓ and the second one is the is the whole walk in my shoes thing. And I mean th this again, like if you're really familiar with your own website, obviously you know where everything is, you know how to find it, you know how to do it. and it can be really difficult for you to imagine somebody who's never seen the website before, and is confused 'cause you're like, but it's right there, it's right there, look, there's a link there, right? And but they actually can't see it, and there's often quite good reasons why they can't see it, and you don't see those reasons anymore because you're so used to it. and and so I often say I I mean I've often had that argument with people, you know, like like they'll say it's but it's right there and I'll say, Yeah, but It's actually not because as soon as you scroll it's disappeared. Or because you're using a technical term in your industry that that actually your visitors maybe don't understand, so they don't click on the thing you want them to click on. Or anything like that, right? So it's some of the time and I I mean, I know people are always looking for the next bright, shiny objects, and we're all talking about AI and all of that. And and you know, it's it's always been the same in all of my years of experience. There are still fundamental things that we have to get right. And ⁓ the website I'd say probably is the most fundamental for sure. And I love what you said about it and it in general as a marketer, you have to think about it's not how you see what you're marketing, it's how your intended audience sees what you're marketing and ensuring that they understand from their perspective. And it's the same thing whether it's content, it's your website, anything that's outward facing. and I love that you touched on AI a little bit. Tell me a little bit about the Ramjack story and walk me through what happened when this Term Well, I I I just ⁓ posted about that on LinkedIn recently, so I guess you saw it. I I so you know, one of the things that marketers are increasingly doing is using AI to help figure out what's going on, which makes absolute sense and I do it too. The problem is that I do find an alarming amount of AI hallucinating and drifting. So ⁓ there's one particular thing that I like to do that I have a very detailed prompt in Claude for. I mean it literally goes on for twelve w pages of word. It's extremely detailed, right? If this is so then do this. If this is so then do you know, right? And and it does some analysis for me. And at one point it came up and said, There is no evidence of Ramjack on this website. Ramjack, R A ⁓ J A C K. And I was like, what? And it said it a couple of times and I I mean at first because it's AI, you know, that we it is proven that people tend to believe it because it's AI. So whatever it says, they go, ⁓ that must be something. So at first I'm going, you know, ⁓ that must be a term that I don't know yet. And I'm just being stupid. But obviously that's ridiculous. Right. So I said to it, wait a minute, what is ramjak? And it said, Ramjak is is in your prompt. It says, if Ramjak, do this. I was like, what? No. I have no idea what that is. And of course I check out the whole prompt and there's no mention of Ramjak anywhere. I believe Ramjak is a the name of a company somewhere in New York. But I mean I have nothing to do with them. So that's if anybody knows that company, this is completely aside from that. Great company guys, we just don't know you. It's okay. But and so I said it was Claude and I said to Claude, Excuse me, ⁓ it's not in the prompt. I have no idea what it is and all of a sudden you're telling me that websites have no evidence of Ramjack. And it said, ⁓ you're right. I completely made that up. And I mean this is really frustrating because we're depending on AI, right? and there are times when that's happened and I mean this was not an isolated incident, it's a good story 'cause it's got a very specific word in it, right? But you know, there is no question that that I'm working with this prompt and applying this prompt multiple times as I as I go through my, you know, work week and so on and and it will drift. And I've said to it, what can I do to stop this? Like in my settings it says, in my preferences, you know, apply the prompt exactly as written, do not divert from what the prompt tells you. And it told me, it actually said to me, that doesn't guarantee that I won't drift. I mean, like, what can we do? And yesterday, there was another one where it said something and I went, Wait a minute, what are you talking about? Where'd you get that from? And it said, Well, I just had a feeling. A feeling This is a machine on a computer, you write a feeling. That's what is that? You know, we we humans have feelings. I mean Yeah, exactly. You as artificial intelligence do not and should not have feelings. Do not propose that you do, and you know you don't. I mean, I think that's one of the biggest things as well as marketers, especially I work in content marketing and it's a lot of writing, it's a lot of websites, it's a lot of copy, it's a lot of this and that. And I think everyone, you I I just You can't be a good marketer if you're not double checking, if you're not verifying, if you're not doing your due diligence. And the problem there is, you know, when it says your website has no ramjack on it, I mean obviously that's a a weird thing to say and you know, I really should have queried it the first time I saw it. But if it's saying something that looks plausible, this is where it gets dangerous, right? Because if it looks plausible and there's no sort of red flags in it, maybe we'll go, Okay, it's telling me the right the truth and I need to act on what it's telling me. ⁓ and that could be complete I mean, I had a client recently, just a very small business. ⁓ and I I you know I'm not working so much with smaller businesses now, but there it there it was and and the you know so Google Analytics four now has an AI assistant. And it came up and was jumping up and down saying, look at all this traffic you're getting from China. ⁓ and of course these visits from China are are like one second visits. And it said you're getting all these visits and particularly from one city in China. This person is is ⁓ an executive leadership coach. Right? She does not do that kind of business. And obviously what she what AI is telling her is that these are bots, I mean you can see they last one less than one second on the site, right? They're all coming out of this place in China. ⁓ it's a bot attack. Google Analytics for is not that great at screening out some types of bots. ⁓ instead of which the AI is jumping up and down saying, Look, look, look, you've got a whole new market here, which is complete rubbish. That's just absolutely insane. And I'm like this is the thing though. I mean it supposing this was a a load of bots out of, I don't know, somewhere in South Carolina somewhere that looks authentic and correct, you you don't want to be led down the wrong path here. Exactly. And so what kind of evidence are you seeing that gets missed when someone just asks a chat bot instead of looking at the data that their website's giving them? Well, I mean, one of the things, you know, I talked about one of my services is really saying to business leaders, Are you looking at the right evidence when you're making decisions? and are you looking at the full picture when you're making decisions? So for example, I mean I used to I used to teach in ⁓ some three different business schools internationally. And I I used to love this example. that I I would give my students as an example, right? Say something like, You're the e commerce manager of a a major e-commerce site. You've just received your monthly sales figures and sales are down by 35%. What's you gonna do? Your boss your boss wants to see you in an hour. What are you gonna do? What right? Yeah. And some of them will say, ⁓ my god, ⁓ my god, I'm going to ⁓ look for bugs, look for you know, did the shopping cart break, did did we run out of stock, you know, did something catastrophic happen? Right? Which okay. But other people will say, ⁓ nothing much, I think I might read the paper and have a cup of coffee. Why? Because the the wider context is that this month is January. Last month was December when everybody was buying gifts on our site, and so our sales are always down by thirty thirty five percent in January. It's not a it's a seasonal thing, it's to be expected, everything's fine. And so that's what I'm saying. I mean the AI would be saying, look, look, look, you've got a big problem, your sales are dropped. And until we we get it to the point where it's very sophisticated and understands I mean that's a very sort of basic example, right? But the human understands the specific business and the specific businesses goals, strategies and context more widely. So I mean another explanation could be that a a new competitor entered the market with a very attractive project pro product at huge discounts. Right. So but again you've got to you've got to know that and you've got to say, could this have an impact on our sales? And if so is this reasonable and things like that, right? so ⁓ That's where I would say that when a an AI to me is not that good yet at recommending the so what. So it's and I don't want to knock it because in many ways it's fantastic, you know, making charts, making dashboards, doing the flags that we need, like, hey, your traffic seriously is down this week and something's gone wrong. I mean, and we need to know that. And if we have a good reason, that's fine, but we need still need to know it. Right? That's all great. But the so what, so is this important? What does it mean for us? What should we do about it? In the context of our specific business, our budgets, our resources. Right, I don't think it's there yet. And I think that's where, you know, the balance is. Use it for the things that hopefully can't go too wrong, where it where it takes away your grunt work. but keep your expertise and really use your expertise. Exactly. And the and that's why the human in the loop is so important. There needs to be that taste. And of course, like you said, only a human would understand, ⁓ this is just seasonality. This isn't the building burning down. But and you know, and AI that's not true. And some sometimes it's it already will. Not there. And that's what I think we can't jump the gun. We can't be assuming that AI is further along than it is. While allowing things to fall by the wayside because we're like, okay, at least that's one less thing I have to think about now. Like you need to verify. ⁓ and actually that leads into my next question about how are you personally verifying your AI output it before you're trusting it enough to act on it? Like what is the process that you follow? well I I don't want to give away too much of specifically what I do because that's my service. Okay. In that sense. But No, I mean y looking at a statement that it makes and saying, Can I see the evidence for that? That's really the question that everybody should ask. What is your evidence for that? Right? Where are you seeing that? How can you back that up? If if it's citing a fact, what's your source? If it's citing something it sees on the website or in a document, show me exactly where you're seeing that. Yeah. ⁓ and that's that's what you need to do. And if it does that then it can I mean that's one of the reasons I love perplexity AI, because it always Tells you its sources where other others don't always. And I find that extremely helpful. 100%. And I think that's the biggest reason, specifically with me creating Boyscale, which is an employee advocacy software. I just saw there's so much happening in the market where people are not verifying. They're putting information out online and they're like hoping no one's gonna t test it or check it or pressure test. And I think if you could have more human voices out there sharing their own expertise and things that they're actually experiencing within their company and at least this in this little niche of thought leadership and employee advocacy that I'm familiar with, I think that is so impactful. And I think that's something that people can genuinely trust when they feel that, you know, this is coming from a human with the expertise in their voice and from their role and perspective. And that's something that currently today AI can't replace that. AI can't s present itself to be an employee at a company and then create all this content about what's happening in the company and how they're and how their role is impacting the company. And I think how to ⁓ yeah. Sorry, can I tell you what's worrying me to some extent in terms of of achieving the right balance. So this piece where now we're hearing about watermarks. ⁓ you know, so people will be able to say, Did AI have a role in writing this? And I I'm just concerned about where the line is there. Because no, I don't want something that is completely AI generated and is absolute rubbish. I mean, yes, we all have seen that now, we all hate it. But at the same time, you know, I do use AI in my writing. And sometimes I'll say to it, I want to write about such and such topic. ⁓ can you give me a start? You're right. Give me a give me a start, let you know, help me spark some ideas. But I'm never going to put anything out that I haven't looked at, that I haven't edited, that I haven't put my own thoughts into. Exactly. And yet And yet it might be now bearing this watermark. Right. It's not clear how you can get rid of the watermark, right? And I question whether that's fair in in yeah. I I don't know, and I'm I'm kind of struggling with this question 'cause I've heard people arguing very strongly on both sides of it. I mean my thing is I I love utilizing AI and like you said it's a tool and the AI projects and the skills that I've created and put time and energy into are nuanced for myself, my business, my clients, and that really does make a difference, right? I think people assume that people everyone's just jumping directly into Chat GPT with no sort of prompting, no background knowledge that you're giving, and you're just like, tell right this. That's for some people that may be the case, but for true marketers, this is an art form. This is a career that we're we've built over time. There's skills that we've built that we're now building into AI. And I feel like the watermark somewhat downplays the effort that someone may have actually put into this piece of content simply because okay, I used Claude to scan through it, right? Or to start me off or to go fix any grammar issues or et cetera. You know what I mean? I I think the watermark is and it also doubles down doubles down on this idea that AI is bad. That's what the watermark means to me. It's like, ⁓ the watermark is now like it's it's your scarlet letter. It's like, who cares? Everyone should be utilizing AI in a way to enhance their life and enhance their workflows. Should it be the one way that you're creating ideas? No. Is it should you be copying and pasting directly without prompting it on your own and having giving it your vision and your human taste? No. But I think the watermark just kind of puts any touch of AI under the same guise of like whether someone that would just lazily on Chat GP. And ask the question and put the answer in versus someone who's utilizing AI in a strategic way. Right. And one argument I have heard that I do find persuasive is is somebody said to me, but it's more about video and audio than it is about writing. In the sense that we we also know that people can create, you know, avatars of people who are real and make them say all sorts of stuff that they never said. And that obviously gets very dangerous very quickly. One hundred percent Yes. That we don't want to that I agree. Using people's likeness and just doing creating content and slap just for the sake of it, totally against it. But I think their use cases, if if done well and tastefully, is extremely helpful in moving people's careers forward and their life and daily workflows. Right. ⁓ But I love this. This was such a good conversation, Philip. But tell the marketers out there how they can reach out. If they're looking to get a little bit more help and more insight and from their evidence and from their website, how can they reach out to you and find out more? So my best way is LinkedIn. So ⁓ nobody else on LinkedIn has my name. So that's great. Very helpful. ⁓ and my website is websites that win dot com. ⁓ and I also have a book that came out recently called Website Wealth. a business leader's guide to driving real value through your analytics. So it's actually about web strategy, stories for business leaders, things to think about, ways to approach it. It's not a techie book. It's not a how to use analytics book. so ⁓ and you can find that at obviously on Amazon and so forth. Philippa congratulations on the book. Thank you for coming on. Really enjoyed our conversation for all you marketers out there. You've heard how you can reach out for any marketers out there exploring the world of employee advocacy and allowing their human voices at their companies to tell the story. Voicale.co, you know what to do. Thank you guys so much for listening today. Have a fabulous day, and we will see you in the next one.