Transcript
Welcome to the Flow State podcast. I am Flow State founder and managing director Stuart P Turner. In this five episode mini series, we are talking about the implications of AI, cutting through the hype to talk about real-world use cases, real examples, and what it means to people in their day-to-day jobs. So, Belinda, thanks for joining, welcome to the show. For those who don’t know you, ABX ABM professional, creator of demand for B2B, is there anything else that people should know about you in your career? No, no, not offhand because I don’t want to say that I’ve had a long and varied career. Illustrious and varied. Illustrious. Prolific, yes. I like it, I like it. Well look, thanks for joining and talking about the implications of AI from all viewpoints.
So look, starting with the big question: is this the start of the robot Revolution and are we in a kind of Terminator style situation right now, or is that all just a bit of hyperbole and we should all calm down? What are your thoughts on that?
I find it really interesting because there is a lot of fear-mongering going on, a bit panicking, a bit of panicking, a bit of people wondering if they are not going to have a job tomorrow, especially in our industry. Yeah, it is quite funny because my husband works in a completely different industry and I happened to mention Chat GPT and he goes, “Oh yeah”. I went, “Okay, well that that is good for that”. Is that the sum total of his thoughts on it? Yeah, yeah, oh yeah. There is a barometer that I think we as marketers, we as content creators, are focused on, whereas the rest of the world isn’t focused on it. They are just worried that people are going to not have their own thoughts, that artificial intelligence is going to push through in the next five years and we are going to become subservient to the robots. There is some sort of robot Overlord exactly, society exactly.
And just in the past I don’t know, month, the proliferation of content generation tools. There is a new tool that I’ve been looking at which trains the models on your own previous content. Yeah, right, okay. So it’s moving at quite a speed, which I think we all have to get on board with. And I think if we try and do things the ultra controlled way, say we have to—every email has to be viewed and vetted by four different people—we’re going to get lost because there are already conversations happening. I do not think that it is going to take the place of copywriters. I think there does need to be a human set of eyes. And we’ve been through this. We’ve been through this with SEO, we’ve been through this with people who tried to game the system and then we were told that the internet is here. User generated content means that professional content creators will no longer exist. Yeah. Now we’ve got streaming services. Now we’ve got people who are hungry for consuming content through so many different types of New Media that it’s actually born more content creators than ever.
Yeah, I think that’s quite an interesting point actually because the point you raise around the panic over change, I’d actually liken this to the Kindle example or e-readers, right? Where people were like, “Look, there’ll be no books ever now that you know Bezos has released the Kindle”. And I’m like, I have a Kindle because it’s very practical, but if you look at like trends like BookTok, people still love real books. Like it’s nice to read a book and not look at a screen all the time and just have that, you know, the experience of going to a Bookshop or a library and, you know, having something in your hands. Like I don’t think that’s ever going to be replaced in our lifetime, right? Like nothing will be completely digitized.
No, and I mean, if you look at all these wellness apps, this is hilarious. You read about a wellness app that says “go for a walk”. Yeah, yeah. It’s like go go outside, read a book. Because “go touch some grass” as the gamers like to like to say it, right? Yeah, that’s the game. I mean technology is moving so fast, but what we’re made of organically we’re not built for. Yeah, yeah, it’s an interesting one, hey? Like I think the other point you raised around the the quality, like I’ve been talking a lot about the quality of the data that goes into these models at the moment. And there’s a halt to use another old school example, you know, of making a copy of a copy, if you know if you used a photocopier, which perhaps a lot of people who will listen to this might not be familiar with. But if you have an original copy of something and you copy it many many times, like you copy the copy, the quality degrades naturally because it’s never a perfect copy. And I think that analogy applies to the AI models that everyone’s using there, right? Because I think it was on Harvard Look, I was one of my the favourite podcasts I was listening to this, they were talking about how if you—there’s been a study I think in the US about how if you continue to put AI against each other and continue to get robots talking to robots—the quality of the output degrades because they’re, you know, they’re just producing a slightly worse outcome every single time. So exactly, you know, that’s where the human factor comes in, right?
So the human factor comes in because you’re—well, I think what we’re missing in all of this is humans created this and it is to automate processes that ease our workloads so that we can continually take that step back. And that’s what I think a lot of people are missing. Yeah. A lot of people are talking about AI. And so I’ve been on the Discord chat for Midjourney. I’ve been looking at those tools and there are some true artists out there and they are making these tools work for them and I don’t think you’re ever going to take that away. Yeah. But what we do especially in B2B and Tech marketing, it’s going to come down to someone who knows what they’re talking about wants to connect with someone else who knows what they’re talking about. It’s yeah, we’ve got 31 touch points in a buyer journey and that’s without us. Like 80% of it is self-guided. So we’re not using it to take over pre-sales, takeover engineers. It’s really just to get that conversation started.
It’s interesting you’re talking about the degradation because robots talking to robots is an intellectual pursuit and that’s up here. People who design robots for the rest of us, we’re still looking for that sale. Yeah, yeah. And look, I think that’s an interesting way to kind of draw it back to real life because there is, as you said, there is a lot of, you know, intellectual sort of waxing lyrical about like where all this could go. But in practical terms, and again, we’ve been speaking a lot about this, we use a lot of these models to do a lot of like the boring analysis that would take people ages and would probably be quite error-prone or bias filled. And it’s interesting you mentioned the buyer journey because like are there areas where you think it’s going to add like a huge amount of value right now across the current B2B buyer journey because that is a complicated process now? Yes, to deliver anything against.
Yes, because so there’s something that we’ve been exploring, I say we meaning me, in my own time. And it’s really taking the idea of a digital twin. People think the digital twin is just a carbon copy or a digital copy, I should say, of what you’ve got. Like we couldn’t do in the studio, like we could do with a warehouse. A warehouse is what people think about most often. But if you think about digital twinning in terms of recreating a scenario whereby you’re bringing in new technology, that’s where this kind of stuff, especially when it comes to image makers, so using a generative AI tool for creating those diagrams rather than having a sales engineer work with a team of people for six weeks saying, “We could do it this way, we could do it,” you can create what-if scenarios. Yeah. The video element of it isn’t there yet. So right now we’re still working with PNGs. But what I’ve discovered is being able to talk to these tools and so, “This is a really complex concept,” and then it gives back to you, “Well, this is how I think it works,” rather than you briefing a creative who doesn’t have that technical background. This this is for me it’s a bridge. It’s bringing these worlds closer together.
Exactly as you just said, the more practical use cases that we’ve seen and when you take a much smaller set of inputs, if you like, and just apply it to a very specific outcome, like exactly as you were saying, if you’re trying to visualize a complex solution, like just explaining how you want that to be visualized, I feel I feel like that’s where you get the best outcome from the models that we’ve tested. As opposed to, you know, the sort of Google or Microsoft style huge data analysis because I always feel like that’s where they may fall over more when you’re trying to do something massive based on a set of data that may have some inherent bias. Whereas, you know, if it’s very specific, like like you’re saying with Midjourney, you know, “I have this one thing I want to know, can you tell me that yes or no?” That seems a safer way to to use it, do you agree with that?
I agree with it because every time we give that type of feedback into the data set that’s learning, that the machine is learning from, that’s where we can we can sort of map that journey to actual intelligence. Because what we’re looking at now is we’re looking at people telling us this is how artificial intelligence is going to develop from machine learning. Yeah. Okay. And we’re looking at a giant scraper. I’m old enough so you get to know what a scraper is and that’s my experience with Chat GPT. It’s not AI, but it will get there because it’s it has that capacity. So if we keep feeding questions rather than sort of strange scenarios because people are bored. Yeah, yeah. And it’s, you know, yeah, yeah.
And it’s funny you say that actually because I think going back to to what you were talking about at the start, this is where the the hype versus reality gap is for me. Where what you have right now is not what I would call the true artificial intelligence. I don’t think anyone designing it would say that either. What you have is a very advanced predictive model that knows from a huge variety of outcomes what the next likeliest thing to happen is. And that’s very different. It’s still very useful, but it’s not as exciting as, you know, like the the Skynet solution, right, where, you know, there’s an actual intelligence has been created. Like we’re miles away from that at the moment in terms of the things that are publicly released. I don’t know about privately.
It’s evolving and it’s learning and we should be helping it because there are so many biases in the data. And I read this great thing yesterday about if you want to market to millennials cut out gender, cut out age from your data sets. So think about the models that are being trained now with these big telcos for example who is saying we can give fantastic CX because we’re using every single data point at our disposal. And yet they’ve got massive segments that the new generation who’ll be buying from them just aren’t interested in. Yeah, yeah. Well I think that’s, I mean again going back to your point around having the the human strategic oversight on this, like that’s exactly where this becomes an issue, right? Where if you’re just feeding in the, “Oh, well the, you know, women aged whatever the cohort is, you know, 35 to 45 will always do this”. And it’s like, well, you know, just because they’ve always done that in the past doesn’t mean that they’re always going to do that in the future, right? It’s my favourite phrase, “We’ve always done it like this before”. The yes, the clarion call of no, we won’t do whatever your idea is because it sounds different and weird, so no.
Yeah, look and that’s what you were saying earlier as well, Belinda, I think what’s actually most exciting to me at the moment—keen to hear what you think about this—is the fact that, as you said, if you put aside like the hype and or the buzzwords, the most useful part of this is making marketing more efficient and less error-prone with better data analysis, better, you know, more efficient processes and just a general sort of way to tie things together that you didn’t have at your disposal before. So, you have you been doing any of that? Have you seen any of that in your current role? Like is there any areas where you’re like, “Oh, this has actually been really useful to sort of start to automate this process or deploy the use of a model” within your, you know, your current sphere of influence?
It works really well for small teams. So if you’re lean and if you’ve got a team that you can’t properly onboard into a product because the product is just it’s too technical, I found this is a really good way to deploy some customer engagement rapidly which matters to the customer, not necessarily us inwardly focusing. Yeah. Okay. And and you can just keep growing that because when we do digital marketing, it’s all about test and learn. What what is the market reacting to? What is it saying? The only way you can do that with is with vast amounts of content. And then that content feeds the data that is typically your manual process to analyze. And so this is where I see the greatest room for expansion.
It was really interesting. I spoke to someone in the US last week and she said the comment was, “I don’t like ABM, I’ve tried ABM and the audience pools were too small”. And I just I thought, “But that that’s not ABM”. ABM is bringing an experience to either a single strategic account or a tiny group of accounts because you want them to feel that. You want them to feel that immersive love. So if you’re looking at an audience pool of a couple of hundred thousand, you’re not doing ABM, you’re doing ads. Yeah, I was going to say this sort of standard demand gen or normal marketing, right? If you’ve gone beyond that small cohort of accounts, I mean, exactly, I wouldn’t even call it demand gen. I’d call it marketing because at least in demand gen you’re segmenting the customers. Yeah, yeah. Right. So that’s quite an interesting an interesting bit of insight from the US though, it was really funny. And it’s just it’s really something that I think we can overcome when we take out some of those erroneous data sets and replace it with this is our own experience or this is this is what the market is responding to, meaning our market targets.
Yeah look, I think I think this is an area that you you know you and I think very similarly around because we were chatting about this earlier about the you know finding intent data and qualification of accounts or people in market, which is obviously an ongoing battle in the in the B2B space to try and understand or stage people are out and how you can best support them in their in their own journey. But that’s where we’ve actually seen sort of most successful use of some of the models in our own processes is analyzing large chunks of social media data or scraping as you were saying like a load of potential market research from various places and then just doing that basic analysis to say like, “Well, you know, what were all these people talking about for the past three months and which were the most prevalent themes or topics or which sites are they linking to and, you know, talking about”. Exactly, it’s pretty straightforward stuff but that’s where I think it is. I mean, maybe just for us it is, but it’s that’s where the most value has been because it actually is helping us to understand our customers much more effectively and say, “Well look, like as you would expect probably 70 to 80% of those people are not relevant right now, but the 10 to 20 that we find, that’s where you can then sort of start to shape up a bit more of an interesting way to reach them, right”.
I mean, how are you have you found that working, you know, within your team at the moment? Interesting because we everything related to demand. Demand is content. When you need the content to drive the engine. So I’m not a designer by any stretch the imagination, but what I’m finding interesting and useful is using generative AI tools to explain concepts not just to the target accounts but also to my team. And to help them understand we’re trying to put up ourselves in the place as someone that we don’t know and that’s where we need to use or automate where possible. Yeah. As marketers because you’re right, I mean, some of the stuff that you guys are doing at Flow State is it is machine learning. It is akin to what Google’s doing. When when you open up Google Ads, this is what a lot of people forget. When you open up Google Ads, Google’s got a widget that says, “This is trending this week or this month” and that’s been there for about five years and no one paid attention to it, but everyone pays attention to GPT. Yeah, yeah, yeah.
So that’s how I’m putting it together. And with some of the stuff that you guys are doing with that balance of automation and manual eyes on the engagement I think is where it all starts to come together. Yeah, I mean look, obviously I agree because, you know, that’s what we’re doing day in day out. But I think it’s interesting what you were saying about using these tools to explain things within the context of someone else’s point of view. Because it I’m interested to know like how how you see that sort of evolving from here, particularly with the visuals. Like, so I haven’t explored that as much, but I know as you were saying some people are doing some amazing things with like Midjourney and the the other generative sort of creative engines. And Adobe, I know I’ve launched a few into Photoshop which looks quite cool to kind of extend photos, producing some weird results at the moment. I think I can see where it’s going. But have you actually managed to reproduce, you know, like because I know you guys are selling a fairly complex solution at the moment, have you produced any sort of useful functional diagrams or explainers that you felt have worked really well and what were what were those?
So, for example, I had I had a four hour briefing session with a graph with a designer. It’s quite a deep deep cut. Brave for that one. Always fun. It was a it was a really big PowerPoint for a webinar and I just kept going over and over in my head and I thought the way he’s this—well, way you speak to designers is, “Here’s something I found on another website, that’s the visual I’m trying to think of”. But they still need to make the connection in their head with what they’re creating. Yeah. And I don’t think it’s a language barrier, I never have. But what I found with these generative image tools is I can say everything that an engineer and I have talked about and it will give me that that image. Yeah, that’s cool. And then you can manipulate how you want the line because line style is also really important. And then you can keep talking to it. That’s why I like the generative part because you can keep adding on. You can say, “No, can you make it simpler?” or, “Can you can you show me the interlock between these two systems for example”. So very specifically I was trying to create a technical drawing of how two systems come together to then connect to a bunch of other systems. Yeah. Okay. And it works beautifully. Nice. So that’s cool. The content is still up to us. We’ve still got to fill in the blanks as it were. But think about that from a marketing time standpoint and think about it from an operational standpoint. We still need a designer to turn it into an animation or turn it into a vector file. Yeah. But that type of use is what we should be looking at as marketers.
As you were saying, in the in the space that, you know, we’re both working in, where everyone has a slightly different requirement. The, you know, the briefing time to say luckily essentially this diagram, you know, six different ways with six different sets of content and six different sort of perspectives for different people in a buying group. Like the ability to do that in, you know, like half an hour I guess instead of four hours is massive, right? I mean that’s a huge resource saving all around. Huge resource saving. And think about what technology marketers are facing globally which is stakeholders getting nervous, Boards of directors getting nervous, people in general getting nervous because they’re told they should be nervous. So all of these technology buying groups are insulating themselves, not necessarily because they don’t have the funds but they just think that they need to they need to batten down the hatches. So us on the other side trying to communicate with them, we’re running really lean teams. There was someone and I don’t remember the name of the company who set up a demand gen center for APAC over the past two years. Amazing piece of work, beautiful structure, but that was 30 people. Yeah. Working round the clock to produce the type of content that we now need to produce with one two people in the space of a week. Yeah. Right. So the pressures on technology marketers in particular are really immense because customers don’t want to talk to us. Yeah. They want to they want to self navigate, they want to self-educate and then they’ll tap us when they’re ready. So we have to continually put that content out there.
Yeah, and look, I think for you guys in particular and anyone who’s in that space of, you know, sort of data data architecture and, you know, digital transformation infrastructure, essentially, is like the space you’re in, right? Like it’s I would say it’s an accurate way to describe that, yeah? You know, the the ways into a platform like the one that you guys currently sell are myriad. Like there’s so many different potential ways to discover it that, you know, as you said, I think that challenge is like a real difficult one at the moment where people are like, as you said, they’ve got money, they just don’t necessarily want to spend it and they think they can self-educate. But yeah, I’ve gone through this process myself with Google’s Cloud platform and with Azure to an extent. And we, you know, we do a load of stuff on like Linux and Debian at the moment and you can kind of self-educate up to a point but then because it’s not my like general area of expertise or day-to-day operation, you know, I’m like, you just hit the point where you’re like, “I don’t want this to become my full-time job to become like a, you know, Google Cloud platforms expert because I just, you know, I don’t care”. I just need some I need someone else to do at that stage. But, you know, my my entry point to that typically is I have a specific use case like I need a virtual machine or a server and it needs to do XYZ things. But yeah, there’s just there’s too much information in those spaces. Google’s like documentation for technical people is amazing. Like there’s guides and, you know, sort of instructions everywhere. But for me as a casual user, there’s too much. Like I need to be like, you know, a sort of simple one-stop guide for, “You want a server for a database and you want to push X amount of data in” and that’s really hard to find. You know how you do that. There’s no like one click install. Well there is a one click install but I don’t think it serves my exact purpose.
So how are you guys bridging that gap based on, you know, your your new use of Midjourney and these models, Melody? Like how are you how are you making that happen? The reason I’m laughing is because you’re you’re reminding me of. I worked on a global client when I was back on agency side and I came to them with this idea and I said, “We need to talk to a different audience group, we need a two-pager on this product”. And the product at the time was the company just acquired a Kubernetes startup. Yeah. Okay. And so I worked with a guy from the startup and he was all over it because that was his that was his business previously is how to explain this to business users. Yeah. Right. Not tech people about the value. So we got that down and then it suddenly went through these multiple layers of a global organization heavily matrixed who then came back to me with a 15-page document. Okay. And no drawings. Just a 15-page document. It could be a bedtime reading to get you into it. Yeah. And and when I think about it that was actually not that long ago. It was probably about five years ago. Yeah. Okay. So have we evolved since then? I think we need to evolve since then. And that’s why everything that’s happening with AI in inverted commas with machine learning is is vital to us understanding what people are consuming because we’re doing it now on a manual level. I don’t think there’s anyone who can say they’re using Sixth Sense that well or they’re using Terminus that well. When I speak to vendors, ABM vendors, they still want to get into a room with a customer. Yeah. Because they’ve, you know, trap you in there. The contract, your target audience. So who are you targeting? Okay, we need to talk to them. So it’s still based on that that anecdotal human-based evidence.
See that that’s funny as well because I think even from a value proposition perspective, we’ve been talking about this a lot with my coach and mentor in the UK, Rich Gregory, Grego. Nice shout out, he wrote, yeah, if you’re listening, thanks Rich. But he, you know, he’s big on all the stuff you were just saying, right? It’s like know your customer, specialize, focus. He tells me to focus a lot, which is an ongoing challenge. But, you know, I went back through off the back of being berated stroke advised by over the past couple of months. I went back through I like experience over the past probably 10 years in the BCB space and was like, “Right, well what, you know, who have we worked with, what kind of industries have we actually specialized in, who do we know really well”? And obviously there’s like take down hardcore technology companies like you guys and then there was Investment Banking of various kinds and insurance slash risk companies. And digging into that was really interesting because that then allowed us to do a bit more evaluation of, “Okay, well, you know, why is that useful to you potentially as a customer”? And we were like, “Well, we’re exactly as you said, we’re about five steps ahead of anyone else in the space because we already know, you know, the guard rails, we know what you kind of can’t do, we understand like the regulation around those industries in the ins and outs of marketing people and who’s just going to slam a door in your face and who isn’t”. So I kind of didn’t really realize that that was just—this sounds stupid to take it now—like that valuable until I went through this exercise and was like, “Oh no, people actually do value that because we don’t have to have seven meetings where we, you know, find out about your industry and get to know like who you are and what you care about as a technology marketer because that’s all just stuff that we know already”.
But that’s where I’ve actually found, you know, talking stuff like Chat is quite helpful because then being able to reiterate that in a more, you know, user-friendly fashion is a challenge when, as you’ll know, when you’re deep inside an industry and you’re kind of in the bubble. So that’s where I think they can be super successful, right? Which sounds like you’re saying as well, you just kind of, you know, you can be like, “Look, if I’ve never heard of digital marketing, how why do I care about what like this company does?” and getting a simple summary of that is is great.
Yeah, exactly. And don’t get—we should probably save this for another episode—but Threads. Oh yeah. So there’s a whole lot of conversations we had though, right? Yeah. So I there is a lot to be said for what people bubble up to the top in terms of trending hashtags, which when you overlaid against the rest of the world isn’t actually trending. Yeah, yeah. Phones by the way is Meta’s new Twitter cologne if you in case you’ve been under a rock, if you’re in marketing or if you’re like a normal person. But it’s interesting, right? Because I I joined just out of curiosity. I just think the platform itself is quite cool. Like it’s like a nicer version Twitter. But I was immediately bombarded with all the kind of crap I was expecting. Like old memes, people posting the same stuff that they post on every other social network and just a whole array of, you know, people joining who are just social media people who are like, “What’s all this about a new social network?”. You’re like, “Okay, we’re not a bit past that having gone through seven or eight massive social networks at this point”.
No, we’re not. And and this is so this is where we’re coming back to what are you using the AI tools for? There’s a new social media Network and we’re it’s still the same format. It’s to stop the rampant biases, ignorance, prejudices, evil presences on Facebook. I mean, come on, is that what we’ve come to? Yeah. It was funny. I joined We Are 8 for about two seconds and then couldn’t get the couldn’t get the mobile interface to work so I went, “Well, forget that”. So it still comes down to the experience and what’s the experience that we’re providing. Yeah.
No, I I agree with you. And I think maybe I’m just too old and, you know, embittered about social having been around since almost the very start at this point where, you know, the utility of most of these networks is now massively superseded by having to fight your way through a huge pile of just crap stuff. And Instagram’s a perfect example of this where it’s like show my age now. But back in the day when Instagram first launched and it was just a photo network akin to Flickr, but, you know, slightly more trendy, it was super cool because it had loads of photographers on it. Encouraged everyone to like actually take pictures and like, yeah, the filters were fun, that was really good. And the minute that Facebook as they were then bought it, you knew that they were just nailing some massive stakes into all of its, you know, essential organs because all they did was ramming their advertising model, shoehorning their crappy algorithm from the Facebook feed. And and like that was it. Now it’s just influencers, crap video and like it’s actually a fight to see stuff that your friends post on there. So yeah, I just don’t see how Friends will be any different. I think it’s going to ride a wave of being clean for a bit and then it all get the The Meta treatment in six to 12 months, Max.
Exactly. It’s it’s going to be clean and shiny and there’ll be lots of B2C marketers who flood to it and then already we’re getting call out going, “Oh B2B marketers, you should be aware of this”. Not yet. Like I’m very I’m very aware of it, I’m hyper aware, thank you. I and this is where I think places who specialize who know what they’re about will actually do well. LinkedIn’s not going to go away. LinkedIn is extremely specialized. And what I what I love on LinkedIn. I’m good I’m prone to a good rant as much as the next person, but I won’t post that on LinkedIn. That’s funny because a lot of people do seem to do that though. Well, if you but if you see that that they sort of get pushed down in the feed or there are people who comment and say, “This is not the right forum” and it’s all very genteel and it’s all very civilized. “This is simply not the right forum”. And and that’s what I like because for me as an advertiser LinkedIn, you can’t get that kind of targeting that you get in LinkedIn. Yeah. And that’s what they’ve got. Whereas Facebook, remember we were talking about demographic segments that suck. That’s what Facebook bases is on. Yeah. So I I think there’s lives in the old girl yet. Let’s go to double meaning. Yeah. And I’ll be interested to see where it goes from here.
Yeah look, I think that’s again like quite a salient point right now because I was reading some thrilling stats from last year on advertising spend across Australia. And something—don’t quote me on this exactly, I’ll find that I’ll find the link—but something like 40 to 45% of media spent digital media spend was wasted last year across Australia. And people have increased their spend this year, I think. Which to me either suggests that people are just not paying attention to the basics of digital advertising or they’re not staying across all these changes across all these networks and, you know, the various layers of technology inside the advertising system that are leaching money away from your budgets. And it all comes down to some basics like channel selection and also a lot of the changes like, you know, the removal of cookies and like the fact that you can’t target as efficiently as you could in some areas. So what are you seeing the same thing now, Melinda? Like have you so much experience any of that yourself or seen it in so much?
And and it’s really funny because LinkedIn has become a delivery mechanism of choice and and not for any any other reason other than what it can deliver in terms of targeting. Retargeting if they don’t accept your cookies is really hard to do. I know that we we tend to get lazy and just accept cookies and then we see those things traveling around us. But the other thing is looking at places like Semrush and getting those reports and trying to figure out where where these people’s watering holes are. I’ve I’ve run a lot of experiments with platforms like Reddit because, you know, you can’t advertise there. But what else can you do there? And I think it’s this is now a good time for us to stop being lazy marketers. Because once upon a time in digital, you could just cover probably about six channels as delivery mechanisms and let them run. But when you’re facing budget restrictions, when you’re trying to do new market entry, and as well as are there any new capabilities to your product that you have to push as opposed to a different one, you’ve really got to look at what channels are performing.
Yeah. And what I’m finding is Google Search still performs. Yeah. It still brings those people to you. Oh totally, yeah. And it’s cost effective. Yeah. I mean, I think that’s a perfect example where, because I’ve been running this. I’ve resisted for as long as possible just paying any media event there’s any money for Flow State for as long as I could. But we’ve turned on like a few pretty tightly targeted Google Ads campaigns recently and they’re actually performing super well. But it’s because, you know, we do what you used to do all the time in the olden days. And we’re like, we only specifically want to turn this one on for like we’ve got a brand awareness one running that’s just all based around, you know, B2B agency and or service provider terms. It’s probably only 20 20 or 30 keywords in there. But that’s performing perfectly as it as intended. You know, it’s driving people to the website, which is the purpose of doing it. And it is doing that and people are chatting to us and, you know, like the grassroots the grass green shoots of, you know, kind of leads are coming through now. And it’s just, you know, because we did the basics of marketing where we were like, “It’s largely a brand awareness campaign, we want to get people to our website and see if our website works”. It’s like it’s pretty straightforward stuff.
You take a little this for granted because this is like you’re you’re hearkening me back to my my beginnings in in digital which was, “What does your business do?”. That simple. Yeah. Because then you can create four or five use cases and put it out there and you get a response immediately. And no, they’re not going to give you their credit card details now. I’ve got to give them time. They’re going to come back. But it’s really funny that this this still performs. Yeah. Exactly. And this is what’s funny, right? Because you like it. I know we’re sort of just teaching it, you know, the tried and true methods. As you said, in times of trying times as we allegedly were experiencing there, right? Economically, Search works, SEO works, content works, and some social media networks. Those are like the things that we’ve been using since I started working in digital marketing. Like whatever that was 15, 16 years ago, they were like they work. They work for a reason and everything else that floats around them like sometimes works and sometimes doesn’t. But that’s just, you know, it should be your go-to shortly for if you’re getting your budgets cut. You’ll be like, “Well, don’t cut Search, don’t cut like any very targeted, you know, kind of defined interest, don’t cut anything it’s first party based”. Like it just seems like people have forgotten the basics because they don’t know the basics. I I honestly believe in performance media, you have to have a really solid foundation to know it. A lot of people will say, “Oh digital, oh yeah, we just we’ve got ad campaigns running”. Yeah, but do you delve into them? Do you understand why you put them together?
It’s really interesting what you’re talking about just now because you’re the you’re probably the only group that I’ve had the pleasure of working with and social. It doesn’t say post every day. I don’t I don’t see the point though. It’s just creating like unless there’s a reason to do it, right? Like you’re just creating extra work. But you say when you’ve got something to say and post when it’s relevant and when it means something coming from you. Not just, “Here’s a post”. Well, you know, it’s funny. I mean, drawing this back to your point about content and AI and how this can be more helpful, I think one of the risks you identified is now that things are easier, the potential risk is you’re just sort of increasing the volume of rubbish stuff that people can generate instead of optimizing, you know, doing less with more. Yes. More with less, not less with more, that’s the wrong way around. Doing more with less. But are you seeing that as well? Like, is it he worried that there’ll be a flood of, you know, now AI generated rubbish floating across the internet that will now go back into these models?
And people have been talking about this for years. They have been talking about the training biased models. So on one hand, I’m really excited by there is a tool out there where you can train models on only, you know, your point of view, your your content, your previous content. But at the same time, what if everyone uses that tool? Yeah. Well exactly, yeah. There is huge scope for bias if you just look at your own CRMs. I mean, my god, how much rubbish is in there? Well exactly, yeah. Whatever. And imagine training a model on that. So I think that’s really important for us to to think about because there is going to be so much rubbish. And if you look at the Facebook algorithms, that’s machine learning at its finest. They are they’re really incredible pieces of of technological advancement. Oh totally. And they were done to promote horrible ethics or lack lack of ethics. Well, it’s just unfortunate that they spend so much time and money telling you what’s popular in a very small sort of field and scope of data, right? Yeah. So people will always seek out their own pockets of information and that’s okay. But when we’re when we’re looking at these big scale biases, there is a real danger. Yeah.
There and look, I think I I agree with you. I think where the value is from my conversation so far—I think you and I have spoken about this as well—is there’s the external, you know, if you can find the right information, know your customer piece. I think that’s really that should be a no-brainer for anyone working in this space to be looking at now. And then internally if you touched on a source of intense frustration between Melinda and the CRM point. Where, you know, looking at your CRM and all your marketing channels and being able to say who is interacting with us the most in a useful manner should be the other. Those two things should go hand in hand, right? Because then you’ve got the what do people want out here and what they’re talking about plus are we actually engaging those people, yes or no? But I haven’t spoken to anyone who’s in a position to do both of those things well now, which is somewhat concerning. I don’t know if they’re out there somewhere and I just I’m not hearing about them. But if you nailed both of those things, then your marketing efforts should kind of be on autopilot from that point because you should be able to steer the ship in the direction of, you know, your customers demonstrating an interest and then understand if you’re engaging them correctly. Is that am I like a way off the mark there or is that an accurate characterization of where we should be going?
I think you’re describing a utopian version. And trust me, I’ve been on this this this personal little hobby horse for quite a while now. It is because there’s technology that can do this. This Terminus, there’s Sixth Sense, there’s Demandbase. But they all rely on your data being able to connect to that. They all rely on can you connect all of your other channels. So it’s it comes again, we’re coming back to what what is the quality of that data. And I think I do have to disagree with you because you’re not going to get that because you’ve got sales who wants to view one dashboard. Marketing wants to view it in a dashboard. Everyone is strapped for time. So until someone gets really comfortable with the fact that it’s not going to be easy, we’re just going to see more dashboards. Yeah. And people are diving into it.
Yeah, that’s an interesting point. I think because we, you know, we I mean I personally bounce around a lot between this of the, as you say, the utopian view of where we could go versus, you know, what people just want to pay us to do at the moment. Which you’re not necessarily right next to each other, but I feel like one can lead to the other with the with the right customers and, you know, desire. But it can feel like a bit of an uphill struggle sometimes. I think I’m going to say something which is highly wise. Okay, we’ll go on. There is that difference with with the various sort of plug and play models that you’ve got going at Flow State. Which is that we’re here, we’re a resource, use us to help explain the data to you so that you’re getting the most out of that data. And a lot of people that I’ve worked with, “Here’s your dashboard”. Really? Yeah, cheers, see you later. Yeah. It’s the same with post every day, end of conversation. Yeah. So that’s what I appreciate. And that’s what I think like if we’re talking about whose jobs are on the line, I think that type of model is where we should be moving to. Rather than automate this, automate everything, post every day and buy and here’s your dashboard. Yeah. Right, okay. So what you’re giving is here is real insight, here is here is the why. And a lot of us when we’re doing enterprise sales specifically, we need to know the why. Yeah, yeah.
Well, I think that’s where, and obviously that’s great to hear, like but I think that’s where the we see same as you. That’s where we see the value of these tools at the moment. Is like it helps you get to that faster because you can answer a lot of those small why questions much more easily and effectively now than we could before. And it’s, you know, I think we’re kind of the major value is both in-house and there’s an agency or a service provider that you can just say, “Look, like exactly as you’re saying, like what, you know, why why is this data telling me this?” or, “Like what what is the, you know, the takeaway from this big body of stuff”? Back to your point about the 15 page, you know, document you got back, you can be like, “Well, what the, you know, what are the key points from this?” or, “What are the what are the things I should be saying if I wanted to present this to this person?”. That’s where we find it really useful. And, you know, as you well know, like we’re big on the the human connection as well because I think that’s why businesses like ours still exist, right? Because people want to work with other fun people who know how to do interesting things and you’re never going to get that from a a robot at the moment, I don’t think.
So, I know we talked a lot about how you can deploy these models to analyze things more effectively, reduce human error. There’s a lot of time saving, all very practical stuff that I think everyone should be thinking about right now given the, you know, belt tightening and or, you know, wallet closing, pick your favourite analogy. What are your big predictions for where you see these models being used most effectively over the next sort of 12 months or so? Are there any big risks that you think are on the horizon that we need to be aware of?
Where I see them being used most effectively, I think we’ve really covered is in is in creating that connection. Not only in the minds of the people who are producing the content, ultimately responsible for it, but also in the minds of your target audience. The risks of the of the data sets. What’s in those data sets that these models are being trained on? Because GPT was launched as a piece of freeware and it was very exciting and suddenly you’ve got millions of people playing with it. Yeah. And if if we’re really going to be strict about what machine learning is going to feed, which everyone hopes is artificial intelligence, this is what it’s learning. Think of it like a child. This is what it’s learning from. So don’t—and they’ve been data experts who have talked about this for years. They have talked about be careful of the biases in your own data. Be careful of that data that you share with other people because then it will go into their pool of data. It’s all of this and that that is a huge risk that I see coming. And I think we as marketers should be grown up about it and not flood GPT without innermost thoughts just to play with it.
I think that’s a very interesting characterization actually because I I’ve been thinking the same. It’s kind of emerging from these conversations, we’re thinking the same thing, right? You’re essentially trying to apply some sort of strange cold logic to the messy nature of human interaction with each other, you know, because because to your point, we all create tons and tons of data every day. It just doesn’t get recorded and endlessly analyzed, whereas now more and more of it is. And yeah, like trying to—I mean, it would be a nightmare to try and unpick that, right? And make something useful out of the the random, you know, day-to-day interactions that you that you have with everyone. All of that’s now digitized. So it’s yeah, it’s an interesting point, I think.
Well look, thanks Melinda for joining. I think we covered a lot of pretty interesting areas there actually. So thank you very much for coming and having a chat. It’s been a pleasure. It’s been wonderful being here, thank you so much. I’m I’m acutely aware that we didn’t go down any rabbit holes so I don’t too deeply. I don’t think which is good. I think we’ve done well. There’s plenty of time for that in future conversations, I think. Wonderful. Look forward to it. Thank you.
Thanks very much for listening. This is the end of the show. A few interesting points raised in this discussion. I hope you found it interesting as well. Fairly positive: the AI is here to enhance our abilities and hopefully not just straight out replace them. And as Melinda noted, a very interesting use case in terms of actually using it to break down and explain very complex concepts and then use them in the creative process to actually help her team, you know, execute more effectively, which I thought was pretty cool. If you would like to find out a bit more about who we are, what we do, please head over to our website theflowstate.io.